Background:Knee osteoarthritis (OA), a leading cause of disability worldwide, can be difficult to define as its development is often insidious and involves different subgroups. We still lack robust prediction models that are able to guide clinical decisions and stratify OA patients according to risk of disease progression.Objectives:This study aimed at identifying the most important features of knee OA progressors. To this end, we used machine learning (ML) algorithms on a large set of subjects and features to develop advanced prediction models that provide high classification and prediction performance.Methods:Participants, features and outcomes were from the Osteoarthritis Initiative. Features were from baseline (1107), including articular knee tissues (135) assessed by quantitative MRI. OA progressors were ascertained by four outcomes: cartilage volume loss in medial plateau at 48 and 96 months (Prop_CV_48M, 96M); Kellgren-Lawrence (KL) grade ≥2; and medial joint space narrowing (JSN) ≥1 at 48 months. Subjects’ numbers were as follows: 1598 for the outcome Prop_CV_96M, 1044 for the Prop_CV_48M, and 1468 for each KL grade ≥2 at 48 months and JSN ≥1 at 48 months. Six feature selection models were used to identify the common features in each outcome. Six classification methods were applied to measure the accuracy of the selected features in classifying the subjects into progressors and non-progressors. Classification of the best features was done using auto-ML interface and the area under the curve (AUC). To prioritize the top features, Sparse Partial Least Square (sPLS) method was used.Results:For the classification of the best common features in each outcome, Multi-Layer Perceptron (MLP) achieved the highest AUC in Prop_CV_96M, KL, and JSN (0.80, 0.88, 0.95), and Gradient Boosting Machine (GBM) for Prop_CV_48M (0.70). sPLS revealed that the baseline top five features to predict knee OA progressors are the joint space width (JSW), mean cartilage thickness of peripheral, medial, and central tibial plateau, and JSN.Conclusion:This is the first time that such a comprehensive study was performed for identifying the best features and classification methods for knee OA progressors. Data revealed that early prediction of knee OA progression can be done with high accuracy and based on only a few features. This study identifies the baseline X-ray and MRI-based features as the most important for predicting knee OA progressors. These results could be used for the development of a tool enabling prediction of knee OA progressors.Acknowledgments:This work was supported in part by the Osteoarthritis Research Unit of the University of Montreal Hospital Research Centre; the Chair in Osteoarthritis, University of Montreal, (both from Montreal, Quebec, Canada); and the Computational Biology Laboratory, Laval University Hospital Research Center, (Québec, Quebec, Canada). A Jamshidi received a bursary from the Canada First Research Excellence Fund through TransMedTech Institute, (Montreal, Quebec, Canada).Disclosure of Interests:Afshin Jamshidi: None declared, Mickaël Leclercq: None declared, Aurelie Labbe: None declared, Jean-Pierre Pelletier Shareholder of: ArthroLab Inc., Grant/research support from: TRB Chemedica, Speakers bureau: TRB Chemedica and Mylan, François Abram Employee of: ArthroLab Inc., Arnaud Droit: None declared, Johanne Martel-Pelletier Shareholder of: ArthroLab Inc., Grant/research support from: TRB Chemedica
Comparison of Diagnostic Performance of Semi-Quantitative Knee Ultrasound and Knee Radiography with MRI: Oulu Knee Osteoarthritis Study