case study x ray

LOGIN 

Annual Report

• Board of Directors
• Nomination Process
• Organizational Structure
• ATS Policies
• ATS Website
• MyATS Tutorial
• ATS Experts
• Press Releases

Member Newsletters

• ATS in the News
• ATS Conference News
• Embargo Policy

ATS Social Media

Breathe easy podcasts, ethics & coi, health equity, industry resources.

• Value of Collaboration
• Corporate Members
• Advertising Opportunities
• Clinical Trials
• Financial Disclosure

In Memoriam

Global health.

• International Trainee Scholarships (ITS)
• MECOR Program
• Forum of International Respiratory Societies (FIRS)
• 2019 Latin American Critical Care Conference

Peer Organizations

Careers at ats, affordable care act, ats comments and testimony, forum of international respiratory societies, tobacco control, tuberculosis, washington letter.

• Clinical Resources
• ATS Quick Hits
• Asthma Center

Best of ATS Video Lecture Series

• Coronavirus
• Critical Care
• Disaster Related Resources
• Disease Related Resources
• Resources for Patients
• Resources for Practices
• Vaccine Resource Center
• Career Development
• Resident & Medical Students
• Junior Faculty
• Training Program Directors
• ATS Reading List
• ATS Scholarships
• ATS Virtual Network

ATS Podcasts

• ATS Webinars
• Professional Accreditation

Pulmonary Function Testing (PFT)

• Calendar of Events

Patient Resources

• Asthma Today
• Breathing in America
• Fact Sheets: A-Z
• Fact Sheets: Topic Specific
• Patient Videos
• Other Patient Resources

Lung Disease Week

Public advisory roundtable.

• PAR Publications
• PAR at the ATS Conference

Assemblies & Sections

• Abstract Scholarships
• ATS Mentoring Programs
• ATS Official Documents
• ATS Interest Groups
• Genetics and Genomics
• Medical Education
• Terrorism and Inhalation Disasters
• Allergy, Immunology & Inflammation
• Behavioral Science and Health Services Research
• Clinical Problems
• Environmental, Occupational & Population Health
• Pulmonary Circulation
• Pulmonary Infections and Tuberculosis
• Pulmonary Rehabilitation
• Respiratory Cell & Molecular Biology
• Respiratory Structure & Function
• Sleep & Respiratory Neurobiology
• Thoracic Oncology
• Joint ATS/CHEST Clinical Practice Committee
• Clinicians Advisory
• Council of Chapter Representatives
• Documents Development and Implementation
• Drug/Device Discovery and Development
• Environmental Health Policy
• Ethics and Conflict of Interest
• Health Equity and Diversity Committee
• Health Policy
• International Conference Committee
• International Health
• Members In Transition and Training
• View more...
• Membership Benefits
• Categories & Fees
• Special Membership Programs
• Renew Your Membership
• Update Your Profile
• ATS DocMatter Community
• Respiratory Medicine Book Series
• Elizabeth A. Rich, MD Award
• Member Directory
• ATS Career Center
• Welcome Trainees
• ATS Wellness
• Thoracic Society Chapters
• Chapter Publications
• CME Sponsorship

Corporate Membership

Clinical cases, professionals.

• Respiratory Health Awards
• Clinicians Chat
• Ethics and COI
• Pulmonary Function Testing
• ATS Resources
• Live from the CCD
• Pediatric Division Directors

Reviewed By Allergy, Immunology & Inflammation Assembly

Submitted by

Jarrod Bruce M.D.

Pulmonary, Allergy, Critical Care and Sleep Medicine

The Ohio State University Medical Center

Columbus, Ohio

Joanna Williams M.D.

Clinical Assistant Professor

Department of Pathology

Nitin Bhatt M.D.

Assistant Professor of Clinical Internal Medicine

Maria Lucarelli M.D.

Associate Professor of Clinical Internal Medicine

Submit your comments to the author(s).

A 61 year old female presented to an outside hospital with increasing shortness of breath and cough. She complained of a non-productive cough for approximately six weeks.  She denied hemoptysis but complained of a low grade fever. Over this time she became progressively more short of breath. At the time of presentation she was able to only walk one hundred meters before she had to stop due to dyspnea.

Her past medical history included obesity, gastroesophageal reflux disease (GERD), peripheral vascular disease, hypertension, and type II diabetes. Family history was negative for lung cancer or chronic obstructive pulmonary disease (COPD). Her medications included aspirin, metoprolol, lisinopril, insulin glargine and esomeprazole. She denied any previous tobacco, alcohol or intravenous drug use. She was currently disabled due to chronic lower back pain.

Physical Exam

She presented to an outside hospital emergency department and was found to have an abnormal chest x-ray. Posterior-anterior (PA) and lateral films were provided ( Figures 1 and 2 ).

Figure 1. Posterior-anterior chest x-ray demonstrating left sided consolidation without volume loss.

Figure 2. Lateral view on chest x-ray demonstrating left upper lobe consolidation.

Subsequently the patient underwent computed tomography (CT) of the chest with pulmonary angiography. Two cuts are provided below ( Figures 3 and 4 ).

Figure 3. Computed tomography image of the chest demonstrating consolidation of the left upper lobe.

Figure 4. Computed tomography image of the chest demonstrating left upper lobe (LUL) consolidation with endobronchial lesion in LUL bronchus.

The patient was admitted to the outside hospital and underwent bronchoscopy for evaluation of left upper lobe (LUL) and lingular collapse. She was electively intubated for the procedure and bronchoscopy revealed an endobronchial lesion in the LUL. The patient was transferred on mechanical ventilation to a tertiary care center for further evaluation of suspected endobronchial malignancy. Repeat bronchoscopy was performed with results as shown in Video 1 .

Bronchoscopy revealed a foreign body in the left upper lobe bronchus. Using cryotherapy, the foreign body was successfully removed and eventually identified as a popcorn kernel ( Figure 5 ).

Figure 5. Foreign body identified as popcorn kernel.

C. Clinically, aspirated foreign bodies may commonly be confused with endobronchial malignancies based on the similar radiographic findings and initial gross appearances. Surrounding granulation tissue and secretions can obscure the margins making it difficult to correctly identify the nature of the obstructing lesion.

Foreign body aspiration is detailed extensively in children but less frequently in adults. The significance of the problem in adults is not minimal. In 2004, choking remained the fourth leading cause of unintentional injury or death in the United States and 4,100 deaths (1.4 deaths per 100,000 population) from unintentional ingestion or inhalation of food or other objects resulting in airway obstruction were reported[1]. The incidence rate in adults increases with age beginning in the sixth decade (2.6 deaths per 100,000 population aged 65-75 y) and rises rapidly after age 70 years (13.6 deaths per 100,000 population older than 75 y). Café coronary syndrome has been defined as the aspiration of a foreign body leading to immediate asphyxiation and death[2]. The incidence has been described at 0.66 per 100000. Since the angles made by the mainstem bronchi with the trachea are identical until 15 years of age foreign bodies are found on either side with equal frequency in children. After age 15 the right mainstem bronchus has a more acute angle off the trachea and foreign bodies are more commonly encountered here. Predisposing factors that have been described in previous reviews include primary neurologic disorders that result in impaired cough and swallowing, poor dentition, trauma, and sedative use[3]. Less common but still pertinent factors described are cultural practices such as not using eating utensils.

With further questioning, this patient remembered an incident of choking while eating popcorn lying on her couch approximately two months prior to presentation. 

B. Infrequently (less than 50% of the time) - In his review, Lan divided patients into clinical categories of acute or chronic aspirated foreign bodies[4]. In acute aspiration, patients can present immediately with a discrete choking episode and respiratory symptoms of dyspnea, coughing, localized wheezing, or absent breath sounds. Penetration syndrome is the presence of intractable cough and choking with or without vomiting that occurs at the time of the event[5]. This has been reported to occur upward of 49% in cases of foreign body aspiration. For patients identified as having chronic aspiration, a discrete episode of choking was not often recalled[4]. In Lan’s review, the duration of foreign body presence in the lung was approximately 25 months. Clinical characteristics associated with chronic foreign body aspiration were refractory asthma symptoms with localized wheezing, delayed resolution of pneumonia with appropriate treatment or recurrent pneumonias. Clinical sequelae in the chronic group included bronchiectasis, hemoptysis or bronchial stricture.

D. - The initial tests to obtain are AP and lateral neck films, inspiratory and expiratory PA chest films and a lateral chest film. Lateral neck views may show swelling or infraglottic opacity[5]. Many of the x-ray findings are dependent upon the length of time that the foreign body has been present and the properties of the foreign body itself. Radiopaque materials such as metal or teeth are easily identified. Organic materials including bones may not be easily seen. Bones from cod, haddock and salmon are radioapaque while trout, mackerel, and herring bones are radiolucent[6]. Some foreign bodies may have a physiologic ball and valve effect on the expiratory film, resulting in hyperinflation during expiration on the affected side[7]. In chronic foreign body aspiration atelectasis is more often seen. Recurrent pneumonias with volume loss are a common presentation of chronic foreign body aspiration[4]. The diagnostic accuracy, sensitivity, and specificity of foreign body detection on chest x-rays was 67%, 68%, and 67%, respectively in one series 34 pediatric cases[8]. Chest computed tomography (CT) in the evaluation of foreign body in adults has been described in the literature but there are no large series that have provided the sensitivity or specificity of CT. CT scans offer the additional ability to detect the foreign body in the lumen of the tracheobronchial tree[9]. Sensitivity may be increased with the use of higher resolution CT scans. Typical findings include hyperlucency, lobar consolidation, bronchiectasis and atelectasis.

B. With regards to removal of the foreign body both flexible bronchoscopy and rigid bronchoscopy can be utilized. In one case series 90% of tracheobronchial foreign bodies were able to be removed using flexible bronchoscopy and 97% using the combination of the two procedures[10]. Cryotherapy can be useful in the removal of soft material as was the case here. Other tools that can be utilized with the flexible bronchoscope include forceps and basket retrieval devices. The nature of the foreign body is important to the clinical presentation and its subsequent removal. The foreign body reaction within the airway tends to be much stronger with vegetative materials than non-vegetative[11]. After becoming lodged vegetable material may swell over hours or days, worsening the obstruction. Oily nuts such as peanuts have been shown to induce significant congestion and edema due to the free fatty acid released from decomposition of the nut itself. Inert objects such as bone or metal often do not cause significant airway obstruction but tend to be sharper and can lend to bleeding on removal.

After the foreign body removal bronchoalveolar lavage (BAL) from the LUL stained positive for a beaded gram positive bacilli that was modified Ziehl-Neelson acid fast stain negative. Poor growth was noted in aerobic culture but granules were seen on examination of the biopsy material.

C. Actinomyces is classically defined by the aforementioned staining characteristics and lack of growth on aerobic media[12]. Culturing material from foreign bodies requires special care. Laboratory personnel should be alerted of the possibility of aerobic and anaerobic organisms along with the possible need for the use of special culture media. Cultures should be allowed to grow for extended periods due to the relatively slow growth of these organisms. The cultures should also be kept after initial identification of more common airway organisms in an attempt to visualize fastidious organisms.

The three members of the genera actinomycetes are Actinomyces, Nocardia and Streptomycoses. The Actinomyces species are nonspore forming anaerobic prokaryotic bacteria. The presence of gram positive bacilli with beaded, branched morphology are suggestive of actinomycoses. It will typically stain acid fast negative. Sulphur granules have classically been associated with these organisms. Risk factors include alcoholism and poor dentition with weaker evidence suggesting immunosuppression as a risk factor[12]. Clinically this is often confused with a malignancy due to its slow growing characteristics. It is often associated with malignancy as it tends to colonize dead tissue which may be seen with necrotic endobronchial malignancies. Diagnosis should include a combination of positive culture, demonstration of sulphur granules in purulent matter from infected tissue, correlation with the clinical and radiological features, and the response to antibiotic treatment.  Actinomyces has been associated with foreign body aspiration in a previous case report[13]. Treatment consists of high dose intravenous penicillin for 2-6 weeks and then oral therapy for 6 to 12 months.

Differentiating amongst the members of the genera can be difficult as all have been associated with human disease. Nocardia species is identified as an aerobic, gram positive rod that stains partially acid fast[14]. This organism rarely forms granules and can be treated with trimethoprim/sulfamethoxazole. Streptomyces species typically cause mycetomas. They will often appear as an aerobic, gram positive rod. Acid fast staining is negative and these organisms often will produce granules. Treatment is also with trimethoprim /sulfamethoxazole.

Conclusion The patient was treated with IV penicillin and transitioned to oral therapy. At time of follow up she had radiographic resolution of her LUL collapse. Her cough had disappeared and she was back to her functional baseline.

• National Safety Council, R.a.S.D. Injury Facts 2008 Edition, ed. N.S. Council. Vol.:8. 2008: Itasca, Ill.
• Wick R.,Gilbert J.D., Byard R.W. Cafe coronary syndrome-fatal choking on food: an autopsy approach. J Clin Forensic Med 2006;13(3):135-8.
• Limper A.H., Prakash U.B. Tracheobronchial foreign bodies in adults. Ann Intern Med 1990;112(8):604-9.
• Lan, R.S. Non-asphyxiating tracheobronchial foreign bodies in adults. Eur Respir J 1994;7(3):510-4.
• Boyd M., Chatterjee A., Chiles C, Chin R. Tracheobronchial Foreign Body Aspiration in Adults. Southern Medical Journal 2009;102(2):171-174.
• Ell S, Sprigg A. The radio-opacity of fishbones — Species variation. Clinical Radiology 1991;44:104-107.
• Esclamado RM. Laryngotracheal foreign bodies in children: A comparison with bronchial foreign bodies. Am J Dis Child 1987;141(3):259-62.
• Svedström E, Kero P. How accurate is chest radiography in the diagnosis of tracheobronchial foreign bodies in children? Pediatr Radiol. 1989;19(8):520-2.
• Zissin R, Shapiro-Feinberg M, Rozenman J, Apter S, Smorjik S, Hertz M. CT findings of the chest in adults with aspirated foreign bodies. Eur Radiol. 2001:11(4):606
• Debeljak A, Sorli J, Music E, Kecelj P. Bronchoscopic removal of foreign bodies in adults: experience with 62 patients from 1974-1998. Eur Respir J.1999;4(4):792-5.
• Takako H. Supplemental experimental findings on foreign body in the bronchus. J Jpn Bronchoesophagol Soc. 1973;24:30-39.
• Mabeza G.F., Macfarlane J. Pulmonary actinomycosis. Eur Respir J. 2003; 21:545-551.
• Walters G, Ware N, Handslip P.Endobronchial actinomycosis associated with aspiration of a shirt button: A 30-year latency. Respiratory Medicine CME 2009 Vol. 2, Issue 1, Pages 18-20, DOI: 10.1016/j.rmedc.2008.10.018.
• Rippon JW. Medical Mycology. Inc. Wonsiewicz MJ, ed. The Pathogenic Fungi and the Pathogenic Actinomycetes. 3rd edn. Philadelphia. W.B. Saunders Co., 1988; pp. 30-52.

The American Thoracic Society improves global health by advancing research, patient care, and public health in pulmonary disease, critical illness, and sleep disorders. Founded in 1905 to combat TB, the ATS has grown to tackle asthma, COPD, lung cancer, sepsis, acute respiratory distress, and sleep apnea, among other diseases.

AMERICAN THORACIC SOCIETY 25 Broadway New York, NY 10004 United States of America Phone: +1 (212) 315-8600 Fax: +1 (212) 315-6498 Email: [email protected]

Privacy Statement | Term of Use | COI Conference Code of Conduct

EmergencyPedia

Free open access medical education, 6 clinical radiology practice cases.

We present 6 practice x-ray cases from Junior Medical Officer teaching at Westmead hospital.  When interpreting films pattern recognition is developed quickly over time but always stick with your solid system so as to  not miss things .  FOAMed images from LITFL.

Lecture Handout Notes – CLICK HERE

Thanks for Dr Andrew Baker for the opportunity to present.

• Where is the Nasogastric Tube?
• Where should it be?

Answer – Look for the “ tip ” 10cm below the Lower Oesophageal Sphincter :

Are you happy to feed the patient without repositioning?

No. Not in this case!

This patient developed an aspiration pneumonitis after enteral feeding into the lung…

• A normally healthy 33 year old man is brought to your Emergency Department with acute shortness of breath after developing chest pain during a run at the park.
• On arrival in hospital, he is pale and sweaty. He was found to be tachycardic and hypotensive.  A CXR was taken immediately after a procedure was performed to stabilise his condition.

Can you describe his Chest X-ray?

• Chest X-ray showing a pneumothorax
• A thoracotomy needle in the left chest
• No evidence tension pneumothorax (deviated heart, mediastinum, trachea)

Can you write down a systematic approach to interpreting a Chest X-ray?

• D emographics (Name, Time Taken)
• –(Quality = Rotation/Adequacy/Penetration)
• A irway (Trachea)
• Mediastinum (>7cm)
• Hilum (Left Should be Higher than Right)
• Lungs (Lung Fields, Fissures)
• C irculation
• E verything Else
• Step Back – Consider ‘Overall Appearance’ (‘gestalt’ comes in here)

What are your treatment options?

Resuscitation Phase

(1) Is there is evidence of tension pneumothorax?

• IV, O2, Monitor
• Decompress the pneunothorax with needle +/- aspiration technique
• Follow the A-G assessment
• Call for help

General  Manangement

(2) Are there minimal symptoms?

• If small (3cm or less at hilum) –  observe (oxygen may increase absorption)
• follow BTS guideline
• consider a chest drain if >3cm

(3) The patient is not improving after chest drain for several days – what should I do?

• Get a CT scan to reveal a possible underlying cause for the pneumothorax
• Involve a Cardiothoracic surgeon in the care (may need surgical intervention)

Approaches to Pneumothorax for reference:

• A 60 year old car driver presents following a head on collision with a bus at 60Kph.
• BP 130/90mmHg
• Respiratory Rate 24/min
• Oxygen Saturations 98% (on 6L oxygen)

Describe and interpret her X-ray

• Widened Mediastinum
• Clavicle Fracture
• Rib Fractures
• ?Right Haemopneumothorax

Outline your management options

• Care of the Trauma patient is delivered by a Team with good communication amount members.  Ultrasound would help you differentiate the chest injuries further in skilled hands
• Start treating the patient with the mantra IV / O2 / Monitor with resuscitation trolley at the bedside
• If shocked (as the picture above would indicate) activate your local transfusion protocols in liaison with haematology
• If Aortic Injury  is a concern caution with over resuscitation
• Chest drain (intercostal catheter insertion) of the correct side – re-expansion of the lung may help tamponade bleeding

If Aortic Injury:

• Thoracotomy for haemothorax – if on going bleeding or massive initial blood loss
• Disposition – admission to ICU
• A 65 year old woman presents to your hospital with gradually increasing breathlessness over the preceding 2 days.
• It is about 10 days since her last chemotherapy (cisplatin).
• Respiratory rate 30/min
• Oxygen Saturations 91% (on 6L Oxygen)
• Temperature 36.8 degrees

What does her Chest X-ray show?

• Large left pleural effusion
• Multiple discrete lung parenchymal lesions typical of metastatic lung disease
• ? Mastectomy
• No Obvious boney mets

Write down 6 differentials in the acutely short of breath patient with malignancy

• Sepsis with metabolic acidosis (in view of recent chemotherapy)
• Pericardial Effuson
• SVC Obstruction
• Pleural Effusion(s)
• Lymphangitis

What are the causes of a Pleural Effusion?

• Transudates – Congestive Heart Failure, Liver Failure, Renal Failure, Nephrotic syndrome, Hypoalbuminaemia, Enteropathy and Dialysis
• Exudates – Lung Ca, TB, Infections (Bacterial), RA, Pancreatitis, Sub-phrenic Abscess, ‘Meig’s’ Syndrome, Dressler’s Syndrome, SLE, Lymphoma, Hypothyroid, PE, Mesothelioma, Yellow Nail Syndrome and Vasculitis

What is Light’s Criteria?

• The ratio of pleural fluid protein to serum protein is greater than 0.5
• The ratio of pleural fluid LDH and serum LDH is greater than 0.6
• Pleural fluid LDH is greater than 0.6 times the normal upper limit for serum. (i.e 0.6 of 200)

A 50 year old regular Emergency Department presenter with a history ETOH excess presents with a few weeks of feeling tired and coughing up small amounts of blood. He has a past history of pancreatitis and appendicectomy

What does the X-ray show?

Answer – X-ray showing a ‘cavitating’ lesion in right chest with soft tissue densities in lower zone.

List your differential diagnosis

• Acute Infective causes – such as TB, Fungi, Aspiration
• Malignancy (Primary and Secondary)
• Other bacterial infections Staphylococcus / Klebsiella
• Wegner’s ( recent name change of this disease due to a Nazi doctor who described it )
• Severe Fibrosis
• PICC lines are occasionally required for venous access IV administration of medication nutrition.
• You are called to the ward for a patient who has just had a PICC line placed.

CXR ( http://www.rtexam.com ) is shown:

Are you happy with the position of the PICC line?

