ATB20-15 - Prospective Evaluation of AI R&D tool for patient stratification (PEAR)
Pear Bio is a London-based computational biology company discovering predictive biomarkers to guide cancer treatment for improved patient outcomes. We have developed a multivariate analysis of 3D cell cultures combining primary tumours, live-cell microscopy, multi-omic readouts, and proprietary computer vision pipelines to analyse time-lapse 3D image data and quantify ex vivo cancer progression against potential treatment options. By comparing therapies side-by-side outside of the patient, the Pear Bio test enables oncologists to make better informed decisions based on personalised differential drug efficacies.
Continued collaboration with ATB has allowed Pear Bio to develop efficient cell extraction protocols from tumours, culture in microfluidics and 3D culture platforms, and dosing of various therapeutic agents. We now seek to expand collection to blood products, so that we may test immunotherapies (treatments that require the presence of immune cells, extracted from a patient’s own blood).
The workflow of Pear Bio’s device:
We culture the primary tumour cells in our proprietary 3D microtumours, which act as a physiologically-relevant proxy for the real patient tumour. The tumour cells are split among multiple 3D cultures: some are treated with approved treatments whilst others serve as negative controls. For 5 days, cells are exposed to different treatments in parallel. Every day, drugs are cleared and re-administered to simulate the drug cycles within a human body. This also allows for combination therapies to be tested. Daily confocal images are acquired in order to visualise cell viability, spread and movement inside the chips and over time. Computer vision algorithms are applied to the images to understand personalised cell behaviours and response to treatments.