Work

Beyond organoids: restoring the value of patient-derived two- dimensional colorectal cancer models

Year: 2026
Type: preprint
Abstract: Background Two-dimensional (2D) cell lines derived from colorectal cancer (CRC) are inexpensive, scalable and well suited to drug testing, but derivation from patient tissue has been inefficient. We asked how often defined, serum-free media yield stable lines from primary tissue and from patient-derived xenografts (PDX), and whether the clinical record predicts which tumors will grow. Methods We analyzed 193 culture attempts on 158 CRC specimens from 136 patients (95 primary-tissue and 98 PDX-derived attempts; 35 specimens by both routes) from a single biobank, 2015 to 2023. Twelve conditions were compared: three defined serum-free media with and without the ROCK inhibitor Y-27632, and six serum-containing historical conditions. A culture counted as established after at least three passages with stable growth. Eleven clinical and histopathological parameters and six molecular markers were tested against establishment with Bonferroni correction. Overall and relapse-free survival by establishment status were analyzed in an exploratory Cox model. Results PDX-derived cultures established more often than primary cultures (57.7% vs 39.1%; odds ratio [OR] 2.13, P = 0.013). Defined media outperformed historical conditions in both routes (P < 0.001); the three defined media did not differ from each other. In paired comparisons, Y-27632 converted 13 failed cultures to successful ones and the reverse occurred 3 times (OR 4.33, P = 0.021). When both routes were attempted, 81.8% of specimens gave a line. In total, 90 authenticated lines were obtained from 74 specimens of 65 patients. No parameter predicted establishment after correction; the lowest P value was for alcohol use (P = 0.017; corrected α = 0.0045). At a median follow-up of 65.8 months, establishment was associated with shorter survival after adjustment for stage and resection status (overall survival hazard ratio [HR] 1.77, 95% CI 1.03–3.06; relapse-free survival HR 1.95, 1.11–3.42). Unadjusted differences were not significant. Conclusions With defined serum-free media, Y-27632 and a PDX backup route, patient-derived 2D CRC lines can be obtained at rates close to those reported for organoids. Clinical, histopathological and molecular data do not predict which tumors will grow, so the panel is not biased toward a clinical subgroup. Whether ex vivo growth marks tumors with poorer outcome needs testing in an independent cohort.
Language: English
Subfield: Oncology
Field: Medicine
Open Access status: green
OpenAlex ID: W7225997413
Cites: 16
Cited by:
Related to: 10