• Chest X-ray showing tip of PICC line in the Superior Vena Cava (SVC) – so in this case we are happy with the position
• Malposition of the tip is common (superior (IJ), across (subclavian), or twisted around). In one study about 8% of PICC lines were badly positioned on initial assessment post insertion
• Feeding / medications should not be commenced unless a good position is identified on the CXR

(1) Image 1 – Click Here

(2) Image 2 – Click Here

(3) Image 3 – Click Here

(4) Image 4 – Click Here

(5) Image 5 – Click Here

(6) Image 6 – Click Here

Share this:

2 thoughts on “ 6 clinical radiology practice cases ”.

Great post! I recommend attributing/citing sources from which figures are obtained. This serves several purposes and allows readers to contact primary authors should a change or typo exist. Thanks for the #FOAMed.

Thanks . If you click on each item it should take you there but we ll add a fuller reference section at end. This was really an intern flipped classroom more than anything else…

Comments are closed.

• Already have a WordPress.com account? Log in now.
• Subscribe Subscribed
• Copy shortlink
• Report this content
• View post in Reader
• Manage subscriptions
• Collapse this bar

Because of a lapse in government funding, the information on this website may not be up to date, transactions submitted via the website may not be processed, and the agency may not be able to respond to inquiries until appropriations are enacted. The NIH Clinical Center (the research hospital of NIH) is open. For more details about its operating status, please visit cc.nih.gov . Updates regarding government operating status and resumption of normal operations can be found at USA.gov .

Are you still there ?

Normal chest x-ray

• Report problem with case

Citation, DOI, disclosures and case data

At the time the case was submitted for publication Frank Gaillard had no recorded disclosures.

Essentially normal chest x-ray in a 50-year-old male. There is a degree of hyperinflation as evidenced by both increased retrosternal airspace and somewhat flattened and depressed diaphragms.

2 articles feature images from this case

• Enlarged pulmonary trunk on chest radiography (differential)
• More black sign

31 playlists include this case

Public playlists

• RX + TCs TÓRAX by Víctor González
• Normal CXRs for getting used to them by Piotr Guzdziol
• Neurology lecture by brianjwright1 ◉
• Egzaminas by Tomas Jurevicius
• Chest - breathlessness 1 by Radiology Registrar PULSE Teaching
• UQ anatomy presentation - Normal Anatomy by Lachlan Inch
• Chest xray by Mark Bolger ◈
• UWSMPH FM2 Block 1 Exam Review by Stuart Feichtinger
• chest new by DOC RAD
• X-ray (non-contrast) by Vasilii
• Pre internship talk by Kenny Sim
• Seen Cases by Carlos Peter
• normal findings by Carlos Peter
• Normal XRs by Rochelle Cabry
• DoctorsLive Session 5 by Osama Al-Jibury
• Trauma - Normal Xrays by Andy
• Radiology test (2d year) by Vasilii
• Normal lung visualisation by Vasilii
• 96. Learning Radiology by Paul-Gabriel Baragan ◉
• Radiografía anatómica by Johan Stev Erazo
• Chest Med Students by Salam Findakly

Unlisted playlists

This case is used in 10 unlisted playlists .

Related Radiopaedia articles

• Normal chest imaging examples

Promoted articles (advertising)

How to use cases.

You can use Radiopaedia cases in a variety of ways to help you learn and teach.

• Add cases to playlists
• Share cases with the diagnosis hidden
• Use images in presentations
• Use them in multiple choice question

Creating your own cases is easy.

• Case creation learning pathway

ADVERTISEMENT: Supporters see fewer/no ads

By Section:

• Artificial Intelligence
• Classifications
• Imaging Technology
• Interventional Radiology
• Radiography
• Central Nervous System
• Gastrointestinal
• Gynaecology
• Haematology
• Head & Neck
• Hepatobiliary
• Interventional
• Musculoskeletal
• Paediatrics
• Not Applicable

Radiopaedia.org

• Editorial Board

Stay up to date with the latest in Practical Medical Imaging and Management with Applied Radiology.

Stay up to date

Create a new print or digital subscription to Applied Radiology.

Renew / Update

Renew or update your current subscription to Applied Radiology.

• Email Alert Manager
• Author Guidelines
• CME Credits
• CRA Credits

Case Studies on X-ray Imaging, MRI and Nuclear Imaging

• First Online: 17 October 2023

Cite this chapter

• Shuvra Sarker 5 ,
• Angona Biswas 5 ,
• Nasim Md Abdullah Al 5 ,
• Md Shahin Ali 6 ,
• Sai Puppala 7 &
• Sajedul Talukder 7  

135 Accesses

The field of medical imaging is an essential aspect of the medical sciences, involving various forms of radiation to capture images of the internal tissues and organs of the body. These images provide vital information for clinical diagnosis, and in this chapter, we will explore the use of X-ray, MRI, and nuclear imaging in detecting severe illnesses. However, manual evaluation and storage of these images can be a challenging and time-consuming process. To address this issue, artificial intelligence (AI)-based techniques, particularly deep learning (DL), have become increasingly popular for systematic feature extraction and classification from imaging modalities, thereby aiding doctors in making rapid and accurate diagnoses. In this review study, we will focus on how AI-based approaches, particularly the use of Convolutional Neural Networks (CNN), can assist in disease detection through medical imaging technology. CNN is a commonly used approach for image analysis due to its ability to extract features from raw input images, and as such, will be the primary area of discussion in this study. Therefore, we have considered CNN as our discussion area in this study to diagnose ailments using medical imaging technology.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime
• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Available as EPUB and PDF
• Durable hardcover edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

General Data Protection Regulation (GDPR): https://en.wikipedia.org/wiki/Medical-imaging (2021). Accessed 12 Feb 2021

Narin, A., Kaya, C., Pamuk, Z.: Automatic detection of coronavirus disease (covid-19) using x-ray images and deep convolutional neural networks. Pattern Anal. Appl. 24 , 1207–1220 (2021)

Article   Google Scholar  

General Data Protection Regulation (GDPR): https://en.wikipedia.org/wiki/X-ray (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://www.fda.gov/radiation-emitting-products/medical-imaging/medical-x-ray-imaging (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://my.clevelandclinic.org/health/diagnostics/21818-x-ray (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://www.iec.ch/blog/invention-magnetic-resonance-imaging-mri (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://www.mayoclinic.org/tests-procedures/mri/about/pac-20384768 (2021). Accessed 12 Feb 2021

Al Nasim, M.A., Al Munem, A., Islam, M., Palash, M.A.H., Haque, M.M.A., Shah, F.M.: Brain tumor segmentation using enhanced u-net model with empirical analysis. In: 2022 25th International Conference on Computer and Information Technology (ICCIT), pp. 1027–1032. IEEE (2022)

Google Scholar  

General Data Protection Regulation (GDPR): https://en.wikipedia.org/

Apostolopoulos, I.D., Mpesiana, T.A.: Covid-19: automatic detection from x-ray images utilizing transfer learning with convolutional neural networks. Phys. Eng. Sci. Med. 43 , 635–640 (2020)

Ozturk, T., Talo, M., Yildirim, E.A., Baloglu, U.B., Yildirim, O., Acharya, U.R.: Automated detection of covid-19 cases using deep neural networks with x-ray images. Comput. Biol. Med. 121 , 103792 (2020)

Hemdan, E.E.-D., Shouman, M.A., Karar, M.E.: Covidx-net: a framework of deep learning classifiers to diagnose covid-19 in x-ray images. arXiv preprint arXiv:2003.11055 (2020)

Wang, L., Lin, Z.Q., Wong, A.: Covid-net: a tailored deep convolutional neural network design for detection of covid-19 cases from chest x-ray images. Sci. Rep. 10 (1), 1–12 (2020)

Nasim, M., Dhali, A., Afrin, F., Zaman, N.T., Karim, N.: The prominence of artificial intelligence in covid-19. arXiv preprint arXiv:2111.09537 (2021)

Kanade, P.B., Gumaste, P.: Brain tumor detection using MRI images. Brain 3 (2), 146–150 (2015)

Saladi, S., Karuna, Y., Koppu, S., Reddy, G.R., Mohan, S., Mallik, S., Qin, H.: Segmentation and analysis emphasizing neonatal MRI brain images using machine learning techniques. Mathematics 11 (2), 285 (2023)

Biswas, A., Islam, M.S.: Brain tumor types classification using k-means clustering and ANN approach. In: 2021 2nd International Conference on Robotics, Electrical and Signal Processing Techniques (ICREST), pp. 654–658. IEEE (2021)

Bhanothu, Y., Kamalakannan, A., Rajamanickam, G.: Detection and classification of brain tumor in mri images using deep convolutional network. In: 2020 6th International Conference on Advanced Computing and Communication Systems (ICACCS), pp. 248–252. IEEE (2020)

Toğaçar, M., Cömert, Z., Ergen, B.: Classification of brain mri using hyper column technique with convolutional neural network and feature selection method. Expert Syst. Appl. 149 , 113274 (2020)

Amin, J., Sharif, M., Gul, N., Yasmin, M., Shad, S.A.: Brain tumor classification based on dwt fusion of mri sequences using convolutional neural network. Pattern Recogn. Lett. 129 , 115–122 (2020)

Munley, M.T., Marks, L.B., Scarfone, C., Sibley, G.S., Patz Jr., E.F., Turkington, T.G., Jaszczak, R.J., Gilland, D.R., Anscher, M.S., Coleman, R.E.: Multimodality nuclear medicine imaging in three-dimensional radiation treatment planning for lung cancer: challenges and prospects. Lung Cancer 23 (2), 105–114 (1999)

Moriguchi, S.M., Koga, K.H., Togni, P.H.A., dos Santos, M.J.: Clinical applications of nuclear medicine. In: Medical Imaging in Clinical Practice, London, pp. 37–62 (2013)

Prieto, E., Martí-Climent, J.M., Domínguez-Prado, I., Garrastachu, P., Díez-Valle, R., Tejada, S., Aristu, J.J., Peñuelas, I., Arbizu, J.: Voxelbased analysis of dual-time-point 18f-fdg pet images for brain tumor identification and delineation. J. Nucl. Med. 52 (6), 865–872 (2011)

Shinoura, N., Nishijima, M., Hara, T., Haisa, T., Yamamoto, H., Fujii, K., Mitsui, I., Kosaka, N., Kondo, T., Hara, T.: Brain tumors: detection with c-11 choline pet. Radiology 202 (2), 497–503 (1997)

Wong, T.Z., van der Westhuizen, G.J., Coleman, R.E.: Positron emission tomography imaging of brain tumors. Neuroimaging Clin. 12 (4), 615–626 (2002)

Wong, K.-P., Feng, D., Meikle, S.R., Fulham, M.J.: Segmentation of dynamic pet images using cluster analysis. IEEE Trans. Nucl. Sci. 49 (1), 200–207 (2002)

Park, Y.-J., Choi, D., Choi, J.Y., Hyun, S.H.: Performance evaluation of a deep learning system for differential diagnosis of lung cancer with conventional CT and FDG PET/CT using transfer learning and metadata. Clin. Nucl. Med. 46 (8), 635–640 (2021)

Papandrianos, N., Papageorgiou, E.: Automatic diagnosis of coronary artery disease in SPECT myocardial perfusion imaging employing deep learning. Appl. Sci. 11 (14), 6362 (2021)

Papandrianos, N.I., Feleki, A., Papageorgiou, E.I., Martini, C.: Deep learning-based automated diagnosis for coronary artery disease using SPECT-MPI images. J. Clin. Med. 11 (13), 3918 (2022)

Puppala, S., Hossain, I., Talukder, S.: Towards federated learning based contraband detection within airport baggage x-rays. In: 2022 IEEE International Conference on Machine Learning and Applied Network Technologies (ICMLANT), pp. 1–6. IEEE (2022)

Talukder, S., Puppala, S., Hossain, I.: Federated learning-based contraband detection within airport baggage x-rays. J. Comput. Sci. Coll. 38 (3), 218–218 (2022)

Talukder, S., Puppala, S., Hossain, I.: A novel hierarchical federated learning with self-regulated decentralized clustering. J. Comput. Sci. Coll. 38 (3), 222–223 (2022)

Hossain, I., Puppala, S., Talukder, S.: Collaborative differentially private federated learning framework for the prediction of diabetic retinopathy. In: 2023 IEEE 2nd International Conference on AI in Cybersecurity (ICAIC), pp. 1–6, IEEE (2023)

General Data Protection Regulation (GDPR): https://github.com/ieee8023/covid-chestxray-dataset (2021). Accessed 12 Feb 2021

Wang, X., Peng, Y., Lu, L., Lu, Z., Bagheri, M., Summers, R.M.: Chestx-ray8: hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2097–2106 (2017)

Toğaçar, M., Ergen, B., Cömert, Z.: Covid-19 detection using deep learning models to exploit social mimic optimization and structured chest x-ray images using fuzzy color and stacking approaches. Comput. Biol. Med. 121 , 103805 (2020)

Badža, M.M., Barjaktarović, M.Č.: Classification of brain tumors from mri images using a convolutional neural network. Appl. Sci. 10 (6), 1999 (2020)

Sahoo, L., Sarangi, L., Dash, B.R., Palo, H.K.: Detection and classification of brain tumor using magnetic resonance images. In: Advances in Electrical Control and Signal Systems: Select Proceedings of AECSS 2019, pp. 429–441. Springer (2020)

Kalaiselvi, T., Padmapriya, S., Sriramakrishnan, P., Somasundaram, K.: Deriving tumor detection models using convolutional neural networks from mri of human brain scans. Int. J. Inf. Technol. 12 , 403–408 (2020)

Menze, B.H., Jakab, A., Bauer, S., Kalpathy-Cramer, J., Farahani, K., Kirby, J., Burren, Y., Porz, N., Slotboom, J., Wiest, R., et al.: The multimodal brain tumor image segmentation benchmark (brats). IEEE Trans. Med. Imaging 34 (10), 1993–2024 (2014)

Multimodal Brain Tumor Segmentation Challenge 2020: Data, https://www.med.upenn.edu/cbica/brats2020/data.html

General Data Protection Regulation (GDPR): https://www.oasis-brains.org/ (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://www.nitrc.org/projects/ibsr (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://www.kaggle.com/datasets (2021). Accessed 12 Feb 2021

The Whole Brain Atlas (WBA), https://www.med.harvard.edu/aanlib/

Shah, F.M., Hossain, T., Ashraf, M., Shishir, F.S., Al Nasim, M.A., Kabir, M.H.: Brain tumor segmentation techniques on medical images-a review. Int. J. Sci. & Eng. Res 10 (2), 1514–1525 (2019)

Anilkumar, B., Kumar, P.R.: Tumor classification using block wise fine tuning and transfer learning of deep neural network and KNN classifier on MR brain images. Int. J. Emerg. Trends Eng. Res 8 (2), 574–583 (2020)

Article   MathSciNet   Google Scholar  

Gopal, N.N., Karnan, M.: Diagnose brain tumor through mri using image processing clustering algorithms such as fuzzy c means along with intelligent optimization techniques. In: 2010 IEEE International Conference on Computational Intelligence and Computing Research, pp. 1–4. IEEE (2010)

General Data Protection Regulation (GDPR): https://www.hopkinsmedicine.org/health/treatment-tests-and-therapies/nuclear-medicine (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://stanfordhealthcare.org/medical-tests/n/nuclear-imaging.html (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://radiopaedia.org/articles/spect-vs-pet (2021). Accessed 12 Feb 2021

General Data Protection Regulation (GDPR): https://radiopaedia.org/articles/single-photon-emission-computed-tomography-spect (2021). Accessed 12 Feb 2021

Nayak, S.R., Nayak, D.R., Sinha, U., Arora, V., Pachori, R.B.: Application of deep learning techniques for detection of covid-19 cases using chest x-ray images: a comprehensive study. Biomed. Signal Process. Control 64 , 102365 (2021)

Ucar, F., Korkmaz, D.: Covidiagnosis-net: deep Bayes-SqueezeNet based diagnosis of the coronavirus disease 2019 (covid-19) from x-ray images. Med. Hypotheses 140 , 109761 (2020)

General Data Protection Regulation (GDPR): https://www.kaggle.com/datasets/tawsifurrahman/covid19-radiography-database (2021). Accessed 12 Feb 2021

Sethy, P.K., Behera, S.K.: Detection of coronavirus disease (covid-19) based on deep features. MDPI AG (2020)

Farooq, M., Hafeez, A.: COVID-ResNet: a deep learning framework for screening of COVID19 from radiographs. arXiv preprint arXiv:2003.14395 (2020)

Tonmoy, H., Shadmani, S.F., Mohsena, A., Abdullah, M.A.N., Faisal, M.S.: Brain tumor detection using convolutional neural network. In: 2019 1st International Conference on Advances in Science, Engineering and Robotics Technology (ICASERT), pp. 1–6. IEEE (2019)

Ghassemi, N., Shoeibi, A., Rouhani, M.: Deep neural network with generative adversarial networks pre-training for brain tumor classification based on mr images. Biomed. Signal Process. Control 57 , 101678 (2020)

Marcus, D.S., Fotenos, A.F., Csernansky, J.G., Morris, J.C., Buckner, R.L.: Open access series of imaging studies: longitudinal mri data in nondemented and demented older adults. J. Cogn. Neurosci. 22 (12), 2677–2684 (2010)

Begum, S.S., Lakshmi, D.R.: Combining optimal wavelet statistical texture and recurrent neural network for tumour detection and classification over mri. Multimed. Tools Appl. 79 , 14009–14030 (2020)

Arsanjani, R., Dey, D., Khachatryan, T., Shalev, A., Hayes, S.W., Fish, M., Nakanishi, R., Germano, G., Berman, D.S., Slomka, P.: Prediction of revascularization after myocardial perfusion SPECT by machine learning in a large population. J. Nucl. Cardiol. 22 , 877–884 (2015)

Berkaya, S.K., Sivrikoz, I.A., Gunal, S.: Classification models for SPECT myocardial perfusion imaging. Comput. Biol. Med. 123 , 103893 (2020)

Download references

Author information

Authors and affiliations.

Research and Development Department, Pioneer Alpha, Dhaka, Bangladesh

Shuvra Sarker, Angona Biswas & Nasim Md Abdullah Al

Department of Biomedical Engineering, Islamic University, Kushtia, Bangladesh

Md Shahin Ali

School of Computing, Southern Illinois University, Carbondale, IL, USA

Sai Puppala & Sajedul Talukder

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Nasim Md Abdullah Al .

Editor information

Editors and affiliations.

Norman, OK, USA

Department of Computer Systems Engineering and Technology, Oregon Institute of Technology, Klamath Falls, OR, USA

Stefan Andrei

Department of Computer Science, Bowie State University, Bowie, MD, USA

Md Kamruzzaman Sarker

Department of Cyber Physical System, Clark Atlanta University, Atlanta, GA, USA

Kishor Datta Gupta

Rights and permissions

Reprints and permissions

Copyright information

© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Sarker, S., Biswas, A., Al, N.M.A., Ali, M.S., Puppala, S., Talukder, S. (2023). Case Studies on X-ray Imaging, MRI and Nuclear Imaging. In: Zheng, B., Andrei, S., Sarker, M.K., Gupta, K.D. (eds) Data Driven Approaches on Medical Imaging. Springer, Cham. https://doi.org/10.1007/978-3-031-47772-0_10

Download citation

DOI : https://doi.org/10.1007/978-3-031-47772-0_10

Published : 17 October 2023

Publisher Name : Springer, Cham

Print ISBN : 978-3-031-47771-3

Online ISBN : 978-3-031-47772-0

eBook Packages : Medicine Medicine (R0)

Share this chapter

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research

• Remote Access
• Save figures into PowerPoint
• Download tables as PDFs

Section I. Chest Radiology

• Download Chapter PDF

Disclaimer: These citations have been automatically generated based on the information we have and it may not be 100% accurate. Please consult the latest official manual style if you have any questions regarding the format accuracy.

Download citation file:

• Search Book

Jump to a Section

• Chest Radiology: Introduction
• When to Order Radiographs
• Radiographic Views
• How to Read a Radiograph
• Systematic Approach to Interpreting a Chest Radiograph
• Normal Radiographic Anatomy
• The Lateral Chest Radiograph
• The AP Portable Chest Radiograph
• Radiographic Patterns of Pulmonary Disease—Increased Pulmonary Opacity
• Radiographic Findings and Clinical Diagnosis
• Hard and Soft Radiographic Signs
• Suggested Reading
• Full Chapter
• Supplementary Content

Chest radiography is the most commonly ordered imaging test in emergency department patients. It can provide considerable diagnostic information for a wide variety of thoracic disorders. Its diagnostic capabilities are based largely on the contrast between the air-filled lungs and pathological processes that cause fluid accumulation within lung tissue.

Although chest CT provides greater anatomical detail of the pulmonary parenchyma and is often used in non-emergency patients with pulmonary disease, the use of chest CT in the ED is limited to certain critical conditions that do not produce distinctive findings on conventional radiography. These include pulmonary embolism and aortic dissection. CT is also used in ED patients with major chest trauma to detect an aortic injury, pneumothorax, hemothorax and pulmonary contusions that may not be evident on the supine portable chest radiographs.

In general, a diagnostic test should be ordered when the disease under consideration produces characteristic findings which help confirm or exclude the suspected disorder. A number of approaches can be used in deciding to order a radiograph.

With a simplistic “geographic” approach, radiographs are obtained of the region where the patient is having symptoms, e.g., a chest radiograph in a patient with chest pain. Such an approach is ill-advised because it can lead to diagnostic errors, as well as excessive and unnecessary testing.

