The Discoid Meniscus Blind Spot: A Structural Gap in Knee OA Prediction Models

Machine learning models predicting knee osteoarthritis progression are commonly trained on WORMS/MOAKS MRI scoring systems, which encode meniscal damage but not meniscal shape. This structural gap means these models cannot account for discoid meniscus, a congenital variant present in 3–5% of the general population and up to 15–20% of Asian populations. Using published scoring-system documentation and cohort demographics, we show this blind spot is quantifiable at scale: applying baseline prevalence estimates to the Osteoarthritis Initiative cohort (n=4,796) suggests on the order of 144–240 participants may carry an unrecorded discoid meniscus, compounded by the cohort's thin representation of populations where the variant is most common. We argue this is a structural property of the scoring instrument itself, not a data-collection gap that more samples would fix.

This is a preprint: it has not been peer reviewed yet, and its findings may change.

Published in Zenodo (CERN European Organization for Nuclear Research). The abstract is shown as published and remains the work of its authors; read the full article at the source.

doi.org/10.5281/zenodo.22859207