Association of sociodemographic, environmental, and health factors with early-onset cancer in a diverse urban cohort.

10592 Background: Early onset cancer (EOC; diagnosed before age 50) is rising globally, yet its drivers are incompletely understood, especially in diverse community-based populations. We examined sociodemographic, environmental, and health factors associated with EOC in the community-based Chicago Multiethnic Prevention And Surveillance Study (COMPASS) cohort spanning 72 Chicago neighborhoods. Methods: This nested case-control study within COMPASS matched EOC cases to controls without cancer 1:2 by age and sex. Standardized mean differences (SMD) compared sample characteristics. Conditional logistic regressions evaluated EOC associations with odds ratios (OR) and 95% confidence intervals (CI), adjusted for race/ethnicity. Meaningful imbalance was defined as SMD>0.2 and statistical significance as p<0.05. Results: Among 13,309 COMPASS participants, 265 EOC cases were matched to 530 controls. While the full sample was well balanced by sex (51.4% male), cases were predominantly female (72.8%) and non-Hispanic Black (72.3%). Breast (28.3%) and cervical (16.2%) cancers were most reported. A greater proportion of cases identified as unemployed/disabled but were also college educated and with incomes ≥$35,000 (SMD=0.382, 0.326, 0.267). Parental cancer history was more common in cases (SMD=0.504). No imbalances were observed in neighborhood-level characteristics, including area deprivation index, particulate matter concentration, and healthcare access strain. Cases reported more recent mammography and prostate-specific antigen screening (SMD 0.385, 0.407) but not pap smears (SMD=0.195). Cigarette use did not differ (SMD=0.141). Cases had higher prevalence of type 2 diabetes (T2D), anxiety, and depression (SMD=0.216, 0.302, 0.242), and lower prevalence of allergies (SMD=0.229). EOC was associated with parental cancer history (OR=3.62, CI 2.31-5.67, p<0.001), T2D (OR=1.65, CI 1.02-2.69, p=0.043), and anxiety (OR=2, CI 1.24-3.23, p=0.005), but not sociodemographic or neighborhood factors. Conclusions: In this diverse cohort, EOC was associated with multiple factors, including family cancer history, breast/prostate screening, and select health comorbidities. No associations were observed with reported neighborhood-level environmental metrics, current tobacco use, or cervical cancer screening. Our results highlight challenges in measuring sociodemographic and environmental contributors to EOC. Our results align with existing hypotheses on cancer development: familial risk, the potentially protective role of allergies, and T2D-related cancer risk. While causality cannot be inferred, the observed EOC-screening and EOC-anxiety associations warrant further investigation with longitudinal, multilevel research to better define EOC drivers and inform EOC prevention and detection strategies.