Using a symptom-based approach to radiograph ordering, the decision to obtain radiographs is based on characteristics of the patient’s symptoms, for example whether the chest pain is severe or mild, pleuritic or pressure-like ( Rothrock 2002 ).

However, a more rational diagnosis-based approach is to first consider the potential disorders that might be present and then to obtain radiography if the suspected disorder has characteristic radiographic findings, such as pneumonia and pneumothorax. This approach is the most likely to yield clinically useful information and to avoid unnecessary testing. Determining which disorders need investigation in an individual patient is ultimately based on the clinical judgment, knowledge and experience of the practitioner.

Two perpendicular views should be obtained whenever possible. The preferred frontal view is a postero-anterior view (PA view) . This view is obtained in the radiology suite with the patient standing and the imaging cassette placed against the patient’s anterior chest wall. The x-ray beam is directed horizontally and traverses the patient from posterior to anterior. The patient’s hands are positioned on the hips, which moves the scapulae laterally and away from the lungs. The patient is instructed to take a full inspiration. The PA view is preferred because the heart and mediastinum are closest to the x-ray imaging cassette and therefore less distorted by magnification.

Get Free Access Through Your Institution

Best of the blogs, pop-up div successfully displayed.

This div only appears when the trigger link is hovered over. Otherwise it is hidden from view.

Please Wait

X-rays and Mom — Case Study into the State of Imaging Technology

Reading Time: 4 minutes read

By Dave Fornell, Editor, Diagnostic & Interventional Cardiology magazine and assistant editor for Imaging Technology News magazine.

While I write a lot about medical imaging technology and how new technology can and should work, it is not often that I get to experience how things actually work in the real world. This past Thanksgiving I received a call from a paramedic explaining that my mom had fractured her leg and I should stop working on the turkey and fixings and rush to the emergency department (ED) at Edward Hospital in Naperville, Ill. She had been walking her dog on wet grass and leaves in a park when her dog took off after another dog and pulled her down. She was whipped around and the change in weight caused her to dislocate her ankle (the bottom of her shoe was facing her when she looked at her feet) and caused a spiral fracture to her fibula.

A bedside screen shot of a Carestream DRX mobile X-ray in the ED of the fractured fibula.

When I got there my mom was already heavily sedated due to the pain and because the ED staff had already put her ankle back in place. The ED doctor ordered a digital radiograph (DR) of her leg to see the extent of damage. They wheeled in a new Carestream DRX mobile X-ray system and I had a live demonstration of how fast these types of systems can snap the pictures. It called up the images immediately on the machine’s screen. The image of the Pott’s fracture with fragments was really interesting as someone who covers radiology, but I also realized from a non-clinical standpoint she was really messed up and in pain. Additionally, she would need reconstructive surgery to put her Humpty Dumpty leg back together again. She was way up the creek without a paddle with it being Thanksgiving and there were no orthopedic surgeons in staff due to the holiday. The day after Thanksgiving was not much better, as we found, since most physicians were out through the following Monday. So the ER splinted the leg, wrapped it in ace bandages and sent her home with heavy pain killers.

Compounding her mobility issues was the fact that she has bilateral knee replacements. Due to the trauma, broken bone and knowing she had these implants that further limited her ability to move around, she was prescribed a prophylactic anticoagulant.

Knowing we would need the images for a surgeon to review, I had the ED burn a CD. However, I was happy to find

The post-surgical X-ray showing the bone repair, which was accessed and copied by the patient using a patient portal.

Edward is among the growing number of hospitals to grant patients access to their health records via a DR Systems Internet image/results distribution system. This technology pulls images and reports from the hospitals’ Epic EMR (electronic medical record) system and makes them available for remote viewing by clinicians outside of the hospital’s picture archiving and communication system (PACS). She also was given login instructions at discharge for a patient portal so she could access her records and images herself on a home computer or smartphone.

We managed to find one orthopedic surgeon in their office on the Friday after Thanksgiving. They thought it was great that we had a CD, but before attempting to open it, they asked which hospital she had been at. Edward was already in a health information exchange, so outlying offices such as this one from a different medical group could access her records remotely in less than a minute. They were able to call up her images and see what meds she was prescribed, which made the office visit go much faster.

She had surgery on Dec. 1, the Tuesday of RSNA 2015. The orthopedic surgeon practiced at Elmhurst Hospital in Elmhurst, Ill., across the county from Naperville. But, thanks to the remote image viewing system, they could get the ED images for reference and planning. The surgeon’s post-surgery DR image showing the reconstruction of the fibula also was available via my mom’s patient portal.

She did what most patients today do with this type of access and posted her X-rays on Facebook. Leveraging the Facebook form of patient engagement, the result was lots of sympathy, flowers and friends volunteering to help her with things around the house and groceries since she cannot walk or drive for at least two months.

While an unfortunate incident and a horrible thing to have happen to my mom, from a professional standpoint, I was happy to see the technology I cover working in the real world as it was intended. The speed in workflow efficiency, speed and ease of access to her imaging at the point of care and remotely, and access to a patient portal are all examples of how the healthcare system should work. In this case, the technology and imaging integration was flawless.

Dave Fornell is the editor of Diagnostic & Interventional Cardiology magazine and assistant editor for Imaging Technology News magazine.

Reposted from Imaging Technology News (ITN)  with permission.

To learn more about the CARESTREAM DRX-Revolution, click here .

RELATED POSTS

Mais do que um fornecedor: visita aos hospitais Shriners Children’s

More than a provider: on-site with shriner’s children’s hospitals.

Llevar a cabo la transformación de la radiografía análoga a la digital

Post a comment cancel reply.

Sign me up for Everything Rad updates as they happen!

This site uses Akismet to reduce spam. Learn how your comment data is processed .

• OpenMed Home
• Cardiology guidelines
• Cardiology resources
• Endocrinology
• Dermatology guidelines
• Dermatology resources
• Neurology guidelines
• Neurology resources
• Respiratory
• Psychiatry Resources
• Psychiatry Guidelines
• Medical education
• Translational Medicine
• Most Recent
• Suggest Resource

Chest X-ray: A case study based tutorial

28 short case studies with questions.

https://www.bris.ac.uk/medical-school/media/secure/ms_res_cxr/

University of Bristol, UK

OpenMed is not responsible for the content of external Internet sites.

Related posts:

• Chest X-ray tutorials (Radiology Masterclass)
• Radiological cases
• Introduction to CXR interpretation
• Chest X-ray atlas and self assessment
• Chest images

How to…

  RSS

  login,   want to join in,   our partners.

• Open access
• Published: 10 September 2024

Interdisciplinary CBT treatment for patients with odontophobia and dental anxiety related to psychological trauma experiences: a case series

• Yngvill Ane Stokke Westad 1 ,
• Gina Løge Flemmen 1 ,
• Stian Solem 1 ,
• Trine Monsen 1 ,
• Henriette Hollingen 1 ,
• Astrid Feuerherm 3 ,
• Audun Havnen 2 , 4 &
• Kristen Hagen 5 , 6 , 7  

BMC Psychiatry volume  24 , Article number:  606 ( 2024 ) Cite this article

Metrics details

While cognitive-behavioural therapy (CBT) is a well-established treatment for odontophobia, research is sparse regarding its effect on patients with dental anxiety related to psychological trauma experiences. This study aimed to evaluate changes in symptoms and acceptability of interdisciplinary Torture, Abuse, and Dental Anxiety (TADA) team treatment for patients with odontophobia or dental anxiety. We also wanted to describe the sample’s oral health status. The TADA teams offer targeted anxiety treatment and adapted dental treatment using a CBT approach.

The study used a naturalistic, case series design and included 20 consecutively referred outpatients at a public TADA dental clinic. Pre- and post-treatment assessments included questionnaires related to the degree of dental anxiety, post-traumatic stress, generalized anxiety, and depression. Patients underwent a panoramic X-ray before treatment. Before dental restoration, patients underwent an oral health examination to determine the mucosal and plaque score (MPS) and the total number of decayed, missing, and filled teeth (DMFT). Patients were referred to dentist teams for further dental treatment and rehabilitation (phase 2) after completing CBT in the TADA team (Phase 1). Results from the dental treatment in phase 2 is not included in this study.

All patients completed the CBT treatment. There were significant improvements in symptoms of dental anxiety, post-traumatic stress, and depression and moderate changes in symptoms of generalized anxiety. Dental statuses were heterogeneous in terms of the severity and accumulated dental treatment needs. The TADA population represented the lower socioeconomic range; 15% of patients had higher education levels, and half received social security benefits. All patients were referred to and started adapted dental treatment (phase 2).

Conclusions

TADA treatment approach appears acceptable and potentially beneficial for patients with odontophobia and dental anxiety related to psychological trauma experiences. The findings suggest that further research, including larger controlled studies, is warranted to validate these preliminary outcomes.

Trial registration

The study was approved by the regional ethical committee in Norway (REK-Midt: 488462) and by the Data Protection Board at Møre and Romsdal County Authority.

Peer Review reports

Patients with mental disorders have a greater risk of oral and dental diseases than the general population. Psychiatric diagnoses are associated with poor dental status, such as carious, missing or filled teeth or surfaces [ 1 ], and patients with severe mental illness are almost three times more likely to lose all of their teeth compared to the general population [ 2 ]. This may be caused by several individual or cumulative factors, such as the inability to perform self-care, diet and lifestyle factors, difficulties in accessing health care services, poor economic status, a negative attitude towards health care providers, shame and anxiety, difficulties cooperating with treatment, and drug use and drug treatment side effects [ 1 , 3 , 4 , 5 , 6 ].

Patients referred for dental anxiety treatment have moderately high levels of comorbid psychological conditions [ 7 ], and this patient group differs with respect to the age of onset, origins, and manifestations [ 8 ]. Individuals with high dental anxiety report more mental health symptoms, poorer oral health, more avoidance behaviour, and more irregular dental visits than those with no or low anxiety [ 9 , 10 , 11 , 12 ]. Furthermore, large variations in oral health and dental treatment needs have been found in patients with dental anxiety and phobia [ 13 , 14 ].

Patients with anxiety disorders, especially post-traumatic stress disorder (PTSD), could be especially prone to developing fears of dental treatment [ 15 ]. The study found that 42.0% of patients with PTSD reported high dental anxiety, compared to 17.6–31.3% in other psychiatric groups, and 4.2% in healthy controls [ 15 ]. Approximately 20% of female patients seeking dental care may have encountered childhood sexual abuse [ 16 ]. Patients who have experienced traumatic events may exhibit distinct psychological and emotional responses that can complicate the treatment process [ 16 , 17 , 18 ]. Furthermore, elements of abuse can resemble the dental treatment environment and make it difficult to tolerate dental treatment [ 17 , 19 , 20 ]. This suggests that it is important for treatment and professionals to be considerate of the patient’s trauma history [ 21 , 22 ].

In 2010, the Norwegian Department of Health concluded that patients who were exposed to torture, sexual abuse, and/or violence in close relationships and/or had odontophobia had inadequate treatment options in the Norwegian public oral health care service [ 23 ]. Based on an overriding goal of ensuring equal access to oral health services regardless of ethnic background, sex, personal finances, and life situations, it was decided to establish interdisciplinary “Torture, Abuse and Dental Anxiety (TADA) teams” nationally. These teams consist of both clinical psychologists and oral health professionals. TADA teams offer anxiety treatment and/or adapted dental treatment based on cognitive-behavioural therapy (CBT) principles.

Previous studies have showed promising results regarding the effectiveness of CBT for odontophobia [ 24 , 25 , 26 ]. However, there is a lack of studies specifically evaluating CBT for patients with odontophobia and dental anxiety who have been exposed to sexual abuse, violence in close relationships, or torture. To our knowledge, there is not any published studies on the effect of dental anxiety treatment in patients with post-traumatic stress symptoms related to abuse or torture in their literature review. However, we found one study that reported an effect of CBT treatment on dental anxiety in patients with post-traumatic stress symptoms triggered by previous dental treatment [ 27 ]. It is uncertain whether findings from that study could be generalized to patients with more extensive and severe trauma experiences originating from torture, abuse, or violence in close relationships.

The aim of this study was therefore to evaluate the change in symptoms from pre-treatment to post-treatment after integrated psychological and dental treatment for a vulnerable patient group who have been exposed to torture, sexual abuse, and/or violence in close relationships and/or who have odontophobia, in a naturalistic case series design, This is important given that the implementation of TADA teams is unique, and the service has not been evaluated [ 28 ].

Participants and procedure

A naturalistic case series design was used. The inclusion criteria for the TADA treatment were: (a) confirming a history of being subjected to torture, abuse, and/or violence in close relationships and/or confirming clinical symptoms of odontophobia (including blood/injection/injury- phobias), (b) being aged 21 years or more at the point of orientation, (c) being willing and having the ability to commit to a treatment plan prepared in collaboration with an interdisciplinary treatment team, and (d) understanding the rationale and treatment principles for the relevant course of treatment. The exclusion criteria were patients who: (a) had an organic disorder such as dementia, delirium, or severe memory problems, or suffered from a severe depressive disorder, mania, or ongoing psychosis at the time of evaluation, and (b) had known cognitive/language delays corresponding to an intellectual disability and were not considered to be able to benefit from the treatment approach because of this.

Patients were invited to the TADA clinic for an orientation with a clinical psychologist (1–2 appointments) after referral. During the orientation, the motivation to commit to therapy was addressed (e.g., willingness to meet at regular intervals for CBT treatment appointments and to gradually expose themselves to feared events). At the time of orientation, patients who confirmed having dental treatment difficulties (e.g., did not seek dental treatment, failed to carry out dental treatment, and/or endured dental treatment with great difficulty), and/or being exposed to sexual abuse/violence/torture, and were willing to commit to CBT treatment, underwent a diagnostic evaluation and were accepted into the TADA treatment program.

After interdisciplinary CBT treatment (phase 1), patients were referred by the first TADA team to the second TADA team (phase 2). Patients referred to the second TADA team were required to attend their first appointment unaccompanied. The first meeting involved reviewing discharge summaries from the first TADA team and developing a treatment plan for dental restoration. The second TADA dentist team (Phase 2) did not function as CBT therapists in this study. If patients did not need full-scale interdisciplinary CBT treatment at the point of orientation, they were referred directly to a TADA dentist team for adapted dental treatment. If needed, the TADA team referred patients to emergency dental treatment before or after the CBT intervention. Both interdisciplinary CBT treatment and dental treatment were delivered free of charge. The TADA dentist and dental nurse involved in phase 1 have their CBT training from continuous guidance and working in collaboration with the CBT trained psychologist. The TADA team involved in phase 2 consist of another dentist and dental nurse with basic training in CBT provided by the TADA psychologist. Both TADA teams participate in annual courses to maintain basic skills in CBT.

Prior to treatment initiation, dental anxiety was assessed with the specific phobia disorder module of the Mini International Neuropsychiatric Interview (MINI) version 7.0.2. [ 29 ] and dental fear and anxiety symptom questionnaires. Patients exposed to torture, sexual abuse, or violence in close relationships were included in the study regardless of whether the diagnostic criteria for odontophobia were met. These patients were further assessed with questionnaires assessing exposure to potentially stressful life events [ 30 ] and related posttraumatic stress symptom severity [ 31 ]. The patients answered their highest level of education completed (primary school, upper secondary school, college/university up to 5 years, or college/university over 5 years). Patients with college/university experience were defined as “higher education”. Furthermore, patients answered their current marital status (single, cohabiting/married, or in a relationship, but not cohabiting). The degree to which personal economy status had affected dental treatment execution was answered as either “not at all”, “to some extent” or “to a large extent”.

The Modified Dental Anxiety Scale (MDAS) [ 32 ] is a brief, self-administered questionnaire consisting of five questions regarding different dentist treatment situations. Each item is scored on a Likert scale ranging from “1” (not anxious) to “5” (extremely anxious). The item scores are summed to produce a total score ranging from 5 to 25. A cut-off score of 19 indicates high dental anxiety [ 33 , 34 ].

The Dental Fear Survey (DFS) [ 35 , 36 ] is a brief measure of dental anxiety and fear that consists of 20 items. Each item is scored on a Likert scale from “1” (never or not at all) to “5” (always or very much). Total DFS scores range from 20 to 100, with increasing scores indicating higher levels of fear. A total score of 20 indicates “no fear,” a score of 21–40 indicates low fear, a score of 41–79 indicates moderate fear, and a score of 80–100 indicates high fear [ 35 , 36 ].

The Stressful Life Events Screening Questionnaire (SLESQ) [ 30 , 37 ] is a 13-item questionnaire assessing lifetime exposure to various traumatic experiences. Each item represents different traumatic experiences and is scored as either “yes” or “no” depending on whether the individual has been exposed to the incident. This questionnaire was used exclusively at pretreatment to screen for exposure to potential traumatic experiences.

The PTSD Checklist for the DSM-5 (PCL-5) [ 31 ] is a 20-item questionnaire assessing 20 PTSD criteria outlined in the Diagnostic and Statistical Manual of Mental Disorders-Fifth Edition (DSM-5). Each item is scored on a Likert scale ranging from “0” (not at all) to “4” (extremely) based on the occurrence of symptoms during the last month. A total cut-off score of 33 has been found to efficiently detect PTSD [ 38 ]. Only patients who were confirmed to have been exposed to potentially traumatic life events completed the PCL-5.

The Patient Health Questionnaire-9 (PHQ-9) [ 39 ] consists of nine items measuring depressive symptoms. Each of the nine DSM-IV criteria is scored on a Likert scale ranging from “0” (not at all) to “3” (nearly every day) with total scores ranging from 0 to 27, with higher scores reflecting greater depression severity. PHQ-9 scores of 5, 10, 15, and 20 represent mild, moderate, moderately severe, and severe depression, respectively.

The Generalized Anxiety Disorder-7 (GAD-7) [ 40 ] is a brief measure for assessing symptoms of generalized anxiety disorder. The measure consists of seven items measuring worry and anxiety symptoms. Each item is scored on a Likert scale, ranging from “0” (not at all) to “3” (nearly every day). A total score above 10 is considered to be within the clinical range. The GAD-7 is also a measure of anxiety symptoms in general [ 41 ].

The mucosal plaque score (MPS) [ 42 ] is designed to evaluate oral health and oral hygiene. The index consists of two measures: a four-point mucosal score (MS) and a four-point plaque score (PS). The scores are combined, and the total score ranges from 2 to 8, with higher scores indicating poorer oral health and oral hygiene.

The decayed, missing, and filled teeth index (DMFT) quantifies a person’s total number of untreated decayed, missing, and filled teeth and is commonly used in oral epidemiology to quantify the extent of caries [ 43 ]. “Decayed” corresponds to primary or secondary caries in dentin, while “Missing” and “Filled” correspond to missing teeth due to caries, root residues/carious teeth beyond repair and filled/restored teeth with no sign of caries in dentin, respectively. 3rd molars were excluded from the DMFT evaluation, except in situations where these functioned as second molars. The index is frequently used to evaluate and monitor oral health and in oral health interventions [ 44 , 45 ].

Oral health and dental status examinations

Before interdisciplinary CBT treatment, the TADA patients underwent a panoramic X-ray (orthopantomography; OPG). OPG provides a panoramic single radiograph image of the teeth, maxilla, mandible, and adjacent tissue. OPG is a frequently employed radiological examination [ 46 ]. The TADA dentist conducted a dental status evaluation when the patients could tolerate the procedure. To evaluate a patient’s oral health status and dental treatment needs, the dentist determined their mucosal and plaque score (MPS) and the total number of decayed, missing, and filled teeth (DMFT).

CBT intervention (phase 1)

The TADA treatment consisted of two phases. In the first phase, patients were offered interdisciplinary CBT treatment before being referred to an other TADA dentist team for further dental treatment and rehabilitation (second phase).

The interdisciplinary CBT treatment team consisted of a dentist, a dental nurse, and a clinical psychologist delivering CBT together. During orientation to TADA treatment, a psychologist prepared the patient for CBT treatment by providing psychoeducation and rationale for exposure therapy, mapped catastrophic thoughts and safety and avoidance behaviours, examined the patient’s motivation for treatment, and clarified the treatment framework (e.g., treatment duration and structure, dental treatment clarification). The CBT treatment team then offered cognitive-behavioural treatment to challenge patients’ catastrophic thoughts and beliefs about dental treatment and find ways to adapt dental treatment to make it feasible. Patients with odontophobia or dental anxiety related to exposure to torture, sexual abuse, or violence in close relationships also received trauma-relevant psychoeducation and were taught skills on how to cope with trauma symptoms to facilitate new remedial learning experiences. The CBT intervention did not include trauma therapy directly focusing on the primary traumatic event. In addition to cognitive restructuring, in-vivo exposure therapies were conducted, tailored to maximize the disconfirmation of each patient’s unique catastrophic beliefs. While these exposure therapies varied somewhat among patients, the majority of CBT sessions included exposure to activities such as using dental mirrors, probes, polishing, administering anaesthesia, tartar cleaning, drilling, filling procedures, and, when necessary, the process of obtaining impressions and extracting root residues or teeth. Throughout the CBT phase, both dental healthcare professionals and a psychologist were typically present.

Anxiolytic drugs were not offered as part of the treatment intervention. The standard CBT treatment consisted of weekly therapy sessions (1–1.5 h) for up to 12 sessions. All exposure sessions were carried out in vivo in the dental office. The extent of the psychologist’s involvement during exposure sessions was evaluated on an individual basis. Additional sessions could be granted if the TADA team expected the patient to benefit from further follow-up.

