How AI May Transform Musculoskeletal Imaging

Role of Artificial Intelligence in Musculoskeletal Interventions

Artificial intelligence (AI) has rapidly emerged as a transformative force in musculoskeletal imaging and interventional radiology. This article explores how AI-based methods-including machine learning (ML) and deep learning (DL)-streamline diagnostic processes, guide interventions, and improve patient outcomes. Key applications discussed include ultrasound-guided procedures for joints, nerves, and tumor-targeted interventions, along with CT-guided biopsies and ablations, and fluoroscopy-guided facet joint and nerve block injections. AI-powered segmentation algorithms, real-time feedback systems, and dose-optimization protocols collectively enable greater precision, operator consistency, and patient safety. In rehabilitation, AI-driven wearables and predictive models facilitate personalized exercise programs that can accelerate recovery and enhance long-term function. While challenges persist-such as data standardization, regulatory hurdles, and clinical adoption-ongoing interdisciplinary collaboration, federated learning models, and the integration of genomic and environmental data hold promise for expanding AI's capabilities. As personalized medicine continues to advance, AI is poised to refine risk stratification, reduce radiation exposure, and support minimally invasive, patient-specific interventions, ultimately reshaping musculoskeletal care from early detection and diagnosis to individualized treatment and rehabilitation.

Deep Learning Models to Detect Anterior Cruciate Ligament Injury on MRI: A Comprehensive Review

Magnetic resonance imaging (MRI) is routinely used to confirm the suspected diagnosis of anterior cruciate ligament (ACL) injury. Recently, many studies explored the role of artificial intelligence (AI) and deep learning (DL), a sub-category of AI, in the musculoskeletal field and medical imaging. The aim of this study was to review the current applications of DL models to detect ACL injury on MRI, thus providing an updated and critical synthesis of the existing literature and identifying emerging trends and challenges in the field. A total of 23 relevant articles were identified and included in the review. Articles originated from 10 countries, with China having the most contributions (n = 9), followed by the United State of America (n = 4). Throughout the article, we analyzed the concept of DL in ACL tears and provided examples of how these tools can impact clinical practice and patient care. DL models for MRI detection of ACL injury reported high values of accuracy, especially helpful for less experienced clinicians. Time efficiency was also demonstrated. Overall, the deep learning models have proven to be a valid resource, although still requiring technological developments for implementation in daily practice.

Improving radiologist detection of meniscal abnormality on undersampled, deep learning reconstructed knee MRI

Background

Accurate interpretation of meniscal anomalies on knee MRI is critical for diagnosis and treatment planning, with artificial intelligence emerging as a promising tool to support and enhance this process through automated anomaly detection.

Purpose

To evaluate the impact of an artificial intelligence (AI) anomaly detection assistant on radiologists' interpretation of meniscal anomalies in undersampled, deep learning (DL)-reconstructed knee MRI and assess the relationship between reconstruction quality metrics and anomaly detection performance.

Materials and Methods

This retrospective study included 947 knee MRI examinations; 51 were excluded for poor image quality, leaving 896 participants (mean age, 44.7 ± 15.3 years; 472 women). Using 8-fold undersampled data, DL-based reconstructed images were generated. An object detection model was trained on original, fully sampled images and evaluated on 1 original and 14 DL-reconstructed test sets to identify meniscal lesions. Standard reconstruction metrics (normalized root mean square error, peak signal-to-noise ratio, and structural similarity index) and anomaly detection metrics (mean average precision, F1 score) were quantified and compared. Two radiologists independently reviewed a stratified sample of 50 examinations unassisted and assisted with AI-predicted anomaly boxes. McNemar's test evaluated differences in diagnostic performance; Cohen's kappa assessed interrater agreement.

Results

On the original images, the anomaly detection model achieved the following: 70.53% precision, 72.17% recall, 63.09% mAP, and a 71.34% F1 score. Comparing performance among the undersampled reconstruction datasets, box-based reconstruction metrics showed better correlation with detection performance than traditional image-based metrics (mAP to box-based SSIM, r = 0.81, P < .01; mAP to image-based SSIM, r = 0.64, P = .01). In 50 participants, AI assistance improved radiologists' accuracy on reconstructed images. Sensitivity increased from 77.27% (95% CI, 65.83-85.72; 51/66) to 80.30% (95% CI, 69.16-88.11; 53/66), and specificity improved from 88.46% (95% CI, 83.73-91.95; 207/234) to 90.60% (95% CI, 86.18-93.71; 212/234) (P < .05).

