British Journal of Radiology最新文献

{"title":"Assessing the Performance of ChatGPT and Bard/Gemini Against Radiologists for PI-RADS Classification Based on Prostate Multiparametric MRI Text Reports.","authors":"Kang-Lung Lee, Dimitri A Kessler, Iztok Caglic, Yi-Hsin Kuo, Nadeem Shaida, Tristan Barrett","doi":"10.1093/bjr/tqae236","DOIUrl":"https://doi.org/10.1093/bjr/tqae236","url":null,"abstract":"<p><strong>Objectives: </strong>Large language models (LLMs) have shown potential for clinical applications. This study assesses their ability to assign PI-RADS categories based on clinical text reports.</p><p><strong>Methods: </strong>One hundred consecutive biopsy-naïve patients' multiparametric prostate MRI reports were independently classified by two uroradiologists, GPT-3.5, GPT-4, Bard, and Gemini. Original report classifications were considered definitive.</p><p><strong>Results: </strong>Out of 100 MRIs, 52 were originally reported as PI-RADS 1-2, 9 PI-RADS 3, 19 PI-RADS 4, and 20 PI-RADS 5. Radiologists demonstrated 95% and 90% accuracy, while GPT-3.5 and Bard both achieved 67%. Accuracy of the updated versions of LLMs increased to 83% (GTP-4) and 79% (Gemini), respectively. In low suspicion studies (PI-RADS 1-2), Bard and Gemini (F1: 0.94, 0.98, respectively) outperformed GPT-3.5 and GTP-4 (F1:0.77, 0.94, respectively), whereas for high probability MRIs (PI-RADS 4-5), GPT-3.5 and GTP-4 (F1: 0.95, 0.98, respectively) outperformed Bard and Gemini (F1: 0.71, 0.87, respectively). Bard assigned a non-existent PI-RADS 6 \"hallucination\" for two patients. Inter-reader agreements (Κ) between the original reports and the senior radiologist, junior radiologist, GPT-3.5, GTP-4, BARD, and Gemini were 0.93, 0.84, 0.65, 0.86, 0.57, and 0.81, respectively.</p><p><strong>Conclusions: </strong>Radiologists demonstrated high accuracy in PI-RADS classification based on text reports, while GPT-3.5 and Bard exhibited poor performance. GTP-4 and Gemini demonstrated improved performance compared to their predecessors.</p><p><strong>Advances in knowledge: </strong>This study highlights the limitations of LLMs in accurately classifying PI-RADS categories from clinical text reports. While the performance of LLMs has improved with newer versions, caution is warranted before integrating such technologies into clinical practice.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142614792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Recurrence-Free Survival Prediction for Anal Squamous Cell Carcinoma After Chemoradiotherapy using Planning CT-based Radiomics Model.","authors":"Shanshan Tang, Kai Wang, David Hein, Gloria Lin, Nina N Sanford, Jing Wang","doi":"10.1093/bjr/tqae235","DOIUrl":"https://doi.org/10.1093/bjr/tqae235","url":null,"abstract":"<p><strong>Objectives: </strong>Approximately 30% of non-metastatic anal squamous cell carcinoma (ASCC) patients will experience recurrence after chemoradiotherapy (CRT), and currently available clinical variables are poor predictors of treatment response. We aimed to develop a model leveraging information extracted from radiation pretreatment planning CT to predict recurrence-free survival (RFS) in ASCC patients after CRT.</p><p><strong>Methods: </strong>Radiomics features were extracted from planning CT images of 96 ASCC patients. Following pre-feature selection, the optimal feature set was selected via step-forward feature selection with a multivariate Cox proportional hazard model. The RFS prediction was generated from a radiomics-clinical combined model based on an optimal feature set with five repeats of nested five-fold cross validation. The risk stratification ability of the proposed model was evaluated with Kaplan-Meier analysis.</p><p><strong>Results: </strong>Shape- and texture-based radiomics features significantly predicted RFS. Compared to a clinical-only model, radiomics-clinical combined model achieves better performance in the testing cohort with higher C-index (0.80 vs 0.73) and AUC (0.84 vs 0.78 for 1-year RFS, 0.84 vs 0.79 for 2-year RFS, and 0.85 vs 0.81 for 3-year RFS), leading to distinctive high- and low-risk of recurrence groups (p < 0.001).</p><p><strong>Conclusions: </strong>A treatment planning CT based radiomics and clinical combined model had improved prognostic performance in predicting RFS for ASCC patients treated with CRT as compared to a model using clinical features only.