Academic Radiology最新文献

{"title":"Enhancing Clinical Utility in Oesophageal Cancer Surgery: A Dialogue on Predictive Modelling for Anastomotic Leakage","authors":"Michail E. Klontzas , Ioannis Rouvelas , Antonios Tzortzakakis","doi":"10.1016/j.acra.2025.04.049","DOIUrl":"10.1016/j.acra.2025.04.049","url":null,"abstract":"","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":"32 6","pages":"Page 3773"},"PeriodicalIF":3.8,"publicationDate":"2025-05-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144167386","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of Posthepatectomy Liver Failure in Narrow Resection Margins HCC: A Model Based on Iodine Map Histogram Analysis of Nontumorous Liver Parenchyma.","authors":"Yuan Xu, Bo Liu, Fukai Li, Jiachen Sun, Yufeng Li, Hong Liu, Tiezhu Ren, Jianli Liu, Junlin Zhou","doi":"10.1016/j.acra.2025.05.002","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.002","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>Posthepatectomy liver failure (PHLF) is a severe postoperative complication. This study aims to develop and validate a model combining iodine map histogram parameters of nontumorous liver parenchyma and clinical characteristics to predict early PHLF in patients with narrow resection margins-hepatocellular carcinoma (NRM-HCC).</p><p><strong>Materials and methods: </strong>A retrospective analysis was conducted on 154 patients with NRM-HCC who underwent hepatectomy at our center, with patients randomly divided into a 7:3 ratio into a training cohort (n=107) and an internal validation cohort (n=47). Iodine map histogram parameters of nontumorous liver parenchyma during the portal venous phase of spectral CT were measured. Standardized Future Residual Liver Volume Ratio (SFLVR) was calculated based on Future Liver Remnant Volume. Based on training cohort data, logistic regression analysis was performed to identify predictors and construct a model for predicting PHLF. The model's performance was evaluated by using receiver operating characteristic curve analysis, calibration curves, and decision curve analyses (DCA).</p><p><strong>Results: </strong>In the training cohort, univariate and multivariate logistic regression analyses identified Albumin-bilirubin score, intraoperative blood loss (L), Kurtosis, and SFLVR as independent risk factors for PHLF. A comprehensive model combining these independent risk factors yielded an area under the curve of 0.87 (95% CI: 0.80-0.94) for predicting PHLF, outperforming each individual risk factor. Calibration curve and DCA demonstrated good consistency and clinical utility of the model in both the training and validation cohorts.</p><p><strong>Conclusion: </strong>A novel comprehensive model combining iodine map histogram parameter Kurtosis of nontumorous liver parenchyma, SFLVR, and clinical features facilitates early prediction of PHLF in NRM-HCC patients.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144175628","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating Large Language Models for Enhancing Radiology Specialty Examination: A Comparative Study with Human Performance.","authors":"Hao-Yun Liu, Shyh-Jye Chen, Weichung Wang, Chung-Hsi Lee, Hsian-He Hsu, Shu-Huei Shen, Hong-Jen Chiou, Wen-Jeng Lee","doi":"10.1016/j.acra.2025.05.023","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.023","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>The radiology specialty examination assesses clinical decision-making, image interpretation, and diagnostic reasoning. With the expansion of medical knowledge, traditional test design faces challenges in maintaining accuracy and relevance. Large language models (LLMs) demonstrate potential in medical education. This study evaluates LLM performance in radiology specialty exams, explores their role in assessing question difficulty, and investigates their reasoning processes, aiming to develop a more objective and efficient framework for exam design.</p><p><strong>Materials and methods: </strong>This study compared the performance of LLMs and human examinees in a radiology specialty examination. Three LLMs (GPT-4o, o1-preview, and GPT-3.5-turbo-1106) were evaluated under zero-shot conditions. Exam accuracy, examinee accuracy, discrimination index, and point-biserial correlation were used to assess LLMs' ability to predict question difficulty and reasoning processes. The data provided by the Taiwan Radiological Society ensures comparability between AI and human performance.</p><p><strong>Results: </strong>As for accuracy, GPT-4o (88.0%) and o1-preview (90.9%) outperformed human examinees (76.3%), whereas GPT-3.5-turbo-1106 showed significantly lower accuracy (50.2%). Question difficulty analysis revealed that newer LLMs excel in solving complex questions, while GPT-3.5-turbo-1106 exhibited greater performance variability. Discrimination index and point-biserial Correlation analyses demonstrated that GPT-4o and o1-preview accurately identified key differentiating questions, closely mirroring human reasoning patterns. These findings suggest that advanced LLMs can assess medical examination difficulty, offering potential applications in exam standardization and question evaluation.</p><p><strong>Conclusion: </strong>This study evaluated the problem-solving capabilities of GPT-3.5-turbo-1106, GPT-4o, and o1-preview in a radiology specialty examination. LLMs should be utilized as tools for assessing exam question difficulty and assisting in the standardized development of medical examinations.