Social deprivation index (SDI) as a predictor of racial and ethnic disparity in a large real world cancer database.

Abstract

e23429 Background: Healthcare access and representation can vary significantly by race, ethnicity and neighborhood deprivation. SDI calculation incorporates factors like poverty, education, employment, housing, transportation and family structure associated with residential zip codes. Here we investigated the association between SDI and race/ethnicity for the most common cancer types in the US. Methods: Patients with Bladder (n = 11,148), Breast (n = 34,285), Colorectal (CRC, n = 48,998), Endometrial (EMCA, n = 26,290), Kidney (n = 5,991), Non-small cell lung cancer (NSCLC, n = 72,397), Melanoma (n = 11044), Pancreatic (n = 19,444) and Prostate (n = 16,170) cancers were included from Caris Life Sciences (Phoenix, AZ) with NGS (NGS-592, WES) sequencing. Race and ethnicity was self-reported. Genetic ancestry (GA) was inferred by 36,962 ancestry-informative variants derived from globally diverse Genome Aggregation Database (gnomAD) reference individuals (n = 4,150). 3-digit residential zipcodes were extracted and paired with publicly available SDI (US Census Bureau). Prevalences were calculated across SDI quartiles (Q1-4). Significance was calculated by chi-square with Benjamini-Hochberg corrections. Results: Across all cancer types SDI Q4 was associated with a greater proportion (%) of Black/African American (BAA) and Hispanic or Latino (H/L) patients (all q < 0.001). When incorporating genetic ancestry data, there was also a greater proportion of patients with African (AFR) ancestry and Indigenous Americans (AMR) ancestry with increased SDI quartile (Q4 vs Q1 q < 0.001) (Table 1). Conclusions: Location-based disadvantage is racialized/ethnicized across cancers as B/AA, H/L patients and those with AFR and AMR ancestry are disproportionately more populated in high deprivation neighborhoods (SDI Q4). This disparity may represent barriers to precision oncology access and can severely bias genomic evidence and should be accounted for in translational research. Proportion of each racial/ethnic characteristic by SDI Quartile by cancer. (Q-values compare Q1 vs Q4, all q<0.001. Headers annotated as "Characteristic: SDI Quartile".). Cancer BAA: Q1% BAA: Q2% BAA: Q3% BAA: Q4% H/L: Q1% H/L: Q2% H/L: Q3% H/L: Q4% AFR: Q1% AFR: Q2% AFR: Q3% AFR: Q4% AMR: Q1% AMR: Q2% AMR: Q3% AMR: Q4% Bladder 13 20 25 42 11 17 19 52 13 21 26 40 12 18 23 47 Breast 15 19 23 43 13 17 25 45 15 20 24 40 13 17 29 42 CRC 13 20 24 43 12 19 24 46 14 21 25 40 12 19 25 44 EMCA 13 20 25 43 12 18 20 50 14 20 25 40 12 20 23 45 Kidney 14 19 20 47 15 17 22 46 17 18 19 46 17 18 22 43 NSCLC 15 19 24 41 10 14 25 51 10 26 26 39 9 16 24 51 Melanoma 12 17 23 49 12 15 23 51 12 19 24 45 11 15 28 45 Pancreatic 13 19 25 43 12 15 23 50 13 19 26 42 13 15 27 46 Prostate 14 16 23 47 13 15 24 48 14 17 25 44 13 16 23 48