Radiology-Artificial Intelligence最新文献

{"title":"Comparative Performance of Federated and Centralized Learning for Brain Tumor Classification and Segmentation.","authors":"Anish Narayan, Chun Goh, Frederick Mariajoseph, Shujie Cui, Enes Makalic, Justin Moore","doi":"10.1148/ryai.260515","DOIUrl":"https://doi.org/10.1148/ryai.260515","url":null,"abstract":"<p><p>Purpose To compare federated learning (FL) and centralized learning (CL) performance for brain tumor classification and segmentation and examine whether specific configurations confer measurable performance advantages. Materials and Methods In this systematic review and meta-analysis, Medline, EMBASE, PubMed, Web of Science, and Scopus were searched from inception to March 4, 2026. Classification accuracy was pooled using random-effects meta-analysis, with heterogeneity quantified using I²; other metrics were summarized descriptively. Paired differences in classification accuracy and Dice similarity coefficient (DSC) were assessed using the Wilcoxon signed-rank test. Mixed-effects meta-regression and subgroup analyses evaluated whether specific configurations were associated with performance differences. Results Of 3,880 records identified, 50 studies were included (26 classification/detection, 26 segmentation; 2 in both). Within-study differences between FL and CL performance were -0.87 percentage points (pp) for classification accuracy (95% CI [-3.32, +1.44]; <i>P</i> = .41, k = 13) and -1.29 pp for segmentation DSC (95% CI [-4.23, +0.81]; Holm-corrected <i>P</i> = .50, k = 11). In unmatched single-arm analyses, pooled classification accuracy was 96.77% for FL (k = 18) versus 96.86% for CL (k = 9); the pooled DSC for segmentation was 84.95% (k = 24) versus 84.61% (k = 11). Heterogeneity was high (I²=93.3-96.5%); the 95% prediction interval for FL accuracy was 86.72-100%. Aggregation strategy, training sample size, task category, and independent and identically distributed (IID) versus non-IID partitioning were not associated with performance differences (all <i>P</i> ≥ .38). Conclusion There was no evidence of paired differences between FL and CL in brain tumor classification accuracy or segmentation DSC, and no specific configuration conferred a measurable performance advantage. © RSNA, 2026.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e260515"},"PeriodicalIF":20.1,"publicationDate":"2026-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149128609","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ARNAI: Artifact Removal Network Based on Autoencoding and Inpainting for Robust Spinal Image Segmentation and Measurement.","authors":"Sang-Jin Park, Jinyoung Choi, Seokwon Kim, Seungeon Song, Insu Park, Dougho Park, Taeyeon Kim, Youjin Lee, Donghoon Yang, Jaeman Cho, Joongwon Yang, Mansu Kim, Heumdai Kwon, Hong Gyu Baek, Dae Chul Cho, Injung Kim","doi":"10.1148/ryai.251134","DOIUrl":"https://doi.org/10.1148/ryai.251134","url":null,"abstract":"<p><p>Purpose To develop an artificial intelligence framework, robust to spinal implants, for automated measurement of spinopelvic parameters on postoperative radiographs. Materials and Methods Lateral lumbar spine radiographs from two institutions (internal dataset [<i>n</i> = 2486]: January 2017-December 2024; external dataset [<i>n</i> = 217]: October 2021-September 2025) were retrospectively reviewed. The <i>Restore, Segment, and Measure</i> (RSM) framework was developed, incorporating a novel <i>Artifact Removal Network based on Autoencoding and Inpainting</i> (ARNAI) to mitigate implant-related artifacts. Segmentation and spinopelvic parameter (including segmental Cobb angle [SCA]) measurement performance relative to expert reference measurements were assessed using the Dice similarity coefficient (DSC), mean absolute error (MAE), Wilcoxon signed-rank tests, and Benjamini-Hochberg correction for multiple comparisons. Results With ARNAI added to the segmentation pipeline, the mean DSC increased to 0.870 from 0.814, with marked gains at L3-L5. For implant-containing radiographs in the internal test set, rejected radiographs decreased by 64.21% (95 to 34), and the L4-L5 SCA MAE decreased to 4.7° from 16.2°-15.6° (~ 70% reduction); this remained significant after correction. For the external test set, rejected radiographs decreased by 65.45% (110 to 38), and the L4-L5 SCA MAE decreased to 9.5°-9.7° from 14.2°-14.5° (~ 33% reduction); however, this improvement did not remain significant after correction, and agreement with expert references remained limited. Conclusion The RSM framework improved automated spinopelvic parameter measurement in implant-containing postoperative radiographs, with the greatest benefit observed for L4-L5 SCA estimation in the internal dataset. In the external dataset, error was reduced but agreement remained limited. © RSNA, 2026.