Faculty Bibliography

BACKGROUND:Alterations in periarticular lipid composition are implicated in musculoskeletal diseases, yet short-term reliability of MRI-based triglyceride composition mapping in the knee is not fully established. PURPOSE/OBJECTIVE:To evaluate 1-week repeatability of proton-density fat fraction (PDFF) and triglyceride fatty-acid composition-saturated (SFA), monounsaturated (MUFA), and polyunsaturated (PUFA)-in periarticular knee tissues. STUDY TYPE/METHODS:Prospective. POPULATION/METHODS:). FIELD STRENGTH/SEQUENCE/UNASSIGNED:3T; 12-echo 3D spoiled gradient-echo acquisition for chemical shift-encoded fat quantification and a proton density-weighted SPACE sequence for segmentation (0.6 mm isotropic). ASSESSMENT/RESULTS:Participants underwent repeated MRI 1 week apart. Femoral and tibial bone marrow, patella, Hoffa's fat pad, prefemoral fat pad, quadriceps fat pad, posterior fat pad, and subcutaneous adipose tissue were segmented and rigidly aligned. Voxelwise spectral fitting was used to estimate PDFF and fatty acid composition, including SFA, MUFA, and PUFA components. Repeatability metrics included bias, within-subject standard deviation (wSD), within-subject coefficient of variation (wCV%), coefficient of repeatability, and intraclass correlation coefficient (ICC). STATISTICAL TESTS/METHODS:Paired t-tests assessed systematic differences (α = 0.05); ICCs used a two-way random-effects, absolute-agreement model (ICC(2,1)). RESULTS:PDFF showed lowest variability across all regions (wCV: 1.5%-5.9%; ICC: 0.33-0.96). SFA demonstrated similar stability (wCV: 2.4%-12.6%; ICC: 0.19-0.87). MUFA exhibited anatomy-dependent reliability (wCV: 4.1%-21.1%; ICC: 0.17-0.97), with highest repeatability in subcutaneous adipose tissue (ICC: 0.97) and Hoffa's fat pad (ICC: 0.85). PUFA displayed the greatest variability (wCV: 3.6%-52.8%; ICC: 0.10-0.94), with the greatest instability in periarticular fat pads. No paired comparisons were significant (all p > 0.05; range p = 0.14-0.98). Regional ordering remained consistent across sessions. DATA CONCLUSION/CONCLUSIONS:A 12-echo chemical shift-encoded MRI protocol provides repeatable PDFF and SFA measurements over 1 week. MUFA reliability varies by tissue, while PUFA remains least stable. EVIDENCE LEVEL/METHODS:2 (Prospective cohort). TECHNICAL EFFICACY STAGE/UNASSIGNED:2 (Reproducibility/feasibility evaluation).

Accurate disease prognosis is essential for patient care but is often hindered by the scarcity of longitudinal data. This study explores deep learning training strategies that utilize large, accessible diagnostic datasets to pretrain models aimed at predicting future disease progression in knee osteoarthritis (OA), Alzheimer's disease (AD), and breast cancer (BC). While diagnostic pretraining improves prognostic task performance, naive fine-tuning for prognosis can cause 'catastrophic forgetting,' where the model's original diagnostic accuracy degrades, a significant patient safety concern in real-world settings. To address this, we propose a sequential learning strategy with experience replay. We used cohorts with knee radiographs, brain MRIs, and digital mammograms to predict 4-year structural worsening in OA, 2-year cognitive decline in AD, and 5-year cancer diagnosis in BC. Our results showed that diagnostic pretraining on larger datasets improved prognosis model performance compared to standard baselines, boosting both the Area Under the Receiver Operating Characteristic curve (AUROC) (e.g., Knee OA external: 0.770 vs 0.747; Breast Cancer: 0.874 vs 0.848) and the Area Under the Precision-Recall Curve (AUPRC) (e.g., Alzheimer's Disease: 0.752 vs 0.683). Additionally, a sequential learning approach with experience replay achieved prognostic performance comparable to dedicated single-task models (e.g., Breast Cancer AUROC 0.876 vs 0.874) while also preserving diagnostic ability. This method maintained high diagnostic accuracy (e.g., Breast Cancer Balanced Accuracy 50.4% vs 50.9% for a dedicated diagnostic model), unlike simpler multitask methods prone to catastrophic forgetting (e.g., 37.7%). Our findings show that leveraging large diagnostic datasets is a reliable and data-efficient way to enhance prognostic models while maintaining essential diagnostic skills.

