An international cancer-research effort has assembled 665 patient-derived laboratory models representing 25 tumor types, creating a linked collection of biological material, molecular measurements and clinical information, the National Institutes of Health announced Wednesday.
The decade-long Human Cancer Models Initiative drew tumor tissue from 2,780 donors. Its models include organoids, which reproduce aspects of an organ’s cellular structure, and neurosphere clusters derived from brain cancers.
Researchers tested whether the laboratory models retained features of the tumors from which they came. In 421 paired sets, NIH reported 97.8% agreement in genetic alterations, 95% concordance in epigenetic features and 92% similarity in RNA-expression patterns.
That validation matters because a model that changes substantially in culture may no longer represent the disease researchers intend to study. Close molecular agreement does not make a model identical to a patient, but it strengthens its usefulness for controlled experiments.
The collection includes 522 models with detailed clinical data, 153 models of rare cancers and 71 models from people of non-European ancestry. NIH said the broader representation addresses gaps in older collections that captured only part of cancer’s biological diversity.
Researchers can use the models to study tumor evolution, genetic vulnerabilities, drug sensitivity and treatment resistance. Glioblastoma models, for example, carried several features associated with resistance to the chemotherapy drug temozolomide.
The models and associated genomic, transcriptomic, epigenomic and clinical data are being distributed through a searchable catalog and the American Type Culture Collection. Access expands the number of laboratories that can test findings against the same documented material.
The resource supports preclinical research; it is not a diagnostic test or evidence that a therapy will work for an individual patient. Drug effects observed in an organoid still require further validation and, where appropriate, clinical trials.
The collection’s value will be measured by reproducibility and discovery: whether outside teams can obtain the models, match the documented features and use them to identify results that hold up across laboratories and ultimately inform safer human studies.