Dental treatment intervention (phase 2)

Only limited dental treatment was carried out in the interdisciplinary CBT phase of the TADA treatment. In this phase, dental treatment was carried out only for the purpose of exposure and for facilitating new learning experiences. In case of acute infections and an immediate need for dental treatment before or during CBT treatment, patients were referred for dental treatment under general anaesthesia before further CBT treatment was provided. Two patients (10%) in this sample underwent dental treatment under general anaesthesia during the CBT intervention phase.

After interdisciplinary CBT treatment, patients were referred to a different TADA dentist team consisting of a dentist and a dental nurse for dental treatment and rehabilitation. This second phase of the treatment was not time limited. These TADA dentist teams were trained in CBT interventions but did not work collaboratively with a psychologist.

Statistical analyses

A repeated-measures ANOVA was conducted to examine changes in symptoms from pre- to posttreatment. The proportion of missing data was 10.5%. To address missing data, the expectation maximization (EM) method in SPSS, version 29, was utilized to replace missing values. The use of the EM algorithm is appropriate when less than 25% of data are missing and the missing data are deemed to be missing at random, which was confirmed to be the case for the present dataset (Little`s MCAR test X² (18.798), df  = 17, p  = .340).

Demographic and clinical characteristics

Twenty-seven patients were referred for TADA treatment during the designated trial period. Of these patients, we were unable to reach four patients on the waiting list to offer them an initial appointment. Furthermore, two patients declined treatment. Of these two patients, one had already managed dental treatment at the time of orientation, and the other did not want TADA treatment. One patient did not meet the inclusion criteria after treatment orientation and evaluation. Consequently, 20 consecutive patients referred to the regional TADA outpatient clinic for adults in the county of Møre and Romsdal, Norway, were included (please see Fig.  1 for the flow chart). Of these 20 patients, 12 were referred by oral health personnel (dentists, dental hygienists, oral surgeons), four were referred by general practitioners, two were referred by psychiatric services, and two referred themselves.

The flow of TADA treatment after referral

The mean time since the last dental treatment was 10.7 years (range = 0–30 years). The study participants had an average age of 41.8 years (range = 21–64 years), 75% were female, and 65% were married or cohabiting. A minority of patients had completed higher education, and half received social security benefits. A significant proportion of individuals (70%) stated that their personal finances, in part or significantly, had affected their ability to pursue dental treatment. Furthermore, the patients had been on a waiting list for a duration of 42 months prior to the start of phase 1 of the TADA treatment.

All patients in this sample met the diagnostic criteria for odontophobia, and all underwent interdisciplinary CBT treatment. No patients were referred directly to the TADA dentist team after treatment orientation. Furthermore, no patients were referred for trauma therapy before or during CBT treatment by the TADA teams. Two patients were granted additional exposure sessions (one and seven sessions).

Ten patients reported that domestic violence and/or abuse experiences were the cause of their dental anxiety. Of the other ten patients, three patients did not report traumatic incidents, while seven did not relate their abuse/violence experiences as the cause, or sustaining cause, of their odontophobia. None of the patients stated that they were survivors of torture experiences. 70% reported a history of sexual abuse, as measured by the stressful life event questionnaire. Furthermore, 65% reported exposure to violence in close relationships. 55% reported being survivors of both sexual abuse and violence in close relationships. Patients exposed to potential stressful life events reported a mean of 6.3 (range = 3–11) potential traumatic experiences.

70% of patients reported having comorbid psychiatric disorders, and six (30%) patients simultaneously received general mental health treatment. Patients did not have to end their ongoing treatments to be included in the study. The most prevalent comorbid diagnoses were mood disorders (35%), attention-deficit/hyperactive disorder (30%), and posttraumatic stress disorder (30%). Table  1 summarizes the sample’s characteristics.

There were no dropouts during the interdisciplinary CBT phase of the TADA treatment program. On average, patients received 10.8 interdisciplinary CBT sessions (SD = 2.6, range = 6–19 sessions). All patients were referred to the TADA dentist team following the completion of the CBT intervention. Additionally, all patients attended further dental appointments and initiated dental treatment and rehabilitation.

Changes in symptoms

There was a significant reduction in the symptoms of dental anxiety from pre- to post-treatment as measured with the MDAS (λ = 0.07, F (1,19) = 262.10, p  < .001, d  = 3.07). There was also a significant reduction in symptoms of dental fear as measured with the DFS (λ = 0.25, F (1,19) = 57.36, p  < .001, d =  2.18).

For the 17 patients who reported having traumatic experiences, there were large reductions in symptoms of post-traumatic stress as measured with the PCL-5 (λ = 0.56, F (1,16) = 12.43, p  = .003, d  = 3.04). For the whole sample, there was an improvement in symptoms of depression as measured with the PHQ-9 (λ = 0.50, F (1,19) = 19.36, p  < .001, d  = 1.00), and there were moderate improvements in symptoms of generalized anxiety as measured with the GAD-7 (λ = 0.74, F (1,19) = 6.60, p  < .001, d  = 0.57). A summary of the analyses is displayed in Table  2 .

Subgroup analyses were conducted to inspect possible effects of ongoing psychological treatment, and to compare possible differences between patients with and without a history of abuse. The results are summarized in supplemental Table S1 . There were no associations between ongoing psychological treatment and changes in MDAS and DFS. However, patients with ongoing psychological treatment showed less improvement in symptoms of depression and anxiety. Patients with a history of abuse reported similar changes in symptoms as patients without such history.

Oral health and dental treatment needs

The average DMFT score for the total sample was 18.8 (range 10–36). The patients in the sample had on average 6.6 decayed teeth, 5.6 missing teeth and 6.7 filled teeth. See Table  3 for the total average DMFT score and MPS. On average, patients had an MPS of 2.8 (range 2–6).

The present study aimed to evaluate the implementation of integrated psychological and dental treatment within the TADA team for a sample of patients exposed to traumatic events and/or diagnosed with odontophobia. Overall, the sample reported positive treatment outcomes. Notably, no patients declined further dental treatment after the CBT intervention, indicating that the treatment was both accepted and tolerated by the participants.It is promising that all patients in this sample completed the interdisciplinary CBT treatment intervention despite their previous psychological trauma experiences, high degree of psychiatric comorbidities, prolonged dental avoidance behaviour, and the absence of anxiolytic drug administration. Additionally, all patients were referred to and started dental treatment and rehabilitation. These results suggest that the treatment approach was acceptable for vulnerable patients with a history of traumatic experiences and patients with odontophobia. This finding is significant given that the implementation of TADA teams is unique, the service has not been evaluated, and characteristics of the specific patient group have not been described in detail [ 28 ].

There were large and significant improvements in all measures of dental fear and phobia after CBT treatment. However, some studies indicated that a relatively large proportion of patients do not show improved dental attendance despite reporting reductions in their dental anxiety level following different treatments [ 47 ]. Our findings are align more closely with a previous meta-analysis on behavioural interventions for dental fear in adults, showing medium to large effect sizes for self-reported dental anxiety after behavioural interventions and post-treatment attendance at dental visits with rates between 33% and 100% within 6 months after treatment [ 25 ]. All patients initiated dental treatment, but the study lacks information concerning long-term dental care attendance. Additionally, consistent with other research indicating wider positive life changes after CBT for dental anxiety treatment, our study found decreased symptoms of depression and generalized anxiety following treatment [ 48 , 49 ].

Most patients in our sample had a history of being exposed to potentially traumatic life experiences and had a high prevalence of comorbid psychiatric diagnoses. The significant reduction in posttraumatic stress symptoms suggest that the treatment was well tolerated and could alleviate PTSD symptoms. Although the treatment did not have a direct focus on altering the primary traumatic experience and related psychopathology, the treatment intervention did focus on managing trauma symptoms as presented in the dental care setting. The purpose of this was to make it possible for the patients to have new and corrective learning experiences with dental treatment and to alter dental-related catastrophic thoughts and behaviours. These results are thus in line with research that indicates that the exposure of patients to corrective information that violates their expectations is central to fear reduction in psychological therapy [ 50 ]. Furthermore, these results support previous findings from qualitative studies of trauma-informed treatment interventions and indicate that interdisciplinary CBT could be potentially beneficial and feasible for patients exposed to psychological trauma caused and/or maintained by reasons other than previous dental treatment experiences [ 20 , 21 , 51 ].

The patients included in this study had a formal diagnosis of dental phobia at treatment entry and had avoided dental treatment for over a decade. The longevity of dental avoidance in our sample was concordant with other findings [ 25 , 52 ]. In summary, we found significant variations in oral health and dental treatment needs as measured by the total MPS and DMFT score. Dental treatment needs were heterogeneous, varying between no/little to many dental treatment needs. We found that the dental status of the sample was in line with a previous study on treatment-seeking patients with odontophobia in Norway [ 13 ] and Sweden [ 14 ]. The Norwegian study found a DMFT mean score of 16.4 in their sample while the Swedish study found an average DMFT score of 18.6, compared to 18.8 in the current study. We also found significant variations in oral health as measured by the total MPS. This is also in line with the previous studies on dental status in treatment-seeking odontophobia patients in Norway [ 13 ] and Sweden [ 14 ]. The variations in the MPS reflect that some patients had a reduced ability for dental-related self-care behaviour, while others had an intact ability to take care of their own oral health despite severe dental anxiety.

Most patients reported having a low socioeconomic background, which could be associated with a heightened risk of dental fear [ 53 ]. Many patients in the sample (70%) stated that their personal economic status, in part or significantly, had affected their ability to receive dental treatment. These findings suggest that a considerable number of patients in the TADA intervention would have faced financial constraints, making it unlikely for them to independently pursue dental treatment due to limited financial resources. The fact that the TADA treatment (both CBT and dental treatment and rehabilitation) was delivered free of charge, therefore, appears to have been important for patients to be able to overcome their dental treatment difficulties. The availability of affordable treatment could play an important role in facilitating access to necessary dental treatment interventions for these patients.

Interdisciplinary CBT treatment was given. Due to limited resources, oral health care personnel are often required to provide anxiety treatment without access, or with limited access, to psychological expertise. The findings in this study suggest that mental health professionals could be a valuable allies for oral health care personnel.

The current case series study must be considered in light of several limitations. The small number of participants and the lack of a control condition makes it impossible to determine whether the findings are unique to TADA treatment and to evaluate the relative efficacy of the treatment received. The study also lacked a long-term follow-up assessment. Furthermore, some patients with dental fear have been subjected to torture [ 54 ]; however, such experiences were not reported by the current sample, making it difficult to generalize the findings to patients with a history of torture. The study also lacked information about substance abuse and previous negative experiences with dental care.

This study indicates that interdisciplinary CBT in the context of TADA teams could be both beneficial and acceptable for odontophobia and dental anxiety related to sexual abuse and violence. The results suggest that mental health professionals could be important allies for oral health professionals when caring for patients with severe dental anxiety and odontophobia. System-oriented interventions could benefit from interdisciplinary collaboration, striving to offer seamless and effective treatment options to vulnerable patient populations. A larger controlled study examining the long-term effects of TADA treatment is warranted.

Data availability

The anonymized datasets used during the current study are available from the corresponding author upon reasonable request.

Abbreviations

Cognitive-Behavioural Therapy

Torture, Abuse, and Dental Anxiety

Posttraumatic stress disorder

Mucosal and Plaque Score

Decayed, Missing, and Filled Teeth

The Modified Dental Anxiety Scale

Dental Fear Survey

Generalized Anxiety Disorder-7

Patient Health Questionnaire-9

PTSD Checklist for DSM-5

Kisely S, Sawyer E, Siskind D, Lalloo R. The oral health of people with anxiety and depressive disorders: a systematic review and meta-analysis. J Affect Disord. 2016;200:119–32.

Article   PubMed   Google Scholar  

Choi J, Price J, Ryder S, Siskind D, Solmi M, Kisely S. Prevalence of dental disorders among people with mental illness: an umbrella review. Aust N Z J Psychiatry. 2022;56(8):949–63.

Torales J, Barrios I, González I. Oral and dental health issues in people with mental disorders. Medwave. 2017;17(8):e7045.

Turner E, Berry K, Aggarwal VR, Quinlivan L, Villanueva T, Palmier-Claus J. Oral health self-care behaviours in serious mental illness: a systematic review and meta-analysis. Acta Psychiatr Scand. 2022;145(1):29–41.

Yazdanian M, Armoon B, Noroozi A, Mohammadi R, Bayat AH, Ahounbar E, Higgs P, Nasab HS, Bayani A, Hemmat M. Dental caries and periodontal disease among people who use drugs: a systematic review and meta-analysis. BMC Oral Health. 2020;20(1):44.

Article   PubMed   PubMed Central   Google Scholar  

Bjørkvik J, Quintero DPH, Vika ME, Nielsen GH, Virtanen JI. Barriers and facilitators for dental care among patients with severe or long-term mental illness. Scand J Caring Sci. 2022;36(1):27–35.

Kani E, Asimakopoulou K, Daly B, Hare J, Lewis J, Scambler S, Scott S, Newton JT. Characteristics of patients attending for cognitive behavioural therapy at one UK specialist unit for dental phobia and outcomes of treatment. Br Dent J. 2015;219(10):501–6. discussion 506.

Article   PubMed   CAS   Google Scholar  

Locker D, Liddell A, Dempster L, Shapiro D. Age of onset of dental anxiety. J Dent Res. 1999;78(3):790–6.

Nermo H, Willumsen T, Rognmo K, Thimm JC, Wang CEA, Johnsen JK. Dental anxiety and potentially traumatic events: a cross-sectional study based on the Tromsø Study-Tromsø 7. BMC Oral Health. 2021;21(1):600.

Hakeberg M, Berggren U, Gröndahl HG. A radiographic study of dental health in adult patients with dental anxiety. Community Dent Oral Epidemiol. 1993;21(1):27–30.

Schuller AA, Willumsen T, Holst D. Are there differences in oral health and oral health behavior between individuals with high and low dental fear? Community Dent Oral Epidemiol. 2003;31(2):116–21.

Halonen H, Nissinen J, Lehtiniemi H, Salo T, Riipinen P, Miettunen J. The association between dental anxiety and psychiatric disorders and symptoms: a systematic review. Clin Pract Epidemiol Ment Health. 2018;14:207–22.

Agdal ML, Raadal M, Skaret E, Kvale G. Oral health and oral treatment needs in patients fulfilling the DSM-IV criteria for dental phobia: possible influence on the outcome of cognitive behavioral therapy. Acta Odontol Scand. 2008;66(1):1–6.

Bohman W, Lundgren J, Berggren U, Carlsson S. Psychosocial and dental factors in the maintenance of severe dental fear. Swed Dent J. 2010;34(3):121.

Google Scholar  

Lenk M, Berth H, Joraschky P, Petrowski K, Weidner K, Hannig C. Fear of dental treatment–an underrecognized symptom in people with impaired mental health. Dtsch Arztebl Int. 2013;110(31–32):517–22.

PubMed   PubMed Central   Google Scholar  

Leeners B, Stiller R, Block E, Görres G, Imthurn B, Rath W. Consequences of childhood sexual abuse experiences on dental care. J Psychosom Res. 2007;62(5):581–8.

Dougall A, Fiske J. Surviving child sexual abuse: the relevance to dental practice. Dent Update. 2009;36(5):294–6. 303 – 294.

Levine PA. Waking the tiger: Healing trauma: the innate capacity to transform overwhelming experiences. North Atlantic Books; 1997.

Larijani HH, Guggisberg M. Improving Clinical Practice: What Dentists Need to Know about the Association between Dental Fear and a History of Sexual Violence Victimisation. Int J Dent 2015, 2015:452814.

Fredriksen TV, Søftestad S, Kranstad V, Willumsen T. Preparing for attack and recovering from battle: understanding child sexual abuse survivors’ experiences of dental treatment. Community Dent Oral Epidemiol. 2020;48(4):317–27.

Kranstad V, Søftestad S, Fredriksen TV, Willumsen T. Being considerate every step of the way: a qualitative study analysing trauma-sensitive dental treatment for childhood sexual abuse survivors. Eur J Oral Sci. 2019;127(6):539–46.

Stalker CA, Russell BDC, Teram E, Schachter CL. Providing dental care to survivors of childhood sexual abuse: treatment considerations for the practitioner. J Am Dent Association. 2005;136(9):1277–81.

Article   Google Scholar  

Norwegian Directorate of Health. Tilrettelagte tannhelsetilbud for mennesker som er blitt utsatt for tortur, overgrep eller har odontofobi (facilitated dental health services for people who have been subjected to torture, abuse or odontophobia). Oslo: Helsedirektoratet (Norwegian Directorate of Health); 2010.

Gordon D, Heimberg RG, Tellez M, Ismail AI. A critical review of approaches to the treatment of dental anxiety in adults. J Anxiety Disord. 2013;27(4):365–78.

Kvale G, Berggren U, Milgrom P. Dental fear in adults: a meta-analysis of behavioral interventions. Community Dent Oral Epidemiol. 2004;32(4):250–64.

Boman UW, Carlsson V, Westin M, Hakeberg M. Psychological treatment of dental anxiety among adults: a systematic review. Eur J Oral Sci. 2013;121(3 Pt 2):225–34.

De Jongh A, Van Der Burg J, Van Overmeir M, Aartman I, Van Zuuren FJ. Trauma-related sequelae in individuals with a high level of dental anxiety. Does this interfere with treatment outcome? Behav Res Ther. 2002;40(9):1017–29.

Bryne E, Hean SCPD, Evensen KB, Bull VH. Exploring the contexts, mechanisms and outcomes of a torture, abuse and dental anxiety service in Norway: a realist evaluation. BMC Health Serv Res. 2022;22(1):533.

Sheehan D, Janavs J, Baker R, Harnett-Sheehan K, Knapp E, Sheehan M. Mini international neuropsychiatric interview. Tampa: University of South Florida; 1994.

Goodman LA, Corcoran C, Turner K, Yuan N, Green BL. Stressful life events screening questionnaire. Washington, DC: US Department of Veterans Affairs; 2013.

Weathers FW, Litz BT, Keane TM, Palmieri PA, Marx BP, Schnurr PP. The ptsd checklist for dsm-5 (pcl-5). Boston, MA: National Center for PTSD; 2013.

Humphris GM, Morrison T, Lindsay SJ. The modified dental anxiety scale: validation and United Kingdom norms. Community Dent Health. 1995;12(3):143–50.

PubMed   CAS   Google Scholar  

Humphris GM, Dyer TA, Robinson PG. The modified dental anxiety scale: UK general public population norms in 2008 with further psychometrics and effects of age. BMC Oral Health. 2009;9:20.

King K, Humphris G. Evidence to confirm the cut-off for screening dental phobia using the modified dental anxiety scale. Soc Sci Dent. 2010;1(1):21–8.

Kleinknecht RA, Klepac RK, Alexander LD. Origins and characteristics of fear of dentistry. J Am Dent Assoc. 1973;86(4):842–8.

Kleinknecht RA, Thorndike RM, McGlynn FD, Harkavy J. Factor analysis of the dental fear survey with cross-validation. J Am Dent Assoc. 1984;108(1):59–61.

Goodman LA, Corcoran C, Turner K, Yuan N, Green BL. Assessing traumatic event exposure: general issues and preliminary findings for the stressful life events screening questionnaire. J Trauma Stress. 1998;11(3):521–42.

Wortmann JH, Jordan AH, Weathers FW, Resick PA, Dondanville KA, Hall-Clark B, Foa EB, Young-McCaughan S, Yarvis JS, Hembree EA, et al. Psychometric analysis of the PTSD checklist-5 (PCL-5) among treatment-seeking military service members. Psychol Assess. 2016;28(11):1392–403.

Spitzer RL, Kroenke K, Williams JB. Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. Primary care evaluation of mental disorders. Patient health questionnaire. JAMA. 1999;282(18):1737–44.

Spitzer RL, Kroenke K, Williams JB, Lowe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166(10):1092–7.

Beard C, Björgvinsson T. Beyond generalized anxiety disorder: psychometric properties of the GAD-7 in a heterogeneous psychiatric sample. J Anxiety Disord. 2014;28(6):547–52.

Henriksen BM, Ambjørnsen E, Axéll TE. Evaluation of a mucosal-plaque index (MPS) designed to assess oral care in groups of elderly. Spec Care Dentist. 1999;19(4):154–7.

Broadbent JM, Thomson WM. For debate: problems with the DMF index pertinent to dental caries data analysis. Community Dent Oral Epidemiol. 2005;33(6):400–9.

Article   PubMed   PubMed Central   CAS   Google Scholar  

Marthaler TM. Changes in dental caries 1953–2003. Caries Res. 2004;38(3):173–81.

Nadanovsky P, Sheiham A. Relative contribution of dental services to the changes in caries levels of 12-year-old children in 18 industrialized countries in the 1970s and early 1980s. Community Dent Oral Epidemiol. 1995;23(6):331–9.

Pandolfo I, Mazziotti S. OPT in Post-treatment Evaluation. In: Orthopantomography. edn. Milano: Springer Milan; 2013: 165–198.

Aartman IH, de Jongh A, Makkes PC, Hoogstraten J. Dental anxiety reduction and dental attendance after treatment in a dental fear clinic: a follow-up study. Community Dent Oral Epidemiol. 2000;28(6):435–42.