Conclusion

AI-assisted meniscal anomaly detection enhanced radiologists' interpretation of undersampled, DL-reconstructed knee MRI. Anomaly detection may serve as a complementary tool alongside other reconstruction metrics to assess the preservation of clinically important features in reconstructed images, warranting further investigation.

The scientific evidence of commercial AI products for MRI acceleration: a systematic review

OBJECTIVES

This study explores the methods employed by commercially available AI products to accelerate MRI protocols and investigates the strength of their diagnostic image quality assessment.

MATERIALS AND METHODS

All commercial AI products for MRI acceleration were identified from the exhibitors presented at the RSNA 2023 and ECR 2024 annual meetings. Peer-reviewed scientific articles describing validation of clinical performance were searched for each product. Information was extracted regarding the MRI acceleration technique, achieved acceleration, diagnostic performance metrics, test cohort, and hallucinatory artifacts. The strength of the diagnostic image quality was assessed using scientific evidence levels ranging from "product's technical feasibility for clinical purposes" to "product's economic impact on society".

RESULTS

Out of 1046 companies, 14 products of 14 companies were included. No scientific articles were found for four products (29%). For the remaining ten products (71%), 21 articles were retrieved. Four acceleration methods were identified: noise reduction, raw data reconstruction, personalized scanning protocols, and synthetic image generation. Only a limited number of articles prospectively demonstrated impact on patient outcomes (n = 4, 19%), and no articles discussed an evaluation in a prospective cohort of > 100 patients or performed an economic analysis. None of the articles performed an analysis of hallucinatory artifacts.

CONCLUSION

Currently, commercially available AI products for MRI acceleration can be categorized into four main methods. The acceleration methods lack prospective scientific evidence on clinical performance in large cohorts and economic analysis, which would help to get a better insight into their diagnostic performance and enable safe and effective clinical implementation.

KEY POINTS

Question There is a growing interest in AI products that reduce MRI scan time, but an overview of these methods and their scientific evidence is missing. Findings Only a limited number of articles (n = 4, 19%) prospectively demonstrated the impact of the software for accelerating MRI on diagnostic performance metrics. Clinical relevance Although various commercially available products shorten MRI acquisition time, more studies in large cohorts are needed to get a better insight into the diagnostic performance of AI-constructed MRI.

Arthroscopy-validated Diagnostic Performance of 7-Minute Five-Sequence Deep Learning Super-Resolution 3-T Shoulder MRI

Background Deep learning (DL) methods enable faster shoulder MRI than conventional methods, but arthroscopy-validated evidence of good diagnostic performance is scarce. Purpose To validate the clinical efficacy of 7-minute threefold parallel imaging (PIx3)-accelerated DL super-resolution shoulder MRI against arthroscopic findings. Materials and Methods Adults with painful shoulder conditions who underwent PIx3-accelerated DL super-resolution 3-T shoulder MRI and arthroscopy between March and November 2023 were included in this retrospective study. Seven radiologists independently evaluated the MRI scan quality parameters and the presence of artifacts (Likert scale rating ranging from 1 [very bad/severe] to 5 [very good/absent]) as well as the presence of rotator cuff tears, superior and anteroinferior labral tears, biceps tendon tears, cartilage defects, Hill-Sachs lesions, Bankart fractures, and subacromial-subdeltoid bursitis. Interreader agreement based on κ values was evaluated, and diagnostic performance testing was conducted. Results A total of 121 adults (mean age, 55 years ± 14 [SD]; 75 male) who underwent MRI and arthroscopy within a median of 39 days (range, 1-90 days) were evaluated. The overall image quality was good (median rating, 4 [IQR, 4-4]), with high reader agreement (κ ≥ 0.86). Motion artifacts and image noise were minimal (rating of 4 [IQR, 4-4] for each), and reconstruction artifacts were absent (rating of 5 [IQR, 5-5]). Arthroscopy-validated abnormalities were detected with good or better interreader agreement (κ ≥ 0.68). The sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve were 89%, 90%, 89%, and 0.89, respectively, for supraspinatus-infraspinatus tendon tears; 82%, 63%, 68%, and 0.68 for subscapularis tendon tears; 93%, 73%, 86%, and 0.83 for superior labral tears; 100%, 100%, 100%, and 1.00 for anteroinferior labral tears; 68%, 90%, 82%, and 0.80 for biceps tendon tears; 42%, 93%, 81%, and 0.64 for cartilage defects; 93%, 99%, 98%, and 0.94 for Hill-Sachs deformities; 100%, 99%, 99%, and 1.00 for osseous Bankart lesions; and 97%, 63%, 92%, and 0.80 for subacromial-subdeltoid bursitis. Conclusion Seven-minute PIx3-accelerated DL super-resolution 3-T shoulder MRI has good diagnostic performance for diagnosing tendinous, labral, and osteocartilaginous abnormalities. © RSNA, 2025 Supplemental material is available for this article. See also the editorial by Tuite in this issue.

