Blog

Molecular Digital Twins, virtual models of an individual patient’s biology that are updated over time to simulate how a cancer may progress and predict how it may respond to treatment, are a new area of research focus for NFCR. We are applying them to one of the most urgent problems in cancer care: colorectal cancer in people in their 20s, 30s, and 40s.

What began as a conversation about digital twins at the 2025 NFCR Global Summit has become a research project with Dr. Ludmil Alexandrov of UC San Diego and Dr. Kimmie Ng of Dana-Farber Cancer Institute, which Dr. Alexandrov presented at the 2026 Summit. Their goal is a twin that can virtually test a young patient’s treatment before it is given, and help answer why colorectal cancer is striking more and more young adults.

Why early-onset colorectal cancer?

Colorectal cancer was long considered a disease of older adults and, thanks to colonoscopy screening, one of the great success stories in cancer prevention. Yet diagnoses in young people are rising sharply, and these early-onset patients face a troubling gap: the treatment guidelines they receive were developed from studies of older adults. Young patients often undergo aggressive therapies, yet respond less well than expected. Something about this cancer is different, and doctors currently lack the tools to know what the best treatment is for these young patients.

Clues written in the genome

Every day, cells in our body accumulate mutations, whether from lifestyle choices, faulty cellular machinery, or simply from growing and dividing. These mutations leave behind patterns called mutational signatures, which act like forensic fingerprints: by reading the genome of a cancer, researchers can identify what exposures helped cause it.

Research from the laboratory of Dr. Ludmil Alexandrov at UC San Diego, building on experimental work from the team of Dr. Hans Clevers, has uncovered one particularly important fingerprint for early-onset colorectal cancer. A common gut bacterium carrying the pks genomic island produces a toxin called colibactin, which can damage the DNA of the cells lining the gut. This damage leaves a distinctive mutational signature, known as signature 88.

By sequencing the genomes of early-onset colorectal cancers and building evolutionary trees of the tumor cells, the team made a striking discovery: the colibactin signature sits at the trunk of the tree, meaning the exposure happened in the very first years of life, with the latest estimates pointing to between 9 and 18 months of age. A substantial proportion of mutations in APC, the gene that delivers the first hit in colorectal cancer development, bear this signature.

Even more striking: in biopsies from children under 14 who had been given a clean bill of health by their gastroenterologists, about 40% carried the colibactin signature. Some children carried thousands of these mutations; one 3-year-old had a mutation burden resembling that of a person in their late 50s. The strongest association found so far is maternal: children whose mothers carry this bacterium are far more likely to be colonized themselves.

The working hypothesis: an early-life exposure to this genotoxic bacterium loads the gut of a young child with mutations, putting them on a fast track. Instead of developing colorectal cancer in their 60s and 70s, they may develop it in their 20s, 30s, or 40s. Proving this will take years of follow-up, but it opens a path toward prevention that simply did not exist before.

From static biomarkers to a living model

Today, most precision oncology relies on static biomarkers: test once, get a positive or negative result, choose a treatment, and wait to see what happens.

A Molecular Digital Twin is fundamentally different. It is a computational, in silico model of an individual patient’s unique biological state, built from a rich baseline (the tumor genome, RNA, tissue histology, imaging, clinical records, and exposure history) and then continuously updated as the patient is treated, with blood-based markers such as circulating tumor DNA (ctDNA), protein biomarkers, and repeat imaging. The twin evolves with the patient, simulating disease trajectories, predicting how the cancer will respond to therapy, and flagging emerging resistance, potentially weeks or months before it is visible on a scan.

Building the first eoCRC treatment twin

With the support of NFCR, Dr. Alexandrov’s team at UC San Diego is partnering with Dr. Kimmie Ng at Dana-Farber Cancer Institute, whose cohort is among the best in the world for early-onset colorectal cancer. Together, they are building a Molecular Digital Twin for 200 early-onset colorectal cancer patients:

• Retrospective foundation: roughly 70 patients with already-collected data are being used to build and train the first version of the twin.
• Prospective monitoring: 150 additional patients will be followed in real time as they receive their first line of treatment, with their twins updated every three months with ctDNA, protein biomarkers, and imaging.
• AI-driven predictions: machine-learning models will digest these streams of data to predict treatment response, anticipate resistance, and suggest clinical actions, which physicians may weigh alongside their own judgment.

The immediate goal is to predict, for the first time, how a young patient will respond to first-line therapy, a question no current guideline can answer. If the twin proves its value, it can justify prospective clinical trials and, ultimately, updated treatment guidelines built specifically for young adults rather than borrowed from older ones.

And beyond treatment: a twin for prevention

A second, more ambitious vision is the exposure twin, monitoring healthy people over time to understand why colibactin colonization happens in the first place, and which children carry the highest risk. If the treatment twin succeeds, Molecular Digital Twins could become a cornerstone not only of how we treat early-onset colorectal cancer, but of how we prevent it.

That is the hope of the eoCRC Molecular Digital Twin: a future where a young patient’s treatment is virtually tested before it is given, and where the cancer that begins in the first years of life can be intercepted before it ever begins.

If you would like to learn more about Molecular Digital Twins, give our team a call at: 1-800-321-CURE (2873).