Vermaire JH, De Jongh A, Aartman IH. Dental anxiety and quality of life: the effect of dental treatment. Community Dent Oral Epidemiol. 2008;36(5):409–16.

Hakeberg M, Berggren U, Carlsson SG, Gröndahl HG. Long-term effects on dental care behavior and dental health after treatments for dental fear. Anesth Prog. 1993;40(3):72–7.

PubMed   PubMed Central   CAS   Google Scholar  

Craske MG, Treanor M, Conway CC, Zbozinek T, Vervliet B. Maximizing exposure therapy: an inhibitory learning approach. Behav Res Ther. 2014;58:10–23.

Erga AH, Kvernenes KV, Evensen KB, Vika ME. Behandling av odontofobi for pasienter med post-traumatiske plager: en litteraturoversikt (treatment of dental phobia in patients with post-traumatic symptoms: a literature review). Nor Tann Tid. 2017;127(8):682–6.

Willumsen T, Vassend O, Hoffart A. A comparison of cognitive therapy, applied relaxation, and nitrous oxide sedation in the treatment of dental fear. Acta Odontol Scand. 2001;59(5):290–6.

Armfield JM, Spencer AJ, Stewart JF. Dental fear in Australia: who’s afraid of the dentist? Aust Dent J. 2006;51(1):78–85.

Høyvik AC, Willumsen T, Lie B, Hilden PK. The torture victim and the dentist: the social and material dynamics of trauma re-experiencing triggered by dental visits. J Rehabil Torture Vict Prev Torture. 2021;31(3):70–83.

Download references

Acknowledgements

The authors thank all the patients who participated in the study, the TADA dental nurses and all TADA dentist teams who participated in the data collection. They also thank Møre and Romsdal County Authority for the encouragement and support for conducting the study protocol.

Open access funding provided by NTNU Norwegian University of Science and Technology (incl St. Olavs Hospital - Trondheim University Hospital). The study and study protocol were founded by Møre and Romsdal County Authority.

Open access funding provided by NTNU Norwegian University of Science and Technology (incl St. Olavs Hospital - Trondheim University Hospital)

Author information

Authors and affiliations.

Molde Competence Clinic for Public Dental Health Service, Møre and Romsdal County Authority, Molde, Norway

Yngvill Ane Stokke Westad, Gina Løge Flemmen, Stian Solem, Trine Monsen & Henriette Hollingen

Department of Psychology, Norwegian University of Science and Technology, Trondheim, Norway

Audun Havnen

Center for Oral Health Services and Research, Mid-Norway (TkMidt), Trondheim, Norway

Astrid Feuerherm

Nidaros Division of Psychiatry, Community Mental Health Centre, St. Olav’s University Hospital, Trondheim, Norway

Molde Hospital, Møre og Romsdal Hospital Trust, Molde, Norway

Kristen Hagen

Bergen Center for Brain Plasticity, Haukeland University Hospital, Bergen, Norway

Department of Mental Health, Norwegian University of Science and Technology, Trondheim, Norway

You can also search for this author in PubMed   Google Scholar

Contributions

YASW was responsible for data collection and drafting and revising the work.KH, SS and AH were responsible for the data analysis and interpretation. TM, GF, HH and AF, KH, SS and AH participated in the data collection, interpretation and/or revision process of the manuscript. All authors gave their final approval of the version to be published.

Corresponding author

Correspondence to Kristen Hagen .

Ethics declarations

Ethics approval and consent to participate.

The study was approved by the regional ethical committee in Middle Norway (REK-Midt: 2022/488462) and by the Data Protection Board at Møre and Romsdal County Authority. Informed written consent were obtained from all participants. The participants were informed that participation in the study was voluntary and that they had the right to withdraw from the study at any time without any consequences for their treatment. All procedures were performed in accordance with the relevant guidelines and regulations.

Consent for publication

Not applicable.

Clinical trial number

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Westad, Y.A.S., Flemmen, G.L., Solem, S. et al. Interdisciplinary CBT treatment for patients with odontophobia and dental anxiety related to psychological trauma experiences: a case series. BMC Psychiatry 24 , 606 (2024). https://doi.org/10.1186/s12888-024-06055-w

Download citation

Received : 13 July 2023

Accepted : 30 August 2024

Published : 10 September 2024

DOI : https://doi.org/10.1186/s12888-024-06055-w

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Odontophobia
• Dental anxiety
• Psychological trauma
• Oral health
• Interdisciplinary treatment

BMC Psychiatry

ISSN: 1471-244X

• General enquiries: [email protected]

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List

Reported Radiation Overexposure Accidents Worldwide, 1980-2013: A Systematic Review

Karen coeytaux.

1 Episight Consulting, Summit, New Jersey, United States of America

2 Plastic and Reconstructive Surgery Department, Percy Military Hospital, Clamart, France

Doran Christensen

3 Radiation Emergency Assistance Center/Training Site (REAC/TS), Oak Ridge, Tennessee, United States of America

Erik S. Glassman

Becky murdock, christelle doucet.

4 Celogos, Paris, France

Conceived and designed the experiments: KC EB DC ESG CD. Performed the experiments: KC BM EB CD. Analyzed the data: KC. Contributed reagents/materials/analysis tools: KC EB DC BM CD. Wrote the paper: KC EB DC ESG BM CD. Data collection: KC BM EB CD. Analysis and interpretation of the data: KC EB DC CD. Drafting article: KC. Critical review of the article: KC EB DC ESG BM CD. Final approval of the version to be published: KC EB DC ESG BM CD. Agreement to be accountable for all aspects of the work: KC EB DC ESG BM CD.

Associated Data

All relevant data are referenced in Table 1 of the paper.

Radiation overexposure accidents are rare but can have severe long-term health consequences. Although underreporting can be an issue, some extensive literature reviews of reported radiation overexposures have been performed and constitute a sound basis for conclusions on general trends. Building further on this work, we performed a systematic review that completes previous reviews and provides new information on characteristics and trends of reported radiation accidents.

We searched publications and reports from MEDLINE, EMBASE, the International Atomic Energy Agency, the International Radiation Protection Association, the United Nations Scientific Committee on the Effects of Atomic Radiation, the United States Nuclear Regulatory Commission, and the Radiation Emergency Assistance Center/Training Site radiation accident registry over 1980-2013. We retrieved the reported overexposure cases, systematically extracted selected information, and performed a descriptive analysis.

297 out of 5189 publications and reports and 194 records from the REAC/TS registry met our eligibility criteria. From these, 634 reported radiation accidents were retrieved, involving 2390 overexposed people, of whom 190 died from their overexposure. The number of reported cases has decreased for all types of radiation use, but the medical one. 64% of retrieved overexposure cases occurred with the use of radiation therapy and fluoroscopy. Additionally, the types of reported accidents differed significantly across regions.

Conclusions

This review provides an updated and broader view of reported radiation overexposures. It suggests an overall decline in reported radiation overexposures over 1980-2013. The greatest share of reported overexposures occurred in the medical fields using radiation therapy and fluoroscopy; this larger number of reported overexposures accidents indicates the potential need for enhanced quality assurance programs. Our data also highlights variations in characteristics of reported accidents by region. The main limitation of this study is the likely underreporting of radiation overexposures. Ensuring a comprehensive monitoring and reporting of radiation overexposures is paramount to inform and tailor prevention interventions to local needs.

Introduction

Radiation overexposure accidents are uncommon, but can have severe long-term health consequences. Radiation is used in various settings. Major sectors include industrial, medical, and military. Some key applications are electricity production, sterilization of material equipment or food, development of nuclear weapons, radiography imaging techniques for welds inspection, radiation therapy, and radiology imaging techniques (e.g. X-ray radiography, fluoroscopy, computed tomography). These last decades, the medical sector in particular experienced a fast growth in the use of ionizing radiation that allowed better diagnostics and treatments [ 1 – 2 ]. In order to guide all facets of prevention, it is critical to understand the reasons and events behind radiation overexposures.

Radiation-related harms have been reported over the years in all sectors, along with those resulting from orphan sources. Harmful effects of radiation overexposure include deterministic effects (e.g. radiation sickness, skin radiation burn, cataracts, infertility) and stochastic effects (e.g. cancer). These effects can take from weeks to years to manifest, with severity depending upon multiple parameters including the total radiation absorbed dose, the radiation dose rate, the volume of body exposed, the parts of the body and tissues involved, the radiation source at stake, as well as personal characteristics of overexposed people (e.g. age, health status) [ 3 ].

Furthermore, local and global overexposures to the body translate into different health outcomes and therefore different treatment needs. Global overexposure of 1 Gray (Gy) or above induces acute radiation syndrome (ARS) characterized by consecutive hematopoietic, gastrointestinal, and neurovascular syndromes [ 3 ]. Local skin overexposures of 3 Gy or more are likely to lead to acute local radiation injuries (LRI), which may be associated with extreme pain. Local organ overexposures of typically 2–8 Gy are likely to result in organ dysfunction (e.g., permanent sterility of ovaries and testes, acute pneumonitis, renal failure, cognitive defect) [ 4 ]. In addition, this type of injuries often progresses over time due to inflammatory waves, inducing the spread of radionecrosis, and requires long-term treatment [ 5 ].

In order to decrease the risk of harm associated with ionizing radiation, its use is often regulated at the country level. At the international level, the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR), the International Atomic Energy Agency (IAEA), and the International Commission on Radiological Protection (ICRP) play a central role. They evaluate radiation risks, provide recommendations, and promote safe use of radiation technologies, which evolve rapidly and gain in complexity. Yet, in an era where resources are scarce, it is essential to identify the most pressing issues in order to better target prevention actions such as training, which can lead to dramatic improvements in the safe use of radiation [ 6 ].

Several non-systematic reviews of radiation overexposure accidents have been published in the literature. The UNSCEAR organization performed the most extensive review, which it considers "a sound basis for conclusions regarding the number of significant radiation accidents that have occurred, the corresponding levels of radiation exposures and number of deaths and injuries, and the general trends for various practices", despite unavoidable underreporting [ 7 ]. Additionally, several voluntary registries of radiation overexposures have been developed. The Radiation Emergency Assistance Center/Training Site (REAC/TS) at the US Department of Energy's (DOE) Oak Ridge Institute for Science and Education (ORISE) maintains the largest US national and international radiation accident registry to our knowledge [ 8 ]. Other registries focus only on the medical sector. For example, the Radiation Oncology Safety Information System (ROSIS) and Safety in Radiation Oncology (SAFRON) register incidents and near incidents in radiation therapy [ 9 – 10 ].

To build further on the existing work and provide new information on characteristics and trends of reported radiation accidents worldwide, we performed a systematic review of reports published between 1980 and 2013 by querying the peer-reviewed literature, national and international reports, and the REAC/TS registry. Our objective was to evaluate the impact of past prevention programs and potential remaining needs for prevention planning. In this report, we present the search strategy and results of our review, and a descriptive analysis of the retrieved radiation overexposure accidents.

A protocol was developed for the conduct of this systematic review and is detailed in S1 PRISMA Checklist and S1 Protocol .

Reported radiation overexposure accidents from 1980 to present were searched using MEDLINE, EMBASE, the IAEA publications, the congress proceedings of the International Radiation Protection Association (IRPA), the collection of UNSCEAR reports, the United States Nuclear Regulatory Commission (US NRC) reports on abnormal occurrences, and the REAC/TS registry.

For the purposes of this review, we elected to use the IAEA definition of accident, which is "Any unintended event, including operating errors, equipment failures or other mishaps, the consequences or potential consequences of which are not negligible from the point of view of protection or safety" [ 11 ]. Our study focused on radiation accidents resulting in one or several people overexposed and meeting our inclusion and exclusion criteria.

Search strategy

We conducted an electronic search in MEDLINE and EMBASE on March 27th 2014 for all relevant articles published since January 1 st 1980, in the English or French languages. The MEDLINE search strategy included a first keyword search in titles using: cause of overexposure (radiation, nuclear) AND type of injury (overexpos*, accident*, injur*).

A second keyword search in titles was performed as follows: cause of interventional radiology overexposure (fluoroscop*, diagnostic, imaging, angioplast*, catheter*, coronar*, arter*, endovasc*, cardiac, stent*, cardiol*, shunt*) AND type of injury (radiodermatitis, "radiation dermatitis", radionecrosis, "radiation necrosis", "radiation injury", "radiation injuries", "radiation effect", "radiation effects", erythem*, "radiation-induced skin", “skin injury", “skin injuries", ulceration) NOT (erythematosus or lupus).

The third search was based on the following Mesh terms: (("Radiodermatitis"[Mesh] OR "Acute Radiation Syndrome"[Mesh] OR ("Radiation Injuries/mortality"[Mesh] OR "Radiation Injuries/radiation effects"[Mesh] OR "Radiation Injuries/radiography"[Mesh] OR "Radiation Injuries/radionuclide imaging"[Mesh] OR "Radiation Injuries/radiotherapy"[Mesh]) OR ("Radiotherapy, Computer-Assisted/adverse effects"[Mesh] OR "Radiotherapy, Computer-Assisted/mortality"[Mesh]) OR "Radiotherapy, Image-Guided"[Mesh] OR "Whole-Body Irradiation/adverse effects"[Mesh] OR "Fluoroscopy/adverse effects"[Mesh])) AND ("case reports"[Publication Type] OR "review"[Publication Type]) AND "humans"[Mesh].

The EMBASE search strategy used the following words of major interest: 'radiation accident'/exp/mj OR 'radiation injury'/exp/mj OR 'acute radiation syndrome'/exp/mj OR 'radiation dermatitis'/exp/mj OR 'radiation sickness'/exp/mj OR 'radiation necrosis'/exp/mj AND [humans]/lim AND ([english]/lim OR [french]/lim) AND [1980–2014]/py AND [embase]/lim NOT ([medline]/lim AND [1980–2014]/py OR 'radiation recall' OR 'pneumonitis'/exp/mj OR 'nephrititis' OR 'efficacy' OR 'trial' OR 'sodium arsenite'/exp/mj OR 'neurofibromatosis'/exp/mj OR 'cisplatin'/exp/mj OR 'vinorelbine'/exp/mj OR 'enteritis'/exp/mj OR 'experimental' OR 'bevacizumab'/exp/mj OR 'chemoradiotherapy'/exp/mj OR 'simvastatin'/exp/mj OR 'cetuximab'/exp/mj OR 'cefotetan'/exp/mj OR 'gemcitabine'/exp/mj).

Reference titles and summaries were screened manually and discarded if not relevant. Selected publications were read in full text for data extraction. Cross-referencing was used to retrieve additional relevant articles.

The IAEA publications (nuclear safety reviews, safety reports series, and non serial publications on radiological accidents from 1980 to 2013), the IRPA congress proceedings(1980–2008), the UNSCEAR reports (1980–2013) and the US NRC reports to congress on abnormal occurrences (1980–2012) were systematically read in full text and extracted when relevant.

Finally, radiation overexposure accidents that occurred on or after January 1 st , 1980 and reported in the REAC/TS registry were retrieved using filters against our inclusion criteria. Full reports were read for retrieved cases and extracted if relevant.

Inclusion and exclusion criteria

A case of radiation overexposure was defined as presenting at least one of the following criteria: (i) unintended global overexposure of 1 Gy or more, (ii) unintended local skin overexposure of 3 Gy or more, (iii) unintended local organ overexposure (e.g. brain, thyroid, prostate) of 5 Gy or more, or (iv) description of clinical presentation providing reasonable index of suspicion for unintended ionizing radiation overexposure (i.e., acute radiation syndrome, radiodermatitis, permanent alopecia, dry or moist desquamation, blister formation, skin ulceration, dermal atrophy, invasive fibrosis, organ failure, radio-necrosis). The thresholds used in this review are based on the literature, keeping in mind that these thresholds are not absolute boundaries [ 3 ], [ 4 ], [ 12 – 18 ]. Cases that met none of these criteria were excluded, as were suicide and criminal acts.

For cases without occurrence date, we used date of first symptoms as first proxy and date of report as second proxy.

Finally, similar cases issued from different reports but with insufficient information to decipher whether they were different or not, were not integrated in the database to prevent duplicates.

Selection process

Two independent researchers screened and reviewed data sources against the inclusion criteria. For selected reports, full-text documents were evaluated and extracted manually by one reviewer and double-checked by a second reviewer. Any divergence between reviewers regarding selection process was resolved through discussion.

Extracted data items

For each accident, select information was extracted into a data sheet. Selected data included date and place of occurrence, number of overexposed people and number of people dying from their overexposure, days between exposure and death, type of overexposed people (i.e., patient, public, or worker if dose was received in the course of employment), estimated global and local dose received, type of source, type of overexposure (i.e., local skin, local organ, or global), and documents in which the accident was reported. Reported symptoms, course of treatment, and treatment outcomes were also recorded when available. Finally, accidents were categorized by sector of occurrence: "industrial" including industrial irradiator, production, and radiography; "radiation therapy" including teletherapy, brachytherapy, and therapeutic nuclear medicine; "fluoroscopy" used to support diagnostic and interventional radiology; "military" (e.g. nuclear testing, submarine accidents), and "orphan sources" for overexposures caused by sources fallen outside of regulatory control [ 19 ]. A category "others" included overexposure accidents resulting from scientific experiments and unknown causes. For cases with incomplete information, missing data were reported as unknown in our extraction sheet.

Quality of selected articles and reports

Only cases published in peer-reviewed journals or reported by official experts in radiation management (e.g., IAEA, US NRC, UNSCEAR, REAC/TS) were considered, to ensure the quality of the study. Furthermore, all sources selected for data extraction addressed our review question, which was to understand the characteristics of reported radiation overexposure accidents worldwide and their evolution between 1980 and 2013. Among these, only sources showing evidence of radiation overexposure, as defined in our inclusion criteria, were considered for extraction.

Our extraction sheet was used to assess the distribution of reported radiation overexposure accidents along recorded items and over time.

Study selection process

Out of 5189 articles and reports identified, 296 met our eligibility criteria and were extracted ( Fig. 1 ). In addition, 194 records of the REAC/TS radiation accidents registry also met our eligibility criteria and were considered for data extraction. 70 of these 194 records were not reported in any other data source considered for this review. For the period 1980–2013, 634 reported radiation overexposure accidents were identified, encompassing 2390 overexposed people, of whom 190 (8%) died from their overexposure.

Search strategy for retrieving reported radiation overexposure accidents worldwide, 1980–2013.

Extracted cases were categorized by sector in which the accident occurred and by type of overexposure ( Table 1 ).

a Scientific experiments and unknown causes.

Characteristics of reported radiation overexposure accidents and evolution over time

Among the 634 reported radiation accidents identified, most of them occurred in the industrial sector (27%) and in the medical sector through the use of radiation therapy (32%) or fluoroscopy (31%) ( Table 1 ). Reported accidents in radiation therapy were greater in terms of number of overexposed people (47%), followed by accidents in the industrial sector (22%), fluoroscopy (17%) and orphan sources (9%). Finally, the number of deaths resulting from radiation overexposure, was the greatest for accidents reported in radiation therapy (51%), followed by those reported in the industrial sector (24%) and accidents involving orphan sources (19%).

Over the 1980–2009 period, the number of reported radiation accidents and the number of overexposed people by decade exhibited an overall downward trend ( Fig. 2a—b ). The same trend held for each sector separately, except for accidents reported in radiation therapy and medical fluoroscopy. The number of reported radiation therapy accidents per decade increased along the three decades, however the number of overexposed people experienced an overall decrease within the same period. Moreover, the number of reported fluoroscopic accidents increased significantly from the decade 1980–1989 to 1990–1999. Then, while reported fluoroscopic accidents decreased during the decade 2000–2009, the number of overexposed people increased. While industrial and orphan sources accidents accounted together for most reported accidents between 1980 and 1989 (60%), their proportion sharply decreased afterwards, reaching 17% of reported accidents between 2000 and 2009. Radiation therapy and fluoroscopy accidents experienced the opposite trend and ultimately accounted for most reported accidents and most of overexposed people between 2000 and 2009 (respectively 80% and 87%).

(a) number of reported radiation accidents (b) number of reported overexposed people. a Partial decade.

Among reported overexposed people, patients represented the largest share ( Fig. 3a ). Additionally, the share of patients among overexposed people increased along the three decades 1980–1989 (44%), 1990–1999 (71%), and 2000–2009 (87%), while the share of public and workers decreased along the same period.

(a) Type of overexposed people (b) Type of overexposure. a Partial decade.

Finally, the types of overexposure in reported accidents also changed ( Fig. 3b ). Among retrieved cases, the share of overexposures with a combined local skin and global overexposure decreased from 40% to 6% between the periods 1980–1989 and 2000–2009. The share of global radiation overexposures remained low (below 4%) since the 1980s. However, the share of local organ or local skin overexposures increased over the same period. Local skin overexposures had the highest share during 1980–2013 (46%).

Profiles of reported radiation overexposures by region

The number of cases retrieved varied greatly across geographic regions ( Fig. 4 ). Altogether, North America, South America, Europe and North & Central Asia accounted for 90% of reported overexposures retrieved for the period 1980–2013. Furthermore, the distribution of sectors involved was different from one region to another. Among the four regions with the highest number of reported cases retrieved, the medical sector accounted for most radiation overexposures reported in North America (663 cases, 91%), Europe (642 cases, 93%), and South America (163 cases, 61%). Within the medical sector, the share of reported cases resulting from fluoroscopic overexposures was higher in North America (297 cases, 41%) compared to Europe (55 cases, 8%) and South America (1 case). In contrast, the leading sector of reported radiation overexposure was the industrial one in North and Central Asia (316 cases, 69%).

