To design new targeted cancer treatments, scientists need lab systems that faithfully mimic the genetic and molecular traits of real tumors. That’s exactly what an international team has achieved: nearly 700 new patient-derived cancer models, representing 25 different cancer types, are now available to researchers worldwide.
Published in Nature, this milestone is the result of a 10-year initiative led by MIT’s Koch Institute, the Broad Institute, Dana-Farber Cancer Institute, the National Cancer Institute, and many other partners.
The study was based on samples from patients, including over 2,700 tumors contributed by people living in the U.S., UK, and Netherlands. Drawing on an approach called organoid technology, scientists expand tumors into three-dimensional structures that can live indefinitely but retain the genetic and molecular hallmarks of the actual cancer.
About one‑third of the samples were successfully converted into organoids or cell lines, creating a powerful resource for studying cancer biology and testing new therapies.
Unlike traditional flat cell cultures, organoids more closely resemble living tissue, offering a powerful tool for drug discovery.
“Since the sequencing of the human genome and the analysis of cancer genomes over the last 20 years, we have had many ideas about cancer targets, but we need experimental systems in the lab to validate those targets and launch drug discovery projects,” said Jesse Boehm, senior author.
These patient-derived models are providing powerful new weapons for cancer researchers to explore. They allow for the discovery of new drug targets, testing therapies in a broader genetic spectrum, and inclusion of rare cancers that are otherwise absent from current collections. Organoids widen the scope of research by reflecting the complexity of real tumors and bringing precision medicine closer to practice.
Ever since the Cancer Genome Atlas and its launch of The Human Cancer Models Initiative in 2016, this has been a crusade to broaden an ever-shrinking pool of patient‑derived cancer cell lines. Only after some time, up to 1,000 models were available; most were from European- or Southeast Asia-derived patients rather than other ethnicities, with far fewer specific cancers represented.
Backed by the U.S. National Cancer Institute and the UUK’sWellcome Trust, scientists across dozens of institutions collected more than 2,700 tumor samples from hospitals in the U.S., UK, and the Netherlands.
Mushriq Al-Jazrawe, scientific director of the High Throughput Sciences (HTS) platform at the Koch Institute and one of the lead authors of the study, said, “A resource of this scale depends on the kind of systematic effort that often happens behind the scenes. I’m especially grateful to the technical and scientific teams across the participating institutes whose careful, expert work turns patient tumor samples into well-characterized models and data that researchers everywhere can use with confidence.”
They ranged from the most frequent cancers like lung and liver (when it is a common group) to 150 fresh rare-type tumors, such as gallbladder or small intestine ones. The researchers then converted around one‑third of the samples into long-lived cell lines and 3D organoids, clusters that mimic real tissue better than earlier technologies. Confirming their fidelity to the originating tumors, rigorous genomic analyses united these collections into a robust new resource for cancer research worldwide.
All these models generated through the Human Cancer Models Initiative (HCMI) have been stored at the American Type Culture Collection (ATCC), together with a wealth of patient data including inherited mutations and treatment history. An irregular shelflike structure at the end of each globule binds to a receptor on the surface, which enlists an organized wave motion.
In a companion Nature paper, Broad Institute researchers profiled more than 300 of the new models using genome and RNA sequencing, plus CRISPR loss‑of‑function screens, uncovering druggable vulnerabilities. These findings expand the Cancer Dependency Map (DepMap), which now covers over 2,000 cancer types. Another Nature study from the Sanger Institute characterized 256 additional organoids.
While the formal HCMI project is winding down, researchers aim to keep building models, especially for pediatric and rare cancers.
“We now have about 2,000, but more work is needed to represent all humans with cancer,” says Boehm.
Al-Jazrawe, who is also a researcher in the Broad Institute’s Cancer Program, said, “A major opportunity now is to carry the lessons of HCMI forward so that we can generate as much insight as possible from these precious tissue donations. Here at HTS, we are continuing the work by developing methods to study patient-derived samples and models reproducibly and at scale, and by providing a platform for close collaboration with clinical and research teams.”
The effort highlights the global value of patient tissue donations for advancing cancer research.
Journal Reference:
- ElHarouni, D., Al-Jazrawe, M., Choi, S. et al. A compendium of next-generation patient-derived models for diverse cancers. Nature (2026). DOI: 10.1038/s41586-026-10806-y