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144175554","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Interpretable Machine Learning Models for Differentiating Glioblastoma From Solitary Brain Metastasis Using Radiomics.","authors":"Xueming Xia, Wenjun Wu, Qiaoyue Tan, Qiheng Gou","doi":"10.1016/j.acra.2025.05.016","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.016","url":null,"abstract":"<p><strong>Purpose: </strong>To develop and validate interpretable machine learning models for differentiating glioblastoma (GB) from solitary brain metastasis (SBM) using radiomics features from contrast-enhanced T1-weighted MRI (CE-T1WI), and to compare the impact of low-order and high-order features on model performance.</p><p><strong>Methods: </strong>A cohort of 434 patients with histopathologically confirmed GB (226 patients) and SBM (208 patients) was retrospectively analyzed. Radiomic features were derived from CE-T1WI, with feature selection conducted through minimum redundancy maximum relevance and least absolute shrinkage and selection operator regression. Machine learning models, including GradientBoost and lightGBM (LGBM), were trained using low-order and high-order features. The performance of the models was assessed through receiver operating characteristic analysis and computation of the area under the curve, along with other indicators, including accuracy, specificity, and sensitivity. SHapley Additive Explanations (SHAP) analysis is used to measure the influence of each feature on the model's predictions.</p><p><strong>Results: </strong>The performances of various machine learning models on both the training and validation datasets were notably different. For the training group, the LGBM, CatBoost, multilayer perceptron (MLP), and GradientBoost models achieved the highest AUC scores, all exceeding 0.9, demonstrating strong discriminative power. The LGBM model exhibited the best stability, with a minimal AUC difference of only 0.005 between the training and test sets, suggesting strong generalizability. Among the validation group results, the GradientBoost classifier achieved the maximum AUC of 0.927, closely followed by random forest at 0.925. GradientBoost also demonstrated high sensitivity (0.911) and negative predictive value (NPV, 0.889), effectively identifying true positives. The LGBM model showed the highest test accuracy (86.2%) and performed excellently in terms of sensitivity (0.911), NPV (0.895), and positive predictive value (PPV, 0.837). The models utilizing high-order features outperformed those based on low-order features in all the metrics. SHAP analysis further enhances model interpretability, providing insights into feature importance and contributions to classification decisions.</p><p><strong>Conclusion: </strong>Machine learning techniques based on radiomics can effectively distinguish GB from SBM, with gradient boosting tree-based models such as LGBMs demonstrating superior performance. High-order features significantly improve model accuracy and robustness. SHAP technology enhances the interpretability and transparency of models for distinguishing brain tumors, providing intuitive visualization of the contribution of radiomic features to classification.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144175626","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving Breast Cancer Diagnosis in Ultrasound Images Using Deep Learning with Feature Fusion and Attention Mechanism.","authors":"Sohaib Asif, Yuqi Yan, Bojian Feng, Meiling Wang, Yuxin Zheng, Tian Jiang, Ruyi Fu, Jincao Yao, Lujiao Lv, Mei Song, Lin Sui, Zheng Yin, Vicky Yang Wang, Dong Xu","doi":"10.1016/j.acra.2025.05.007","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.007","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>Early detection of malignant lesions in ultrasound images is crucial for effective cancer diagnosis and treatment. While traditional methods rely on radiologists, deep learning models can improve accuracy, reduce errors, and enhance efficiency. This study explores the application of a deep learning model for classifying benign and malignant lesions, focusing on its performance and interpretability.</p><p><strong>Materials and methods: </strong>In this study, we proposed a feature fusion-based deep learning model for classifying benign and malignant lesions in ultrasound images. The model leverages advanced architectures such as MobileNetV2 and DenseNet121, enhanced with feature fusion and attention mechanisms to boost classification accuracy. The clinical dataset comprises 2171 images collected from 1758 patients between December 2020 and May 2024. Additionally, we utilized the publicly available BUSI dataset, consisting of 780 images from female patients aged 25 to 75, collected in 2018. To enhance interpretability, we applied Grad-CAM, Saliency Maps, and shapley additive explanations (SHAP) techniques to explain the model's decision-making. A comparative analysis with radiologists of varying expertise levels is also conducted.</p><p><strong>Results: </strong>The proposed model exhibited the highest performance, achieving an AUC of 0.9320 on our private dataset and an area under the curve (AUC) of 0.9834 on the public dataset, significantly outperforming traditional deep convolutional neural network models. It also exceeded the diagnostic performance of radiologists, showcasing its potential as a reliable tool for medical image classification. The model's success can be attributed to its incorporation of advanced architectures, feature fusion, and attention mechanisms. The model's decision-making process was further clarified using interpretability techniques like Grad-CAM, Saliency Maps, and SHAP, offering insights into its ability to focus on relevant image features for accurate classification.