Radiology resident competency in orthopedic trauma detection in simulated on-call scenarios

PURPOSE

To evaluate radiology residents' ability to accurately identify three specific types of orthopedic trauma using radiographic imaging within a simulated on-call environment.

METHODS

We utilized the Wisdom in Diagnostic Imaging Emergent/Critical Care Radiology Simulation (WIDI SIM) to assess residents' preparedness for independent radiology call. The simulation included 65 cases, with three focusing on orthopedic trauma: sacral ala, femoral neck, and pediatric tibial/Toddler's fractures. Faculty graded residents' responses using a standardized 10-point rubric and categorized errors as observational (failing to identify key findings) or interpretive (incorrect conclusions despite correct identification of findings).

RESULTS

321 residents evaluated sacral ala fracture radiographs and received an average score of 1.29/10, with 8.71 points lost to observational errors. Only 6% produced effective reports (scores ≥ 7), while 80% made critical errors (scores < 2). For femoral neck fracture CT images (n = 316 residents), the average score was 2.48/10, with 6.71 points lost to observational errors. 25% produced effective reports, and 66% made critical errors. Pediatric tibial/Toddler's fracture radiographs (n = 197 residents) yielded an average score of 2.94/10, with 6.60 points lost to observational errors. 29% generated effective reports, while 71% made critical errors.

CONCLUSION

Radiology residents demonstrated significant difficulty in identifying these orthopedic trauma cases, with errors primarily attributed to observational deficiencies. These findings suggest a need for targeted educational interventions in radiology residency programs to improve the identification of these fractures.

Deep Learning Superresolution for Simultaneous Multislice Parallel Imaging-Accelerated Knee MRI Using Arthroscopy Validation

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.

Artificial intelligence in commercial fracture detection products: a systematic review and meta-analysis of diagnostic test accuracy

Conventional radiography (CR) is primarily utilized for fracture diagnosis. Artificial intelligence (AI) for CR is a rapidly growing field aimed at enhancing efficiency and increasing diagnostic accuracy. However, the diagnostic performance of commercially available AI fracture detection solutions (CAAI-FDS) for CR in various anatomical regions, their synergy with human assessment, as well as the influence of industry funding on reported accuracy are unknown. Peer-reviewed diagnostic test accuracy (DTA) studies were identified through a systematic review on Pubmed and Embase. Diagnostic performance measures were extracted especially for different subgroups such as product, type of rater (stand-alone AI, human unaided, human aided), funding, and anatomical region. Pooled measures were obtained with a bivariate random effects model. The impact of rater was evaluated with comparative meta-analysis. Seventeen DTA studies of seven CAAI-FDS analyzing 38,978 x-rays with 8,150 fractures were included. Stand-alone AI studies (n = 15) evaluated five CAAI-FDS; four with good sensitivities (> 90%) and moderate specificities (80-90%) and one with very poor sensitivity (< 60%) and excellent specificity (> 95%). Pooled sensitivities were good to excellent, and specificities were moderate to good in all anatomical regions (n = 7) apart from ribs (n = 4; poor sensitivity / moderate specificity) and spine (n = 4; excellent sensitivity / poor specificity). Funded studies (n = 4) had higher sensitivity (+ 5%) and lower specificity (-4%) than non-funded studies (n = 11). Sensitivity did not differ significantly between stand-alone AI and human AI aided ratings (p = 0.316) but specificity was significantly higher the latter group (p < 0.001). Sensitivity was significant lower in human unaided compared to human AI aided respectively stand-alone AI ratings (both p ≤ 0.001); specificity was higher in human unaided ratings compared to stand-alone AI (p < 0.001) and showed no significant differences AI aided ratings (p = 0.316). The study demonstrates good diagnostic accuracy across most CAAI-FDS and anatomical regions, with the highest performance achieved when used in conjunction with human assessment. Diagnostic accuracy appears lower for spine and rib fractures. The impact of industry funding on reported performance is small.

