Abstract B029: High-throughput screening of patient-derived organoids and biomarker identification: Combining lab automation, data analysis automation, and organoid-specific drug metrics

Abstract Patient-derived tumor organoids (PDOs) represent human tumors better than other in vitro models and are positioned to replace traditional cell lines in the future. However, significant limitations in PDO analysis have prevented their use in high-throughput screening (HTS) applications, namely accurate representation of PDO drug response and scalability. Most PDO assays for drug screening fall into two categories: biochemical assays or live-imaging assays. Currently, only biochemical assays are scalable and high-throughput compatible, and readouts from both assays fail to reflect accurate drug response because they quantify proxy parameters like ATP or organoid size. More importantly, drug metric calculations from these readouts, such as growth rate inhibition (GR) or normalized drug response (NDR)1,2, utilize assumptions from cell lines that may grossly misrepresent PDOs and confound results. Therefore, to address these limitations, we have developed an HTP-compatible protocol to facilitate the adoption of PDOs for HTS applications3. Furthermore, we have develop an PDO-specific drug metric, normalized organoid growth rate (NOGR), that more accurately reflects their drug response.4,5Our protocol combines live-imaging techniques with AI for computer vision and data processing to visualize and quantify PDO drug response3. By quantifying individual PDO growth rate through label-free detection, we developed a new PDO-specific drug metric, which we coined the NOGR metric4. Using this HTS protocol, we screened 10 distinct drug combinations focused on the repurposed drug candidate Auranofin, a Thioredoxin reductase inhibitor5. 10 PDOs sourced from healthy lung, non-small cell lung cancer, and pancreatic ductal adenocarcinoma were used for screening. Our developed NOGR metric enabled us to distinguish cytostatic from cytotoxic drug responses and identify drug candidates that synergistically enhanced the efficacy of Auranofin in a tumor-selective manner. Moreover, through our method, we identified CA12 as a predictive biomarker for Auranofin sensitivity and enhanced efficacy with other therapeutics. This study marks a significant advancement in PDO-based HTS screening by addressing the current challenges in accuracy and scalability. The adoption of our method facilitates PDO use and the gradual replacement of cell lines to accelerate novel drug research. References: 1. M. Hafner, et al. Nature Methods, 13 (2016) 2. A. Gupta, et al. Communications Biology, 3 (2020) 3. M. Le Compte, et al. JoVE 190 (2022) 4. C. Debe, et al. Communications Biology, Accepted (2024) 5. C. Deben, et al. Journal of Experimental & Clinical Cancer Research 43 (2024). Citation Format: Abraham Lin, Maxim Le Compte, Edgar Cardenes De La Hoz, Tyler Gilcrest, Geert Roeyen, Jeroen Hendriks, Filip Lardon, Evelien Smits, Christophe Deben. High-throughput screening of patient-derived organoids and biomarker identification: Combining lab automation, data analysis automation, and organoid-specific drug metrics [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Functional and Genomic Precision Medicine in Cancer: Different Perspectives, Common Goals; 2025 Mar 11-13; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2025;85(5 Suppl):Abstract nr B029.