Abstract PO-093: Evaluating dependencies by rapid image-based ex vivo cancer biosensors

Abstract

Successful mapping of cancer dependencies requires conducting genetic and drug screens on a diversity of models. However, the difficulty in generating long-term models of many cancers limits the share of patient samples that can be studied. Such long-term models have likely also lost the cellular heterogeneity present in the original tumor due to in vitro propagation. To overcome these limitations, we are developing image-based ex vivo cancer biosensors from early patient material. Using freshly received gastroesophageal cancer ascites, we are optimizing perturbation methods and utilizing single-cell transcriptomics and label-free microscopy to infer a subpopulation-specific vulnerability profile. We show that label-free microscopy can infer cell identity and viability in heterogeneous early patient samples. Additionally, early drug perturbation recapitulates observations made in established gastroesophageal cancer organoids. Successful implementation of ex vivo biosensors will expand the cancer dependency space by making perturbational studies accessible to more diverse samples, and by identifying and validating hits in a more immediate setting to the original tumor. Citation Format: Mushriq Al-Jazrawe, Csaba Molnar, Niklas Rindtorff, Pinar Eser, Sean Misek, Maria Alimova, Oana Ursu, William Colgan, Adel Attari, Natalie Tsang, Paula Keskula, Carmen Rios, Moony Tseng, Anne Carpenter, James McFarland, Adam Bass, Samuel Klempner, Jesse Boehm. Evaluating dependencies by rapid image-based ex vivo cancer biosensors [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-093.