Survival Disparities Among People with HIV and Diffuse Large B-Cell Lymphoma in the Era of Novel Antiretroviral Therapy

Background Diffuse large B-cell lymphoma (DLBCL) is the most frequent AIDS defining malignancy in people with HIV (PWH). Historically, HIV-positive (HIV+) DLBCL patients (pts) had poor outcomes compared to HIV-negative (HIV-) DLBCL pts, largely due to aggressive disease and suboptimal HIV therapies. More recently, novel combined anti-retroviral therapies (ART) have allowed for safe and concurrent management of lymphoma and HIV. However, limited data exists on how this progress has affected survival over time. We aimed to evaluate, on a national level, the trends in overall survival (OS) outcomes for HIV+ compared to HIV- DLBCL pts.

Methods The National Cancer Database was used to identify pts with DLBCL diagnosed from 2004-2018. Only pts with reported HIV status were included. Diagnostic years (yrs) were grouped based on ART availability from 2004-2007 (three drug therapy), 2008-2012 (integrase inhibitors), and 2013-2018 (second generation integrase inhibitors). Kaplan-Meier and Cox regression analyses were performed to compare OS only for pts who received multiagent chemotherapy. Propensity score matching was then performed with 1:1 matching of HIV+ to HIV- pts. Propensity score matching and multivariate analysis were performed with adjustment for age, stage, sex, race, ethnicity, median income, and insurance status. Survival analysis was repeated in a landmarked cohort to limit immortal time bias as a surrogate for early death, excluding patients with death or last-contact within 6-months of diagnosis.

Results Of 146,249 DLBCL pts identified, 6% were HIV+ (N=9,373). Compared with the 136,876 HIV- pts at diagnosis, HIV+ pts were more likely to be younger with median age of 47 yrs (interquartile range [IQR] 40-55) vs. 68 yrs (IQR 57-77), male (78% vs. 54%), of Black race (37% vs. 7%) or Hispanic ethnicity (17% vs. 7%), have B-symptoms (51% vs. 30%) and stage IV disease (48% vs. 37%) (all p<0.001). HIV+ pts were more likely to be uninsured (12% vs. 4%), reside in the lowest income areas (32% vs. 16%), and more likely to receive treatment at academic centers (52% vs. 38%) (all p<0.001). Furthermore, HIV+ pts were significantly more likely to have no documented treatment for their DLBCL (23% vs. 17%, p<0.001). With a median follow up of 3.5 yrs, median OS for all pts treated with multiagent chemotherapy was 8.8 yrs (95% CI 8.6-8.9 yrs). On propensity matched analysis, comparing HIV+ (N=5,112) to HIV- (N=5,112) pts, those with HIV+ DLBCL had significantly worse median OS (6.7 vs. 12.5 yrs), along with shorter 1-year (69% vs. 82%%), 3-year (57% vs. 71%%), and 5-year OS (53% vs. 67%) (all p<0.001). When examining outcomes by access to HIV therapies, the median OS for HIV+ DLBCL pts improved throughout each successive period (2004-2007: median 4.0 yrs, 2008-2012: median 5.9 yrs, 2013-2018: median not reached) (all p<0.001). Furthermore, HIV+ pts had significantly increased early death with an overall 6-month mortality of 18% vs. 10% compared to HIV- pts (p<0.001). This trend persisted through each successive time period, but improved slightly with access to second generation integrase inhibitors (2004-2007: 20% vs.11%, 2008-2012: 20% vs. 6%, and 2013-2018: 15% vs. 6%) (p<0.001). Overall, we found that HIV infection remained a significant predictor of increased risk of death for people with DLBCL on Cox regression throughout the entire time period (hazard ratio [HR] 1.56, 95% CI 1.46-1.70, p<0.001). On landmark analysis among those who were alive at 6 months, HIV+ DLBCL pts continued to have shorter 1-year (85% vs. 91%), 3-year (70% vs. 77%), and 5-year OS (64% vs. 70%) as compared with HIV- DLBCL pts. This trend remained within each successive diagnostic period, though the absolute difference in 5-year OS shortened between HIV+ and HIV- pts (2004-2007: 5-year OS, 59% vs. 69%, p<0.001, 2008-2012: 5-year OS 70% vs. 76%, p<0.01, 2013-2018: 5-year OS, 74% vs. 78%, p<0.001).

Conclusions We present the largest study to date evaluating HIV-associated DLBCL outcomes through the era of novel ART. These data highlight that survival disparities persist among patients with HIV-associated DLBCL despite limiting the analysis to patients who survive at least 6 months. There is a high risk of early mortality among PWH and DLBCL, even in the novel ART era. Further research is necessary to delineate the cause-specific mortality to mitigate early as well as late mortality among this vulnerable population.

Narkhede:TG Therapeutics: Membership on an entity's Board of Directors or advisory committees, Research Funding; Genmab: Research Funding; Genetech: Research Funding; Roche: Research Funding; Gilead: Research Funding; Gilead/Forty-seven: Research Funding; EUSA pharmaceuticals;: Research Funding; ADC Therapeutics: Membership on an entity's Board of Directors or advisory committees; Seagen Inc.: Research Funding. Shah:AbbVie: Membership on an entity's Board of Directors or advisory committees; ADCT: Research Funding; BeiGene: Research Funding; Astrazeneca: Research Funding; Seattle Genetics: Research Funding; Epizyme: Research Funding. Stephens:Genentech: Consultancy; Novartis: Research Funding; Mingsight: Research Funding; Arqule: Research Funding; Karyopharm: Research Funding; JUNO: Research Funding; Acerta: Research Funding; Epizyme: Consultancy; Newave: Research Funding; TG Therapeutics: Consultancy; Lilly: Consultancy; Celgene: Consultancy; Beigene: Consultancy; CSL Behring: Consultancy; AstraZeneca: Consultancy; AbbVie: Consultancy. Hu:ADC Therapeutics: Consultancy; Lymphoma Research Foundation: Research Funding; Bristol Meyers Squibb: Consultancy; Novartis: Consultancy; Morphosys AG: Research Funding; Repare Therapeutics: Research Funding; Caribou Biosciences: Research Funding; CRISPR Therapeutics: Research Funding; Celgene: Research Funding; Genentech: Research Funding. Goyal:Sutro Biopharma: Research Funding; Viracta Therapeutics: Research Funding; SeaGen: Research Funding; UpToDate: Patents & Royalties; 2nd.MD: Consultancy.