Musculoskeletal (MSK) imaging was among the first radiology subspecialties to adopt artificial intelligence (AI), with applications now spanning the entire MSK workflow, from image acquisition to reporting. Deep learning-based reconstruction protocols can accelerate MRI by reducing scan times and artefacts, improving accessibility in high-volume and resource-limited settings. Furthermore, AI interpretation tools have demonstrated strong performance in fracture detection, assessment of meniscal and ligament tears, bone tumour characterization and automated quantification of measurements, supporting greater diagnostic consistency across radiologists with varying experience levels. Large language models (LLMs) extend AI's impact beyond image analysis by simplifying reports for patients, automating classification systems, and streamlining clinical communication. Despite these advances, important challenges remain. Integration of AI into already established clinical workflows can be complex, and requires robust technical solutions, regulatory compliance, and strategies to maintain radiologist oversight. Questions of liability, cost-effectiveness, and the role of AI in medical education further underscore the need for careful implementation. AI is poised to fundamentally reshape MSK radiology by enhancing efficiency, improving diagnostic accuracy, and enabling more patient-centred communication. To fully realize this potential, adoption must balance innovation with safety, equity, and sustainability, ensuring AI remains a trusted assistive tool that strengthens rather than replaces radiologist expertise.

Accurate disease prognosis is essential for patient care but is often hindered by the lack of long-term data. This study explores deep learning training strategies that utilize large, accessible diagnostic datasets to pretrain models aimed at predicting future disease progression in knee osteoarthritis (OA), Alzheimer's disease (AD), and breast cancer (BC). While diagnostic pretraining improves prognostic task performance, naive fine-tuning for prognosis can cause 'catastrophic forgetting,' where the model's original diagnostic accuracy degrades, a significant patient safety concern in real-world settings. To address this, we propose a sequential learning strategy with experience replay. We used cohorts with knee radiographs, brain MRIs, and digital mammograms to predict 4-year structural worsening in OA, 2-year cognitive decline in AD, and 5-year cancer diagnosis in BC. Our results showed that diagnostic pretraining on larger datasets improved prognosis model performance compared to standard baselines, boosting both the Area Under the Receiver Operating Characteristic curve (AUROC) (e.g., Knee OA external: 0.77 vs 0.747; Breast Cancer: 0.874 vs 0.848) and the Area Under the Precision-Recall Curve (AUPRC) (e.g., Alzheimer's Disease: 0.752 vs 0.683). Additionally, a sequential learning approach with experience replay achieved prognostic performance comparable to dedicated single-task models (e.g., Breast Cancer AUROC 0.876 vs 0.874) while also preserving diagnostic ability. This method maintained high diagnostic accuracy (e.g., Breast Cancer Balanced Accuracy 50.4% vs 50.9% for a dedicated diagnostic model), unlike simpler multitask methods prone to catastrophic forgetting (e.g., 37.7%). Our findings show that leveraging large diagnostic datasets is a reliable and data-efficient way to enhance prognostic models while maintaining essential diagnostic skills.

BACKGROUND/UNASSIGNED:Large language models (LLMs) have shown promising abilities in text-based clinical tasks but they do not inherently interpret medical images such as knee radiographs. PURPOSE/UNASSIGNED:To develop a human-artificial intelligence interactive diagnostic approach, named radiology generative pretrained transformer (RadGPT), aimed at assisting and synergizing with human users for the interpretation of knee radiological images. MATERIALS AND METHODS/UNASSIGNED:A total of 22 512 knee roentgen ray images and reports were retrieved from Massachusetts General Hospital; 80% of these were used for model training and 10% were used for model testing and validation, respectively. Fifteen diagnostic imaging features (eg, osteoarthritis, effusion, joint space narrowing, osteophyte) were selected to label images based on their high frequency and clinical relevance in the retrieved official reports. Area under the curve scores were calculated for each feature to assess the diagnostic performance. To evaluate the quality of the generated medical text, historical clinical reports were used as the reference text. Several metrics for text generation tasks are applied, including BiLingual Evaluation Understudy, Recall-Oriented Understudy for Gisting Evaluation, Metric for Evaluation of Translation with Explicit Ordering, and Semantic Propositional Image Caption Evaluation. RESULTS/UNASSIGNED:RadGPT, in collaboration with human users, achieved area under the curve scores ranging from 0.76 for osteonecrosis to 0.91 for arthroplasty across 15 diagnostic categories for knee conditions. Compared with the baseline LLM method, RadGPT achieved higher scores, specifically 0.18 in BiLingual Evaluation Understudy score, 0.30 in Recall-Oriented Understudy for Gisting Evaluation-L, 0.10 in Metric for Evaluation of Translation with Explicit Ordering, and 0.15 in Semantic Propositional Image Caption Evaluation, which is significantly higher than the baseline LLM method, demonstrating good linguistic overlap and clinical consistency with the reference reports. CONCLUSION/UNASSIGNED:RadGPT has achieved advanced results in knee roentgen ray image feature recognition, illustrating the potential of LLMs in medical image interpretation. The study establishes a training protocol for developing artificial intelligence-assisted tools specifically focusing on the diagnosis and interpretation of knee radiological images.

