Radiology-Artificial Intelligence最新文献

Chronic low back pain is a global health issue with considerable socioeconomic burdens and is associated with changes in lumbar paraspinal muscles (LPMs). In this retrospective study, a deep learning method was trained and externally validated for automated LPM segmentation, muscle volume quantification, and fatty infiltration assessment across multisequence MR images. A total of 1302 MR images from 641 participants across five centers were included. Data from two centers were used for model training and tuning, while data from the remaining three centers were used for external testing. Model segmentation performance was evaluated against manual segmentation using the Dice similarity coefficient (DSC), and measurement accuracy was assessed using two one-sided tests and intraclass correlation coefficients (ICCs). The model achieved global DSC values of 0.98 on the internal test set and 0.93 to 0.97 on external test sets. Statistical equivalence between automated and manual measurements of muscle volume and fat ratio was confirmed in most regions (P < .05). Agreement between automated and manual measurements was high (ICCs > 0.92). In conclusion, the proposed automated method accurately segmented LPM and demonstrated statistical equivalence to manual measurements of muscle volume and fatty infiltration ratio across multisequence, multicenter MR images. Keywords: MR-Imaging, Muscular, Volume Analysis, Segmentation, Vision, Application Domain, Quantification, Supervised Learning Type of Machine Learning, Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning Algorithms Supplemental material is available for this article. © RSNA, 2025.

Purpose To develop and evaluate an open-source deep learning model for detection and localization of breast cancer on MRI scans. Materials and Methods In this retrospective study, a deep learning model for breast cancer detection and localization was trained on the largest breast MRI dataset to date. Data included all breast MRI examinations conducted at a tertiary cancer center in the United States between 2002 and 2019. The model was validated on sagittal MRI scans from the primary site (n = 6615 breasts). Generalizability was assessed by evaluating model performance on axial data from the primary site (n = 7058 breasts) and a second clinical site (n = 1840 breasts). Results The primary site dataset included 30 672 sagittal MRI examinations (52 598 breasts) from 9986 female patients (mean age, 52.1 years ± 11.2 [SD]). The model achieved an area under the receiver operating characteristic curve of 0.95 for detecting cancer in the primary site. At 90% specificity (5717 of 6353), model sensitivity was 83% (217 of 262), which was comparable to historical performance data for radiologists. The model generalized well to axial examinations, achieving an area under the receiver operating characteristic curve of 0.92 on data from the same clinical site and 0.92 on data from a secondary site. The model accurately located the tumor in 88.5% (232 of 262) of sagittal images, 92.8% (272 of 293) of axial images from the primary site, and 87.7% (807 of 920) of secondary site axial images. Conclusion The model demonstrated state-of-the-art performance on breast cancer detection. Code and weights are openly available to stimulate further development and validation. Keywords: Computer-aided Diagnosis (CAD), MRI, Neural Networks, Breast Supplemental material is available for this article. See also commentary by Moassefi and Xiao in this issue. © RSNA, 2025.