</p><p><strong>Advances in knowledge: </strong>The use of radiomics from planning CT is promising in assisting in personalized management in ASCC. The study outcomes support the role of planning CT-based radiomics as potential imaging biomarker.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142614818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A framework for developing Generic Implant Safety Procedures (GISPs) for scanning patients with medical implants and devices in MRI.","authors":"Jonathan P Ashmore, Sarah J Prescott, John McLean, Daniel J Wilson, Geoff Charles-Edwards, Peter Wright, David Grainger, Gareth J Barker, Alexandra J Lipton, Rachel Watt, Deepa Gopalan, Mark R Radon","doi":"10.1093/bjr/tqae232","DOIUrl":"https://doi.org/10.1093/bjr/tqae232","url":null,"abstract":"<p><p>UK guidelines for MR safety recommend that MRI departments refer to the implant manufacturer for advice regarding the MRI safety of scanning patients with an implantable medical device prior to scanning [1]. This process of assuring safety can be time consuming, leading to delays and potential cancellations of a patient's MRI. Furthermore, at times the implant cannot be identified, or the implant manufacturers cannot provide up to date MRI safety information. The purpose of generic implant safety procedures (GISPs) is to define a process for managing patients with certain types of implants where the risk from scanning is low. This process incorporates scope for an evidence-based risk-benefit decision to scan some groups of patients under locally-approved conditions, without seeking to identify the exact make and model of the implant and subsequent assurance of MR safety from the implant manufacturer. This publication provides best practice recommendations from a multi-professional working group for the development of these procedures. It is supported by The Institute of Physics and Engineering in Medicine, The Society of Radiographers, The Royal College of Radiologists, The British institute of Radiology, The British Association of MR Radiographers, The International Society of Magnetic Resonance in Medicine British and Irish Chapter and the NHS Scotland MRI Physics Group.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142614263","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A generalized health index: automated thoracic CT-derived biomarkers predict life expectancy.","authors":"Cameron Beeche, Tong Yu, Jing Wang, David Wilson, Pengyu Chen, Emrah Duman, Jiantao Pu","doi":"10.1093/bjr/tqae234","DOIUrl":"https://doi.org/10.1093/bjr/tqae234","url":null,"abstract":"<p><strong>Objective: </strong>To identify image biomarkers associated with overall life expectancy from low-dose computed tomography and integrate them as an index for assessing an individual's health.</p><p><strong>Methods: </strong>Two categories of CT image features, body composition tissues and cardiopulmonary vasculature characteristics, were quantified from LDCT scans in the Pittsburgh Lung Screening Study cohort(n = 3,635). Cox proportional-hazards models identified significant image features which were integrated with subject demographics to predict the subject's overall hazard. Subjects were stratified using composite model predictions and feature-specific risk stratification thresholds. The model's performance was validated extensively, including 5-fold cross-validation on PLuSS baseline, PLuSS follow-up examinations, and the National Lung Screening Trial (NLST).</p><p><strong>Results: </strong>The composite model had significantly improved prognostic ability compared to the baseline model (p < 0.01) with AUCs of 0.774 (95% CI: 0.757-0.792) on PLuSS, 0.723 (95% CI: 0.703-0.744) on PLuSS follow-up, and 0.681 (95% CI: 0.651-0.710) on the NLST cohort. The identified high-risk stratum were several times more likely to die, with mortality rates of 79.34% on PLuSS, 76.47% on PLuSS follow-up, and 46.74% on NLST. Two cardiopulmonary structures (intrapulmonary artery vein ratio, intrapulmonary vein density) and two body composition tissues (SM density, bone density) identified high-risk patients.</p><p><strong>Conclusions: </strong>Body composition and pulmonary vasculatures are predictive of an individual's health risk; their integrations with subject demographics facilitate the assessment of an individual's overall health status or susceptibility to disease.</p><p><strong>Advances in knowledge: </strong>CT-computed body composition and vasculature biomarkers provide improved prognostic value. The integration of CT biomarkers and patient demographic information improves subject risk stratification.