Impact of Race and Social Determinants of Health on Outcomes in Patients with Aggressive B-Cell Lymphomas Treated with Chimeric Antigen Receptor T-Cell (CART) Therapy

Background:

Health care disparities, driven by multiple social, economic, and/or environmental factors lead to inequalities in health outcomes. With CAR T-cell therapy (CART), differences in response rates have been identified across race. However, data that address the convergence of race with other social determinants of health on outcomes with CART remain limited. We examined the impact of interactions between race, insurance status, clinical trial utilization, and age on access to care and outcomes in patients (pts) treated with CART for aggressive B-cell lymphomas (B-NHL).

Methods:

Adult pts with relapsed/refractory B-NHL treated with CD19 CART between 2015- 2021 across 13 US academic centers were identified. Insurance type, demographic and clinical data were collected and analyzed via Chi-squared and Kaplan-Meier analysis. Cox multivariable regression was used to determine impact of race/ethnicity and other variables on survival. Variables used in the multivariable analysis (MVA) included race, age, insurance type, prior autologous stem cell transplantation (autoSCT), CART on clinical trial, use of bridging therapy, IPI at CART infusion, and elevated LDH pre-CART. Race (Caucasian, African American (AA), Asian or Other with the latter cohort excluded from analyses) and ethnicity (Hispanic vs non-Hispanic) were self-reported by pts.

Results:

A total of 466 adult pts were included. Table 1 outlines demographic, clinical, and health care utilization parameters by race. 406 (87%) were Caucasian, 34 (7%) AA and 26 (6%) Asian; 9 (2%) were Hispanic (all Caucasian). Caucasians were older compared to AAs and Asians (median age 59 vs 55 vs 55 yrs; p=0.004). Median number of lines of therapy prior to CART were similar across race (p=0.44). Caucasians were more likely to have a prior autoSCT (p= 0.04). For Caucasians vs AAs vs Asians: rates of bridging were 44% vs 61% vs 46% (p=0.17) and rates of utilization of clinical trials for CART access were 29% vs 15% vs 19% (p=0.12), respectively. Median time from last relapse/progression pre-CART to CART was 2 vs 1.8 vs 1.2 mo (p=0.9) and median time from apheresis to CART was 1.1 vs 1.1 vs 0.9 mo (p=0.2), respectively. D180 ORR were 51% vs 46% vs 19% (p =0.01) with D180 CR rates of 43% vs 42% vs 19% (p = 0.17), respectively. No difference in grade >/= 3 CRS or ICANS was observed across race.

Median follow-up after CART was 12.7 mo. Median PFS (mPFS) was longer for Caucasians (11.5 mo) than for AAs (3.5 mo, HR 1.56 [1.03-2.4], p=0.04) or Asians (2.7 mo, HR 1.7 [1.02-2.67], p=0.04). Differences in median OS (mOS) were not statistically different: mOS was 25.4 mo in Caucasians vs 30.4 mo in AAs (HR 1.0 [0.59 – 1.69], p=0.99) vs 16.8 mo in Asians (HR 1.42 [0.84 – 2.42], p=0.19).

Rates for administration of therapy after CART failure were non-significantly higher for Caucasians vs AAs or Asians, (46% vs 25% vs 28%, p = 0.05). On MVA, race, insurance type, use of bridging, and elevated LDH pre-CART impacted PFS (p < 0.001 – 0.031, Table 2). As compared to Caucasians, AAs (but not Asians) had a worse PFS (HR 1.72 [1.05 – 2.82], p = 0.03). Race did not impact OS. Medicare, lack of bridging, and receipt of further therapy post CART failure positively impacted OS (p = 0.006 - 0.044).

Insurance type did not impact median time from last relapse/progression to CART. However, there were significant differences in mPFS and mOS between payer groups: mPFS was 15.9 vs 4.2 vs 6.0 vs 0.9 mo (p<0.001) and mOS was 31.2 vs 12.8 vs 21.5 vs 3.2 mo (p<0.001) for Medicare (n=206) vs Medicaid (n=33) vs private insurance (n=219) vs self-pay (n=7).

Conclusion:

In our cohort, AAs and Hispanics were greatly underrepresented which underscores the presence of health care disparities across race and ethnicity for relapsed/refractory B-NHL. With CART, Caucasians had similar complete response rates but longer mPFS as compared to AAs and Asians warranting exploration of factors that may predict for early progression in racial minorities. No difference in OS was observed by race perhaps attributed to the effects of next line therapy after CART. Insurance type impacted both PFS and OS with best outcomes seen in pts with Medicare. Results do not capture pts facing insurance denial and/or poor access that would likely augment trends for race and socioeconomic disparities identified in our CART population. The interaction between race and insurance status and their relative contributions to access and outcomes with CART should be further explored.