Novel classification of knee osteoarthritis severity based on spatiotemporal gait analysis.

Objective: To describe a novel classification method for knee osteoarthritis (OA) based on spatiotemporal gait analysis.

Methods: Gait analysis was initially performed on 2911 knee OA patients. Females and males were analyzed separately because of the influence of body height on spatiotemporal parameters. The analysis included the three stages of clustering, classification and clinical validation. Clustering of gait analysis to four groups was applied using the kmeans method. Two-thirds of the patients were used to create a simplified classification tree algorithm, and the model's accuracy was validated by the remaining one-third. Clinical validation of the classification method was done by the short form 36 Health Survey (SF-36) and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) questionnaires.

Results: The clustering algorithm divided the data into four groups according to severity of gait difficulties. The classification tree algorithm used stride length and cadence as predicting variables for classification. The correct classification accuracy was 89.5%, and 90.8% for females and males, respectively. Clinical data and number of total joint replacements correlated well with severity group assignment. For example, the percentages of total knee replacement (TKR) within 1 year after gait analysis for females were 1.4%, 2.8%, 4.1% and 8.2% for knee OA gait grades 1-4, respectively. Radiographic grading by Kellgren and Lawrence was found to be associated with the gait analysis grading system.

Conclusions: Spatiotemporal gait analysis objectively classifies patients with knee OA according to disease severity. That method correlates with radiographic evaluation, the level of pain, function, number of TKR.