BACKGROUND:The methods for predicting time-to-total knee replacement (TKR) do not provide enough information to make robust and accurate predictions. PURPOSE/OBJECTIVE:Develop and evaluate an artificial intelligence-based model for predicting time-to-TKR by analyzing longitudinal knee data and identifying key features associated with accelerated knee osteoarthritis progression. METHODS:A total of 547 subjects underwent TKR in the Osteoarthritis Initiative over nine years, and their longitudinal data was used for model training and testing. 518 and 164 subjects from Multi-Center Osteoarthritis Study and internal hospital data were used for external testing, respectively. The clinical variables, magnetic resonance (MR) images, radiographs, and quantitative and semi-quantitative assessments from images were analyzed. Deep learning (DL) models were used to extract features from radiographs and MR images. DL features were combined with clinical and image assessment features for survival analysis. A Lasso Cox feature selection method combined with a random survival forest model was used to estimate time-to-TKR. RESULTS:Utilizing only clinical variables for time-to-TKR predictions provided the estimation accuracy of 60.4% and C-index of 62.9%. Combining DL features extracted from radiographs, MR images with clinical, quantitative, and semi-quantitative image assessment features achieved the highest accuracy of 73.2%, (p=.001) and C-index of 77.3% for predicting time-to-TKR. CONCLUSIONS:The proposed predictive model demonstrated the potential of DL models and multimodal data fusion in accurately predicting time-to-TKR surgery that may help assist physicians to personalize treatment strategies and improve patient outcomes.

Background Deep learning (DL) methods can improve accelerated MRI but require validation against an independent reference standard to ensure robustness and accuracy. Purpose To validate the diagnostic performance of twofold-simultaneous-multislice (SMSx2) twofold-parallel-imaging (PIx2)-accelerated DL superresolution MRI in the knee against conventional SMSx2-PIx2-accelerated MRI using arthroscopy as the reference standard. Materials and Methods Adults with painful knee conditions were prospectively enrolled from December 2021 to October 2022. Participants underwent fourfold SMSx2-PIx2-accelerated standard-of-care and investigational DL superresolution MRI at 3 T. Seven radiologists independently evaluated the MRI examinations for overall image quality (using Likert scale scores: 1, very bad, to 5, very good) and the presence or absence of meniscus and ligament tears. Articular cartilage was categorized as intact, or partial or full-thickness defects. Statistical analyses included interreader agreements (Cohen κ and Gwet AC2) and diagnostic performance testing used area under the receiver operating characteristic curve (AUC) values. Results A total of 116 adults (mean age, 45 years ± 15 [SD]; 74 men) who underwent arthroscopic surgery within 38 days ± 22 were evaluated. Overall image quality was better for DL superresolution MRI (median Likert score, 5; range, 3-5) than conventional MRI (median Likert score, 4; range, 3-5) (P < .001). Diagnostic performances of conventional versus DL superresolution MRI were similar for medial meniscus tears (AUC, 0.94 [95% CI: 0.89, 0.97] vs 0.94 [95% CI: 0.90, 0.98], respectively; P > .99), lateral meniscus tears (AUC, 0.85 [95% CI: 0.78, 0.91] vs 0.87 [95% CI: 0.81, 0.94], respectively; P = .96), and anterior cruciate ligament tears (AUC, 0.98 [95% CI: 0.93, >0.99] vs 0.98 [95% CI: 0.93, >0.99], respectively; P > .99). DL superresolution MRI (AUC, 0.78; 95% CI: 0.75, 0.81) had higher diagnostic performance than conventional MRI (AUC, 0.71; 95% CI: 0.67, 0.74; P = .002) for articular cartilage lesions. DL superresolution MRI did not introduce hallucinations or erroneously omit abnormalities. Conclusion Compared with conventional SMSx2-PIx2-accelerated MRI, fourfold SMSx2-PIx2-accelerated DL superresolution MRI in the knee provided better image quality, similar performance for detecting meniscus and ligament tears, and improved performance for depicting articular cartilage lesions. © RSNA, 2025 Supplemental material is available for this article. See also the editorial by Nevalainen in this issue.