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142614279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Refining MRI Protocols for Endometriosis: A Comparative Study of Abbreviated and Full MRI Sequences.","authors":"Arvin Arian, Nafise Karimi, Nasrin Ahmadinejad, Sina Azadnajafabad, Sina Delazar","doi":"10.1093/bjr/tqae230","DOIUrl":"https://doi.org/10.1093/bjr/tqae230","url":null,"abstract":"<p><strong>Objectives: </strong>Endometriosis is a significant cause of chronic abdominal pain and infertility in females, often overlooked due to its resemblance to other abdominopelvic pathologies. This study aims to compare the diagnostic performance and agreement rate between an abbreviated MRI protocol (aMRI) and a full MRI protocol (fMRI) for detecting pelvic endometriosis.</p><p><strong>Methods: </strong>We retrospectively analyzed 446 consecutive MRI exams, including both full (fMRI) and abbreviated (aMRI) protocols, performed for suspected pelvic endometriosis. An expert radiologist assessed the presence of endometriosis at 14 distinct anatomical sites. Each MRI protocol was interpreted in random order, with a minimum two-week interval between sessions to minimize recall bias. Agreement between the protocols was evaluated using kappa statistics.</p><p><strong>Results: </strong>The average age of the patients was 34.13 years. The highest incidences of endometriosis were found in the ovaries (88.8%) and the rectouterine pouch (65%). The MRI protocols demonstrated perfect agreement (kappa coefficient = 1) for the ovaries, bladder, uterus, and cesarean section scar. High agreement was also observed in the rectum and uterine ligaments (kappa coefficients of 0.98 and 0.97). Detection of malignant transformation in existing ovarian endometriomas showed substantial concordance with a kappa coefficient of 0.66.</p><p><strong>Conclusions: </strong>An abbreviated non-contrast MRI protocol exhibits diagnostic accuracy comparable to that of a comprehensive protocol in detecting pelvic endometriosis, with similar confidence and reproducibility.</p><p><strong>Advances in knowledge: </strong>This study demonstrates that an abbreviated MRI protocol is as effective as a full protocol in diagnosing pelvic endometriosis, potentially allowing for quicker, cost-effective imaging without compromising diagnostic accuracy.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142614991","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Application value of strain elastography and shear wave elastography in patients with type 2 diabetic peripheral neuropathy: a prospective observational study.","authors":"Siqi Zheng, Miao Zhu, Gaoxiang Fan, Xueting Yang, Min Bai","doi":"10.1093/bjr/tqae227","DOIUrl":"https://doi.org/10.1093/bjr/tqae227","url":null,"abstract":"<p><strong>Objective: </strong>To evaluate the value of conventional ultrasound (US), strain elastography (SE), and shear wave elastography (SWE) in detecting diabetic peripheral neuropathy (DPN) of the tibial nerve (TN), and to establish a predictive model for the diagnosis of DPN.</p><p><strong>Methods: </strong>32 healthy participants, 34 diabetic patients without DPN, and 36 diabetic patients with DPN were recruited for this study. The TN at the ankle and popliteal fossa were selected for examination. US was used to measure the cross-sectional area (CSA) and perimeter of the TN. Additionally, SE was employed to measure the strain ratio (SR) between the TN and the surrounding adipose tissue, and SWE was used to measure the Shear Wave Velocity (SWV) of the TN.</p><p><strong>Results: </strong>The CSA, perimeter, SR and SWV of the TN at the ankle were significantly higher in the DPN group compared to both the Non-DPN group and control group (P < 0.05). Similarly, the TN at the popliteal fossa showed these differences. At the ankle, the CSA, perimeter, SR, and SWV of the TN in patients without DPN were significantly higher than those in the control group (P < 0.05). At the popliteal fossa, the SR and SWV of the TN in patients without DPN were significantly higher than those in the control group (P < 0.05). However, the CSA and perimeter of the TN in patients without DPN did not show a statistically significant difference compared to the control group. The area under the curve (AUC) for the diagnosis of DPN using SWE is significantly greater than that of SE and US.