Search

While multiple factors impact disease, artificial intelligence (AI) studies in medicine often use small, non-diverse patient cohorts due to data sharing and privacy issues. Federated learning (FL) has emerged as a solution, enabling training across hospitals without direct data sharing. Here, we present FL-PedBrain, an FL platform for pediatric posterior fossa brain tumors, and evaluate its performance on a diverse, realistic, multi-center cohort. Pediatric brain tumors were targeted due to the scarcity of such datasets, even in tertiary care hospitals. Our platform orchestrates federated training for joint tumor classification and segmentation across 19 international sites. FL-PedBrain exhibits less than a 1.5% decrease in classification and a 3% reduction in segmentation performance compared to centralized data training. FL boosts segmentation performance by 20 to 30% on three external, out-of-network sites. Finally, we explore the sources of data heterogeneity and examine FL robustness in real-world scenarios with data imbalances., (© 2024. The Author(s).)

Background & Aims: Perianal fistulizing Crohn's disease (PFCD)-associated anorectal and fistula cancers are rare but often devastating diagnoses. However, given the low incidence and consequent lack of data and clinical trials in the field, there is little to no guidance on screening and management of these cancers. To inform clinical practice, we developed consensus guidelines on PFCD-associated anorectal and fistula cancers by multidisciplinary experts from the international TOpClass consortium., Methods: We conducted a systematic review by standard methodology, using the Newcastle-Ottawa Scale quality assessment tool. We subsequently developed consensus statements using a Delphi consensus approach., Results: Of 561 articles identified, 110 were eligible, and 76 articles were included. The overall quality of evidence was low. The TOpClass consortium reached consensus on 6 structured statements addressing screening, risk assessment, and management of PFCD-associated anorectal and fistula cancers. Patients with long-standing (>10 years) PFCD should be considered at small but increased risk of developing perianal cancer, including squamous cell carcinoma of the anus and anorectal carcinoma. Risk factors for squamous cell carcinoma of the anus, notably human papilloma virus, should be considered. New, refractory, or progressive perianal symptoms should prompt evaluation for fistula cancer. There was no consensus on timing or frequency of screening in patients with asymptomatic perianal fistula. Multiple modalities may be required for diagnosis, including an examination under anesthesia with biopsy. Multidisciplinary team efforts were deemed central to the management of fistula cancers., Conclusions: Inflammatory bowel disease clinicians should be aware of the risk of PFCD-associated anorectal and fistula cancers in all patients with PFCD. The TOpClass consortium consensus statements outlined herein offer guidance in managing this challenging scenario., (Copyright © 2024 The Author(s). Published by Elsevier Inc. All rights reserved.)

This dataset is composed of cervical spine CT images with annotations related to fractures; it is available at https://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/., Competing Interests: Disclosures of conflicts of interest: H.M.L. No relevant relationships. E.C. No relevant relationships. T.R. No relevant relationships. F.C.K. Consultant for MD.ai and GE HealthCare; member of the Radiological Society of North America (RSNA) Machine Learning Steering Committee member and the Society for Imaging Informatics in Medicine Machine Learning Education Subcommittee (both unpaid). L.M.P. Associate editor for Radiology: Artificial Intelligence; patents planned, issued, or pending: US-20220051060-A1, “Methods for creating privacy-protecting synthetic data leveraging a constrained generative ensemble model,” and US-20220051402-A1, “Systems for automated lesion detection and related methods.” J.T. Provided expert witness deposition for Phillips, Spallas, & Angstadt in October 2022, unrelated to this article. R.L.B. Support from RSNA to author. E.G. No relevant relationships. K.W.Y. No relevant relationships. M.H. No relevant relationships. A.L.S. Author’s lab receives funding from the National Institutes of Health (NIH); member of the advisory board for the National Cancer Institute Imaging Data Commons. J.S. No relevant relationships. D.B. No relevant relationships. S.A. No relevant relationships. A.P.L. No relevant relationships. M.I.G.A. No relevant relationships. J.O.J. No relevant relationships. J.J.P. Author’s lab receives funding from the NIH, which pays this author’s salary. M.L. No relevant relationships. H.D. No relevant relationships. E.A. No relevant relationships. A.Y. No relevant relationships. Y.M. No relevant relationships. J.K.C. Grants or contracts from GE HealthCare and Genentech; technology licensed to Boston AI; consulting fees from Siloam Vision; deputy editor of Radiology: Artificial Intelligence. A.E.F. Standing director, liaison for information technology, of RSNA board of directors; member of RSNA News editorial board., (© 2023 by the Radiological Society of North America, Inc.)

