An automated workflow based on hip shape improves personalized risk prediction for hip osteoarthritis in the CHECK study

Abstract

s u m m a r yObjective: To design an automated workflow for hip radiographs focused on joint shape and tests its prognostic value for future hip osteoarthritis. Design: We used baseline and 8-year follow-up data from 1,002 participants of the CHECK-study. The primary outcome was definite radiographic hip osteoarthritis (rHOA) (KellgreneLawrence grade 2 or joint replacement) at 8-year follow-up. We designed a method to automatically segment the hip joint from radiographs. Subsequently, we applied machine learning algorithms (elastic net with automated parameter optimization) to provide the Shape-Score, a single value describing the risk for future rHOA based solely on joint shape. We built and internally validated prediction models using baseline demographics, physical examination, and radiologists scores and tested the added prognostic value of the Shape-Score using Area-Under-the-Curve (AUC). Missing data was imputed by multiple imputation by chained equations. Only hips with pain in the corresponding leg were included. Results: 84% were female, mean age was 56 (±5.1) years, mean BMI 26.3 (±4.2). Of 1,044 hips with pain at baseline and complete follow-up, 143 showed radiographic osteoarthritis and 42 were replaced. 91.5% of the hips had follow-up data available. The Shape-Score was a significant predictor of rHOA (odds ratio per decimal increase 5.21, 95%-CI (3.74e7.24)). The prediction model using demographics, physical examination, and radiologists scores demonstrated an AUC of 0.795, 95%-CI (0.757e0.834). After addition of the Shape-Score the AUC rose to 0.864, 95%-CI (0.833e0.895). Conclusions: Our Shape-Score, automatically derived from radiographs using a novel machine learning workflow, may strongly improve risk prediction in hip osteoarthritis.

Subregional statistical shape modelling identifies lesser trochanter size as a possible risk factor for radiographic hip osteoarthritis, a cross-sectional analysis from the Osteoporotic Fractures in Men Study

“…Hip shape variations, in the form of developmental dysplasia of the hip (DDH) 6 , femoro-acetabular impingement (FAI) syndrome comprising cam and pincer morphologies measured geometrically 7 and similar morphologies measured via statistical shape modelling (SSM), have strong associations with hip OA 8,9 . Better knowledge of these shape variations and their origins could potentially offer new pathways to prediction 10,11 and prevention of hip OA, the later based on interventions that mediate the effects of hip shape 12 .…”

Artificial Learning and Machine Learning Decision Guidance Applications in Total Hip and Knee Arthroplasty: A Systematic Review

“… Author, year Pathology/Surgery ML algorithms Prediction outputs Patients in testing set (n) Avg. age %Female Data source Alam et al., 2019 [ 14 ] THA ANN, regression PROs/outcomes 10,000 — — Multicenter Bini et al., 2019 [ 45 ] TJA Cluster analysis PROs/outcomes — 63 68 Single institution Fontana et al., 2019 [ 32 ] TJA Regression, SVM, decision tree PROs/outcomes 2744 63 — Patient database Galivanche et al., 2019 [ 47 ] THA Boosting Adverse event/other complication 34,982 — — ACS-NSQIP database Gielis et al., 2020 [ 48 ] THA Regression PROs/outcomes 1044 55.9 87.3 CHECK cohort Gong et al., 2014 [49] TJA ANN, regression Adverse event/other complication — 69.6 53.3 Single institution Harris et al., 2019 [ 50 ] TJA Regression Adverse event/other complication, cardiovascular complication, postoperative mortality — 65.7 59.4 ACS-NSQIP database Hirvasniemi et al., 2019 [ 33 ] THA Regression PROs/outcomes 19...…”

“…TKA and total hip arthroplasty revisions and reoperations are also modeled with AI/ML algorithms in some studies, [ 15 , 16 , 21 , 64 ] as well as hospital readmissions [ 20 , 21 , 26 , 27 ]. In the postoperative period, AI/ML tools offer surgeons the ability to predict patients’ outcomes after surgery, including functional outcomes and PRO scores [ 14 , 32 , 33 , 43 , 45 , [48] , [53] , [54] , [57] , [58] , [59] , [61] ]. Postoperative pain has also been shown to be predicted with AI/ML, [ 43 , 53 , 56 , 55 ] including identification of patients at high risk for prolonged postoperative opioid prescriptions.…”

Convolutional Neural Network-Based Automatic Measurement of Joint Space Width to Predict Radiographic Severity and Progression of Knee Osteoarthritis

“…Furthermore, this class of models was recently shown to outperform another decision tree-based segmentation technique, BoneFinder 7,35 . Speci cally, our deep learning-based bone segmentation approach is superior to the existing approaches in a way that it produces continuous contour of the tibial plateau and femoral condyle rather than discrete landmarks 7,35,36 , and is capable of accurately identifying the relevant tibial contour for JSW measurements. This allows preservation of pixel-level boundary information in the tibiofemoral joint, hence bene cial to the extraction of ne-grained morphological details such as multiple JSWs.…”