</p><p><strong>Conclusion: </strong>The proposed deep learning model offers superior accuracy in classifying benign and malignant lesions in ultrasound images, outperforming traditional models and radiologists. Its strong performance, coupled with interpretability techniques, demonstrates its potential as a reliable and efficient tool for medical diagnostics.</p><p><strong>Data availability: </strong>The datasets generated and analyzed during the current study are not publicly available due to the nature of this research and participants of this study, but may be available from the corresponding author on reasonable request.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144175568","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Night Float vs. Traditional Call in Interventional Radiology: Impacts on Resident Wellness and Nighttime Clinical Service.","authors":"Oussama Metrouh, Julie Bulman, Spencer Degerstedt, Sarah Schroeppel DeBacker, Muneeb Ahmed, Jeffrey Weinstein","doi":"10.1016/j.acra.2025.05.008","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.008","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>To evaluate the impact of an Interventional Radiology resident night float system on resident well-being and clinical workflow.</p><p><strong>Materials and methods: </strong>This study spanned a period of 12 months, 6 months in which residents worked on night using a traditional call system without guaranteed post-call days, and 6 months of a night float system in which a night float dedicated resident was taking a week of night calls, with no daytime duties. \"IR short communication notes\", and \"Full patient consult notes\", documented by the on-call- resident between 6 PM and 7 AM were reviewed. The overall number of notes and the number of notes documented per resident were compared between the two periods. Additionally, current and alumni residents, advanced practice providers, and faculty physicians were surveyed about their perception of the night float system as it compares to the old call system.</p><p><strong>Results: </strong>The volume of notes increased significantly with the implementation of the night float system, from 127 to 375 (p=0.03). Over the 6-month night float period, the median number of short communication notes per resident increased from 4.5 (4-6.5) to 9.5 (9-13), p=0.04, while full patient consult notes rose from 16 (8.75-19.5) to 53 (36.25-61.5), p=0.002. Survey results from residents showed a better perception of their wellness under night float (25% were positively affected, compared to 0% during the traditional call system). A perceived overall improvement in patient care was reported by 100% of IR faculty and APPs with night float.</p><p><strong>Conclusion: </strong>Night float implementation enhanced IR resident productivity and wellness as evidenced by increased documented clinical encounters and positive feedback.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144163580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bridging the Gap between Injections and Surgery: Meta-Analysis of Genicular Artery Embolization in Knee Osteoarthritis.","authors":"Rada Abussa, Aleksandar Jeremic","doi":"10.1016/j.acra.2025.05.011","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.011","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>To evaluate the clinical efficacy and safety of genicular artery embolization (GAE) for chronic knee osteoarthritis (OA), incorporating recent trials and long-term outcomes.</p><p><strong>Materials and methods: </strong>A systematic review identified peer-reviewed studies of GAE in knee OA, including randomized controlled trials (RCTs) and prospective series. Data on pain scores, function, follow-up, and adverse events were extracted. Pooled pain reductions (VAS, WOMAC) were analyzed using random-effects models. Forest and funnel plots visualized treatment effects and publication bias.</p><p><strong>Results: </strong>Fourteen studies (510 patients, 567 knees) met inclusion criteria, including three sham-controlled RCTs and several prospective series. GAE consistently reduced pain in open-label studies, with a pooled pre-post pain reduction of ∼30 points (0-100 scale) at 6-12 months. Functional scores (WOMAC, KOOS) also improved. About 78-92% of patients achieved clinically meaningful improvement (≥50% pain reduction or ≥10-15 point change) by 12 months. However, sham-controlled RCTs yielded mixed results: one showed early benefit, while two found no significant difference versus placebo at 4-12 months. Heterogeneity was moderate (I² ∼74%). Minor adverse events included transient skin discoloration (∼10-30%) and groin hematoma (∼2-3%). One case of vasculitis and one deep vein thrombosis were reported; no major complications occurred.</p><p><strong>Conclusion: </strong>GAE appears effective and safe for chronic knee OA, though its benefit over placebo remains uncertain. This meta-analysis, incorporating recent evidence and 2-year data, underscores GAE's promise-but highlights the need for larger, rigorously designed RCTs to confirm efficacy, refine patient selection, optimize techniques, and guide clinical use.