Introduction: Radiology education is essential for nonradiologist specialists and practitioners as well as for radiologists. We conducted a survey to gather the opinions of Turkish medical students from first to sixth grade regarding their radiology curricula, radiology education content, and perceptions of various imaging modalities and to assess the amount, adequacy, and homogeneity of radiology education in various schools., Methods: Turkish medical students were reached by student ambassadors from 10 different schools of medicine via social media and email. They were provided with a 20-question survey-via the SurveyMonkey platform-related to their radiology curriculum and their perceptions of the radiology education at their schools and of different imaging modalities. Subjective parameters were scaled by a 4-point Likert scale and the results are reported by percentages of students., Results: A total of 988 medical students (F/M: 61%/39%) from 41 different medical schools participated in this survey. Of those, 57% were preclinical students (≤ third year of medical school), while 43% were clinical students (> third year). More than half of the students (51%) stated that the amount of radiology education included in their curriculum was too little , while 44% of them stated it was just right and only 5% stated it was too much . Only 31% of the participants stated that they were able to review radiology images on their own. When asked about their level of confidence in identifying the position of lines and tubes, pneumonia, pneumothorax, and pleural effusion on chest radiographs, 41%, 39%, 41%, and 41% of the participants, respectively, stated that they were not confident. Thirty-five percent of the participants had not received any training in comparing normal to abnormal imaging of bone fractures, pneumonia, pleural effusion, subdural hemorrhage, or pneumothorax. The majority of the Turkish medical students in this survey had never heard (57%) nor used (64%) the American College of Radiology Appropriateness Criteria., Conclusions: The radiology curriculum in Turkey differs among various schools and most students stated that preclinical radiology course content was inadequate. Further studies and improvements must be conducted to provide high-quality, equitable radiology education that begins during preclinical training with respect to the students' opinions., Competing Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article., (© The Author(s) 2023.)

Background: Radiology education in Turkey is mainly taught during clinical years of medical school and often lacks main principles. Exposure to the fundamentals of radiology at an early stage of medical education may drastically help students generate a better understanding of radiology and expand their interest in the specialty. With the Principles of Radiology Course that we provided, pre- and post-session tests, and assessment survey at the end of the course, we aimed to evaluate the effectiveness of such an online course among Turkish medical students., Methods: A total of nine online sessions on imaging modalities principles was developed by radiology professors. Each session was given through Zoom by radiologists from different U.S.-institutions to Turkish medical students from state (n = 33) and private (n = 8) universities. Pretests and post-tests were given to participants via Qualtrics before and after each session, respectively. Paired two-sample t-tests were conducted to detect the variance and p=-.05 was used as the significance level. An evaluation survey was distributed at the end of the course to collect their feedback through SurveyMonkey., Results: A total of 1,438 predominantly Turkish (99.32%) medical students engaged with this course. An average of 506 students completed both pre-test and post-test. There was a statistically significant (p < .001) increase in the scores in post-test (mean[range]:7.58[5.21-8.53]) relative to pre-test (mean[range]:5.10[3.52-8.53]). Four hundred and thirty-nine participants (F/M:63.33%/35.54%) completed the end-of-course survey. A total of 71% and 69.70% of the participants strongly agreed that the course would be useful in their clinical practice and had increased their understanding of radiology. They also reported that their level of confidence in the subjects had increased 68% and reached a weighted average of 3.09/4. The survey revealed that 396 (90.21%) of the participants strongly or somewhat agree that introductory principles and concepts should be presented in earlier years of medical education. Compared to in-person education, 358 (81.55%) found the course extremely or very convenient., Conclusion: Online lecture series consisting of the principles of the radiological imaging modalities can be offered to Turkish medical students to enhance their grasp of the various imaging modalities and their correct clinical application., (© 2022. The Author(s).)