Artificial intelligence for automated detection and measurements of carpal instability signs on conventional radiographs

OBJECTIVES

To develop and validate an artificial intelligence (AI) system for measuring and detecting signs of carpal instability on conventional radiographs.

MATERIALS AND METHODS

Two case-control datasets of hand and wrist radiographs were retrospectively acquired at three hospitals (hospitals A, B, and C). Dataset 1 (2178 radiographs from 1993 patients, hospitals A and B, 2018-2019) was used for developing an AI system for measuring scapholunate (SL) joint distances, SL and capitolunate (CL) angles, and carpal arc interruptions. Dataset 2 (481 radiographs from 217 patients, hospital C, 2017-2021) was used for testing, and with a subsample (174 radiographs from 87 patients), an observer study was conducted to compare its performance to five clinicians. Evaluation metrics included mean absolute error (MAE), sensitivity, and specificity.

RESULTS

Dataset 2 included 258 SL distances, 189 SL angles, 191 CL angles, and 217 carpal arc labels obtained from 217 patients (mean age, 51 years ± 23 [standard deviation]; 133 women). The MAE in measuring SL distances, SL angles, and CL angles was respectively 0.65 mm (95%CI: 0.59, 0.72), 7.9 degrees (95%CI: 7.0, 8.9), and 5.9 degrees (95%CI: 5.2, 6.6). The sensitivity and specificity for detecting arc interruptions were 83% (95%CI: 74, 91) and 64% (95%CI: 56, 71). The measurements were largely comparable to those of the clinicians, while arc interruption detections were more accurate than those of most clinicians.

CONCLUSION

This study demonstrates that a newly developed automated AI system accurately measures and detects signs of carpal instability on conventional radiographs.

CLINICAL RELEVANCE STATEMENT

This system has the potential to improve detections of carpal arc interruptions and could be a promising tool for supporting clinicians in detecting carpal instability.

Reading Times of Common Musculoskeletal MRI Examinations: A Survey Study

BACKGROUND

The workload of musculoskeletal radiologists has come under pressure. Our objective was to estimate the reading times of common musculoskeletal MRI examinations.

METHODS

A total of 144 radiologists were asked to estimate reading times (including interpretation and reporting) for MRI of the shoulder, elbow, wrist, hip, knee, and ankle. Multivariate linear regression analyses were performed.

RESULTS

Reported median reading times with interquartile range (IQR) for the shoulder, elbow, wrist, hip, knee, and ankle were 10 (IQR 6-14), 10 (IQR 6-14), 11 (IQR 7.5-14.5), 10 (IQR 6.6-13.4), 8 (IQR 4.6-11.4), and 10 (IQR 6.5-13.5) min, respectively. Radiologists aged 35-44 years reported shorter reading times for the shoulder (β coefficient [β] = B-3.412, p = 0.041), hip (β = -3.596, p = 0.023), and knee (β = -3.541, p = 0.013) than radiologists aged 45-54 years. Radiologists not working in an academic/teaching hospital reported shorter reading times for the hip (β = -3.611, p = 0.025) and knee (β = -3.038, p = 0.035). Female radiologists indicated longer reading times for all joints (β of 2.592 to 5.186, p ≤ 0.034). Radiologists without musculoskeletal fellowship training indicated longer reading times for the shoulder (β = 4.604, p = 0.005), elbow (β = 3.989, p = 0.038), wrist (β = 4.543, p = 0.014), and hip (β = 2.380, p = 0.119). Radiologists with <5 years of post-residency experience indicated longer reading times for all joints (β of 5.355 to 6.984, p ≤ 0.045), and radiologists with 5-10 years of post-residency experience reported longer reading time for the knee (β = 3.660, p = 0.045) than those with >10 years of post-residency experience.

CONCLUSIONS

There is substantial variation among radiologists in reported reading times for common musculoskeletal MRI examinations. Several radiologist-related determinants appear to be associated with reading speed, including age, gender, hospital type, training, and experience.