e17544 Background: High-grade serous ovarian cancer (HGSOC) can be separated by gene expression profiling into four molecular subtypes with clear correlation of the clinical outcome. However, these gene signatures have not been implemented in clinical practice to stratify patients for targeted therapy. This is mainly due to a lack of easy, cost-effective and reproducible methods, as well as the high heterogeneity of HGSOC. Hence, we aimed to examine the potential of unsupervised matrix assisted laser desorption/ionization imaging mass spectrometry (MALDI-IMS) to stratify patients, which might benefit from targeted therapeutic strategies. Methods: Molecular subtyping of paraffin-embedded tissue samples from 279 HGSOC patients was performed by NanoString analysis (ground truth labeling). Next, we applied MALDI-IMS, a novel technology to identify distinct mass profiles on the same paraffin-embedded tissue sections paired with machine learning algorithms to identify HGSOC subtypes by proteomic signature. Finally, we devised a novel strategy to annotate spectra of stromal origin. Results: We elucidated a MALDI-derived proteomic signature (135 peptides) able to classify HGSOC subtypes. Random forest classifiers achieved an area under the curve (AUC) of 0.983. Furthermore, we demonstrated that the exclusion of stroma associated spectra provides tangible improvements to classification quality (AUC = 0.988). False discovery rates (FDR) were reduced from 10.2% to 8.0%. Finally, novel MALDI-based stroma annotation achieved near-perfect classifications (AUC = 0.999, FDR < 1.0%). Conclusions: Here, we present a concept integrating MALDI-IMS with machine learning algorithms to classify patients according to distinct molecular subtypes of HGSOC. This has great potential to assign patients for targeted therapies.
332 Two-year prognosis estimation of advanced high grade serous ovarian cancer patients using machine learning
