Glioblastoma Pseudoprogression Detection

To deepen my skills in biomedical imaging I am currently working as a researcher in Prof. Davatzikos’ lab at the University of Pennsylvania. My work leverages generative machine learning for analyzing brain mpMRI data aiming to distinguish true glioblastoma recurrence from pseudo-progression.

After studying relevant literature, I realized that tumor pseudo-progression is a very under-documented setting, with very few confirmed cases, as compared to tumor recurrence cases. Thus, I reformulated the question of classifying suspicious lesions as glioblastoma recurrence or pseudo-progression (discriminative modeling) into learning the distribution of true recurrence (generative modeling). This way, I can identify pseudo-progression cases as the outliers of the data distribution.

This work is currently in progress.