Abstract

Imagen Therapeutics is a leading provider of end-to-end in vitro pharmacology services with applications for both patient diagnostics and oncology drug development. Our unique platform harnesses the power of high-content imaging to visualise treatment outcomes with the largest collection of patient-derived cell (PDC) models. PDCs bridge the gap between conventional cancer lines and costly, time-consuming patient-derived xenograft (PDX) models. Testing PDCs in two-dimensional environments is an excellent first step for high-throughput compound testing that is crucial to the drug development life cycle. However, spheroids have the potential to provide a more representative model of in vivo responses. Critical to the development of these three-dimensional assays is reproducibility and quality control measures to translate this into a high-content screening assay. We have investigated alternative methods for gaining granularity, intensity and texture information from brightfield images to analyse spheroid response to compound treatment. We have used these features and machine learning to predict which compound each spheroid has been treated with. High-content screening offers direct patient benefit with our predictRx in vitro diagnostic tool while accelerating and de-risking drug development with predictTx drug screening offering for biopharma. Authors: R.A. Howard-Jones, G.J. Griffiths, C. Noakes, D.I. James, R.J. Errington, D. Parthimos, R. Edward, S. Singh, G. Muckle