This review explored a wide array of information sources. It included the assessment of peer-reviewed literature, reports from key national and international organizations in radiation management, and the review of the largest international radiation accidents registry, REAC/TS, for the period 1980–2013.

This review showed that a limited and decreasing number of worldwide radiation accidents have been reported by decade since 1980 and that these accidents can be dramatic, as observed previously [ 271 ]. Furthermore, this review suggested that the characteristics of reported radiation overexposure accidents differ over time and across regions. These new findings are important for future interventions in radiation protection.

Overall downward trends

Our results indicated that the number of reported radiation accidents and overexposed people have decreased over the 1980–2013 period. This likely reflects the impact of continuous efforts in radiation protection. Along the years, the IAEA in conjunction with new ICRP recommendations, has updated safety standards [ 303 ]. In 1989, it introduced the Safety Standards Series, which includes Safety Fundamentals, Safety Requirements, and Safety Guides [ 304 ]. These standards are used to improve regulation and radiation protection worldwide [ 305 ]. For example, ample documentation including lessons learned from accidents in industrial radiography, guidance on safe work practices and training material, have been developed to promote safety prevention in the industrial sector [ 306 – 308 ]. The downward trend in reported radiation accidents could also reflect a decrease in the reporting level. However, this explanation is less likely as the decrease is not observed in all sectors.

Singularity of the medical sector

This review suggested that the medical sector accounted for most reported radiation overexposures along 1980–2013. Consistently, most reported cases involved patients and a local skin overexposure component.

Unlike other sectors, reported radiation therapy accidents increased along the three last decades, while the number of reported overexposed people tended to decrease. Still, for each decade, radiation therapy accidents represented the highest share of overexposed people.

This upward trend in reported radiation therapy accidents could result from improved reporting or growing use of radiation therapy. UNSCEAR estimated that 5.1 million of radiation therapy treatments were delivered annually worldwide over the 1997–2007 period compared to 4.7 million annually for the period 1991–1996 [ 158 ]. The introduction of new technologies (e.g., gamma knife, intensity modulated radiation therapy or IMRT) produced new types of accidental exposures and could also have contributed to this upward trend observed in reported radiation therapy accidents [ 309 – 310 ].

Furthermore, the average number of people involved per reported accident decreased from the 1980’s to the 2000–09 decade. This figure typically varies according to the cause of the accident. While some errors such as errors in treatment site or dose administered, affect a single patient, other errors such as software issue, calibration or treatment programming errors, can affect multiple patients before the issue comes to light. Error reporting systems such as ROSIS allow learning from the past through knowledge of near-misses, incidents or accidents and constitute essential prevention tools [ 311 ].

As the use of radiation therapy is expected to grow even more in the future, it is crucial to ensure high quality assurance standards in order to avoid the multiple possible errors in the course of treatment and thus optimize all benefits of radiation therapy [ 309 – 311 ].

This review also brings attention to medical fluoroscopy, which ranked second in number of reported accidents for 1980–2013. Corresponding reported radiation accidents increased significantly from the 1980s to the 1990s, and then decreased. This was consistent with the literature [ 17 – 18 ], [ 312 ]. This increase could reflect the expanded use of fluoroscopy since the 1990s. Although fluoroscopy was initially used primarily for diagnostic procedures, it then became widely used during therapeutic interventions (e.g. coronary angioplasty), as it provided a less invasive and costly solution than classic surgery [ 313 ]. Potential serious adverse effects of fluoroscopy, however, were rapidly encountered and acknowledged. In 1994, the FDA issued a warning following the reporting of several injuries resulting from the prolonged use of fluoroscopic procedures [ 312 ], [ 314 ]. Thereafter, the risks of cumulative radiation exposure through fluoroscopy have been documented. Also, important efforts to track and decrease patients' overall exposure to imaging radiation following the ALARA (As Low As Reasonably Achievable) safety principle were initiated worldwide and some programs have been implemented successfully at the sub-national level [ 315 – 317 ].

Of note, the increase in number of reported overexposed people through fluoroscopy from 1990–1999 to 2000–2009 despite the decrease in reported accidents over the same period, is primarily due to a single accident involving 206 patients. Its cause was an error in resetting a CT scan, which went undetected for 18 months [ 318 ]. This accident also accounted for about two thirds of reported fluoroscopic overexposure cases in North America.

Geographic differences

Our exploratory analysis suggests that the causes of reported radiation overexposure accidents differ across regions. One possible explanation relates to variations in radiation equipment and radiation use across countries. The review conducted by UNSCEAR for the period 1991–2007 emphasized that the level of X-ray equipment, radiological examinations, and radiation therapy differ greatly from one country to another and tend to be concentrated in a limited set of countries [ 158 ]. For example, during 1997–2007 three-quarters of all radiation therapy treatments were received in countries, which have at least one physician for every 1,000 people in the general population. Still, these countries only represented 24% of the overall surveyed population [ 158 ].

Additionally, differences in reporting by country could also account for the observed differences in causes of overexposure. For example, Baeza highlighted the lack of reports from developing countries when going through ROSIS [ 311 ].

Thus, these geographic differences should be monitored and accounted for in prevention strategies. Identifying specific needs and practical challenges in the implementation of the safety standards is cornerstone to adjust prevention efforts adequately and efficiently reduce the incidence of radiation injuries.

Challenges in reporting radiation injuries

The reporting of radiation overexposure injuries faces unique challenges even when a well-established regulation and a solid reporting system network are in place. Indeed, the latency period before the appearance of radiation-related adverse effects varies from days to years [ 5 ]. Thus, people can easily go undiagnosed or be misdiagnosed. Furthermore, radiation injuries are uncommon, which can contribute to diagnostic errors [ 17 ], [ 201 ]. Additionally potential lack of knowledge or access to reporting systems and fear of legal liabilities can be other causes of underreporting [ 7 ].

Another major difficulty rises for radiation overexposure accidents resulting from orphan sources or medical fluoroscopy. In these contexts, the exposed subjects often cannot directly relate their injuries to their radiation exposure as they are not aware that they have been exposed to radiation. Even among medical professionals performing procedures assisted by fluoroscopy, a lack of awareness of risks associated with radiation imaging can still exist [ 209 ], [ 244 ], [ 319 ]. Additionally, patients noticing their skin lesions usually seek advice from a dermatologist, but without necessarily providing information about their history of prior fluoroscopy. They might think this information irrelevant or simply forget it. This makes the diagnostic of radiation injury even more challenging for dermatologists [ 13 ], [ 16 ], [ 209 ], [ 215 ]. Thus, reported radiation accidents are undoubtedly underestimated.

Limitations

This review offers a solid basis of reported radiation overexposure accidents to inform radiation protection planning. Still, it has several limitations.

Our literature search only included publications written in English and French languages, which might introduce some publication bias. Moreover, our review was limited to sources of information that were publicly available, with the exception of the REAC/TS registry, which has limited public access. Hence, our review does not capture cases that are exclusively reported in private databases (e.g. hospital registries). Thus, our review likely underestimates the number of reported radiation accidents. Additionally, when the date of overexposure was not reported (125 cases out of 2390), date of first symptoms was used as first proxy (11 cases) and date of report as second proxy (114 cases). This might introduce some bias as symptoms can appear months or years following the overexposure. Finally, reporting country was also used as a proxy for country of occurrence when not mentioned explicitly (116 cases out of 2376 with some localization information).

Despite these limitations, this review captured reported radiation accidents in systematic and consistent way, enabling valuable analysis to support future prevention actions.

This systematic study updates and broadens the view of reported radiation overexposure accidents. It indicates that reported radiation overexposure accidents are rare and decreased from 1980–1989 to 2000–2009. However, their potential dramatic outcomes stress the importance of radiation protection regulations. This review suggests the greater share of the medical sector in reported overexposures, for which the use of radiation has become central and is expected to grow even more in the future. Thereby, it confirms the importance of quality assurance programs in radiation therapy and medical fluoroscopy.

Finally, this review suggests that the characteristics of reported accidents vary by geography and over time, and are thereby likely to require different interventions. A close reporting and monitoring of radiation overexposure accidents is of great value to inform and prioritize prevention interventions adequately and ultimately reducing further the incidence of these accidents.

Supporting Information

S1 prisma checklist, s1 protocol, acknowledgments.

We thank Dr Andre Ulmann, MD, PhD, Chairman of HRA Pharma, France, for providing valuable and extensive comments on the document.

Funding Statement

This work was supported by HRA Pharma, 15 rue Beranger, 75003 Paris, France. Celogos and Episight Consulting provided support in the form of salaries for authors KC and CD. The specific roles of these authors are articulated in the “Author Contributions” section. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data Availability

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• My Bibliography
• Collections
• Citation manager

Save citation to file

Email citation, add to collections.

• Create a new collection
• Add to an existing collection

Add to My Bibliography

Your saved search, create a file for external citation management software, your rss feed.

• Search in PubMed
• Search in NLM Catalog
• Add to Search

Current abdominal X-rays practice in accident and emergency

Affiliation.

• 1 University Hospital Wales, Heath Park Way, Cardiff, CF14 4XW, United Kingdom. Electronic address: [email protected].
• PMID: 37573181
• DOI: 10.1016/j.jmir.2023.07.018

Introduction: Previous literature reviews revealed that abdominal X-rays (AXR) performed for the accident and emergency department (A&E), had low sensitivity, high further imaging and non-alignment rate to the Royal College of Radiologists (RCR) guidelines. A study was performed to investigate the current practice with the aim of making recommendations to improve practice, which can reduce patients' radiation exposures, while can re-routing resources to other priorities.

Methods: A study was performed in one of the UK's largest A&Es, in accordance with the RCR guidelines. All the AXR requests from A&E, regardless of the patient's age, within a 28-day period, were retrospectively assessed. Non-A&E patients and abandoned examinations due to uncooperative patients were excluded. The total number of AXR requests received by the A&E imaging department was 169, with 28/169 falling into the exclusion criteria.

Results: Of the 141 included requests, five unjustified requests were correctly rejected. The remaining 136 requests were accepted and performed, though only 115/136 (84.6%) of these were justified. The most common justified and unjustified indications were obstruction and renal stones, respectively. Only 4% of reported AXR had pathological abnormalities, while 45/136 patients had further imaging.

Conclusions: The small proportion of significant findings echoed previous studies, suggesting an AXR overuse. Over 80% of non-compliant requests were performed, and awareness of the justification guidelines can be increased by clinical governance, posters, or an algorithm previously presented. The 32.4% further imaging rate recorded in this study, as opposed to the 73.7% reported in previous literature, merits attention.

Implications to practice: Stopping the overuse of AXR can minimise the radiation dose received and relieve the mounting pressure in imaging and reporting, which can serve other patients who would benefit from the services otherwise.

Keywords: Abdominal X-ray; Abdominal pain; Accident & emergency; Overuse; Radiation protection.

Copyright © 2023. Published by Elsevier Inc.

PubMed Disclaimer

Similar articles

• Utilization of abdominal radiography in the emergency department: Appropriateness, interpretation, radiation protection and costs. García García P, Del Campo Del Val L, Salmerón Béliz I, Paz Calzada E, Alonso Rodríguez C, García Castañón P, Rodríguez Carnero P. García García P, et al. Radiologia (Engl Ed). 2024 Jul-Aug;66(4):307-313. doi: 10.1016/j.rxeng.2023.01.012. Epub 2024 Jul 17. Radiologia (Engl Ed). 2024. PMID: 39089791
• Appropriateness of adult plain abdominal radiograph requesting in a regional Emergency Department. Fernandez M, Craig S. Fernandez M, et al. J Med Imaging Radiat Oncol. 2019 Apr;63(2):175-182. doi: 10.1111/1754-9485.12847. Epub 2019 Jan 10. J Med Imaging Radiat Oncol. 2019. PMID: 30628194
• Abdominal radiograph requesting in the setting of acute abdominal pain: temporal trends and appropriateness of requesting. Morris-Stiff G, Stiff RE, Morris-Stiff H. Morris-Stiff G, et al. Ann R Coll Surg Engl. 2006 May;88(3):270-4. doi: 10.1308/003588406X98586. Ann R Coll Surg Engl. 2006. PMID: 16719997 Free PMC article.
• The use of plain abdominal x rays in the emergency department. Smith JE, Hall EJ. Smith JE, et al. Emerg Med J. 2009 Mar;26(3):160-3. doi: 10.1136/emj.2008.059113. Emerg Med J. 2009. PMID: 19234001 Review.
• Diagnostic management strategies for adults and children with minor head injury: a systematic review and an economic evaluation. Pandor A, Goodacre S, Harnan S, Holmes M, Pickering A, Fitzgerald P, Rees A, Stevenson M. Pandor A, et al. Health Technol Assess. 2011 Aug;15(27):1-202. doi: 10.3310/hta15270. Health Technol Assess. 2011. PMID: 21806873 Free PMC article. Review.
• Radiology education for medical students: a qualitative exploration of educational topics, teaching methods and future strategies. Harthoorn FS, Scharenborg SWJ, Brink M, Peters-Bax L, Henssen D. Harthoorn FS, et al. BMC Med Educ. 2024 Aug 19;24(1):891. doi: 10.1186/s12909-024-05879-0. BMC Med Educ. 2024. PMID: 39160535 Free PMC article.
• Search in MeSH

Related information

Linkout - more resources, full text sources.

• Elsevier Science

• Citation Manager

NCBI Literature Resources

MeSH PMC Bookshelf Disclaimer

The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 02 September 2024

Obtaining parallax-free X-ray powder diffraction computed tomography data with a self-supervised neural network

• H. Dong 1 , 2 ,
• S. D. M. Jacques   ORCID: orcid.org/0000-0002-7275-5272 2 ,
• K. T. Butler   ORCID: orcid.org/0000-0001-5432-5597 1 , 3 ,
• O. Gutowski 4 ,
• A.-C. Dippel 4 ,
• M. von Zimmerman   ORCID: orcid.org/0000-0002-9320-6846 4 ,
• A. M. Beale   ORCID: orcid.org/0000-0002-0923-1433 1 , 2 , 5 &
• A. Vamvakeros   ORCID: orcid.org/0000-0002-4745-0602 2 , 6  

npj Computational Materials volume  10 , Article number:  201 ( 2024 ) Cite this article

1 Altmetric

Metrics details

• Characterization and analytical techniques
• Imaging techniques

In this study, we introduce a method designed to eliminate parallax artefacts present in X-ray powder diffraction computed tomography data acquired from large samples. These parallax artefacts manifest as artificial peak shifting, broadening and splitting, leading to inaccurate physicochemical information, such as lattice parameters and crystallite sizes. Our approach integrates a 3D artificial neural network architecture with a forward projector that accounts for the experimental geometry and sample thickness. It is a self-supervised tomographic volume reconstruction approach designed to be chemistry-agnostic, eliminating the need for prior knowledge of the sample’s chemical composition. We showcase the efficacy of this method through its application on both simulated and experimental X-ray powder diffraction tomography data, acquired from a phantom sample and an NMC532 cylindrical lithium-ion battery.

Similar content being viewed by others

Deep denoising for multi-dimensional synchrotron X-ray tomography without high-quality reference data

Towards end-to-end structure determination from x-ray diffraction data using deep learning

Highly sensitive 2D X-ray absorption spectroscopy via physics informed machine learning

Introduction.

Deep learning, an advanced subset of machine learning, has become a game-changer across a diverse array of fields, including image recognition and text translation 1 , 2 , 3 , 4 , 5 , 6 . Unlike traditional ‘hand-crafted’ algorithms that operate on fixed principles, deep learning harnesses flexible neural networks that evolve based on exposure to different datasets. This dynamic, data-driven, learning process allows deep learning models to continually refine their performance, driving significant advancements in complex tasks where flexibility and adaptability are of utmost importance.

One of the key areas that deep learning has made a significant impact is in the field of tomographic image reconstruction 7 , 8 , 9 . Traditionally, tomographic image reconstruction has relied on either direct methods, like the filtered back projection (FBP) algorithm 10 , or iterative methods that depend on prior knowledge and fine-tuning. However, these methods face their own limitations, especially when it comes to scalability, handling noise and angular undersampling data, computational demand, and the necessity for absolute values in certain applications 11 , 12 , 13 . Deep neural networks (DNNs) have emerged as a compelling solution, offering the potential to surpass the performance of these traditional physics-based approaches 14 , 15 , 16 .

In recent years, there has been a burgeoning interest in the application of DNNs in tomography, notably in enhancing the quality of real-space reconstructed images generated from sinograms. Besides, some innovative applications even leverage supervised learning and generative models to automatically map from sinogram to real space 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 . Despite certain bottlenecks such as handling large images and the computational cost of large networks, the promise of deep learning in this sphere is quite palpable.

The advent of X-ray (powder) diffraction computed tomography (XRD-CT), a specialised form of tomography, has added a new dimension to the mix. This technique uses a pencil beam scanning method to yield reconstructed images corresponding to a sample’s cross-section 25 , 26 , 27 . What sets XRD-CT apart is its ability to resolve chemical species of similar density, a task that conventional X-ray CT often struggles with. As such, XRD-CT has found applications in a wide array of fields ranging from material science to cultural heritage conservation, as well as biological samples 28 , 29 . More importantly, XRD-CT has become an invaluable tool to investigate, non-destructively, functional materials and devices, such catalytic reactors 30 , 31 , 32 , 33 , 34 , 35 , 36 , fuel cells 37 , 38 and secondary/rechargeable batteries in custom-made laboratory cells 38 , 39 , 40 , 41 , 42 , 43 as well as in commercially available and industrially relevant cylindrical form 44 , 45 , 46 , under static or operating conditions (in situ/operando studies). These studies have shown that the spatially-resolved diffraction patterns in the XRD-CT data can yield unique physicochemical information regarding these complex materials systems and their evolving solid-state chemistry. Recently, the method has also been demonstrated for five-dimensional (5D) experiments, where the dimensions correspond to three spatial, one chemical (diffraction) and one temporal or imposed operating condition (e.g. temperature, pressure and potential) 47 , 48 , 49 .

Given the prowess of deep learning and the unique capabilities of XRD-CT, the combination of these two could potentially revolutionise tomographic image reconstruction. Deep learning methods could not only accelerate the XRD-CT on both data acquisition and analysis but also enhance it by addressing challenges like image super-resolution using high-resolution region-of-interest CT scans, data denoising, as well as single-crystal diffraction, self-absorption and parallax artefacts. This combination, if realised, could unlock new possibilities, including higher spatial and temporal resolution in chemical imaging and better handling of complex data sets, paving the way for breakthroughs in various fields.

One major obstacle that prohibits the scale-up of the XRD-CT technique and its widespread adoption to study large samples is the parallax artefact. In wide-angle scattering-based CT experiments, it is generally assumed that the X-rays, whether scattered or diffracted, arrive at the same detector element when measured at any given scattering angle 2 θ across the sample’s thickness, as depicted in Fig. 1a . This assumption holds when the sample thickness is relatively small, typically on the order of a few millimetres. However, for thicker samples, this assumption becomes invalid. In these cases, diffracted X-rays measured at a specific 2 θ angle are detected by multiple detector elements due to the significantly varying distances between components within the sample and the detector. This phenomenon, known as the parallax effect, exhibits a tan(2 θ ) dependency 25 , 27 . The parallax effect is further illustrated in Fig. 1b . As a result of this effect, artefacts may manifest as shifts in peak position, peak broadening, or even peak splitting 50 .

a Schematic representation of a 2D XRD pattern collected during the XRD-CT scanning of a small sample when there is no parallax artefact. b 2D XRD pattern collected during the XRD-CT scanning of a large sample with parallax artefacts present; the X-rays scattered/diffracted along the sample at certain 2 θ angles arrive at different detector elements, leading to peak broadening and peak splitting.

In our previous work, we developed a reconstruction approach, termed the “direct least-squares reconstruction” (DLSR) algorithm, which overcomes the parallax artefact in XRD-CT data 51 . Conventionally, the XRD-CT sinogram data are reconstructed one by one, typically using the filtered back projection algorithm, yielding an XRD-CT reconstructed volume. The next step involves the analysis of all the local diffraction patterns in this reconstructed XRD-CT volume which can be single, multi-peak fitting or full profile analysis using methods such as LeBail or Rietveld. The DLSR was implemented using the TOPAS software 52 version 7 and combines the reconstruction and full profile analysis steps into a single step. To clarify, the sinogram XRD-CT (projection) data are fitted, and the results are real-space maps corresponding to the various properties of the model that are being refined (e.g. scale factor, lattice parameter and crystallite size maps for each phase). This approach yields parallax artefact-free images but has some severe limitations:

It requires a priori knowledge about the chemistry of the sample before reconstruction

It requires the construction of a robust physical model that models all chemistry accurately in the sinogram data; minor components being overlooked during the inspection of diffraction patterns when preparing the physical model will not be part of the final results

DLSR in its TOPAS version 7 implementation suffers from scalability; even XRD-CT images that are nowadays trivially obtained in experiments (e.g. 256 × 256) cannot be handled due to RAM requirements and the data have to be rebinned losing spatial resolution.