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e251134"},"PeriodicalIF":20.1,"publicationDate":"2026-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149128581","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Checklist for Artificial Intelligence in Medical Imaging (CLAIM): Explanation, Elaboration, and Examples.","authors":"Tugba Akinci D'Antonoli, Lisa C Adams, Amine Amyar, Ken Chang, Gunvant Chaudhari, Salvatore Claudio Fanni, Anthony A Gatti, Ricardo A Gonzales, Merel Huisman, Michail E Klontzas, Mahsa Mayeli, Mana Moassefi, Dogan S Polat, Ali Tejani, Charles E Kahn, John Mongan","doi":"10.1148/ryai.260835","DOIUrl":"https://doi.org/10.1148/ryai.260835","url":null,"abstract":"<p><p>The Checklist for Artificial Intelligence in Medical Imaging (CLAIM) provides a structured framework for transparent and reproducible reporting of AI studies in medical imaging. Since its introduction in 2020, CLAIM has been widely adopted by researchers, reviewers, and journal editors, but variability in its interpretation has limited consistent application. In 2024, the CLAIM Steering Committee published an updated checklist developed through a structured Delphi consensus process involving 72 experts across imaging-related medical specialties, AI science, journal editing, and biostatistics. This article provides a detailed explanation and elaboration of each of the 44 items in the CLAIM 2024 Update, clarifying the intent, common misinterpretations, and appropriate implementation of each item. Illustrative examples from the published literature demonstrating adherence to each item are provided in an accompanying Supplement and through a user-friendly online tool at https://rsna.github.io/claim/. The scope of this work spans the full AI study lifecycle covered by CLAIM, from study design and data sourcing to model development, evaluation, and reporting of results. This resource is intended to support authors, reviewers, and editors in the accurate and consistent application of CLAIM, thereby improving the quality, transparency, and reproducibility of AI research in medical imaging.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e260835"},"PeriodicalIF":20.1,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149082090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deep Learning Analysis of Dual-Modality US Videos for the Characterization of Superficial Lymphadenopathy.","authors":"Yangyang Zhu, Ran Cao, Zhibin Zhu, Jinxia Zhang, Haina Zhao, Lin Chen, Fei Ouyang, Wenhan Su, Yifang Li, Ying Duan, Qingqing Liang, Li Qiu, Dean Ta, Fang Nie","doi":"10.1148/ryai.260131","DOIUrl":"https://doi.org/10.1148/ryai.260131","url":null,"abstract":"<p><p>Purpose To develop and validate a deep learning (DL) model based on dual-modality US videos to differentiate benign from malignant superficial lymphadenopathy (LA) and stratify malignant subtypes. Materials and Methods This multicenter study included patients with pathologically confirmed LA from five centers (June 2019-December 2025). The Dual-Modality US Video Lymphadenopathy Diagnostic Network (DMUVL-DiagNet) integrated B-mode and color Doppler flow imaging videos with basic clinical information. A retrospective video dataset (<i>n</i> = 1616) was used for training and internal testing, while external testing used a prospective dual-modality US video set (<i>n</i> = 454) and a static US image set (<i>n</i> = 1956). Model performance was compared against clinical baseline and US-based DL models and an artificial intelligence (AI)-assisted reader study was conducted. Areas under the receiver operating characteristic curves (AUCs) were calculated to evaluate diagnostic performance. The study was prospectively registered at the Chinese Clinical Trial Registry (ChiCTR2400090592). Results Overall, 4026 patients (median age, 52.6 [IQR 41-62] years; 2175 males) were included. DMUVL-DiagNet consistently outperformed clinical and US-based models across all test sets, achieving AUCs of 0.91-0.95 for benign versus malignant classification and 0.87-0.91 for lymphoma versus metastasis classification. The model showed good generalizability (subgroup AUCs, 0.85-0.93) and surpassed the average performance of all radiologist experience groups in both tasks. Additionally, DMUVL-DiagNet assistance improved junior radiologist AUCs from 0.72 to 0.88 and from 0.58 to 0.77 for the two tasks, respectively (all <i>P</i> < .001). Conclusion DMUVL-DiagNet enabled accurate characterization of superficial LA and improved junior radiologist diagnostic performance. © The Authors 2026. Published by the Radiological Society of North America under a CC BY 4.0 license.