Background: The risk profile for posterior fossa ependymoma (EP) depends on surgical and molecular status [Group A (PFA) versus Group B (PFB)]. While subtotal tumor resection is known to confer worse prognosis, MRI-based EP risk-profiling is unexplored. We aimed to apply machine learning strategies to link MRI-based biomarkers of high-risk EP and also to distinguish PFA from PFB., Methods: We extracted 1800 quantitative features from presurgical T2-weighted (T2-MRI) and gadolinium-enhanced T1-weighted (T1-MRI) imaging of 157 EP patients. We implemented nested cross-validation to identify features for risk score calculations and apply a Cox model for survival analysis. We conducted additional feature selection for PFA versus PFB and examined performance across three candidate classifiers., Results: For all EP patients with GTR, we identified four T2-MRI-based features and stratified patients into high- and low-risk groups, with 5-year overall survival rates of 62% and 100%, respectively (P < .0001). Among presumed PFA patients with GTR, four T1-MRI and five T2-MRI features predicted divergence of high- and low-risk groups, with 5-year overall survival rates of 62.7% and 96.7%, respectively (P = .002). T1-MRI-based features showed the best performance distinguishing PFA from PFB with an AUC of 0.86., Conclusions: We present machine learning strategies to identify MRI phenotypes that distinguish PFA from PFB, as well as high- and low-risk PFA. We also describe quantitative image predictors of aggressive EP tumors that might assist risk-profiling after surgery. Future studies could examine translating radiomics as an adjunct to EP risk assessment when considering therapy strategies or trial candidacy., (© The Author(s) 2021. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)

BACKGROUND Metabolic syndrome affects 35% of the adult population in developed countries associated with non-alcoholic steatohepatitis, insulin resistance, and cardiovascular events. Fatty infiltration of the pancreas, or pancreatic steatosis, is a risk factor for acute pancreatitis, pancreatic malignancies, and diabetes mellitus, yet its relationship with metabolic syndrome is not well defined. METHODS We performed a single-centered retrospective observational study of 322 healthy subjects (subjects volunteering to be kidney transplant donors, mean age=46.3±13.5, 163 men and 159 women) in the last 2 years (July 2018-February 2020) from our institution. Pancreatic steatosis and hepatosteatosis were confirmed by computed tomography. RESULTS Pancreatic steatosis was present in 26.3% (85/322) of the subjects, and this finding correlated with age, body mass index (BMI), male sex, a family history of diabetes, creatinine, cystatin C, uric acid, low-density lipoprotein (LDL) cholesterol, triglycerides, glycemia, hemoglobin, transverse body diameter, and subcutaneous fat thickness levels by univariable logistic regression. On multiple linear regression only age (95% CI 1.01, 1.06), BMI (95% CI 1.01, 1.19), male sex (95% CI 1.49-5.99), uric acid (95% CI 1.01, 1.76), and subcutaneous fat thickness levels (95% CI 1.21-2.36) remained independently associated with pancreatic steatosis. CONCLUSION Pancreatic steatosis is common and associated with obesity, elevated serum uric acid, subcutaneous fat thickness, and male sex. Future studies are needed to evaluate if there are specific clinical consequences to the presence of pancreatic steatosis., Competing Interests: CONFLICT OF INTEREST The authors declare no conflict of interest related to this work., (© 2021 The Author(s).)