It typically requires laborious data pre-processing to decrease memory requirements (to what is realistically available) and yield stable reconstructions. For example, one needs to create a separate binary mask for each crystalline phase present by analysing the FBP reconstructed XRD-CT volume (which contains parallax artefacts) and/or subtract the background from the sinogram data (in order to make it linear/ use a simple background model)

Therefore, our motivation was to develop a method that overcomes all these limitations of the DLSR approach and yields parallax artefact-free XRD-CT images.

We developed a self-supervised parallax XRD-CT data reconstruction architecture by integrating a forward operator that can transfer an XRD-CT volume without parallax artefacts to the sinograms with parallax artefacts. A schematic representation of the method is shown in Fig. 2a . We use an artificial neural network which acts as an XRD-CT volume generator i.e. it creates a stack of real-space XRD-CT images. The input to the generator is a random non-zero constant. Next, the generated images are converted into sinograms with the addition of parallax artefacts using a differentiable forward operator. The forward operator contains two parts, the first part adds artificial parallax artefacts into the images by taking into account the geometry of the experimental setup, and the second part is the Radon transformation that converts the images to sinograms. This generated XRD-CT sinogram volume is then compared with the experimental sinogram dataset using a designated loss function. Based on this comparison, the weights of the generator network are updated accordingly.

a The self-supervised ParallaxNet flow chart. The generator network takes a single-digit number as input and outputs a volume (stack of images), where the third dimension corresponds to the scattering angles (diffraction dimension). A forward operator is applied to convert these XRD-CT images into sinograms containing parallax artefacts. A loss function is then used to compare the differences between the generated sinograms and the experimental sinograms. b The Single Digit to Volume (SD2Vol) generator architecture with a single constant as input. CONV represents 3-D convolutional layers, and FC represents fully connected layers. The filter numbers and layer sizes are shown above each layer. Here n represents the number of translation steps, and m represents the volume size of the output image. The ReLU function is used to connect the layers, except the Leaky ReLU function is used on the last fully connected layer to adapt possible negative values generated.

Building on the recently developed SD2I architecture 53 , which is a lightweight and scalable CNN architecture that utilises a single number input for CT image reconstruction and addresses angular undersampling artefacts, we introduce the Single Digit to Volume (SD2Vol) network architecture for enhanced volumetric reconstructions. By transitioning from 2D to 3D convolutional layers and reducing layer parameters, SD2Vol offers both tailored 3D capability and greater efficiency. Figure 2b illustrates the design of the SD2Vol network employed for 3D image reconstruction from a sinogram volume (stack).

The SD2Vol network begins with a single seed input value (specifically, 1 is used in this work), and is followed by a decoder to reconstruct the image based on the single number. After the input layer, three fully connected layers with 32 nodes and another fully connected layer where the number of nodes equal to the total number of voxels in the reconstructed volume are used. Next, the output of the final fully connected layer is reshaped to a 3D volume followed by four 3D convolutional layers. All activation functions are chosen as ReLU, except for the last layer we used the Leaky ReLU activation function. This architecture can be scaled up, as it allows to reconstruct volumes with n x n x m sizes reaching 550 × 550 × 51 or 100 × 100 × 4010 (i.e. projection data with parallax artefacts). The maximum size of a volume that ParallaxNet can reconstruct with SD2Vol is presented in Supplementary Fig. 1 , and Supplementary Tables 1 – 4 .

We use a joint loss function with the mean squared error (MSE) and the structural similarity index measure (SSIM) in this architecture. The loss function compares the real experimental sinograms with the generated sinograms and updates the weights in the generator to give a generated sinogram volume that better resembles the experimental sinogram volume on the iteration. Normally, the architecture can yield high-quality reconstructions after 1000 iterations. We tested various loss functions, including MSE, mean absolute error (MAE), SSIM, and a joint loss function that combines MSE and SSIM. These tests were conducted with ParallaxNet on the simulated XRD-CT dataset used in this work. The results are presented in Supplementary Table 6 . Based on these findings, we chose the joint loss function that combines MSE and SSIM as the loss function to be used for ParallaxNet. The reconstruction process of this self-supervised method can be expressed as:

Here, G ( a ) is the generated reconstruction image by sending a random constant ‘ a ’ into the generator. The constant does not change while training. The L MSE and L SSIM represent the MSE loss and SSIM loss respectively, and their sum is adjusted by the fraction λ . According to Supplementary Table 6 , we used λ  = 10 −4 for all the reconstructions presented in this work.

The MSE loss is defined as 54 :

The ntr, npr and nim represent the number of translation steps, the number of projections, and the volume size (number of channels), respectively. Then, the SSIM loss can be expressed as 55 :

The μ x , μ y are the average of all pixels in the input sinograms and the σ x , σ y represent their standard deviations. L is the dynamic range of the input images. K 1 and K 2 are two constants that are set as 0.01 and 0.03.

To account for the varying signal strengths across each chemical (diffraction/scattering angle) channel and facilitate easier training of the generator, all sinogram channels along the 2 θ axis are normalised based on the maximum value of each channel. Additionally, to ensure that the output images from the generator maintain consistent relative intensity, they are divided by the same normalisation factors used for the sinograms before applying the forward operator. The forward operator then processes the images at their actual intensity scale, yielding generated sinograms with accurate intensities. Subsequently, these generated sinograms are multiplied by the normalisation factors, and the loss is calculated in comparison to the normalised input reference sinograms

A circular mask is applied to the images during training to filter out signals outside the CT reconstruction area. A 3D grid is calculated considering the experimental setup and specifically the 2 θ diffraction angles (1D vector), the sample-to-detector distance, the translation step size and the X-ray wavelength. Starting with a tomographic angle of 0°, the forward operator accounts for nT voxels across the sample’s thickness and simulates the parallax effect with an nT 2 θ axis vector yielding the 3D grid. The modelling of the 3D grid is based on a relationship between the new 2θ axis, its offset from the centre of rotation, and the distance from the sample to the detector, as defined by Scarlett et al. 50 :

The sinogram volume with parallax can be created from parallax-free images by rotating the 3D grid, interpolating the XRD-CT data over it, and calculating the 3D Radon transform at each CT angle. The pseudocode for creating the 3D grid based on the experimental setup can be found in Algorithm 1, and the pseudocode for the forward operator can be found in Algorithm 2 in the Supporting Information.

Simulated XRD-CT data

To test the performance of ParallaxNet on XRD-CT data with parallax artefacts, we first use a simulated XRD-CT dataset with noiseless and zero-background XRD patterns of a Ni fcc structure (ICSD: 64989) 53 . When testing the performance of algorithms designed to solve inverse problems, it’s crucial to ensure the forward projector used to generate simulated data is different from the forward projector the algorithm employs to solve the inverse problem. This differentiation helps maintain the rigour and validity of the evaluation. Being conscious of this, we coded different forward models for testing our approach with the simulated XRD-CT data. Specifically, we used an A matrix (ray tracing) calculated from astra-toolbox 56 as the forward projector to produce the simulated Ni XRD-CT dataset and a custom Radon using image rotation with bilinear interpolation as the forward projector for our ParallaxNet algorithm. This approach ensures a more unbiased assessment of ParallaxNet’s capabilities.

We take into account the non-constant sample-to-detector distance for large samples by creating a 3D grid, where each pixel represents a distinct 2 θ axis. The XRD-CT data, both the simulated and the experimental presented in the following sections, are interpolated using this 3D grid and subsequently, their 3D Radon transform is calculated. The simulated data were created using a sample-to-detector distance of 1000 mm, a translation step size of 0.2 mm, and a 100 keV X-ray energy. The simulated XRD-CT data consists of 121 translation steps, 121 projections covering 0-180° angular range and 2000 scattering angles, which form a sinogram volume with the size 121 × 121 × 2000. ParallaxNet was used to reconstruct the whole dataset in ca. 7.4 h and 5000 epochs (iterations).

Figure 3a shows the results obtained from the sequential Rietveld analysis of both FBP and ParallaxNet reconstructed volumes using the TOPAS software which is guided by in-house developed Python scripts. As presented in Fig. 3a , the maps of Ni scale factor, crystallite size, and lattice parameter a from ParallaxNet are almost identical to the ground truth maps while the FBP significantly diverge; this is apparent when one observes specifically the lattice parameter and crystallite size maps. This suggests that the diffraction peak positions and shapes reconstructed by ParallaxNet closely match the ground truth patterns. The differences between the crystallite size and lattice parameter maps and their ground truth values are presented in Supplementary Fig. 2 . Their accuracy is also confirmed by the R wp maps from the Rietveld analysis and the distribution of lattice parameters for all pixels, as shown in Supplementary Fig. 3 . The mean image and diffraction patterns, along with selected channels of maps, are shown in Supplementary Figs. 4 and 5 . By visually inspecting the maps reconstructed by FBP and ParallaxNet, we can conclude that ParallaxNet accurately reconstructs the signals in the correct positions and addresses the parallax artefacts present in the simulated data. The distribution of lattice parameters obtained by the three methods is shown in Supplementary Fig. 6 . In a pixel-wise analysis, we select one region of interest, as depicted in Fig. 3b corresponding to one particle. It can be seen that in the simulated dataset this particle has one value for lattice parameter and crystallite size. As illustrated in Fig. 3c , the parallax artefact causes the FBP reconstructed patterns to exhibit significant shifts in peak positions, broadening, and some instances of splitting when reconstructed using the conventional 0–180° CT acquisition. However, the ParallaxNet can accurately reconstruct the volume without these artefacts and the reconstructed diffraction peaks are well aligned with the ground truth data.

a Parallax XRD-CT simulations. This figure shows the results obtained from the sequential Rietveld analysis of the reconstructed XRD-CT data with FBP and ParallaxNet and their ground truth value. b The mean image of the simulated Ni XRD-CT dataset with a marked region of interest. c Selected peaks from the mean diffraction pattern using the region of interest.

Experimental XRD-CT data

Next, we evaluate the efficacy of the method using experimental XRD-CT data. The first dataset is a custom-made phantom consisting of four glass pipettes filled with different powder samples 51 . The mean image of the phantom XRD-CT sample can be found in Fig. 4a and Supplementary Fig. 7a , which provides a view of the cross-section containing the powder samples within the four glass pipettes. This dataset contains two crystalline MgO (ICSD 9863), one SiC (ICSD 603798) and one TiO 2 rutile (ICSD 33837) phases, respectively. Unlike the dataset presented in our DLSR work, here we utilise the full size of the image dataset with ParallaxNet without any image rebinning/resizing. This is because the ParallaxNet method boasts better scalability compared to the DLSR-TOPAS approach. The XRD-CT sinogram volume dataset comprises 269 translation steps, 300 projections, and 670 selected diffraction channels. We split the dataset into three batches, and each batch contains 250 channels. To mitigate the edge effect between the batches, we incorporated an overlap of 40 channels for each batch. As a result, the three batches are defined with channel numbers 0–250, 210–460, and 420–670, respectively. The three batches are merged afterwards by taking the average of the overlapped channels.

a The mean image of the Phantom XRD-CT dataset with three marked regions. b A selected peak of the average diffraction pattern in Region A. c A selected diffraction peak from the average diffraction pattern in Region B. d A selected peak from the average diffraction pattern in Region C. This figure shows the ParallaxNet can solve the peak broadening artefacts, and peak positions reconstructed by the ParallaxNet with 0–180° scans are aligned with the FBP reconstructed with 0–360° scans. In contrast, the FBP reconstructed diffraction peaks over the 0–180° scan range exhibit significant diffraction peak shifting which was caused by the parallax artefacts.

For each batch, we ran 5000 epochs, which took ca. 5.28 h excluding the initialisation time. To improve image quality and reduce the number of required epochs, we pre-trained the generator using FBP images from 180° projections. This preliminary step required only 1000 iterations and was completed in 4 min for each batch. In total, the image reconstruction with ParallaxNet took ca. 16.43 h. The ParallaxNet training was performed using a workstation equipped with an NVIDIA Quadro RTX8000 GPU, Intel Xeon W-2155 CPU at 3.30 GHz and using PyTorch version 1.13.1.

Selected reflections corresponding to each of the three phases are shown in Fig. 4 . It can be clearly observed that the XRD-CT reflections reconstructed by FBP from both 180 and 360° scans exhibit significant peak broadening artefacts. Additionally, the diffraction peaks generated by FBP with a 180° scan range exhibit peak shifting artefacts. In contrast, the diffraction peaks generated with 360° scan range using FBP are in good alignment with the ones obtained by the ParallaxNet with 180° scan range. This demonstrates that Parallax effectively reduces various artefacts brought about by parallax and also that it simply requires a 0–180° scan range to reconstruct parallax artefact-free data.

The Rietveld analysis of the reconstructed volumes further demonstrates the efficacy of our method. From the lattice parameter maps shown in Fig. 5a , it can be clearly seen that the lattice parameter maps from the ParallaxNet with 180° scans have almost identical values with those from the maps of the FBP derived from the 360° scan range. FBP images reconstructed using a 360° scan range should not exhibit any peak shifting, even when significant parallax artefacts are present; to clarify, the centroid position of the peaks will be in the correct position as if it were a dataset without parallax. As such, the lattice parameter maps of the ParallaxNet can be considered close to the ground truth. It is also worth mentioning that the ParallaxNet maps exhibit less noise compared to those from the FBP. The distribution of the lattice parameters is shown in histograms in Supplementary Fig. 8 .

a Lattice parameter maps associated with the four components shown in Fig. 4a . b Top row: Scale factor maps (normalised) associated with four components. Button row: Crystallite size maps associated with four components.

Figure 5b presents the scale factor and crystallite size maps obtained from the Rietveld analysis. A key observation is that the parallax artefact significantly affects the crystallite sizes obtained by conventional approaches. Specifically, on both 180 and 360° XRD-CT scans, it leads to broadened diffraction peaks and reduced crystallite values when using the FBP reconstruction algorithm. The maps suggest that ParallaxNet has, to a certain extent, solved the peak broadening artefact instigated by parallax. This correction is particularly pronounced for the two MgO components, where their crystallite sizes offer mutual validation. Based on these observations, we can deduce that ParallaxNet can correctly solve the parallax artefact on the real phantom experimental dataset. The R wp map for all Rietveld analyses can be further observed in Supplementary Fig. 9 . In Supplementary Figs. 10 and 11 and Supplementary Table 7 we provide comparisons between the lattice parameters and crystallite sizes for all three phases obtained using region-of-interest parallax-free diffraction patterns from the sinogram data and ParallaxNet reconstructed diffraction patterns.

NMC532 cylindrical Li-ion battery

In addition to the phantom dataset presented in the previous section, the efficacy of the method was evaluated with a second experimental dataset. Specifically, a dataset acquired from a commercially available and industrially relevant 10440 NMC532 Li-ion battery was used 44 . This dataset consisted of 521 translation steps, 1000 projections and 1800 channels of sinograms. To train this big dataset, we divided it into batches of 55 channels, and each batch took ca. 8 h to process using 5000 epochs. To address this large dataset, we first selected an XRD diffraction peak from the Cu phase ((111) reflection) and reconstructed only the images without parallax corresponding to this peak within the 55-channel range. Then, we performed Rietveld analysis on this 521 × 521 × 55 XRD-CT dataset to get the chemical information of the Cu phase presented in this Li-ion battery dataset. The original dataset was performed using a 0–360° scan range, but for testing the ParallaxNet, we only used the part of the data corresponding to the 0–180° scan range so for each batch, the size of the reference sinogram was 521 × 500 × 55. For comparison, the XRD-CT data were also reconstructed using the FBP algorithm using both the 180 and 360° ranges and were analysed using the Rietveld method i.e. on the 55 selected channels of the Cu XRD peak (hkl reflection (111)). The ParallaxNet also utilised the FBP with 180° projections to pre-train the generator for faster convergence.

Figure 6a displays the average image from the selected 55 channels of the Cu XRD (111) peak, highlighting a region of interest. Supplementary Fig. 12 illustrates two more regions. Both Figs. 6b and S 12 depict the average XRD peaks from the marked regions. As illustrated in these figures, ParallaxNet can accurately reconstruct the Cu peak, producing a significantly sharper peak. The peak positions align with the XRD-CT data reconstructed by the FBP from the full 360° scan. This result shows that the ParallaxNet correctly removed both the peak shifting and broadening artefacts caused by the Parallax on this Cu XRD peak. It is important to note here that the centres of the peaks obtained from the FBP 360° scan align with those of the ParallaxNet-reconstructed peaks. However, it becomes evident that the peak shape cannot be effectively described using a single peak shape model. This observation is distinctly apparent across all peaks illustrated in Supplementary Fig. 12 , with a particularly noticeable manifestation in the Cu diffraction peak from Fig. 6b .

a The mean image of the Cu phase of the Li-ion battery dataset with a marked region of interest. b Selected peak from average diffraction pattern from the region of interest. c The lattice parameter a maps obtained by the Rietveld method for a Li-ion battery dataset. d The distribution of lattice parameters for the maps is shown in ( a ).

This observation bears significance, as attempting to fit these peaks using a single model, such as Gaussian or pseudo-Voigt models—commonly employed in XRD data analysis—can potentially yield inaccurate data interpretations. Such an approach might result in artificial shifts of the peaks, given that the employed model does not adequately capture the intricacies of the data’s true behaviour. This is especially crucial when high precision is required for the calculated lattice parameter values, e.g . in the order of <10 −3  Å such as when attempting to capture shifts in the Cu peak introduced by temperature gradients in these battery systems 57 .

Figure 6c displays the lattice parameter maps obtained through Rietveld analysis. As seen in the FBP with the 180° scan range lattice parameter map, the parallax artefact results in unevenly distributed lattice parameter values (as determined by the Rietveld analysis) across different positions of the same material (Cu phase). However, both the FBP with the 360° scan range and the ParallaxNet results with the 180° scan range yield lattice parameter maps that are evenly distributed across all positions. The histogram depicting the distribution of lattice parameters for the three maps is shown in Fig. 6d . The mean values of the lattice parameters inside the signal area for the FBP with 360° scan range, the FBP with 180° scan range, and the ParallaxNet result are 3.6098, 3.6088, and 3.6103 Å, respectively. The scale factor maps and the R wp from the Rietveld analysis maps can be found in Supplementary Figs. 13 and 14 .

It was, therefore, demonstrated that ParallaxNet can accurately reconstruct XRD-CT images/diffraction patterns of this experimental Li-ion dataset and that it is possible to extract meaningful chemical information from just a single peak of the XRD pattern. Subsequently, a broader range of diffraction channels was chosen, encompassing 555 out of the 1800 channels from the original dataset. These channels span a native 2θ value range from 1.203 to 4.877°. We divided these 555 channels into 11 segments, each containing 55 channels, consistent with the phantom dataset approach. To mitigate edge effects between the reconstructed images of each batch, we incorporated a 5-pixel overlap on either side of each segment. We then averaged the overlapping sections to produce the final XRD-CT image volume with dimensions of 521 × 521 × 555. To reconstruct this expanded dataset, ParallaxNet required 90 h of training time, which includes both initialisation and pre-training with FBP. Since each batch is independent, we utilised three NVIDIA Quadro RTX8000 GPUs to process these 11 batches in parallel using PyTorch. In the end, it took ca. 33 real-world h to complete this dataset. It is worth noting that this represents the most extreme scenario encountered in real-world experimental datasets, and the DLSR method cannot handle a dataset of this magnitude.

Figure 7a shows the NMC532 phase of the reconstructed dataset highlighting a region of interest. Supplementary Fig. 15 illustrates two more regions. Both Figs. 7b and S 15 depict the average XRD peaks from the marked regions. Other NMC532 peaks of these three regions are also shown in Supplementary Fig. 16 . These figures confirm that ParallaxNet can accurately reconstruct the same peak positions as those derived from the FBP reconstructed with the 360° scan range. Furthermore, ParallaxNet effectively addresses the issue of peak broadening artefacts, producing peaks that are sharper and narrower compared to those in the FBP images.

a The mean image of the NMC532 phase of the Li-ion battery dataset with a marked region of interest. b Selected peak from average diffraction pattern in the region of interest. c Top row: crystallite size maps of the NMC532 phase. Mid row: lattice parameter a maps of the NMC532 phase. Bottom row: lattice parameter c maps. All maps are obtained by Rietveld analysis on the Li-ion battery XRD-CT dataset.

Maps obtained from the Rietveld analysis of the NMC532 phases are shown in Fig. 7c and Supplementary Fig. 17 . The crystallite sizes obtained with the three different methods on the top row of Fig. 7c indicate the crystallite sizes calculated from the ParallaxNet reconstructed volume are larger than both FBP methods with 180° and 360° scans, respectively. The average crystallite sizes raised from ca. 91 nm (for FBP with 180° scans) and 92 nm (for FBP with 360° scans) to 137 nm (for the ParallaxNet), which also supports the conclusion we drew from visual inspection: the diffraction peaks are sharper and narrower than those produced by conventional methods. The R wp maps of the Rietveld analysis are shown in Supplementary Fig. 18 . The lattice parameter maps of the NMC532 phase also indicate that the ParallaxNet can correctly reconstruct the evenly distributed lattice parameter maps which align with the FBP with 360° scan. The peak shifting artefact in the images reconstructed by the FBP with a 180° scan range has been effectively eliminated by ParallaxNet. The distribution of the lattice parameters of the NMC532 is presented in Supplementary Fig. 19 .

