Computer‐Aided Detection AI Reduces Interreader Variability in Grading Hip Abnormalities With MRI

Abstract

Background Accurate interpretation of hip MRI is time‐intensive and difficult, prone to inter‐ and intrareviewer variability, and lacks a universally accepted grading scale to evaluate morphological abnormalities. Purpose To 1) develop and evaluate a deep‐learning‐based model for binary classification of hip osteoarthritis (OA) morphological abnormalities on MR images, and 2) develop an artificial intelligence (AI)‐based assist tool to find if using the model predictions improves interreader agreement in hip grading. Study Type Retrospective study aimed to evaluate a technical development. Population A total of 764 MRI volumes (364 patients) obtained from two studies (242 patients from LASEM [FORCe] and 122 patients from UCSF), split into a 65–25–10% train, validation, test set for network training. Field Strength/Sequence 3T MRI, 2D T2 FSE, PD SPAIR. Assessment Automatic binary classification of cartilage lesions, bone marrow edema‐like lesions, and subchondral cyst‐like lesions using the MRNet, interreader agreement before and after using network predictions. Statistical Tests Receiver operating characteristic (ROC) curve, area under curve (AUC), specificity and sensitivity, and balanced accuracy. Results For cartilage lesions, bone marrow edema‐like lesions and subchondral cyst‐like lesions the AUCs were: 0.80 (95% confidence interval [CI] 0.65, 0.95), 0.84 (95% CI 0.67, 1.00), and 0.77 (95% CI 0.66, 0.85), respectively. The sensitivity and specificity of the radiologist for binary classification were: 0.79 (95% CI 0.65, 0.93) and 0.80 (95% CI 0.59, 1.02), 0.40 (95% CI –0.02, 0.83) and 0.72 (95% CI 0.59, 0.86), 0.75 (95% CI 0.45, 1.05) and 0.88 (95% CI 0.77, 0.98). The interreader balanced accuracy increased from 53%, 71% and 56% to 60%, 73% and 68% after using the network predictions and saliency maps. Data Conclusion We have shown that a deep‐learning approach achieved high performance in clinical classification tasks on hip MR images, and that using the predictions from the deep‐learning model improved the interreader agreement in all pathologies. Level of Evidence 3 Technical Efficacy Stage 1 J. Magn. Reson. Imaging 2020;52:1163–1172.

Deep learning for brain disorders: from data processing to disease treatment

“…Conversely, specific patterns can be identified for each patient to propose individual follow-up and treatment [144,154]. Additionally, a study outside the scope of this review proposed to use saliency maps for clinical guidance, to improve inter-radiologist diagnosis variability [238]. Finally, some authors strove to provide a framework that is interpretable because of its architecture [143,196] or constraints [200].…”

Automated delineation of orbital abscess depicted on CT scan using deep learning

“…As compared with the manual delineation, the use of computerized schemes may enable a more accurate, objective, and efficient quantification of an orbital abscess and ultimately improve pediatric care. As demonstrated by other investigations, [21][22][23][24] a computer approach may potentially reduce the inter-and intra-reader variability in assessing and quantifying diseases.…”

Editorial for “Computer‐Aided Detection AI Reduces Interreader Variability in Grading Hip Abnormalities With MRI”

“…Deep learning for magnetic resonance imaging (MRI) has been explored from the field of image acquisition to image reconstruction and image postprocessing . The study by Tibrewala et al in this issue of the Journal of Magnetic Resonance Imaging focuses on the application of deep learning in musculoskeletal imaging and extends it into hip OA detection.…”

“…One of the crucial points in developing a CADe with deep learning is the access to a large, reliable and validated clinical image database with annotations. In the study of Tibrewala et al, a large cohort of patients with femoroacetabular impingement and hip OA were included and the abnormalities such as cartilage lesions, bone marrow edema‐like lesions, and subchondral cyst were evaluated for the purpose of deep learning classification. In this retrospective study, the input data for processing and network training were carefully designed by using the convolutional neural network (CNN) where a binary classification prediction was as an output.…”

“…The deep neural networks will be very useful when accurate decision‐making is needed based on complicated datasets . Although Tibrewala et al evaluated the MRNet architecture for a binary classification and provided a good picture of the performance on hip MRI, as stated in the study, a deep multiview classification is indeed needed, which might have an impact on further stratifying lesion types. Besides that, 2D coronal hip MR imaging datasets were used in this study as input for network training, multiview or 3D MR imaging scans might be beneficial to obtain highly accurate classification of hip abnormality.…”