Purpose: We aimed to investigate the value of magnetic resonance imaging (MRI)-based parameters in differentiating between progressive massive fibrosis (PMF) and lung cancer., Methods: This retrospective study included 60 male patients (mean age, 67.0±9.0 years) with a history of more than 10 years working in underground coal mines who underwent 1.5 T MRI of thorax due to a lung nodule/mass suspicious for lung cancer on computed tomography. Thirty patients had PMF, and the remaining ones had lung cancer diagnosed histopathologically. The sequences were as follows: coronal single-shot turbo spin echo (SSH-TSE), axial T1- and T2-weighted spin-echo (SE), balanced turbo field echo, T1-weighted high-resolution isotropic volume excitation, free-breathing and respiratory triggered diffusion-weighted imaging (DWI). The patients' demographics, lesion sizes, and MRI-derived parameters were compared between the patients with PMF and lung cancer., Results: Apparent diffusion coefficient (ADC) values of DWI and respiratory triggered DWI, signal intensities on T1-weighted SE, T2-weighted SE, and SSH-TSE imaging were found to be significantly different between the groups (p < 0.001, for all comparisons). Median ADC values of free-breathing DWI in patients with PMF and cancer were 1.25 (0.93-2.60) and 0.76 (0.53-1.00) (× 10-3 mm2/s), respectively. Most PMF lesions were predominantly iso- or hypointense on T1-weighted SE, T2-weighted SE, and SSH-TSE, while most malignant ones predominantly showed high signal intensity on these sequences., Conclusion: MRI study including SE imaging, specially T1-weighted SE imaging and ADC values of DWI can help to distinguish PMF from lung cancer.

The coronavirus disease 2019 (COVID-19) pandemic is a global health care emergency. Although reverse-transcription polymerase chain reaction testing is the reference standard method to identify patients with COVID-19 infection, chest radiography and CT play a vital role in the detection and management of these patients. Prediction models for COVID-19 imaging are rapidly being developed to support medical decision making. However, inadequate availability of a diverse annotated data set has limited the performance and generalizability of existing models. To address this unmet need, the RSNA and Society of Thoracic Radiology collaborated to develop the RSNA International COVID-19 Open Radiology Database (RICORD). This database is the first multi-institutional, multinational, expert-annotated COVID-19 imaging data set. It is made freely available to the machine learning community as a research and educational resource for COVID-19 chest imaging. Pixel-level volumetric segmentation with clinical annotations was performed by thoracic radiology subspecialists for all COVID-19-positive thoracic CT scans. The labeling schema was coordinated with other international consensus panels and COVID-19 data annotation efforts, the European Society of Medical Imaging Informatics, the American College of Radiology, and the American Association of Physicists in Medicine. Study-level COVID-19 classification labels for chest radiographs were annotated by three radiologists, with majority vote adjudication by board-certified radiologists. RICORD consists of 240 thoracic CT scans and 1000 chest radiographs contributed from four international sites. It is anticipated that RICORD will ideally lead to prediction models that can demonstrate sustained performance across populations and health care systems., (© RSNA, 2021 See also the editorial by Bai and Thomasian in this issue.)

The Coronavirus disease 2019 (COVID-19) presents open questions in how we clinically diagnose and assess disease course. Recently, chest computed tomography (CT) has shown utility for COVID-19 diagnosis. In this study, we developed Deep COVID DeteCT (DCD), a deep learning convolutional neural network (CNN) that uses the entire chest CT volume to automatically predict COVID-19 (COVID+) from non-COVID-19 (COVID-) pneumonia and normal controls. We discuss training strategies and differences in performance across 13 international institutions and 8 countries. The inclusion of non-China sites in training significantly improved classification performance with area under the curve (AUCs) and accuracies above 0.8 on most test sites. Furthermore, using available follow-up scans, we investigate methods to track patient disease course and predict prognosis.

