Science 359, 920–926 (2018).

Co-clinical trials evaluate drug responses from patients enrolled in a parallel clinical trial using lab-based approaches, which traditionally rely on mouse models or xenografts. Valeri and colleagues created an organoid biobank from 71 patients with metastatic cholangiocarcinoma, colorectal or gastroesophageal cancer that were enrolled in phase I/II clinical trials, to evaluate molecular profiles and drug responses.

Histological and transcriptomic assessment confirmed similarities between organoids and original tumours, whereas intratumour heterogeneity was evident between primary cancer and metastases. Drug screening confirmed cancer organoid responsiveness if known mutations in signalling pathways were targeted with their respective inhibitors. The authors found a high clinical predictive value when comparing organoid data to 21 cases of patient responses, such as sensitivities to paclitaxel, cetuximab and others. To assess the response to regorafenib, organoids from responsive or resistant patients were xenografted into mice, and showed reduced vasculature and tumour volume only in sensitive patient-xenografts. Grafts from metastasis-derived organoids were used to assess the initial response to regorafenib and progression after the emergence of resistance, indicative of tumour evolution. Clinical and preclinical responses to a treatment for chemorefractory colorectal cancer, confirming the ability of organoids to recapitulate intra-patient heterogeneity.

Overall, the authors reported 100% sensitivity, 93% specificity and a positive predictive value of 88% for patient drug responses, highlighting the efficiency of organoids in providing molecular insights and accurate predictions for drug responses.