AOSSM 2023 Annual Meeting Recordings no CME-Which Osteochondritis Dissecans Subjects Will Heal Non-Operatively? An Application of Machine Learning Methods to the ROCK Cohort

Which Osteochondritis Dissecans Subjects Will Heal Non-Operatively? An Application of Machine Learning Methods to the ROCK Cohort

Back to course

Video Transcription

Video Summary

In this video, Thomas, a third-year medical student at Stanford, discusses a study on osteochondritis dissecans (OCD) subjects and their healing with non-operative management. The study aimed to develop a machine learning algorithm to predict which OCD lesions would heal without surgical treatment. The study used the ROC cohort, consisting of 26 institutions, to train the algorithm. They selected variables such as lesion width, length, and location on MRI scans as predictors for healing. The results showed that larger lesions were less likely to heal, while lesions in non-weight-bearing zones had a higher chance of healing. The algorithm achieved better accuracy than previous efforts, with a support vector machine having the best performance. Future steps include incorporating more data and applying the same techniques to OCD of the elbow. Thomas thanks the ASSM, co-authors, mentors, and the ROC group for their support.

Asset Caption

Thomas Johnstone, BS

Keywords

Osteochondritis dissecans

Non-operative management

Machine learning algorithm

ROC cohort

Lesion healing