KNOAP2020 Banner

Motivation

Knee osteoarthritis causes a large economic burden on the society and reduces life quality of an individual [Salmon et al. 2016]. Therefore, methods that are able to identify subjects and knees that will develop the disease in the future are important. Usually such methods are optimized for specific imaging datasets and it is unclear how different methods would perform on previously unseen data. Therefore, we are organizing a challenge to objectively compare methods that use MRI, X-ray image data, and clinical risk factors for prediction of incident symptomatic radiographic knee osteoarthritis.

The prediction method is not pre-specified and can be based for example on deep learning, hand-crafted image features, or semi-quantitative image features.

Objective

The aim of this challenge is to use an independent test set (MRI and X-ray image data along with clinical risk factors at baseline) to objectively compare methods for prediction of incident symptomatic radiographic knee osteoarthritis within 78 months (6.5 years).

Outcome

In this challenge, incident symptomatic radiographic knee osteoarthritis is the outcome variable. Symptomatic knee osteoarthritis is defined as knee pain and a definite tibiofemoral osteophyte in the same knee. Knee pain was assessed by questionnaire “Did you experience pain in or around left, right, or both knees during most days in the past month?”. It should be noted that the knees with the knee pain and osteophytes also fulfilled the combined (clinical and radiographic) ACR criteria [Altman et al. 1986] (knee pain + definite osteophyte and one of the following: age >50 years, morning stiffness \<30 minutes, crepitus on active knee motion), because all study subjects in the test set were over 50 years old at baseline. Incident symptomatic radiographic knee osteoarthritis is defined when symptomatic radiographic knee osteoarthritis was present at 2.5 and/or 6.5 years and not present at baseline.

References