AUTHORS
Yinting Liu, Jingkun Li2, Yan Zhuang3, Haibo Cheng4, Jing Luo5, Yunxuan Deng6, Iris Luo7, Samantha M Thomas8, Jingke Yang, MD, PhD9, Jennifer K Plitcha10
AFFILIATIONS
1 University of Nebraska Medical Center, Omaha, Nebraska, United States
2 Harbin Medical University, Harbin, China
3 Duke University, Durham, North Carolina, United States
4 Virginia Polytechnic Institute & State University, Roanoke, Virginia, United States
5 College of Liberal Arts & Sciences, Tempe, Arizona, United States
6 Duke University, Durham, North Carolina, United States
7 University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States
8 Duke University, Durham, North Carolina, United States
9 Academy of Clinical Research and Study, Washington, DC, United States
10 Duke University, Durham, North Carolina, United States
ABSTRACT
Background: Combined-modality therapy with multiagent chemotherapy and radiotherapy (RT) remains an accepted approach for selected patients with early-stage diffuse large B-cell lymphoma (DLBCL). However, racial and ethnic differences in receipt and timing of RT, and whether the association between combined-modality therapy and survival varies across groups, remain incompletely characterized. We evaluated treatment receipt, chemotherapy-to-RT interval, and race-specific survival associations among non-Hispanic (NH) White, NH Black, and Hispanic patients with stage I-II DLBCL.
Methods: We identified adults with stage I-II DLBCL in the National Cancer Database who received multiagent chemotherapy from 2004-2022. Treatment selection was analyzed using facility-clustered multivariable logistic regression comparing chemotherapy alone with chemotherapy followed by external-beam RT. Among patients receiving RT, chemotherapy-to-RT interval was modeled on the log scale. Survival analyses used a modern 2018-2021 cohort and a 12-month landmark to reduce guarantee-time bias. Race-specific propensity scores and overlap weights were estimated separately within each group; stabilized inverse probability of treatment weighting (IPTW) was used as a sensitivity analysis. We estimated overall survival (OS), 36-month OS differences, and restricted mean survival time (RMST) differences. Heterogeneity was evaluated in a pooled overlap-weighted Cox model with treatment-by-race/ethnicity interaction terms. Because proportional-hazards assumptions were not uniformly satisfied, fixed-horizon OS and RMST estimates were emphasized.
Results: The treatment-selection cohort included 72,686 patients: 61,722 NH White patients (median age, 67 years; 55.9% male), 5,059 NH Black patients (median age, 57 years; 52.8% male), and 5,905 Hispanic patients (median age, 59 years; 53.0% male). Compared with NH White patients, NH Black patients were less likely to receive chemotherapy plus RT (adjusted odds ratio [aOR], 0.84; 95% CI, 0.77-0.90), corresponding to an adjusted absolute difference of -3.6 percentage points (95% CI, -5.1 to -2.1). Hispanic patients were also less likely to receive chemotherapy plus RT (aOR, 0.87; 95% CI, 0.78-0.97), with an adjusted difference of -2.7 percentage points (95% CI, -4.9 to -0.6). Median chemotherapy-to-RT intervals were 104, 127, and 124 days for NH White, NH Black, and Hispanic patients, respectively. Relative to NH White patients, adjusted geometric mean intervals were 7% longer for NH Black patients (ratio, 1.07; 95% CI, 1.05-1.09) and 4% longer for Hispanic patients (ratio, 1.04; 95% CI, 1.02-1.07).
The modern survival cohort included 15,744 patients and 1,920 deaths: 13,109 NH White, 1,091 NH Black, and 1,544 Hispanic patients. All post-overlap-weighting absolute standardized mean differences were <0.10. In the pooled analysis, chemotherapy plus RT was associated with lower mortality (hazard ratio, 0.72; 95% CI, 0.62-0.83). Adjusted 36-month OS differences favoring chemotherapy plus RT were +4.0 percentage points in NH White, +2.0 in NH Black, and +3.9 in Hispanic patients; corresponding RMST differences were +1.06, +0.39, and +0.96 months. There was no evidence that the treatment-associated survival relationship differed by race or ethnicity (global interaction P=0.783). Stabilized IPTW yielded concordant conclusions.
Conclusions: NH Black and Hispanic patients with early-stage DLBCL were less likely to receive chemotherapy plus RT and had longer chemotherapy-to-RT intervals than NH White patients. Combined-modality therapy was associated with better survival overall, but this association did not differ significantly across racial and ethnic groups. These findings identify treatment access and delivery, rather than demonstrated differences in treatment-associated survival, as the principal areas of disparity.
DISCLOSURES
Yinting Liu: No relevant conflicts of interest to declare.