Does ethnicity matter? A two-statewide population based analysis of outcomes of primary mediastinal B-cell lymphoma in Hispanics.

Abstract

e19568 Background: Primary mediastinal B-cell lymphoma (PMBCL) is considered a rare distinct clinicopathological and molecular subtype of diffuse large B-cell lymphoma (DLBCL), mainly affecting females in their 4th and 5th decades [1,2,3]. Since the 1960s, there has been a significant increase in the number of immigrants to the US; and the states of Florida (FL) and Texas (TX) are well known for their Hispanic (HI) enriched population. Since there is scarce data of PMBCL in HI; we researched demographics, treatment patterns and outcomes in H compared to Non-Hispanics (NH) from FL and TX. Methods: This is a retrospective study of patients diagnosed with PMBCL from the Texas Cancer Registry (TCR) and the Florida Cancer Data System (FCDS), with years of analysis 2006-2017. Patients were identified by the International Classification of Diseases for Oncology Third Edition (ICD-O-3). Key variables collected included gender, race, ethnicity, dates at diagnosis and death, primary payer at diagnosis, stage, type of treatment, poverty index, and vitality status. The significance of variation in the distribution of categorical outcomes with ethnicity (H, NH) was assessed with Fisher’s Exact tests or Pearson’s Chi-square tests as appropriate. Survival distributions were described with Kaplan-Meier curves, and the significance of variation in median survival with ethnicity was assessed with log-rank testing. All statistical testing was two-sided with a significance level of 5%. Results: We had 2 cohorts; cohort A (A) includes 19 H vs. 89 NH from TX, cohort B (B), 27 H vs. 121 NH from FL. The median age at diagnosis in A was 33 years (y) in H vs 40 y in NH ( p-value = 0.091); while median age in B was 36 y in H vs 46 y in NH ( p-value = 0.005). Regarding poverty index; in A, most of H were in the 10-20% bracket, and most NH were in the 5-10% bracket. On the contrary, in B, most of H were in the 5-10%, and most NH were in the 10-20% bracket. Lack of insurance was more prevalent in H in A vs. NH, 42% vs. 12%, respectively, where it was less prevalent in H vs. NH in B, 8% vs. 5%. Multiple agents of chemotherapy were delivered to 74% of H, 70% of NH vs. 82% of H and 65% of NH with p-value = 0.292 and 0.635 in A and B, respectively. In A, the NH had better median survival (m) than H, not reached vs. 10.3 y, respectively. In B, H had better m than the NH, 10.7 y vs. not reached, respectively. The 5 y survival probability in A was even among H and NH; 0.702 [95% CI 0.49-1] and 0.732 [95% CI 0.578 - 0.9280], respectively. While 5 y survival probability in B was higher in H compared to NH, 0.923 [95% CI 0.675 -0.869] and 0.766 [95% CI 0.434 - 0.615], respectively. The overall survival (OS) difference in A and B, H vs. NH was not statistically significant. Conclusions: Despite some identified demographic differences in patients diagnosed with PMBCL, comparing H vs. NH on each cohort, there was no statistically significant difference in the overall survival probability.