PURPOSE/UNASSIGNED:Accurately predicting the expected duration of time until total knee replacement (time-to-TKR) is crucial for patient management and health care planning. Predicting when surgery may be needed, especially within shorter windows like 3 years, allows clinicians to plan timely interventions and health care systems to allocate resources more effectively. Existing models lack the precision for such time-based predictions. A survival analysis model for predicting time-to-TKR was developed using features from medical images and clinical measurements. METHODS/UNASSIGNED:From the Osteoarthritis Initiative dataset, all knees with clinical variables, MRI scans, radiographs, and quantitative and semiquantitative assessments from images were identified. This resulted in 895 knees that underwent TKR within the 9-year follow-up period, as specified by the Osteoarthritis Initiative study design, and 786 control knees that did not undergo TKR (right-censored, indicating their status beyond the 9-year follow-up is unknown). These knees were used for model training and testing. Additionally, 518 and 164 subjects from the Multi-Center Osteoarthritis Study and internal hospital data were used for external testing, respectively. Deep learning models were utilized to extract features from radiographs and MR scans. Extracted features, clinical variables, and image assessments were used in survival analysis with Lasso Cox feature selection and a random survival forest model to predict time-to-TKR. RESULTS/UNASSIGNED:The proposed model exhibited strong discrimination power by integrating self-supervised deep learning features with clinical variables (eg, age, body mass index, pain score) and image assessment measurements (eg, Kellgren-Lawrence grade, joint space narrowing, bone marrow lesion size, cartilage morphology) from multiple modalities. The model achieved an area under the curve of 94.5 (95% CI, 94.0-95.1) for predicting the time-to-TKR. CONCLUSIONS/UNASSIGNED:The proposed model demonstrated the potential of self-supervised learning and multimodal data fusion in accurately predicting time-to-TKR that may assist physicians to develop personalize treatment strategies.

Different pathologies of the hip are characterized by the abnormal shape of the bony structures of the joint, namely the femur and the acetabulum. Three-dimensional (3D) models of the hip can be used for diagnosis, biomechanical simulation, and planning of surgical treatments. These models can be generated by building 3D surfaces of the joint's structures segmented on magnetic resonance (MR) images. Deep learning can avoid time-consuming manual segmentations, but its performance depends on the amount and quality of the available training data. Data augmentation and transfer learning are two approaches used when there is only a limited number of datasets. In particular, data augmentation can be used to artificially increase the size and diversity of the training datasets, whereas transfer learning can be used to build the desired model on top of a model previously trained with similar data. This study investigates the effect of data augmentation and transfer learning on the performance of deep learning for the automatic segmentation of the femur and acetabulum on 3D MR images of patients diagnosed with femoroacetabular impingement. Transfer learning was applied starting from a model trained for the segmentation of the bony structures of the shoulder joint, which bears some resemblance to the hip joint. Our results suggest that data augmentation is more effective than transfer learning, yielding a Dice similarity coefficient compared to ground-truth manual segmentations of 0.84 and 0.89 for the acetabulum and femur, respectively, whereas the Dice coefficient was 0.78 and 0.88 for the model based on transfer learning. The Accuracy for the two anatomical regions was 0.95 and 0.97 when using data augmentation, and 0.87 and 0.96 when using transfer learning. Data augmentation can improve the performance of deep learning models by increasing the diversity of the training dataset and making the models more robust to noise and variations in image quality. The proposed segmentation model could be combined with radiomic analysis for the automatic evaluation of hip pathologies.