Colorectal cancer (CRC) is the third most commonly diagnosed cancer in the world and the second leading cause of cancer-related death. In countries with widespread population screening programs, both rates of new cases and deaths have steadily declined, showing that catching this cancer early truly saves lives.
Current screening guidelines recommend either colonoscopy (a camera exam of the colon) or fecal immunochemical testing (FIT) - a simple stool test that looks for hidden blood - for most people above a certain age. However, treating everyone the same regardless of their personal risk has produced mixed results: low participation rates, limited detection of early lesions, and heavy strain on colonoscopy facilities.
The concept of risk-adapted screening proposes a smarter approach: direct people with high personal risk straight to colonoscopy, while guiding lower-risk individuals through FIT first. The challenge is finding an accurate way to calculate each person's individual risk. Traditional tools that rely only on age, sex, and lifestyle factors have limited accuracy in separating high-risk from low-risk individuals.
Researchers wondered whether genetic information could improve this risk calculation. Over the past few decades, large genetic studies called genome-wide association studies (GWAS) have identified hundreds of common DNA variants that each nudge a person's CRC risk slightly up or down. Could combining all these tiny effects into a single score improve how we identify who most needs intensive screening?
A polygenic risk score (PRS) is a single number that summarizes how much a person's inherited DNA variants collectively influence their risk for a disease. It works by adding up the small effects of hundreds or even thousands of genetic variants identified through GWAS, with each variant weighted by the strength of its association with the disease.
The idea is powerful because no single gene variant causes most CRC cases - rather, it is the combined influence of many common variants that tips a person's risk higher or lower. A PRS captures this cumulative genetic burden in one easy-to-interpret number, making it practical to use in a clinical or screening setting.
Most PRS research for CRC has been done in European populations. This study specifically focused on developing and validating a PRS for East Asian (EAS) populations, which are underrepresented in genetic research yet represent a huge portion of the global population. Using genetic data from Chinese and Japanese cohorts, the researchers tried multiple methods to build the most predictive PRS possible.
The final optimized score, called PRS121, combined 115 variants from an existing well-validated East Asian PRS with 6 additional variants discovered through a new multi-study genetic analysis. This combined approach achieved the best predictive performance among all models tested.
The researchers used a multi-stage approach. They first built candidate PRS models using three different statistical methods applied to GWAS data from Chinese and Japanese populations, totaling over 12,000 CRC cases and 164,000 controls. They then tested how well each candidate score predicted CRC in the China Kadoorie Biobank (CKB) - a large prospective cohort of over 100,000 people followed for an average of 14 years.
The combined PRS121 approach - taking an established 115-variant score and supplementing it with 6 new independently-informative variants - outperformed all other models with a C-index of 0.602. The C-index measures how well a model distinguishes between people who develop disease and those who do not, with 0.5 meaning no predictive power and 1.0 meaning perfect prediction.
PRS121 was then applied in the TARGET-C trial, a real-world CRC screening study involving 2,821 participants across six provinces in China. This trial compared colonoscopy, FIT-only screening, and risk-adapted screening strategies. PRS121 classified participants as high-risk (top 20% of PRS distribution) or low-risk.
In the risk-adapted strategy, high-risk individuals received direct referral for colonoscopy, while low-risk individuals were guided through FIT first - with FIT-positive results then triggering a follow-up colonoscopy. This design aimed to catch more cancers while being smarter about allocating limited colonoscopy resources.
In the China Kadoorie Biobank analysis, people in the top 20% of PRS values had a 2.69-fold higher CRC risk compared to those in the bottom 20%. This is a meaningful difference - identifying people at nearly triple the risk using only a DNA analysis taken at one point in time.
The combination of genetic risk and lifestyle factors proved even more revealing. Lifestyle was scored using five components: diet, alcohol, smoking, physical activity, and body weight. People with both high genetic risk and an unfavorable lifestyle had the highest overall risk - a hazard ratio of 3.32, meaning their risk was more than three times that of people with low genetic risk and a healthy lifestyle.
An important finding was that even for people with high genetic risk, adopting a favorable lifestyle substantially reduced absolute CRC risk. The benefit of healthy living was actually more pronounced among the high-genetic-risk group than among those with low genetic risk, highlighting that lifestyle modification matters most for those most genetically susceptible.