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144163570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"MRI Morphological Features for Assessing Lymph Node Restaging in Rectal Cancer After Neoadjuvant Chemoradiotherapy.","authors":"Hengxiao Hu, Jing Xu, Xiaowen Xie, Chenyi Xie, Kuanhong Wang, Yingying Guo, Liujun Yi, Xin Chen","doi":"10.1016/j.acra.2025.05.004","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.004","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>Accurate lymph node (LN) restaging after neoadjuvant chemoradiotherapy (nCRT) is critical for guiding subsequent treatment decisions in patients with rectal cancer. The currently recommended European Society of Gastrointestinal and Abdominal Radiology (ESGAR) method has shown limited accuracy. This study aimed to evaluate the diagnostic performance of specific MRI morphological features for LN restaging post-nCRT and to develop a reliable, clinically applicable imaging-based assessment method.</p><p><strong>Materials and methods: </strong>In this retrospective multi-center study, one training cohort and two external validation cohorts were included. MRI morphological features were assessed for their association with pathological LN status using χ² tests and multivariable logistic regression. Diagnostic performance was evaluated using the area under the receiver operating characteristic curve (AUC). Interobserver agreement was evaluated using Cohen κ coefficient and the intraclass correlation coefficient (ICC [1,1]).</p><p><strong>Results: </strong>Among 238 patients in the training cohort, 64 (26%) had pathologically confirmed positive LNs. Multivariate analysis revealed that signal homogeneity (odds ratio [OR], 3.23; P=.02), interruption of vessels (OR, 6.97; P=.001), and tail sign (OR, 4.02; P=.002) were independent predictors of positive LN status. The developed \"Three-features method\" (positive if any two features were present) achieved an AUC of 0.83, with 69% sensitivity, 86% specificity, and 82% overall accuracy-significantly outperforming the ESGAR method (AUC=0.68). External validation showed consistent diagnostic performance (AUC=0.74 and 0.77).</p><p><strong>Conclusion: </strong>The integration of signal homogeneity, interruption of vessels, and tail sign into a combined MRI-based \"Three-features method\" significantly improves the accuracy of LN restaging after nCRT compared to ESGAR method.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144163573","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Radiologists Making a Difference Beyond Radiology.","authors":"Richard B Gunderman","doi":"10.1016/j.acra.2025.05.015","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.015","url":null,"abstract":"","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144163582","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multimodal Diffusion MRI Synergized with VI-RADS for Precision Grading of Bladder Urothelial Carcinoma: A Prospective Diagnostic Model Validation.","authors":"Xiaoxian Zhang, Shaoyu Wang, Mengzhu Wang, Lifeng Wang, Shouning Zhang, Xuejun Chen, Chunmiao Xu","doi":"10.1016/j.acra.2025.05.014","DOIUrl":"https://doi.org/10.1016/j.acra.2025.05.014","url":null,"abstract":"<p><strong>Rationale and objectives: </strong>To evaluate the diagnostic performance of diffusion tensor imaging (DTI), diffusion kurtosis imaging (DKI), mean apparent propagator (MAP), neurite orientation dispersion and density imaging (NODDI), and the Vesical Imaging-Reporting and Data System (VI-RADS) in discriminating the pathological grade of bladder urothelial carcinoma (UCB).</p><p><strong>Materials and methods: </strong>This prospective study enrolled patients with pathologically confirmed UCB between May 2023 and December 2023. Preoperative MRI protocols included spin-echo echo-planner imaging (SE-EPI) and conventional DWI. Quantitative parameters from SE-EPI (DTI, DKI, MAP, NODDI) and apparent diffusion coefficient (ADC) values were measured. Group comparisons between low-grade and high-grade UCB were performed using t-tests or Mann-Whitney U tests. Receiver operating characteristic (ROC) analysis and DeLong's test were used to evaluate diagnostic performance.</p><p><strong>Results: </strong>A total of 50 patients with UCB (low-grade/ high-grade = 16/34) were included. VI-RADS score and mean kurtosis (MK) derived from DKI emerged as independent predictors for differentiating low-grade and high-grade UCB (area under the curve (AUC): 0.692 and 0.865, respectively). The combination of VI-RADS and DKI-MK achieved superior diagnostic performance (AUC: 0.915, sensitivity: 0.941) compared to VI-RADS alone (AUC: 0.692, sensitivity: 0.471; p < 0.001) or ADC alone (AUC: 0.787, sensitivity: 0.813; p < 0.05).</p><p><strong>Conclusion: </strong>Integrating VI-RADS with DKI-MK significantly enhances preoperative assessment of UCB pathological grading, demonstrating higher accuracy and sensitivity than VI-RADS or ADC alone. This approach improves diagnostic objectivity by combining qualitative imaging criteria with quantitative diffusion metrics, offering potential clinical utility for treatment stratification.</p>","PeriodicalId":50928,"journal":{"name":"Academic Radiology","volume":" ","pages":""},"PeriodicalIF":3.8,"publicationDate":"2025-05-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144163576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}