Deep Learning for Histo-genomics of ovarian cancers in guiding PARP-I treatment eligibility








About project

In this study, I investigate the clinical value of predicting BRCA1/2 mutation status in high-grade serous ovarian cancer directly from H&E stained Whole Slide Images using self-supervised learning methods. Other than that, there are two main clinical outcomes of this project. Firstly, it can be used to filter out patients that do not have a mutation in the BRCA1/2 gene and remove the need to perform genetic tests resulting in reduced stress, starting non-BRCA1/2-targeted therapies earlier, and reduced testing costs. Secondly, if a patient has a BRCA1/2 mutation, they can be recommended for the poly-ADP ribose polymerase Inhibitor (PARP-I) treatment which is known to be beneficial.

For this project, I collaborate with medical researchers from LUMC (Leiden University Medical Center) and several groups within the NKI (Netherlands Cancer Institute).


Our publications & software