Germline multigene panel testing identified pathogenic or likely pathogenic variants in 15.2% of patients with colorectal cancer (CRC) — roughly one in seven — including patients diagnosed at age 50 or older, according to a systematic review and meta-analysis published in The Lancet Gastroenterology & Hepatology.
Lead and corresponding author Alessandro Mannucci, MD, of Vita-Salute San Raffaele University and IRCCS San Raffaele Hospital in Milan, told GI & Hepatology News the findings support offering testing to a broader group of patients.
“Germline multigene panel testing is currently underused in colorectal cancer, and more patients should be offered it,” Dr. Mannucci said. “The diagnostic yield remained substantial even after we excluded patients diagnosed before age 50, which indicates that the benefit of testing extends well beyond the age thresholds that guide current practice.”
Current guidelines recommend routine germline evaluation for patients diagnosed before age 50 and conditional testing thereafter, the authors noted.
The meta-analysis included 21 studies identified in searches of four databases from inception through Aug. 2, 2025. Eligible studies offered germline testing to consecutive patients with CRC using next-generation sequencing panels of at least five genes, without selection based on family history, mismatch repair status, or other hereditary criteria. The primary outcome was the proportion of patients with at least one pathogenic or likely pathogenic variant in a cancer predisposition gene.
The primary analysis pooled 12 studies including 6,925 patients. Cohorts limited to patients diagnosed before age 50 were excluded from that analysis but included in age-stratified analyses, which pooled 16 studies (7,444 patients) for early-onset disease and 11 studies (5,243 patients) for those diagnosed at 50 or older. Meta-regression assessed the effects of cohort median age at diagnosis, panel size, family history, and publication year.
Results by age at diagnosis
Across the primary analysis, 15.2% of patients had at least one pathogenic or likely pathogenic germline variant. High-penetrance variants were identified in 7.9%, including high-penetrance CRC predisposition variants in 5.4%.
About 1.7% of patients carried low-penetrance variants — APC p.I1307K or a single MUTYH variant — that the authors noted lack consensus management implications; excluding those, the overall yield was 13.0%.
Among patients diagnosed before age 50, the pooled diagnostic yield was 17.7%, with high-penetrance variants detected in 14.2%.
When patients diagnosed before age 50 were excluded, the yield for any pathogenic or likely pathogenic variant was 13.0% among those diagnosed at age 50 or older (10.5% after excluding low-penetrance variants), and 6.5% had high-penetrance variants. In this group, high-penetrance variants in genes predominantly associated with CRC were found in 3.8%.
Meta-regression showed that the yield of high-penetrance variants declined with increasing age at diagnosis but remained above 5% — a pre-test probability the authors noted has been used to recommend testing — until approximately age 67.
Dr. Mannucci cautioned against treating that figure as a strict cutoff.
“The 67-year figure deserves careful framing,” he said. “It comes from a meta-regression of cohort-level yield against cohort median age. That analysis estimates a between-study association rather than the relationship between an individual patient’s age and the probability of a germline finding.”
He said the estimate should therefore be viewed as a guide for further study rather than a definitive testing cutoff.
“The confidence interval is more informative than the point estimate,” he explained. “The predicted yield for high-penetrance variants crosses the 5% clinical testing threshold somewhere between roughly 59 and 75 years of age.”
Panel size and gene categories
The number of genes included in a panel had little effect on overall diagnostic yield. Larger panels did not significantly increase detection of pathogenic or likely pathogenic variants overall or of high-penetrance variants.
“A second takeaway is that yield did not depend on how many genes were tested, provided the panel included the principal high-penetrance genes,” Dr. Mannucci said. “Testing more patients matters more than testing more genes.”
In patients diagnosed at 50 or older, however, larger panels were associated with a statistically significant increase in detection of variants of any penetrance, though not of high-penetrance variants.
The study also found different age-related patterns across gene categories. Variants in genes predominantly associated with CRC declined with age, whereas high-penetrance genes traditionally associated with other cancers were more stable across age groups. Those non-CRC genes accounted for a 1.5% diagnostic yield overall and were driven largely by BRCA1, BRCA2, and CHEK2.
Dr. Mannucci said some of these findings may simply reflect how common the variants are in the general population, although the genes may also make a small contribution to CRC risk. In the paper, the authors noted that the 1.5% yield was nearly three times the roughly 0.5% background carrier rate reported in population sequencing studies.
He added that the value of broader testing extends beyond the patient with CRC.
“The value of a positive result is not limited to the patient in front of you,” he said. “Identifying a pathogenic variant opens cascade testing across the family, so a single decision to test can benefit many relatives, and the benefit carries forward across generations. When testing is viewed in that light, the yield we report is clinically meaningful.”
Limitations and next steps
The authors noted between-study heterogeneity that may partly reflect differences in panel composition. Most cohorts had a median age at diagnosis below the U.S. population median, likely because many studies were conducted at tertiary or academic centers, and data from patients older than 80, community practices, and lower-income settings were limited. The analysis relied on each study’s original variant classifications, and several recently identified predisposition genes were not tested in earlier studies, which the authors said could mean yields were underestimated. Formal cost-effectiveness analyses are needed, they wrote, and cascade testing of relatives, patient preferences, and psychological outcomes were beyond the scope of the study.
Dr. Mannucci said the findings support broader access to testing rather than a specific age at which testing should stop.
“Implementation still requires clinical judgment, incorporating family history, personal cancer history, tumor characteristics, and validated risk models,” he said.
He identified prospective evaluation of broader testing strategies, formal cost-effectiveness analyses, and studies in older patients and patients treated outside academic centers as key next steps.
Dr. Mannucci reported no competing interests. Coauthors reported relationships with Myriad Genetics, Ambry Genetics, Natera, GlaxoSmithKline, Janssen, Nouscom, UpToDate, and Sanofi Pasteur, including royalties for the PREMM risk model. The study was supported by the Italian Ministry of University, EU–Next Generation EU, the National Cancer Institute, and other institutional or research grants.