</p><p><strong>Conclusion: </strong>US, SE, and SWE could be used to diagnose DPN, and they also have good diagnostic value for sub-clinical DPN. Among these methods, SWE has demonstrated superior diagnostic efficacy. Compared to examining the TN in the popliteal fossa, the ankle level offers a better site for examination.</p><p><strong>Advances in knowledge: </strong>For diabetic peripheral neuropathy, US, SE, and SWE are all promising diagnostic methods with high clinical utility.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142589375","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dual Energy Metal Artifact Reduction for Iodine-125 Seed Identification in Postimplant CT after Prostate Brachytherapy.","authors":"Chiaki Suzuki, Kosuke Matsubara, Yuta Ujihara, Kenta Isogai","doi":"10.1093/bjr/tqae225","DOIUrl":"https://doi.org/10.1093/bjr/tqae225","url":null,"abstract":"<p><strong>Objectives: </strong>To assess the metal artifact reductions of dual-energy computed tomography (DECT) high-energy virtual monochromatic images (VMI) combined with the single energy metal artifact reduction (SEMER) (CANON MEDICAL SYSTEMS, Otawara, Japan) processing techniques for iodine (I)-125 seed identification in postimplant computed tomography (CT) after prostate brachytherapy.</p><p><strong>Methods: </strong>Dual-energy acquisition with fast tube voltage switching was performed on a prostate phantom with simulated seeds and six clinical cases treated with I-125 prostate brachytherapy. The images were retrospectively reconstructed at VMI energy levels of 65-200 keV and with and without SEMAR (SEMAR and non-SEMAR images). To estimate seed swelling, the caliber of iodine-125 seed was calculated as the full width at half maximum. The metal artifacts were evaluated using the artifact index (AI). The dose distributions were calculated and were compared among the high-energy VMI (SEMAR and non-SEMAR images) and low-energy VMI (SEMAR images).</p><p><strong>Results: </strong>The blooming artifacts decreased at higher energy levels. In addition, the SEMAR process markedly reduced AI, which helped reduce overestimation of high dose ranges in the treatment planning dose map.</p><p><strong>Conclusion: </strong>The locations and number of iodine-125 seed were clearly identified in the dose distribution map of the treatment planning using 200keV VMI with SEMAR.</p><p><strong>Advance in knowledge: </strong>The high-energy VMI of the dual energy CT in combination with SEMAR is appropriate for the postimplant planning process of I-125 prostate brachytherapy.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142567623","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Patterns of metabolic response in patients receiving commercial CAR T-cells for relapsing/refractory aggressive B cells lymphoma.","authors":"Laetitia Vercellino, Siham Betterki, Estelle Blanc, Eric de Kerviler, Caterina Cristinelli, Pascal Merlet, Catherine Thieblemont, Véronique Meignin, Roberta Di Blasi","doi":"10.1093/bjr/tqae178","DOIUrl":"10.1093/bjr/tqae178","url":null,"abstract":"<p><p>CAR T-cells is an innovative treatment for relapsed/refractory aggressive B cell lymphomas, initially proposed as third-line therapy and beyond, now allowed as soon as second-line treatment for patients with early relapse after first-line treatment. FDG PET/CT remains the modality of choice to evaluate response to this therapeutic strategy, to detect or confirm treatment failure, and allow for salvage therapy if needed. Correct classification of patients regarding response is thus of the utmost importance. In many cases, metabolic response follows classical known patterns, and Deauville score and Lugano criteria yield accurate characterization of patient status. However, given its specific mode of action, it can result in delayed response or atypical patterns of response. We report here a few examples of response from our experience to illustrate the existence of tricky cases. These atypical cases require multidisciplinary management, with clinical, biological, imaging, and pathological work-up.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":"1755-1764"},"PeriodicalIF":1.8,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142139374","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"\"Artificial intelligence Reading Digital Mammogram: Enhancing Detection and Differentiation of Suspicious Microcalcifications\".","authors":"Sahar Mansour, Omnia Mokhtar Nada, Mennat-Allah Samir Mohammed Abd El Galil, Sherif Nasser Taha, Ola Magdy","doi":"10.1093/bjr/tqae220","DOIUrl":"https://doi.org/10.1093/bjr/tqae220","url":null,"abstract":"<p><strong>Purpose: </strong>To investigate the impact of artificial intelligence (AI) on enhancing the sensitivity of digital mammograms in the detection and specification of grouped microcalcifications.