Supplemental material is available for this article., Competing Interests: Disclosures of Conflicts of Interest: E.C. disclosed no relevant relationships. F.C.K. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: consultant for MD.ai; employed by DASA as Head of AI. Other relationships: disclosed no relevant relationships. S.B.H. disclosed no relevant relationships. C.C.W. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution received grant from NIBIB and has MIDRC grant pending; author receives royalties from Elsevier. Other relationships: disclosed no relevant relationships. M.P.L. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: board member of Nines Radiology, SegMed, and BunkerHill; author has stock/stock options in Nines Radiology, SegMed, and BunkerHill. Other relationships: disclosed no relevant relationships. L.M.P. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: editorial board member of Radiology: Artificial Intelligence. Other relationships: disclosed no relevant relationships J.K.C. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution has grants pending with GE and Genentech; author receives support for travel/accommodation/meeting expenses from IBM; deputy editor of Radiology: Artificial Intelligence. Other relationships: disclosed no relevant relationships. R.L.B. Activities related to the present article: author paid consulting fee or honorarium from RSNA. Activities not related to the present article: employed by Stanford University. Other relationships: disclosed no relevant relationships. G.S. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: board member of MD.ai (no compensation); consultant for MD.ai (no compensation); stock/stock options in MD.ai; editorial board member of Radiology: Artificial Intelligence. Other relationships: disclosed no relevant relationships. A.S. Activities related to the present article: employee of MD.ai. Activities not related to the present article: employed by MD.ai. Other relationships: disclosed no relevant relationships. S.S.H. disclosed no relevant relationships. E.A. disclosed no relevant relationships. M.L. disclosed no relevant relationships. P.K. disclosed no relevant relationships. K.A.M. disclosed no relevant relationships. D.C.N.R. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: speaker for Bayer de Mexico (author gave one lecture for the International Day of Radiology and received speaker training from Bayer). Other relationships: disclosed no relevant relationships. J.W.S. disclosed no relevant relationships. P. Germaine disclosed no relevant relationships. E.C.L. disclosed no relevant relationships. T.A. disclosed no relevant relationships. P. Gupta disclosed no relevant relationships. M.J. disclosed no relevant relationships. F.U.K. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: editorial board member of Radiology: Cardiothoracic Imaging; employed at UT Southwestern Medical Center. Other relationships: disclosed no relevant relationships. C.T.L. disclosed no relevant relationships. S.S. disclosed no relevant relationships. J.W.R. disclosed no relevant relationships. C.C.B. disclosed no relevant relationships. J.M. Activities related to the present article: disclosed no relevant relationships. Activities not related to the present article: institution has grant from GE Healthcare; editorial board member of Radiology: Artificial Intelligence. Other relationships: disclosed no relevant relationships., (2021 by the Radiological Society of North America, Inc.)

Objectives: To evaluate the use and benefits of tumor navigation during performing robotic assisted radical prostatectomy (RARP)., Patients and Methods: Borders of the visible tumor(s) was/were and surrounding structures marked on multiparametric prostate magnetic resonance imaging (mpMRI) and 68 Ga-labeled prostate-specific membrane antigen ligand using positron emission computed tomography (Ga68 PSMA-PET/CT). Three dimensional (3D) reconstruction of the images were done that were transferred to virtual reality (VR) headsets and Da Vinci surgical robot via TilePro. Images were used as a guide during RARP procedures in five cases. Indocyanine green (ICG) guided pelvic lymph node dissection (n = 2) and Martini Klinik Neurosafe technique (n = 2) were also applied., Results: Mean patient age was 60.6 ± 3.7 years (range, 56-66). All VR models were finalized with the agreement of radiologist, urologist, nuclear physician, and engineer. Surgeon examined images before the surgery. All VR models were found very useful particularly in pT3 diseases. Pathological stages included pT2N0 (n = 1), pT3aN0 (n = 1), pT3aN1 (n = 2), and pT3bN1 (n = 1). Positive surgical margins (SMs) occurred in two patients with extensive disease (pT3aN1 and pT3bN1) and tumor occupied 30% and 50% of the prostate volumes. Mean estimated blood loss was 150 ± 86.6 cc (range, 100-300). Mean follow-up was 3.4 ± 1.7 months (range, 2-6). No complication occurred during perioperative (0-30 days) and postoperative (30-90 days) periods in any patient., Conclusions: 3D reconstructed VR models by using mpMRI and Ga68 PSMA-PET/CT images can be accurately prepared and effectively applied during RARP that might be a useful tool for tumor navigation. Images show prostate tumors and anatomy and might be a guide for the console surgeon. This is promising new technology that needs further study and validation., (© 2020 The Authors. BJUI Compass published by John Wiley & Sons Ltd on behalf of BJU International Company.)