In this paper, we introduced an XRD-CT reconstruction approach, ParallaxNet, designed to reconstruct images from XRD-CT data containing parallax artefacts. The ParallaxNet strategy employs a 3D self-supervised neural network generator framework, SD2Vol, together with a customised forward projector to produce parallax artefact-free images/diffraction patterns. This is achieved through an iterative approach by comparing the difference between the generated sinogram volume and the input reference sinogram volume. We evaluated ParallaxNet’s performance using three datasets: a simulated XRD-CT dataset, an experimental XRD-CT dataset acquired using a phantom object and an experimental XRD-CT dataset recorded on an NMC532 cylindrical Li-ion battery.

For all three datasets, this new approach accurately reconstructed the peak positions using only a 0°–180° angular range, eliminating the need for a 0°–360° scan which halves the required acquisition time (i.e. half the number of projections). Furthermore, the reconstructed peaks were sharper and narrower than those produced by traditional FBP methods, both with 180° and 360° scans. It should also be noted that in this work it was also shown that simply using a 360° scan approach as a means to remove parallax artefacts is insufficient and should be avoided as it leads to peaks with shapes that cannot be modelled with a single profile (e.g. Gaussian peak). This was clearly demonstrated with the experimental XRD-CT presented in this work, an example being the Cu component in the cylindrical Li-ion battery. ParallaxNet overcomes the peak shape problems associated with the 0°–360 ° scan approach and also presents distinct advantages over the previously developed DLSR methodology, which is to the best of our knowledge the only alternative solution to removing parallax artefacts in XRD-CT data, addressing several inherent limitations:

Firstly, ParallaxNet operates without requiring a priori knowledge about the chemical composition of the sample being measured.

DLSR also requires the identification of all phases and the construction of a robust physical model; this can potentially lead to some minor components being overlooked during the inspection of diffraction patterns in the sinogram data.

Furthermore, ParallaxNet is more scalable. In this work, we applied the ParallaxNet on the full size experimental phantom XRD-CT datasets. However, the DLSR can only be applied on the scaled-down version of the same dataset(s) as shown in the DLSR paper (e.g. with 121 × 121 image sizes for the Li-ion battery). Moreover, ParallaxNet does not need as many RAM requirements as the DLSR approach, especially when DLSR is used in conjunction with TOPAS version 7.

ParallaxNet does not require any data preprocessing. ParallaxNet can be applied to the raw sinograms, but DLSR needs the manually created masks for each phase and background subtraction on the sinograms in order to use a simple background model.

We have demonstrated that the conventional method of employing a 0°–360° scan with FBP to eliminate parallax artefacts and obtain precise lattice parameter values can be precarious. For instance, the FBP reconstruction of the Cu peaks in the 0°–360° scan revealed peaks that cannot be modelled using a single peak shape model (such as Gaussian or pseudo-Voigt). This could potentially result in wrong lattice parameter values and misinterpretation of the data, especially if lattice parameter values with high precision are required to be extracted from the data.

At this stage, while ParallaxNet presents a promising approach to XRD-CT image reconstruction with parallax artefact, it is not without its limitations. A significant constraint is the extended computational time required for large datasets. For instance, the Li-ion dataset, with sinogram dimensions of 521 × 500 × 540, demanded a staggering 90 h of computational time. Currently, due to GPU memory constraints, there’s a necessity to divide datasets into smaller batches for processing. On the other hand, it should be noted that currently, this is the only available method to correct for parallax artefacts XRD-CT images that are larger than 256 × 256 pixels. This is important as, nowadays, significantly larger XRD-CT images (512 × 512 or larger) can be acquired relatively easily at synchrotron radiation facilities, especially at fourth-generation synchrotrons such as the ESRF. We hope our developed approach will encourage the synchrotron community to see the potential of machine learning approaches also as a means to yield higher quality experimental data in terms of artefact removal, not just for optimising data acquisition, and invest more resources in the field.

Moving forward, there is potential to explore a more streamlined generator, which could significantly minimise the computational resources needed and address some of these challenges. Last but not least, it should be noted that the developed method can be applied to other X-ray scattering-based computed tomography data suffering from parallax artefacts, such as pair distribution computed tomography.

The powder samples measured in this work were SiC (nanopowder, <100 nm particle size, 594911-100G, Sigma-Aldrich), TiO 2 Rutile (204757-25G, Sigma-Aldrich) and MgO (307742-500 G, Sigma-Aldrich). The three powder samples were mounted into separate glass pipettes with an outer diameter of ca. 7.5 mm supported by quartz wool from both ends. Two pipettes were prepared using the same MgO powder sample. The four glass pipettes containing the powder samples were mounted onto a 3D-printed sample holder designed for the parallax experiment. Photographs of the experimental setup can be found in our previous work 51 .

XRD-CT measurements

XRD-CT measurements of the phantom sample were performed at beamline station P07 (EH2) at PETRA III, DESY, using a 103.5 keV ( λ  = 0.11979 Å) monochromatic X-ray beam focused to a spot size of 20 × 3 μm ( H  ×  V ). 2D powder diffraction patterns were collected using a Pilatus3 X CdTe 2 M hybrid photon counting area detector. The 3D-printed sample holder was mounted directly on the rotation stage. The rotation stage was mounted perpendicularly to a hexapod; the hexapod was used to translate the sample across the beam. The XRD-CT scans were measured by performing a series of zigzag line scans in the z (vertical) direction using the hexapod and rotation steps. Two XRD-CT scans were performed, in both cases the number of translation steps were 300 with a 80 μm step size and a 10 ms exposure time per point. The first XRD-CT scan was performed over a 0°–180° range while the second over a 0°–360° range, both using 300 angular steps. The second sample was a pristine (as-received) 10440 Li-ion NMC532 Trustfire cylindrical battery 44 , and it was scanned using the same beamline and experimental setup using a 73.89 keV (λ = 0.16779 Å) monochromatic X-ray beam focused to the same spot size of 20 × 3 μm. An XRD-CT dataset was acquired using 521 translation steps with a 20 μm step size and a 10 ms exposure time per point. The XRD-CT scan was performed over a 0°–360° range using 1000 angles in total.

XRD-CT data analysis

The detector calibration was performed using a CeO 2 standard. Every 2D diffraction image was calibrated and azimuthally integrated to a 1D powder diffraction pattern with a 10% trimmed mean filter using the pyFAI software package, nDTomo software and in-house developed scripts 58 , 59 , 60 . The integrated diffraction patterns were reshaped into sinograms and centred; the air scatter signal was subtracted from the data. For the conventional data analysis approach, the XRD-CT images (i.e. reconstructed data volume) were reconstructed using the FBP algorithm. A pseudo-voigt peak shape function was used for the refinements after the analysis of the CeO 2 pattern. Rietveld analysis was performed on the reconstructed diffraction patterns with the TOPAS software version 7 52 on a voxel-by-voxel basis. Rietveld analysis was first performed using the summed diffraction pattern of each XRD-CT dataset (i.e. to provide a good starting model) before running the voxel-by-voxel Rietveld analysis to provide the spatially-resolved physico-chemical information. The parameters refined were the scale factor, lattice parameter and crystallite size for each phase. A 2nd-order Chebyshev polynomial was used to model the background as it was fairly flat in all reconstructed patterns. A workstation with an Intel Xeon W-2155 CPU, an NVidia Quadro RTX8000 GPU, and 128 g GB of RAM was used to perform the ParallaxNet and full profile analysis on all datasets presented in the paper.

Data availability

The integrated XRD-CT data presented in this work, both simulated and experimental, have been made publicly available through an open-access repository and can be found here: https://zenodo.org/record/8344637 .

Code availability

The code developed in this work is available from the authors upon reasonable request.

Esteva, A. et al. Deep learning-enabled medical computer vision. Npj Digit. Med. 4 , 1–9 (2021).

Article   Google Scholar  

LeCun, Y., Bengio, Y. & Hinton, G. Deep learning. Nature 521 , 436–444 (2015).

Article   CAS   PubMed   Google Scholar  

Jiao, L. & Zhao, J. A survey on the new generation of deep learning in image processing. IEEE Access 7 , 172231–172263 (2019).

Voulodimos, A., Doulamis, N., Doulamis, A. & Protopapadakis, E. Deep learning for computer vision: a brief review. Comput. Intell. Neurosci. 2018 , 7068349 (2018).

Article   PubMed   PubMed Central   Google Scholar  

Cambria, E. & White, B. Jumping NLP curves: a review of natural language processing research. IEEE Comput. Intell. Mag. 9 , 48–57 (2014).

Olsson, F. A literature survey of active machine learning in the context of natural language processing. SICS Technical Report T2009:06, SSN: 1100–3154 (2009).

Ahishakiye, E., Bastiaan Van Gijzen, M., Tumwiine, J., Wario, R. & Obungoloch, J. A survey on deep learning in medical image reconstruction. Intell. Med. 01 , 118–127 (2021).

Wang, G., Ye, J. C. & De Man, B. Deep learning for tomographic image reconstruction. Nat. Mach. Intell. 2 , 737–748 (2020).

Zhang, M., Gu, S. & Shi, Y. The use of deep learning methods in low-dose computed tomography image reconstruction: a systematic review. Complex Intell. Syst. 8 , 5545–5561 (2022).

Bracewell, R. N. & Riddle, A. C. Inversion of fan-beam scans in radio astronomy. Astrophys. J. 150 , 427 (1967).

Beister, M., Kolditz, D. & Kalender, W. A. Iterative reconstruction methods in X-ray CT. Phys. Med. 28 , 94–108 (2012).

Article   PubMed   Google Scholar  

Pan, X., Sidky, E. Y. & Vannier, M. Why do commercial CT scanners still employ traditional, filtered back-projection for image reconstruction? Inverse Probl. 25 , 123009 (2009).

Chetih, N. & Messali, Z. Tomographic image reconstruction using filtered back projection (FBP) and algebraic reconstruction technique (ART). in 2015 3rd International Conference on Control, Engineering & Information Technology (CEIT) 1–6 (IEEE, 2015).

Solomon, J., Lyu, P., Marin, D. & Samei, E. Noise and spatial resolution properties of a commercially available deep learning-based CT reconstruction algorithm. Med. Phys. 47 , 3961–3971 (2020).

Higaki, T. et al. Deep learning reconstruction at CT: phantom study of the image characteristics. Acad. Radiol. 27 , 82–87 (2020).

Yaqub, M. et al. Deep learning-based image reconstruction for different medical imaging modalities. Comput. Math. Methods Med. 2022 , e8750648 (2022).

Zhu, B., Liu, J. Z., Cauley, S. F., Rosen, B. R. & Rosen, M. S. Image reconstruction by domain-transform manifold learning. Nature 555 , 487–492 (2018).

Rodriguez, A. F., Blass, W. E., Missimer, J. H. & Leenders, K. L. Artificial neural network Radon inversion for image reconstruction. Med. Phys. 28 , 508–514 (2001).

Paschalis, P. et al. Tomographic image reconstruction using artificial neural networks. Nucl. Instrum. Methods Phys. Res. 527 , 211–215 (2004).

Article   CAS   Google Scholar  

Argyrou, M., Maintas, D., Tsoumpas, C. & Stiliaris, E. Tomographic image reconstruction based on artificial neural network (ANN) techniques. in 2012 IEEE Nuclear Science Symposium and Medical Imaging Conference Record (NSS/MIC) 3324–3327. https://doi.org/10.1109/NSSMIC.2012.6551757 (2012).

Würfl, T., Ghesu, F. C., Christlein, V. & Maier, A. Deep learning computed tomography. in Medical Image Computing and Computer-Assisted Intervention—MICCAI 2016 (eds Ourselin, S., Joskowicz, L., Sabuncu, M. R., Unal, G. & Wells, W.) 432–440 (Springer International Publishing, 2016).

Ge, Y. et al. ADAPTIVE-NET: deep computed tomography reconstruction network with analytical domain transformation knowledge. Quant. Imaging Med. Surg. 10 , 415–427 (2020).

Ma, G., Zhu, Y. & Zhao, X. Learning image from projection: a full-automatic reconstruction (FAR) net for computed tomography. IEEE Access 8 , 219400–219414 (2020).

Yang, X. & Schroer, C. Strategies of deep learning for tomographic reconstruction. in 2021 IEEE International Conference on Image Processing (ICIP) 3473–3476. https://doi.org/10.1109/ICIP42928.2021.9506395 (2021).

Harding, G., Kosanetzky, J. & Neitzel, U. X-ray diffraction computed tomography. Med. Phys. 14 , 515–525 (1987).

Kleuker, U., Suortti, P., Weyrich, W. & Spanne, P. Feasibility study of X-ray diffraction computed tomography for medical imaging. Phys. Med. Biol. 43 , 2911 (1998).

Beale, A. M., Jacques, S. D., Gibson, E. K. & Di Michiel, M. Progress towards five dimensional diffraction imaging of functional materials under process conditions. Coord. Chem. Rev. 277 , 208–223 (2014).

Omori, N. E., Bobitan, A. D., Vamvakeros, A., Beale, A. M. & Jacques, S. D. M. Recent developments in X-ray diffraction/scattering computed tomography for materials science. Philos. Trans. R. Soc. Math. Phys. Eng. Sci. 381 , 20220350 (2023).

CAS   Google Scholar  

Mürer, F. K. et al. 3D maps of mineral composition and hydroxyapatite orientation in fossil bone samples obtained by X-ray diffraction computed tomography. Sci. Rep. 8 , 10052 (2018).

Vamvakeros, A. et al. Real-time multi-length scale chemical tomography of fixed bed reactors during the oxidative coupling of methane reaction. J. Catal. 386 , 39–52 (2020).

Middelkoop, V. et al. 3D printed Ni/Al2O3 based catalysts for CO2 methanation-a comparative and operando XRD-CT study. J. CO2 Util. 33 , 478–487 (2019).

Matras, D. et al. Operando and postreaction diffraction imaging of the La–Sr/CaO catalyst in the oxidative coupling of methane reaction. J. Phys. Chem. C 123 , 1751–1760 (2019).

Wragg, D. S. et al. Mapping the coke formation within a zeolite catalyst extrudate in space and time by operando computed X-ray diffraction tomography. J. Catal. 401 , 1–6 (2021).

Matras, D. et al. Effect of thermal treatment on the stability of Na-Mn-W/SiO 2 Catalyst for the oxidative coupling of methane. Faraday Discuss . 229 , 176–196 (2020).

Matras, D. et al. In situ X-ray diffraction computed tomography studies examining the thermal and chemical stabilities of working Ba 0.5 Sr 0.5 Co 0.8 Fe 0.2 O 3- δ membranes during oxidative coupling of methane. Phys. Chem. Chem. Phys. 22 , 18964–18975 (2020).

Vamvakeros, A. et al. Real-time tomographic diffraction imaging of catalytic membrane reactors for the oxidative coupling of methane. Catal. Today 364 , 242–255 (2020).

Martens, I. et al. Imaging heterogeneous electrocatalyst stability and decoupling degradation mechanisms in operating hydrogen fuel cells. ACS Energy Lett. 6 , 2742–2749 (2021).

Li, T. et al. Design of next-generation ceramic fuel cells and real-time characterization with synchrotron X-ray diffraction computed tomography. Nat. Commun. 10 , 1497 (2019).

Sottmann, J. et al. Chemical structures of specific sodium ion battery components determined by operando pair distribution function and X-ray diffraction computed tomography. Angew. Chem. Int. Ed. 56 , 11385–11389 (2017).

Tonin, G. et al. Operando investigation of the lithium/sulfur battery system by coupled X-ray absorption tomography and X-ray diffraction computed tomography. J. Power Sources 468 , 228287 (2020).

Liu, H. et al. Quantifying reaction and rate heterogeneity in battery electrodes in 3D through operando X-ray diffraction computed tomography. ACS Appl. Mater. Interfaces 11 , 18386–18394 (2019).

Finegan, D. P. et al. Spatially resolving lithiation in silicon–graphite composite electrodes via in situ high-energy x-ray diffraction computed tomography. Nano Lett. 19 , 3811–3820 (2019).

Finegan, D. P. et al. Spatial quantification of dynamic inter and intra particle crystallographic heterogeneities within lithium ion electrodes. Nat. Commun. 11 , 1–11 (2020).

Vamvakeros, A. et al. Cycling rate-induced spatially-resolved heterogeneities in commercial cylindrical Li-ion batteries. Small Methods 5 , 2100512 (2021).

Petz, D. et al. Lithium distribution and transfer in high-power 18650-type Li-ion cells at multiple length scales. Energy Storage Mater. 41 , 546–553 (2021).

Matras, D. et al. Emerging chemical heterogeneities in a commercial 18650 NCA Li-ion battery during early cycling revealed by synchrotron X-ray diffraction tomography. J. Power Sources 539 , 231589 (2022).

Vamvakeros, A. et al. 5D operando tomographic diffraction imaging of a catalyst bed. Nat. Commun. 9 , 1–11 (2018).

Matras, D. et al. Multi-length scale 5D diffraction imaging of Ni–Pd/CeO2–ZrO2/Al2O3 catalyst during partial oxidation of methane. J. Mater. Chem. A 9 , 11331–11346 (2021).

Sottmann, J. et al. 5D total scattering computed tomography reveals the full reaction mechanism of a bismuth vanadate lithium ion battery anode. Phys. Chem. Chem. Phys. 24 , 27075–27085 (2022).

Scarlett, N. V. Y., Rowles, M. R., Wallwork, K. S. & Madsen, I. C. Sample-displacement correction for whole-pattern profile fitting of powder diffraction data collected in capillary geometry. J. Appl. Crystallogr. 44 , 60–64 (2011).

Vamvakeros, A. et al. DLSR: a solution to the parallax artefact in X-ray diffraction computed tomography data. J. Appl. Crystallogr. 53 , 1531–1541 (2020).

Coelho, A. A. TOPAS and TOPAS-Academic: an optimization program integrating computer algebra and crystallographic objects written in C++. J. Appl. Crystallogr. 51 , 210–218 (2018).

Dong, H. et al. A scalable neural network architecture for self-supervised tomographic image reconstruction. Digit. Discov . https://doi.org/10.1039/D2DD00105E (2023).

Dong, J., Fu, J. & He, Z. A deep learning reconstruction framework for X-ray computed tomography with incomplete data. PloS ONE 14 , e0224426 (2019).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Wang, Z., Bovik, A. C., Sheikh, H. R. & Simoncelli, E. P. Image quality assessment: from error visibility to structural similarity. IEEE Trans. Image Process 13 , 600–612 (2004).

van Aarle, W. et al. The ASTRA Toolbox: a platform for advanced algorithm development in electron tomography. Ultramicroscopy 157 , 35–47 (2015).

Heenan, T. M. M. et al. Mapping internal temperatures during high-rate battery applications. Nature 617 , 507–512 (2023).

Vamvakeros, A. et al. Removing multiple outliers and single-crystal artefacts from X-ray diffraction computed tomography data. J. Appl. Crystallogr. 48 , 1943–1955 (2015).

Vamvakeros, A. & Dong, H. nDTomo Software Suite . https://doi.org/10.5281/zenodo.7139214 (2019).

Ashiotis, G. et al. The fast azimuthal integration Python library: pyFAI. J. Appl. Crystallogr. 48 , 510–519 (2015).

Download references

Acknowledgements

Finden acknowledges funding through the Innovate UK Analysis for Innovators (A4i) programme (Project No. 106003). We acknowledge DESY (Hamburg, Germany), a member of the Helmholtz Association HGF, for the provision of experimental facilities. Parts of this research were carried out at PETRA III. A.V. acknowledges financial support from the Royal Society as a Royal Society Industry Fellow (IF\R2\222059).

Author information

Authors and affiliations.

Department of Chemistry, University College London, 20 Gordon Street, London, WC1H 0AJ, UK

H. Dong, K. T. Butler & A. M. Beale

Finden Limited, Merchant House, 5 East St Helens Street, Abingdon, OX14 5EG, UK

H. Dong, S. D. M. Jacques, A. M. Beale & A. Vamvakeros

SciML, Scientific Computer Division, Rutherford Appleton Laboratory, Harwell, OX11 0QX, UK

K. T. Butler

Deutsches Elektronen-Synchrotron DESY, Notkestraße 85, 22607, Hamburg, Germany

O. Gutowski, A.-C. Dippel & M. von Zimmerman

Research Complex at Harwell, Rutherford Appleton Laboratory, Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0FA, UK

A. M. Beale

Dyson School of Design Engineering, Imperial College London, London, SW7 2DB, UK

A. Vamvakeros

You can also search for this author in PubMed   Google Scholar

Contributions

H.D. and A.V. developed ParallaxNet with contributions and discussions with K.T.B., S.D.M.J. and A.M.B. O.G., A.C.D. and M.Z. were responsible for P07 instrumentation and setup at the PETRA III, DESY. The XRD-CT data were analysed by H.D. and A.V. H.D. and A.V. are responsible for writing the manuscript with feedback given by all contributors.

Corresponding author

Correspondence to A. Vamvakeros .

Ethics declarations

Competing interests.

Keith Butler is Deputy Editor at npj Computational Materials.

Inclusion & ethics

All research scientists who have contributed to this work are co-authors of the paper.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supporting information, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Dong, H., Jacques, S.D.M., Butler, K.T. et al. Obtaining parallax-free X-ray powder diffraction computed tomography data with a self-supervised neural network. npj Comput Mater 10 , 201 (2024). https://doi.org/10.1038/s41524-024-01389-1

Download citation

Received : 02 October 2023

Accepted : 20 August 2024

Published : 02 September 2024

DOI : https://doi.org/10.1038/s41524-024-01389-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.