</p><p><strong>Methods and materials: </strong>The study is a retrospective analysis of grouped microcalcifications for 447 patients. Grouped microcalcifications detected were correlated with AI, which was applied to the initial mammograms. AI provided a heat map, demarcation, and quantitative evaluation for abnormalities according to the degree of suspicion of malignancy. Histopathology was the standard for confirmation of malignancy.</p><p><strong>Results: </strong>AI showed a high correlation percentage of 67.5% between the red color of the color hue bar and malignant microcalcifications (p value <0.001). The scoring of probable cancer was suggested (ie, more than 50% abnormality scoring) in 39.5% of true cancer lesions. The diagnostic performance of mammography for grouped microcalcifications revealed a sensitivity of 94.7% and a negative predictive value of 82.1%. False negatives were only 12 out of 228 that proved malignant calcifications. The agreement of cancer probability between standard mammograms and examinations read by AI presented a Kappa value of -0.094 and a p value of < 0.001.</p><p><strong>Conclusions: </strong>The used AI system enhanced the sensitivity of mammograms in detecting suspicious microcalcifications, yet an expert human reader is required for proper specification.</p><p><strong>Advances in knowledge: </strong>Grouped calcifications could be early breast cancer on a mammogram. The morphology and distribution are correlated with the nature of breast diseases. AI is a potential decision support for the detection and classification of grouped microcalcifications and thus positively affects the control of breast cancer.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142543852","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Ovarian-Adnexal Reporting and Data System (O-RADS) MRI Scoring: Diagnostic Accuracy, Interobserver Agreement, and Applicability to Machine Learning.","authors":"Hüseyin Akkaya, Emin Demirel, Okan Dilek, Tuba Dalgalar Akkaya, Turgay Öztürkçü, Kübra Karaaslan Erişen, Zeynel Abidin Tas, Sevda Bas, Bozkurt Gülek","doi":"10.1093/bjr/tqae221","DOIUrl":"https://doi.org/10.1093/bjr/tqae221","url":null,"abstract":"<p><strong>Objectives: </strong>To evaluate the interobserver agreement and diagnostic accuracy of ovarian-adnexal reporting and data system magnetic resonance imaging (O-RADS MRI) and applicability to machine learning.</p><p><strong>Material and methods: </strong>Dynamic contrast-enhanced pelvic MRI examinations 471 lesions were retrospectively analyzed and assessed by three radiologists according to O-RADS MRI criteria. Radiomic data were extracted from T2, and post-contrast fat-suppressed T1-weighted images. Using these data, an artificial neural network (ANN), support vector machine, random forest, and naive Bayes models were constructed.</p><p><strong>Results: </strong>Among all readers, the lowest agreement was found for the O-RADS 4 group (kappa: 0.669 (95% confidence interval [CI] 0.634-0.733)), followed by the O-RADS 5 group (kappa: 0.709 (95% CI 0.678-0.754)). O-RADS 4 predicted a malignancy with an area under the curve (AUC) value of 74.3% (95% CI 0.701-0.782), and O-RADS 5 with an AUC of 95.5% (95% CI 0.932-0.972),(p < 0.001). Among the machine learning models, ANN achieved the highest success, distinguishing O-RADS groups with an AUC of 0.948, a precision of 0.861, and a recall of 0.824.</p><p><strong>Conclusion: </strong>The interobserver agreement and diagnostic sensitivity of the O-RADS MRI in assigning O-RADS 4-5 were not perfect, indicating a need for structural improvement. Integrating artificial intelligence into MRI protocols may enhance their performance.</p><p><strong>Advances in knowledge: </strong>Machine learning can achieve high accuracy in the correct classification of O-RADS MRI. Malignancy prediction rates were 74% for O-RADS 4 and 95% for O-RADS 5.</p>","PeriodicalId":9306,"journal":{"name":"British Journal of Radiology","volume":" ","pages":""},"PeriodicalIF":1.8,"publicationDate":"2024-10-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142543853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}