Boston researchers and international collaborators have unveiled a landmark cancer research resource: 665 next-generation patient-derived models representing 27 common and rare cancer types. The collection, described in a study published in Nature, is being made available to researchers worldwide together with extensive clinical and molecular information. Its creators say the resource is the largest coordinated release of validated patient-derived cancer models to date and could significantly accelerate the discovery of cancer vulnerabilities, therapeutic targets and treatment strategies.
The models were developed through the Human Cancer Models Initiative, an international effort involving the National Cancer Institute, Cancer Research UK, the Wellcome Sanger Institute and Hubrecht Organoid Technology. The initiative aims to create 1,000 patient-derived models that accurately reproduce the biology of human tumors in laboratory systems. Approximately 2,800 patients from the United States, the United Kingdom, Italy and the Netherlands consented to provide tumor tissue and associated clinical information for the project.
Unlike many traditional laboratory cancer models, the new collection was designed to preserve the biological features of the tumors from which they originated. Patient-derived models can include three-dimensional organoids and spheroids, as well as two-dimensional cell lines. These systems are grown under conditions tailored to the specific cancer type, helping maintain the genetic, molecular and cellular characteristics of the original tumor. The models that passed rigorous quality-control procedures were subjected to standardized genomic sequencing and molecular profiling.
This validation process addresses a major weakness of earlier cancer models. Cells grown in laboratories can gradually acquire genetic or biological changes, a phenomenon often described as “drift,” which may make them increasingly different from the patient’s tumor. Such changes can undermine experiments designed to predict how a cancer will respond to a drug or how a genetic alteration contributes to disease. The HCMI models were selected for their ability to remain faithful to the original samples and to retain stable biological behavior over extended periods.
The collection includes cancers affecting both adults and children, with examples ranging from colorectal, pancreatic, lung and brain cancers to much rarer malignancies. More than 20 percent of the models represent rare cancer types, some of which previously had only one or two experimental models available to researchers worldwide. Expanding representation of these diseases could be particularly important because rare cancers often lack the large patient populations and research infrastructure that support studies of more common tumors.
Clinical context is another defining feature of the resource. Among the models are 168 derived from patients who had already received treatment, including immunotherapy, targeted therapy, chemotherapy and radiotherapy. Another 318 models were generated from samples collected before treatment. Linking the laboratory models to treatment history and patient outcomes may allow researchers to investigate why some tumors resist therapy, identify molecular features associated with response and test potential combinations of drugs in systems that reflect real-world disease.
The models and their associated data are being distributed through the American Type Culture Collection. Researchers will be able to access not only the physical biological materials but also information such as genomic sequencing results, clinical annotations and molecular measurements generated using consistent methods. According to the investigators, this unified structure is essential because it allows findings from different laboratories to be compared more reliably than when researchers use unrelated models created under different conditions.
The resource has already contributed to the expansion of the Cancer Dependency Map, or DepMap, a large-scale effort managed by the Broad Institute that uses CRISPR gene-editing technology to identify genes on which cancer cells depend. By incorporating the HCMI models, investigators have broadened DepMap’s coverage of genetic and molecular cancer subtypes. The new models also include gene-expression patterns and cellular states that were not consistently represented in earlier patient-derived systems, potentially revealing vulnerabilities that had remained invisible in previous screens.
The scientific importance of the collection extends beyond the immediate experiments it enables. With hundreds of carefully characterized models connected to clinical and genomic data, researchers can perform large-scale studies of tumor evolution, drug resistance, cancer dependencies and interactions between genetic alterations. The dataset may also provide valuable training material for computational tools and artificial-intelligence systems designed to predict treatment response or prioritize drug targets. The investigators describe the release as a major change in the experimental infrastructure available to cancer biology, particularly because it combines standardized models with deep patient-level information.
The Nature study, titled “A Compendium of Next-Generation Patient-Derived Models for Diverse Cancers,” was led by investigators including Keith Ligon of Dana-Farber Cancer Institute, Jesse Boehm of the Massachusetts Institute of Technology, Mathew Garnett of the Wellcome Sanger Institute, David Tuveson of Cold Spring Harbor Laboratory and collaborators from institutions across the United States and Europe. The HCMI was funded primarily by the National Cancer Institute and the Wellcome Trust. By making the models broadly accessible, the initiative aims to give researchers the experimental systems needed to translate cancer genome discoveries into new therapies more quickly.
Subject of Research:
Next-generation patient-derived cancer models, including organoids, spheroids and cell lines, for studying tumor biology, treatment response and therapeutic vulnerabilities.
Article Title:
A Compendium of Next-Generation Patient-Derived Models for Diverse Cancers
Web References:
https://www.atcc.org/hcmi
https://depmap.org/portal/
https://doi.org/10.1038/s41586-026-10806-y
References:
Nature article, DOI: 10.1038/s41586-026-10806-y
Keywords:
Cancer research, patient-derived models, organoids, cancer biology, precision medicine, drug discovery, tumor modeling, Cancer Dependency Map, CRISPR screening, genomics, rare cancers, Dana-Farber Cancer Institute, Human Cancer Models Initiative
