Abstract
Background:
3D-spatial interaction between cancer cells influences tumor behavior, making it essential to replicate tumor structures for predicting patient outcomes.
Methods:
We collected data from three multicenter prospective studies to evaluate the ability to establish Patient-Derived Organoids (PDOs) from different biological samples and timepoints, correlating their characteristics and drug sensitivity with clinical outcomes.
Results:
From 184 patients (17 tumor types), 249 samples were collected: 149 (59.8%) from tumor tissue, 61 (24.5%) from peritoneal fluids, 39 (15.7%) from peripheral blood. Success rates for PDO establishment were 39.5%, 34.4%, and 25.6%, respectively. PDOs reproduced pathological and immunohistochemical patterns of source tumors, with pathogenic variants confirmed in 84% (21/25). In a series of 13 baseline and sequential PDOs from 9 patients undergoing treatment, responses to therapy mirrored patient responses during therapy.
Conclusions:
PDOs preserve tumor features, reflect disease progression, and predict treatment responses, providing valuable models to complement molecular testing in precision medicine.