The international effort was conducted through the Human Cancer Models Initiative, a 10-year program supported primarily by the U.S. National Cancer Institute and the Wellcome Trust. Led by scientists at MIT’s Koch Institute, the Broad Institute, Dana-Farber Cancer Institute, the National Cancer Institute, and partner institutions worldwide, the initiative collected more than 2,700 tumor samples from consenting patients in the United States, the United Kingdom, and the Netherlands. Approximately one-third of the samples were successfully converted into models capable of surviving and multiplying in laboratory conditions.

Most of the resulting models are organoids, three-dimensional structures made from tumor cells and grown in specialized culture media. Unlike conventional cancer cell lines, which typically form a flat layer on the surface of a laboratory dish, organoids develop within a supportive, gelatin-like scaffold that allows cells to organize into tissue-like structures. This three-dimensional environment can reproduce aspects of tumor architecture and cell behavior that are often lost when cancer cells are adapted to traditional two-dimensional culture.

The need for such models became clear after the Cancer Genome Atlas revealed the extraordinary genetic diversity of human tumors. Although thousands of patient samples had been sequenced, researchers had only about 1,000 established patient-derived cancer cell lines available for experiments. Those models were also disproportionately derived from patients of European or Southeast Asian ancestry, while many rare cancers and genetically unusual tumors were poorly represented. A limited model collection makes it difficult to determine whether a potential drug target is broadly relevant or applies only to a narrow subset of patients.

To create the new models, scientists developed tissue-specific culture conditions that provide cancer cells with the nutrients, signaling molecules, and physical support needed for long-term growth. Establishing a stable organoid or cell line can take as long as a year. Once a model was established, researchers compared it with the original tumor using several layers of molecular analysis, including DNA sequencing, RNA-expression profiling, and examination of epigenomic modifications. These tests helped determine whether the cultured cells retained the mutations, gene-activity patterns, and chemical changes that influence how cancer cells behave.

The collection includes models from common cancers such as lung, liver, and pancreatic tumors, as well as roughly 150 rare cancer types, including gallbladder and small-intestinal tumors. Each model has been deposited at the American Type Culture Collection, a nonprofit organization that distributes biological research materials. In addition to the cancer cells themselves, the associated data include information about the patient’s inherited genetic variants, known as germline mutations, and the treatments the patient received. This clinical context may allow researchers to investigate why tumors respond to certain therapies, develop resistance, or recur after treatment.

The models have already been incorporated into large-scale functional studies. In a companion Nature study, researchers at the Broad Institute analyzed more than 300 models using high-throughput DNA sequencing and RNA sequencing. They also performed CRISPR loss-of-function screens on more than 100 models. In these experiments, individual genes are systematically disrupted to reveal which ones cancer cells depend on for survival. If disabling a gene selectively kills cancer cells while leaving normal cells less affected, that gene may represent a potential therapeutic vulnerability.

The resulting data have been added to the Cancer Dependency Map, or DepMap, a research platform that connects cancer genotypes with cellular dependencies and possible drug targets. The resource now contains information on more than 2,000 cancer models. A separate companion study from the Wellcome Sanger Institute characterized another 256 organoids generated through the initiative, expanding the molecular and functional information available to researchers investigating tumor biology.

Scientists involved in the project emphasize that the collection is a major step rather than a final catalog of human cancer diversity. The formal HCMI program is winding down, but participating researchers hope to continue producing models from additional patient samples, especially pediatric and rare cancers. Because tumors can evolve during treatment and differ substantially between patients, even a collection of thousands of models cannot represent every clinically relevant cancer state. The researchers argue that continued tissue donation and international collaboration will be essential for building experimental systems that more accurately reflect the people who ultimately need new therapies.

Subject of Research: Patient-derived cancer models, cancer organoids, cancer genomics, drug discovery, and therapeutic vulnerabilities

Article Title: A compendium of next-generation patient-derived models for diverse cancers

News Publication Date: 5-Aug-2026

Web References: https://doi.org/10.1038/s41586-026-10806-y

References: Nature article, “A compendium of next-generation patient-derived models for diverse cancers”; Human Cancer Models Initiative; Cancer Dependency Map

Keywords: Cancer research, patient-derived models, organoids, cancer cell lines, drug development, drug discovery, genomics, CRISPR, Cancer Dependency Map, tumor biology, precision medicine, rare cancers

Organoids are three-dimensional clusters of living cells grown from tissue taken from patients. Unlike conventional cancer cell lines, which can acquire extensive genetic and behavioral changes after prolonged culture, patient-derived organoids are designed to preserve key characteristics of the original tumor. They can reproduce aspects of tumor architecture, genetic variation, cell behavior and treatment response, allowing scientists to investigate cancer biology in a controlled laboratory environment.

