Disparities in pan-cancer patients undergoing germline cancer risk assessment by self-reported race/ethnicity and ancestry.

Abstract

10508 Background: Disparities in access to germline testing for cancer patients (pts) have been demonstrated; however, disparities in post-testing care are unknown. We sought to evaluate germline findings and subsequent genetic counseling/care in cancer pts undergoing tumor-germline sequencing to explore differences by self-reported ancestry. Methods: Pan-cancer pts were prospectively consented to tumor-normal sequencing via a custom NGS panel (MSK-IMPACT) from 1/2015-12/2019 inclusive of germline analysis up to 88 genes. Germline analysis was performed as a research non-billable test in 97.5% of cases. Referral to clinical genetics service (CGS) was recommended for all pts with a new positive (likely pathogenic/pathogenic) germline variant (PV). Ancestry was defined using self-reported Federal definitions of race/ethnicity and designations of Ashkenazi Jewish (AJ) ancestry. Pts were categorized into mutually exclusive groups: AJ, White, Non-White (Asian, Black/African American, Hispanic, Other), and unknown. All pts self-identifying as Hispanic were classified as such, regardless of race. Abstracted data on germline findings and downstream CGS follow-up were compared across groups using non-parametric statistical tests. Results: Among the 15,775 pts in this cohort (59.6% White, 15.7% AJ, 20.5% Non-White [6.9% Asian, 6.8% Black, 6.7% Hispanic, 0.1% Other], and 4.2% unknown), 2663 (17%) had a PV. AJ pts had the highest rates of PV (n = 683, 27.6%), and Non-White pts had a lower proportion of PV (n = 433, 13.6%) compared to Whites (n = 1451, 15.5%), p < 0.01, with differences mostly due to increased prevalence of moderate/low penetrance variants in White and AJ pts . These findings were consistent across multiple tumor types. Prior knowledge of the PV (424/2663, 16%) was more common in Non-White (19.9%) and AJ (19.2%) than White pts (13.4%), p < 0.01. Among 2239 pts with new PV, all were referred to CGS, and 1652 (73.8%) pts were seen. Non-White pts had lower rates of completing visits (67.7%) than White (73.7%) and AJ pts (78.8%), p < 0.01, with the lowest rates occurring in Black (63%) and Hispanic (68.1%) pts. All pts without a visit (n = 587) received a close out letter including 139 pts (6.2% of pts with new PV) who had no documentation of receipt of results in the medical record. Higher rates of non-disclosure were observed in Non-White (6.7%) compared to White (5.4%) and AJ (3.4%) pts with new PV, p = 0.032; non-disclosure did not vary by gene penetrance. There was a non-significant trend towards lower rates of cascade testing at CGS in Asian and Black pts with ongoing analysis. Conclusions: Even when traditional barriers to genetic testing were minimized, Non-White pts were less likely to receive recommended cancer genetics follow-up for subsequent cancer risk counseling, with potential implications for oncological care, cancer risk reduction, and at-risk family members.