1568P Establishment of a platform to predict radiation sensitivity in organoids derived from esophageal cancer patients

To increase the effectiveness of the treatment, radiotherapy (RT) is frequently combined with chemotherapy, immunotherapy, or radiosensitizer medications. Identifying a patient’s radiosensitivity can be important for determining patient’s therapy strategy. But there isn’t a reliable and pertinent radiation sensitivity prediction model for assessing radiation sensitivity. Here, we propose the cancer organoid-based platform for predicting the individual RT response evaluation and recurrence. Eighteen Esophageal cancer biopsy tissues were dissociated to single cells and then disposed with BME on pillar plate (MBD co., Korea) to make the cancer organoids. After forming the organoids, these were exposed to radiation doses of 2, 4 and 8Gy. We are staining live organoid with calcein AM after radiation and obtain organoid viability. From viability, AUC (area under curve) and growth rate were calculated to predict patient’s radiation sensitivity. With these two variables, the patient's cancer stage score was also used to predict patient’s radiation sensitivity. CODRP (cancer organoid-based diagnosis reactivity prediction), radiosensitivity predication index, were calculated by patient’s AUC, growth rate, and cancer stage. Radiation sensitivity prediction model using each single-factor (patient’s AUC, growth rate and cancer stage) showed about 62% specificity, 72% sensitivity and positive predictive value (PPV) was 69%. However, when applying multi-factor model (CODRP), the specificity was 81.8%, the sensitivity was 85.7%, and the PPV was 90%. In the CODRP model, the radiation sensitive group showed 11.1% recurrence rate, while the radiation resistive group showed 72.7% recurrence rate. Therefore, there was a significant difference in recurrence-free survival rate between the sensitive group and the non-responder group based on the CODRP model. Our proposed CODRP models using patient derived cancer organoid successfully predict clinical response of patient’s radiation treatment. Thus, this platform has promise as a novel prognostic indicator for people with esophageal cancer and can support precision treatment.