The HCMI collection includes organoids derived from pancreatic, breast, endometrial, colorectal, bladder, ovarian, head and neck, and lung cancers, among other malignancies. Scientists at Cold Spring Harbor Laboratory (CSHL) and Northwell Health led one of the largest contributing teams, providing more than 150 models to the international compendium. Collaborators at the University of Verona in Italy and Hubrecht Institute in the Netherlands also played major roles, including the development and distribution of dozens of pancreatic and colorectal cancer models.

The resource goes beyond living organoid cultures. It links many models to matched tissue from the original tumors, detailed clinical information and molecular data, including DNA and RNA sequencing, transcriptomic profiles and epigenetic measurements. Of the 665 models, 522 include comprehensive clinical data, while 153 represent rare cancers. The collection also contains 71 models from participants of non-European ancestry, addressing a long-standing problem in biomedical research: the underrepresentation of diverse patient populations in preclinical studies.

Researchers created 43 models from pediatric or adolescent patients, and approximately 23 percent of successful models came from rare cancer types. This breadth could make the collection especially valuable for studying cancers that are difficult to obtain in sufficient numbers for conventional experiments. Rare tumor subtypes often have few established cell lines and limited clinical trial data, meaning that organoids may provide an important experimental bridge between individual patient samples and broader biological discoveries.

To ensure that the organoids remained faithful to their source tumors, the HCMI teams established standardized production and quality-control procedures. At CSHL’s Genome Center, researchers used cancer hotspot sequencing to monitor key genetic alterations and assess the identity and quality of developing models. Organoids can also be expanded and cryopreserved, enabling laboratories to perform repeated experiments on the same tumor-derived material without requiring a new biopsy or continuously collecting fresh tissue.

The models are already being incorporated into drug-discovery research. Because organoids can be grown in multiwell plates, scientists can expose them to many drugs or drug combinations at different concentrations and measure effects on cell survival, proliferation and morphology. This approach could help identify therapies that are more likely to work against a particular tumor while revealing resistance mechanisms. Researchers emphasize, however, that organoid drug testing is not yet a substitute for clinical trials and must be interpreted alongside patient data and other biological models.

The HCMI organoids have also expanded the Cancer Dependency Map, or DepMap, a major research program designed to identify genetic and molecular vulnerabilities that cancer cells depend on for survival. By testing diverse organoid models, investigators can examine whether a potential dependency is shared across tumors or limited to a particular genetic background, tissue type or patient population. Such information may help researchers prioritize therapeutic targets and develop more precise strategies for tumors that do not respond to existing treatments.

The initiative was built around close coordination between clinical teams, hospital biospecimen repositories and laboratory scientists. Northwell Health, which treats more than 19,000 cancer patients annually, helped provide access to patient samples and established a pipeline connecting clinical care with experimental research. The project’s leaders say that patient consent and participation were fundamental to creating the resource. By making these models and their associated data available worldwide, the HCMI aims to accelerate cancer research, improve the reproducibility of preclinical experiments and move the field closer to personalized treatment decisions.

Subject of Research: Patient-derived organoid models for cancer research, precision medicine and therapeutic discovery

Article Title: A compendium of next-generation patient-derived models for diverse cancers

News Publication Date: 5 August 2026

Web References: Cold Spring Harbor Laboratory: https://www.cshl.edu/ ; Northwell Health: https://www.northwell.edu/ ; Nature article: https://www.nature.com/articles/s41586-026-10843-7

References: Nature, “A compendium of next-generation patient-derived models for diverse cancers.” DOI: 10.1038/s41586-026-10806-y

Image Credits: Hardik Patel/Cold Spring Harbor Laboratory

Keywords: Organoids, cancer research, patient-derived models, personalized medicine, cancer genomics, translational research, transcriptomics, epigenomics, drug screening, tumor biology