Purpose To develop a transformer-based deep learning model-MR-Transformer-that leverages ImageNet pretraining and three-dimensional spatial correlations to predict the progression of knee osteoarthritis to total knee replacement using MRI. Materials and Methods This retrospective study included 353 case-control matched pairs of coronal intermediate-weighted turbo spin-echo (COR-IW-TSE) and sagittal intermediate-weighted turbo spin-echo with fat suppression (SAG-IW-TSE-FS) knee MRI scans from the Osteoarthritis Initiative database, with a follow-up period up to 9 years, and 270 case-control matched pairs of coronal short-tau inversion recovery (COR-STIR) and sagittal proton-density fat-saturated (SAG-PD-FAT-SAT) knee MRI scans from the Multicenter Osteoarthritis Study database, with a follow-up period up to 7 years. Performance of the MR-Transformer to predict the progression of knee osteoarthritis was compared with that of existing state-of-the-art deep learning models (TSE-Net, 3DMeT, and MRNet) using sevenfold nested cross-validation across the four MRI tissue sequences. Results Among the 353 Osteoarthritis Initiative case-control pairs, 215 were women (mean age, 63 years ± 8 [SD]); among the 270 Multicenter Osteoarthritis Study case-control pairs, 203 were women (mean age, 65 years ± 7). The MR-Transformer achieved areas under the receiver operating characteristic curve (AUCs) of 0.88 (95% CI: 0.85, 0.91), 0.88 (95% CI: 0.85, 0.90), 0.86 (95% CI: 0.82, 0.89), and 0.84 (95% CI: 0.81, 0.87) for COR-IW-TSE, SAG-IW-TSE-FS, COR-STIR, and SAG-PD-FAT-SAT, respectively. The model achieved a higher AUC than that of 3DMeT for all MRI sequences (P < .001). The model showed the highest sensitivity of 83% (95% CI: 78, 87) and specificity of 83% (95% CI: 76, 88) for the COR-IW-TSE MRI sequence. Conclusion Compared with the existing deep learning models, the MR-Transformer exhibited state-of-the-art performance in predicting the progression of knee osteoarthritis to total knee replacement using MRI scans. Keywords: MRI, Knee, Prognosis, Supervised Learning Supplemental material is available for this article. © RSNA, 2025.

Purpose To develop a deep learning model that uses a single inspiratory chest CT scan to perform parametric response mapping (PRM) and predict functional small airways disease (fSAD). Materials and Methods In this retrospective study, predictive and generative deep learning models for PRM using inspiratory chest CT were developed using a model development dataset with fivefold cross-validation, with PRM derived from paired respiratory CT as the reference standard. Voxelwise metrics, including sensitivity, area under the receiver operating characteristic curve (AUC), and structural similarity index measure, were used to evaluate model performance in predicting PRM and generating expiratory CT images. The best-performing model was tested on three internal test sets and an external test set. Results The model development dataset of 308 individuals (median age, 67 years [IQR: 62-70 years]; 113 female) was divided into the training set (n = 216), the internal validation set (n = 31), and the first internal test set (n = 61). The generative model outperformed the predictive model in detecting fSAD (sensitivity, 86.3% vs 38.9%; AUC, 0.86 vs 0.70). The generative model performed well in the second internal (AUCs of 0.64, 0.84, and 0.97 for emphysema, fSAD, and normal lung tissue, respectively), the third internal (AUCs of 0.63, 0.83, and 0.97), and the external (AUCs of 0.58, 0.85, and 0.94) test sets. Notably, the model exhibited exceptional performance in the preserved ratio impaired spirometry group of the fourth internal test set (AUCs of 0.62, 0.88, and 0.96). Conclusion The proposed generative model, using a single inspiratory CT scan, outperformed existing algorithms in PRM evaluation and achieved comparable results to paired respiratory CT. Keywords: CT, Lung, Chronic Obstructive Pulmonary Disease, Diagnosis, Reconstruction Algorithms, Deep Learning, Parametric Response Mapping, X-ray Computed Tomography, Small Airways Supplemental material is available for this article. © The Author(s) 2025. Published by the Radiological Society of North America under a CC BY 4.0 license. See also the commentary by Hathaway and Singh in this issue.