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e260131"},"PeriodicalIF":20.1,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149082096","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Deep Learning Model to Predict Breast Cancer Recurrence Using Longitudinal Mammograms and Clinical Data.","authors":"Chunyao Lu, Xin Wang, Tianyu Zhang, Xinglong Liang, Luyi Han, Miguel Braga, Carla Sitges, Zahra Aghdam, Yuan Gao, Antonio Portaluri, Tao Tan, Ritse Mann","doi":"10.1148/ryai.250941","DOIUrl":"https://doi.org/10.1148/ryai.250941","url":null,"abstract":"<p><p>Purpose To develop a deep learning model that integrates longitudinal mammograms and static clinical data for predicting the recurrence risk and recurrence subtype of breast cancer. Materials and Methods In this retrospective study, the data of patients examined via imaging from January 2004 to December 2020 were included. A multimodal deep learning model incorporating pretreatment mammographic features and clinical data were developed. A multitask framework was implemented to jointly predict recurrence occurrence (binary classification) and recurrence subtype (local/regional/distant). The model was trained and evaluated within a single institutional cohort. Results A total of 3,923 patients and 19,684 examinations (mean age ± SD, 52.98 years ± 10.98; 3923 female) were included. In the test cohort, the model achieved an area under the receiver operating characteristic curve (AUC) of 0.80 (95% CI: 0.74, 0.85) and an area under the precision-recall curve (AUPRC) of 0.537 (95% CI: 0.432, 0.638) for overall recurrence, outperforming unimodal approaches. For recurrence subtype prediction among recurrencepositive patients, the model achieved a balanced accuracy of 65.1% (95% CI: 60.2%, 71.1%) and an AUC of 0.72 (95% CI: 0.68, 0.78). Conclusion The proposed multimodal deep learning framework improves both recurrence prediction and recurrence subtype discrimination over existing models. ©RSNA, 2026.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e250941"},"PeriodicalIF":20.1,"publicationDate":"2026-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148912824","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DeepVEST: Deep Learning-based Vessel Segmentation and Erasure in Breast MRI for Improved Lesion Assessment.","authors":"Marco Cantone, Tianyu Zhang, Claudio Marrocco, Luyi Han, Antonio Portaluri, Jarek van Dijk, Carla Sitges Puigivila, Miguel Braga, Noemi Schmidt, Alessandro Bria, Ritse Mann","doi":"10.1148/ryai.250630","DOIUrl":"https://doi.org/10.1148/ryai.250630","url":null,"abstract":"<p><p>Purpose To investigate whether selective removal of vascular structures can improve lesion visibility and interpretability in maximum intensity projection (MIP) images derived from dynamic contrast-enhanced MRI. Materials and Methods A retrospective analysis was conducted using breast MRI scans from the Duke-Breast-Cancer-MRI (Duke) dataset for model development. DeepVEST, a deep learning method for automatic vessel segmentation and removal, was developed. A reader study with five breast radiologists was conducted to evaluate the impact of vessel-removed MIPs on lesion assessment, using images from the Duke and Advanced-MRI-Breast-Lesions (AMBL) datasets. Segmentation performance was evaluated against manually annotated vessel segmentations using the Dice similarity coefficient (DSC), and perceived usefulness and quality were measured through reader study metrics. Interreader agreement on vessel removal quality and artifact impact was calculated using the Gwet agreement coefficient (AC1). Results DeepVEST achieved a DSC of 0.611 for vessel segmentation. In the reader study (150 assessments, 600 responses), vessels partially or fully obscured lesion margins in 91 of 150 assessments (60.7%). Vessel removal effectiveness was rated 3.820 ± 0.749 on a 5-point Likert scale. Artifacts were reported in 45 of 150 assessments (30.0%), with a low average severity score of 0.493 ± 0.900 (scale 0-5), indicating minimal image quality impact. Substantial interreader agreement was observed for vessel removal quality (AC1 = 0.703) and artifact evaluation (AC1 = 0.736). Conclusion DeepVEST enabled automatic vessel segmentation and removal in breast MRI MIP images, showing high effectiveness with minimal artifacts. Most readers found it helpful for lesion assessment in challenging cases. ©RSNA, 2026.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e250630"},"PeriodicalIF":20.1,"publicationDate":"2026-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148913304","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assuring Transparency of AI Software after Deployment.","authors":"Irvine Sihlahla","doi":"10.1148/ryai.260811","DOIUrl":"https://doi.org/10.1148/ryai.260811","url":null,"abstract":"","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":"8 5","pages":"e260811"},"PeriodicalIF":20.