Rationale and Objective: To evaluate computed tomography (CT) and positron emission tomography-computed tomography (PET-CT) imaging manifestations of lymphomas secondarily involving the adrenal gland., Materials and Methods: Seven patients (Five men, two women; median age [range], 66 years [34-75 years]) with pathologically proven adrenal lymphoma were assessed retrospectively. Clinical findings, prior history of lymphoproliferative malignancy, CT (n = 7) and fludeoxyglucose positron emission tomography-computed tomography ( 18 F-FDG-PET-CT) (n = 6) features were analyzed., Results: Six cases were diffuse large B-cell lymphoma, and one case was peripheral T-cell lymphoma. The longest diameter of the lesions ranged from 3.2 to 6.6 cm (median 4.3 cm). Six lesions were well-defined and one lesion was ill-defined. In five cases, an adreniform shape was preserved. No lesions contained fat, calcification or hemorrhage. Two lesions had necrosis on CT. Median (range) unenhanced CT density of six lesions was 31.8 (29.2-35.2) Hounsfield units. Following administration of IV contrast media (n = 6), three lesions enhanced homogenously whereas three enhanced heterogeneously. The median increase in attenuation was 35.1 Hounsfield units. Two patients had 15-minute delayed CT and they both demonstrated limited wash-out consistent with nonadenoma. Six patients had fludeoxyglucose positron emission tomography-computed tomography( 18- F-FDG-PET-CT) and all lesions were fludeoxyglucose (FDG) avid with a median SUV max of 18.6 (range: 10.3-49.2)., Conclusion: Secondary adrenal lymphomas usually manifest as, large (>3 cm), well-defined, homogenously or slightly heterogeneously enhancing masses on CT with preserved adreniform shape. These lesions tend to show limited wash-out and high fludeoxyglucose (FDG) uptake., (Copyright © 2018 The Association of University Radiologists. All rights reserved.)

Rationale and Objectives: To determine if combination of washout and noncontrast data from delayed adrenal computed tomography (CT) improves diagnostic performance, and demonstration of an optimizing analytical framework., Materials and Methods: This retrospective study consisted of 97 adrenal lesions, in 96 patients, with pathologically proven adrenal lesions (75 benign; 22 malignant), who had undergone noncontrast, portal- and approximate 15-minute delayed-phase CT. Lesion CT attenuations (Hounsfield units [HU]) during each phase, and "absolute" and "relative" percent enhancement washouts (APEW and RPEW) were assessed. The optimum combination of sequential parameters and thresholds was determined by recursive partitioning analysis; resultant diagnostic performance was compared to commonly applied single-parameter criteria for malignancy (noncontrast > 10 HU, APEW < 60%, RPEW < 40%)., Results: The above single-parameter criteria yielded sensitivities, specificities, and accuracies for malignancy of 100.0%, 41.3%, and 54.6%; 97.9%, 61.3%, and 69.1%; and 96.6%, 74.7%, and 78.4%, respectively. Recursive partitioning analysis identified noncontrast ≥24.75 HU, with subsequent APEW ≤63.49%, as the optimum sequential parameter-threshold combination, which yielded increased sensitivity, specificity, and accuracy of 100.0%, 85.3%, and 90.7%, respectively. Discrimination using the combined sequential classifier yielded statistically significant improvements in accuracy when compared to the above conventional single-parameter criteria (all P ≤ .039)., Conclusion: Sequential application of noncontrast and washout criteria from delayed contrast-enhanced adrenal CT can improve diagnostic performance beyond that of commonly applied single-parameter criteria. Validation of the sequential ordering and refinement of the specific threshold values warrant further study., (Copyright © 2018 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.)