Performance of family history-based colorectal cancer screening criteria by race and age at diagnosis in the Disparities and Cancer Epidemiology (DANCE) study

Importance: Family history (FH) and age are the primary criteria employed for early colorectal cancer (CRC) risk stratification. We evaluated how well these criteria identify individuals diagnosed with CRC across age and racial groups. Objective: To evaluate the performance of FH and age based screening criteria for identifying individuals with CRC, with attention to differences by race and age at diagnosis. Design, Setting, and Participants: This case control and case only analysis used data from the Disparities and Cancer Epidemiology (DANCE) cohort, a population based study of invasive CRC cases diagnosed from 2013 to 2022, recruited through the Metropolitan Detroit Cancer Surveillance System and the Louisiana Tumor Registry. Analyses included 1,158 non-Hispanic Black (NHB) and non-Hispanic White (NHW) CRC cases and 1,434 cancer-free controls from the Inflammation Health and Lung Epidemiology (INHALE) study, enrolled from the same Detroit catchment area. Data were analyzed in 2025. Exposures: Self reported cancer FH among first-degree (FD) relatives and grandparents, summarized into three FH-based screening criteria: at least one FD relative with CRC (colon early-screening criterion), any FH of Lynch syndrome related cancers, and meeting NCCN criteria for Lynch syndrome genetic testing. Main Outcomes and Measures: Proportion of cases meeting each FH based screening criterion stratified by race and age at diagnosis (<45, 45 - 49, 50 - 64, and <65 years); case only odds ratios for younger age at diagnosis; and case control odds ratios for CRC associated with each criterion, with race-by-age interaction tested. Results: Cancer FH burden differed by age at diagnosis across both racial groups. First degree (FD) CRC FH was highest among NHB CRC cases diagnosed before age 45 (22.6%) and lowest in those diagnosed at ages 45-49 (8.2%), while NHW participants reported more CRC FH with older age at diagnosis (p interaction=0.011). In case control analyses, having at least one FD relative with CRC was associated with higher odds of CRC before age 45 among NHB (OR=1.44, 95% CI 1.09 - 1.89) but not NHW individuals. The proportion of cases diagnosed before age 45 with a FD CRC FH was low, though markedly higher in NHB than NHW individuals (22.6% vs. 4.0%). While the proportion was slightly higher when including FH of any Lynch syndrome-related cancers (NHB: 24.5%, NHW: 10.0%), the proportion of controls with a FD FH of these cancers also increased. Conclusions and Relevance: Current family history-based criteria fail to identify the majority of individuals diagnosed with CRC before age 45, with performance varying substantially by race, highlighting the urgent need for more equitable and effective approaches to early-onset CRC risk stratification.