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148913235","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"What LUNA25 Teaches Us about AI for Lung Cancer Screening.","authors":"Eduardo Moreno Júdice de Mattos Farina, Gilberto Szarf","doi":"10.1148/ryai.260657","DOIUrl":"10.1148/ryai.260657","url":null,"abstract":"","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":"8 5","pages":"e260657"},"PeriodicalIF":20.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148680003","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Image Quality in the Era of Artificial Intelligence: Understanding the Limitations of AI-based Image Reconstruction and Postprocessing.","authors":"Jana G Delfino, Jason L Granstedt, Frank W Samuelson, Robert Ochs, Krishna Juluru","doi":"10.1148/ryai.260100","DOIUrl":"10.1148/ryai.260100","url":null,"abstract":"<p><p>Artificial intelligence (AI) is being deployed in radiology for image reconstruction and postprocessing to produce images that appear sharper, smoother, and more detailed. However, incorporating AI also introduces new failure modes and can exacerbate the disconnect between the perceived quality of an image and its diagnostic information content. Understanding the limitations of AI-enabled image reconstruction and postprocessing is essential for the safe and effective use of this technology. Therefore, the purpose of this report is to raise awareness of the limitations of AI-based image reconstruction and postprocessing to help users realize the benefits of the technology while minimizing risks. Accordingly, this report reviews approaches to image quality assessment, discusses the regulatory framework relevant to AI-enabled imaging devices, describes AI-specific failure modes, and outlines strategies to mitigate associated risks. <b>Keywords:</b> Artificial Intelligence, Image Quality Assessment, Hallucinations, Artifacts, Interpretation, Image Reconstruction, Denoising, Superresolution, Postacquisition Processing, Perception, Reconstruction Algorithms, Image Postprocessing © RSNA, 2026.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e260100"},"PeriodicalIF":20.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148881961","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Predetermined Change Control Plan Adoption and Documentation Transparency in U.S. Food and Drug Administration-cleared Radiology Artificial Intelligence/Machine Learning Devices.","authors":"Ketan Dayma, Palak Patel, Kenneth Hildreth, Tamara Jamaspishvili","doi":"10.1148/ryai.260385","DOIUrl":"10.1148/ryai.260385","url":null,"abstract":"<p><p>Purpose To characterize adoption of Predetermined Change Control Plan (PCCP) and documentation transparency among U.S. Food and Drug Administration (FDA)-cleared radiology artificial intelligence (AI)/machine learning (ML) devices (2015-2025). Materials and Methods A cross-sectional systematic scoping review was conducted with linked data from FDA AI/ML-enabled device databases through April 2026. Completeness of PCCP documentation was scored by two independent observers (intraclass correlation coefficient, 0.93) using an eight-point rubric (possible scores, 0-8) derived from the FDA's final PCCP guidance (December 2024). Identified PCCP devices underwent manual verification against FDA regulatory summaries. Results Among FDA-listed AI/ML device submissions, 1080 of 1394 (77.5%) were radiology submissions, and nearly all cleared via 510(k) review. Across all FDA panels, 170 devices were cleared with a PCCP. Radiology led AI/ML-specific PCCP adoption (34 of 37 radiology PCCP devices [91.9%]). Among the PCCP-cleared radiology AI devices, 22 of 34 (65%) were cleared in 2025 alone following the final FDA guidance. Discrepancies between the FDA's public database and individual summaries required manual adjudication for nine of 34 (27%) devices. PCCP documentation scores ranged from 0 to 8 (mean, 5), with most modifications focused on data retraining, compatibility expansion, and algorithm optimization. Continuous monitoring of device performance and predefined drift triggers for retraining were absent from public summaries. Conclusion PCCP adoption increased in radiology after issuance of the final FDA guidance, yet public lifecycle controls, particularly monitoring performance metrics and trigger thresholds, were limited. Standardized PCCP reporting of lifecycle controls are suggested as a condition of PCCP authorization to enable systematic postmarket monitoring as this pathway scales. <b>Keywords:</b> Health Policy and Practice, Informatics, Computer Aided Diagnosis (CAD), QA/QC, Safety, Technical Aspects, Legal Issues, Application Domain <i>Supplemental material is available for this article.</i> © RSNA, 2026 See also commentary by Sihlahla in this issue.</p>","PeriodicalId":29787,"journal":{"name":"Radiology-Artificial Intelligence","volume":" ","pages":"e260385"},"PeriodicalIF":20.1,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148621601","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}