The HCMI Resource: A Decade of Cataloging Cancer's Molecular Complexity

A decade-long effort led by the National Cancer Institute (NCI) has delivered one of the largest publicly available collections of patient-derived cancer models. The Human Cancer Models Initiative (HCMI) now offers 665 next-generation laboratory models—including organoids and neurosphere clusters—derived from 2,780 tumor donors across 25 cancer types. The resource, published in Nature, is designed to help researchers worldwide study tumor biology, drug sensitivity, and resistance.

The core advance: the catalog’s models show remarkably high molecular fidelity to the tumors they came from. Scientists compared 421 paired tumor-model sets and found 97.8% agreement in genetic alterations, 95% concordance in epigenetic features, and 92% similarity in RNA expression patterns. This validation confirms that the lab-grown organoids retain the key biological traits of the original cancers, even after extended growth in culture.

“By linking patient-derived models with detailed molecular and clinical data, HCMI has established a framework for advancing both our understanding of cancer and outcomes for patients,” said Anthony Letai, director of NCI. The collection includes 522 models with clinical annotations, 153 rare cancer models, and 71 models from individuals of non-European ancestry—expanding the diversity that has often been lacking in prior repositories.

All models and associated genomic, transcriptomic, and epigenomic data are being distributed to the scientific community through the American Type Culture Collection (ATCC). Researchers can search and filter by cancer type, treatment history, demographics, and molecular features via the HCMI Searchable Catalog.

Advertisement

Why This Model Compendium Matters for Cancer Science

Closing the Tumor-Model Fidelity Gap

Before HCMI, existing patient-derived cancer models captured only a slice of the disease’s biological diversity, and their molecular resemblance to original tumors was often uncertain. By demonstrating near-identical genetic and gene-expression profiles across hundreds of paired samples, the initiative provides a level of validation that the research community has lacked. This means scientists can now use organoids with much greater confidence that results will reflect actual patient biology, potentially improving the predictability of preclinical drug testing.

A Direct Boost for Drug Resistance Research

The detailed characterization of each model opens a direct path to studying therapeutic resistance. In glioblastoma models, for example, researchers detected a range of genetic features linked to temozolomide resistance—inherited abnormalities, oncogene amplification, and mutational signatures from prior treatment. Because the organoids did not drift genetically under culture conditions, they offer a stable platform to screen for drugs that might overcome resistance mechanisms in brain cancer and other tumor types.

Building a More Inclusive Research Foundation

Cancer models have historically under-represented rare tumor subtypes and patients from non-European backgrounds. HCMI deliberately addressed this by including 153 rare-cancer models and 71 models from diverse ancestries. Combined with the freely searchable database and ATCC distribution, the effort lowers the logistical and cost barriers for academic labs and smaller biotech firms to pursue studies that reflect the full spectrum of human cancer.

What This Means for Researchers and Drug Developers

For researchers and drug developers, the HCMI catalog translates directly into experimental opportunity:

  • Explore the HCMI Searchable Catalog to identify organoid and neurosphere models that match your target cancer type, genetic alteration, or patient treatment history.
  • Use the models’ genomic and drug-sensitivity data to design preclinical studies that are tightly aligned with real tumor biology, particularly for drug resistance and precision oncology programs.
  • Leverage the rare cancer and ancestry-diverse models to investigate understudied patient populations—an area where traditional cell line banks have fallen short.