Purpose To develop and optimize a federated learning (FL) framework across multiple clients for biparametric MRI prostate segmentation and clinically significant prostate cancer (csPCa) detection. Materials and Methods A retrospective study was conducted using Flower FL (Flower.ai) to train a nnU-Net-based architecture for MRI prostate segmentation and csPCa detection using data collected from January 2010 to August 2021. Model development included training and optimizing local epochs, federated rounds, and aggregation strategies for FL-based prostate segmentation on T2-weighted MR images (four clients, 1294 patients) and csPCa detection using biparametric MR images (three clients, 1440 patients). Performance was evaluated on independent test sets using the Dice score for segmentation and the Prostate Imaging: Cancer Artificial Intelligence (PI-CAI) score, defined as the average of the area under the receiver operating characteristic curve and average precision, for csPCa detection. P values for performance differences were calculated using permutation testing. Results The FL configurations were independently optimized for both tasks, showing improved performance at 1 epoch (300 rounds) using FedMedian for prostate segmentation and 5 epochs (200 rounds) using FedAdagrad, for csPCa detection. Compared with the average performance of the clients, the optimized FL model significantly improved performance in prostate segmentation (Dice score, increase from 0.73 ± 0.06 [SD] to 0.88 ± 0.03; P ≤ .01) and csPCa detection (PI-CAI score, increase from 0.63 ± 0.07 to 0.74 ± 0.06; P ≤ .01) on the independent test set. The optimized FL model showed higher lesion detection performance compared with the FL-baseline model (PI-CAI score, increase from 0.72 ± 0.06 to 0.74 ± 0.06; P ≤ .01), but no evidence of a difference was observed for prostate segmentation (Dice scores, 0.87 ± 0.03 vs 0.88 ± 03; P > .05). Conclusion FL enhanced the performance and generalizability of MRI prostate segmentation and csPCa detection compared with local models, and optimizing its configuration further improved lesion detection performance. Keywords: Federated Learning, Prostate Cancer, MRI, Cancer Detection, Deep Learning Supplemental material is available for this article. © RSNA, 2025.

Purpose To improve the generalizability of pathologic complete response prediction following neoadjuvant chemotherapy using deep learning-based retrospective pharmacokinetic quantification of early treatment dynamic contrast-enhanced MRI. Materials and Methods This multicenter retrospective study included breast MRI data from four publicly available datasets of patients with breast cancer acquired from May 2002 to November 2016. Pharmacokinetic quantification was performed using a previously developed deep learning model for clinical multiphasic dynamic contrast-enhanced MRI datasets. Radiomic analysis was performed on pharmacokinetic quantification maps and conventional enhancement maps. These data, together with clinicopathologic variables and shape-based radiomic analysis, were subsequently applied for pathologic complete response prediction using logistic regression. Prediction performance was evaluated by area under the receiver operating characteristic curve (AUC). Results A total of 1073 female patients with breast cancer were included. The proposed method showed improved consistency and generalizability compared with the reference method, achieving higher AUC values across external datasets (0.82 [95% CI: 0.72, 0.91], 0.75 [95% CI: 0.71, 0.79], and 0.77 [95% CI: 0.66, 0.86] for datasets A2, B, and C, respectively). For dataset A2 (from the same study as the training dataset), there was no significant difference in performance between the proposed method and reference method (P = .80). Notably, on the combined external datasets, the proposed method significantly outperformed the reference method (AUC, 0.75 [95% CI: 0.72, 0.79] vs AUC, 0.71 [95% CI: 0.68, 0.76]; P = .003). Conclusion This work offers an approach to improve the generalizability and predictive accuracy of pathologic complete response for breast cancer across diverse datasets, achieving higher and more consistent AUC scores than existing methods. Keywords: Tumor Response, Breast, Prognosis, Dynamic Contrast-enhanced MRI Supplemental material is available for this article. © RSNA, 2025 See also commentary by Schnitzler in this issue.

Although there are relatively few diverse, high-quality medical imaging datasets on which to train computer vision artificial intelligence models, even fewer datasets contain expertly classified observations that can be repurposed to train or test such models. The traditional annotation process is laborious and time-consuming. Repurposing annotations and consolidating similar types of annotations from disparate sources has never been practical. Until recently, the use of natural language processing to convert a clinical radiology report into labels required custom training of a language model for each use case. Newer technologies such as large language models have made it possible to generate accurate and normalized labels at scale, using only clinical reports and specific prompt engineering. The combination of automatically generated labels extracted and normalized from reports in conjunction with foundational image models provides a means to create labels for model training. This article provides a short history and review of the annotation and labeling process of medical images, from the traditional manual methods to the newest semiautomated methods that provide a more scalable solution for creating useful models more efficiently. Keywords: Feature Detection, Diagnosis, Semi-supervised Learning © RSNA, 2025.