Towards patient-centric drug discovery: Drug action in malignant pleural effusions and ascites using high content imaging and deep learning.

e21633 Background: Many anticancer drugs found to be active in preclinical development later do not show desired effect clinically. This suggests that currently used preclinical models do not fully recapitulate the complexity of the disease. The study of drug activity in primary samples could provide a more immediate picture of a molecule’s activity in a patient. Factors that have so far hampered the use of primary tissue samples for drug discovery and development include access in sufficient quantity as well as robust analytical methods. We hypothesised that malignant pleural effusions and ascites (MPAs) of solid tumour patients are a promising model system to study preclinical drug activity. MPAs are easily accessible and contain cancer cells as well as recruited immune cells. Following previous successes in studying drug action in primary tissues of patients with haematological malignancies with automated microscopy (Snijder et al 2017, Lanc Haem, NCT03096821) we describe advances in using high content imaging and deep learning-based image analysis to study drug action in MPAs of solid tumour patients. Methods: MPAs from patients with metastatic breast, pancreatic and ovarian cancer (at least n = 10 of each) were collected. The response of EpCam+/CD45- and CD45+ cells against small molecule drugs was evaluated using high content microscopy. Drug response was quantified at single cell resolution using regional convolutional neural networks (R-CNNs) comprising object detection and single cell classification. Results: MPAs contain both cancer cells and recruited myeloid and lymphoid immune cells with varying activation. Ex vivo drug responses from each patient sample were measured and the EC50 of each molecule determined by curve fitting. Sensitivity mirrored drug approvals for some indications, and also revealed drugs with potential off label use. On target and off-target response curves, along with integrative scores are used to visualize the effects. Conclusions: Single-cell phenotypic analysis of MPAs enables the study of anticancer drug action in a setting that is one step closer to the clinic than cell line or outgrown organoid models of solid tumor. While initial response patterns can be observed that mirror current approvals, further biological and clinical validation must occur to understand in how far these data can be used for drug discovery and translational research purposes.