Cancer cell lines remain foundational for cancer research and drug discovery, yet they incompletely capture tumour diversity, lack linked patient context, and undergo adaptation to in vitro culture. Patient-derived tumour organoids are three-dimensional cultures established directly from tumour tissue and represent a powerful complementary model. We aimed to generate and systematically characterise a large panel of clinically annotated tumour organoids to define gene dependencies across multiple cancer types.

Tumour specimens from colorectal, oesophageal, ovarian, pancreatic, and gastric cancers were used to derive 256 patient-derived organoids as renewable, genetically stable models. Each organoid and its matched tumour sample underwent whole-genome and transcriptome sequencing. Genome-wide CRISPR–Cas9 loss-of-function screens were performed across 164 organoids to map gene dependencies. Integrative analyses were conducted to associate genetic and clinical features with dependency patterns, including evaluation of paired pre- and post-treatment samples.

The organoid models retained key genomic and transcriptional features of their source tumours and remained genetically stable during propagation. Functional screening revealed genomic and clinical markers of dependency across common and rare subtypes and identified essential genes specific to organoid models. Analysis of longitudinal paired samples illuminated targetable vulnerabilities emerging during tumour evolution following treatment. In colorectal cancer organoids, functional and pharmacological interrogation of the EGFR–RAS–MAPK pathway demonstrated differential effects among KRAS variant alleles, highlighting variant-specific biological consequences.

We present the first systematic map of gene dependencies in patient-derived organoids across multiple cancer types. This renewable organoid resource, together with its integrated genomic and functional datasets, will be made publicly available, expanding model diversity and mechanistic insight to advance precision oncology.