Meeting Banner
Abstract #3998

Radiomic Features on Quantitative Susceptibility Mapping Classify Amyotrophic Lateral Sclerosis Patients from Mimics

Anja Samardzija1, Thanh Nguyen2, Elizabeth Sweeney2, Kailyn Lee2, Ilhami Kovanlikaya2, Yi Wang 2, Andrew Schweitzer2, and Apostolos Tsiouris2
1Electrical and Computer Engineering, Cornell University, Highlands, NJ, United States, 2Weill Cornell Medicine, New York City, NY, United States

We trained a Random Forest classification model to classify amyotrophic lateral sclerosis (ALS) patients from those with mimicking clinical presentations based on QSM radiomic features extracted from the primary motor cortex. In a validation set, the model has 0.8 accuracy, 0.75 specificity of 0.75 and 0.84 sensitivity, which is superior to models using the mean QSM value as cutoff with 0.59 accuracy, 0.94 specificity and 0.14 sensitivity.

This abstract and the presentation materials are available to members only; a login is required.

Join Here