In the TARGET-C trial, PRS121 was also a stronger discriminator than conventional risk factors like age, sex, family history, BMI, and the Asia-Pacific Colorectal Screening (APCS) clinical score. When PRS was added to these conventional predictors, predictive accuracy improved meaningfully across all comparisons.
When it came to actually finding advanced neoplasia (AN) - which includes both CRC and advanced pre-cancerous polyps called advanced adenomas - the PRS-based strategy significantly outperformed FIT-only screening. The detection rate increased from 13.1% with FIT alone to 16.7% with the PRS-based strategy.
Most impressively, the PRS-based strategy found 21.1% of advanced neoplasia cases that FIT completely missed. These were patients who had FIT-negative results (meaning no blood in their stool) but whose high genetic risk score correctly flagged them for colonoscopy - revealing cancer or pre-cancer that would otherwise have been overlooked.
When PRS and the APCS clinical score were combined under an "any high-risk" strategy, the proportion of FIT-missed AN cases detected rose to 37.9%. This combined approach achieved the highest sensitivity (87.2%) of all strategies tested, catching the most cancers overall - though at the cost of requiring more colonoscopies.
The trade-off between sensitivity and specificity was clear: the PRS-based strategy caught more cancers but also directed more people to colonoscopy who did not have AN. The number of colonoscopies needed to find one AN case increased slightly from 2.52 (FIT-only) to 2.75 (PRS-based) - a modest increase for a substantial gain in detection.
The study revealed a fundamental complementarity between PRS and FIT. FIT excels at finding cancers that are already bleeding, but pre-cancerous polyps and early cancers often do not bleed enough to trigger a positive FIT result. PRS, by contrast, identifies people who are genetically predisposed to develop CRC regardless of current bleeding status.
This complementarity explains why combining the two approaches catches so many more cases. FIT is better at specificity - correctly clearing people without cancer - while PRS provides higher sensitivity, meaning it catches more of the cancers that do exist. Together, they cover each other's blind spots.
The researchers note that this study was the first real-world trial to evaluate PRS-based risk-adapted screening in a CRC screening program. Most prior PRS studies only used cross-sectional data or theoretical modeling. The TARGET-C trial provided actual evidence of how PRS performs when deployed in a community screening context.
An important practical advantage of PRS121 is its cost efficiency: it uses only 121 genetic variants, unlike genome-wide PRS methods that require millions of variants. This makes it feasible to implement in clinical or public health screening programs without prohibitive genotyping costs.
This research supports moving colorectal cancer screening toward a more precision medicine approach - one where the type and frequency of screening is matched to an individual's actual risk profile rather than applied uniformly to everyone above a certain age.
For high-risk individuals identified by PRS, a strategy of direct colonoscopy referral can catch cancer earlier and potentially prevent deaths. For lower-risk individuals, a FIT-first approach conserves colonoscopy resources while still providing a safety net for those who develop bleeding symptoms.
The findings also have implications for lifestyle counseling: people who learn they carry a high genetic risk for CRC may be more motivated to adopt healthier behaviors, and the data confirm this motivation would translate into meaningful real-world risk reduction. Communicating genetic risk alongside lifestyle guidance could be a powerful prevention strategy.
Important limitations include the fact that the study was conducted in an East Asian population, so PRS121 may not directly transfer to other ethnic groups. Validation in diverse populations and prospective studies measuring actual impact on long-term CRC mortality are needed before wide clinical implementation.
This study provides strong evidence that polygenic risk scores can meaningfully improve colorectal cancer screening beyond what conventional risk factors alone can achieve. The PRS121 score identified individuals at substantially elevated risk with a relatively simple genetic test, and this information translated into better real-world cancer detection.
The most striking finding was that one in five advanced neoplasia cases detected by PRS would have been completely missed by FIT alone. For patients in that group, a genetic risk-guided referral to colonoscopy may catch pre-cancerous lesions before they become invasive cancer.
Looking ahead, the researchers call for validation across diverse ethnic populations and long-term randomized trials to demonstrate whether PRS-based screening actually reduces CRC mortality rates. Cost-effectiveness analyses specific to different healthcare systems will also be essential for guiding implementation decisions.