Brief Report: Real-World Outcomes in Patients Living With Human Immunodeficiency Virus and Lung Cancer Treated With Immune Checkpoint Inhibitors.

Introduction
Lung cancer is the most common non–acquired immunodeficiency syndrome (AIDS)–defining malignancy in people living with human immunodeficiency virus (HIV) (PLWH). After widespread adoption of combination antiretroviral therapy for HIV and AIDS, lung cancer became the most common cancer in PLWH1,2 and is projected to remain so through 2030.3 The risk of lung cancer in PLWH is approximately three times higher than in people without HIV (PLWoH),1 even when controlling for higher rates of smoking in PLWH2; HIV infection is an independent risk factor for lung cancer.4,5 In a study using six sites of the HIV/AIDS Cancer Match study, age-adjusted all-cause mortality (hazard ratio [HR], 1.85) and cancer-specific mortality (HR, 1.38) were significantly higher for PLWH and lung cancer compared with PLWoH, even when time to first-treatment was controlled (HR 1.28).6
Immune checkpoint inhibitors (ICIs) have driven improved survival in advanced-stage lung cancer over the last two decades. ICIs are a novel class of drugs that restore antitumor T-cell responses by blocking the inhibition of T-cell activation in the tumor microenvironment and beyond.7,8 Antibodies against (PD-1) and PD-L1 have transformed the landscape of therapy for patients with NSCLC, and are now part of the standard treatment armamentarium in both stage III9,10 and stage IV NSCLC.11,12 However, because of their mechanism of action, ICIs have unique toxicities, or immune-related adverse effects (irAEs), for which predictive biomarkers currently do not exist.
PLWH were initially excluded from clinical trials of ICIs, because of concerns for safety and efficacy in this population.13 However, research including mechanistic studies, retrospective reports, and prospective phase I and II trials have ignited calls to include PLWH in ICI trials.14 A growing body of evidence has resulted in increasing PLWH inclusion in ICI clinical trials,13 but real-world evidence of outcomes and incidence of irAEs remains sparse. In the retrospective Cancer Therapy Using Checkpoint Inhibitors in People Living With HIV-International (CATCH-IT) Consortium, safety and efficacy of ICIs were similar in a cohort of PLWH compared with a matched cohort of PLWoH.15 The CATCH-IT study found no significant difference in irAE rates (19% versus 24%) or overall survival (OS) between PLWH and PLWoH with metastatic NSCLC. Data for the CATCH-IT study were drawn from 33 participating academic cancer centers and comprehensive cancer institutes across the United States, Europe, and Australia. These findings from academic centers provide important evidence, but validation in broader, real-world health care settings is needed. Using population-level health care claims data that represents patients seen at a broader range of clinical institutions, we sought to understand the survival outcomes and incidence of irAEs in PLWH with lung cancer treated with ICIs compared with those without HIV.

Methods

Data Source
This was a retrospective cohort study using the Merative MarketScan database, which comprises the health care experience of more than 46 million individuals with self-insured commercial medical insurance and Medicare supplemental insurance in the United States annually. The database integrates deidentified patient-level health data, including medical inpatient, outpatient, and pharmaceutical claims from participating employers and health plans. This study was deemed exempted from the Institutional Review Board because of absence of identifiable data.

Study Population and Design
Adults aged 18 years and older diagnosed with lung cancer and treated with ICIs between 2015 and 2021 were identified using International Classification of Diseases, Ninth Revision, Clinical Modification codes and categorized into the following two cohorts on the basis of their HIV status: PLWH and PLWoH (Fig. 1).
OS was defined as the time interval between the initiation of ICIs and the date of death or the date of last follow-up for survivors, whichever came first. Death was determined using established operational definitions for claims databases, combining in-hospital death indicators with cases in which no medical claims occurred within 365 days of the last claim. irAEs were determined using International Classification of Diseases, Ninth Revision, Clinical Modification codes.

Statistical Analysis
We performed Kaplan–Meier survival analysis to estimate OS in both cohorts. The survival function was calculated from the start of ICI treatment to the time of death or the last follow-up. We used the log-rank test to assess statistical differences in survival between the cohorts. In addition, the rate of irAEs was evaluated for both groups. We conducted statistical analyses in R version 4.5.1 (R Core Team, The R Foundation for Statistical Computing, Vienna, Austria) using the survival and survminer packages and visualized survival curves with 95% confidence intervals, p values, and risk tables.
To further analyze survival outcomes, we compared OS between patients who experienced an irAE and those who did not, stratified by HIV status. Differences in median OS and statistical significance were reported.

Results

Patient Characteristics
A total of 21,259 patients with lung cancer treated with ICIs were identified in the Merative MarketScan database between 2015 and 2021, of whom 105 (0.49%) were PLWH and 21,154 (99.51%) were PLWoH. PLWH were more likely to be male compared with PLWoH, were younger at treatment initiation, had more comorbidities, and were more likely to have gone to the emergency department (Table 1). There were no differences in type of ICI received between the two groups (Supplementary Table 1). At the end of follow-up (censor date December 31, 2021), 74 PLWH (70.48%) and 15,189 PLWoH (71.80%) had died.

Overall Survival Outcomes
The median OS was comparable between PLWH and PLWoH (343 days versus 364 days, p = 0.57). Kaplan–Meier survival analysis revealed overlapping survival curves with no statistically significant difference between the two groups throughout the follow-up period (Fig. 2A). The survival curves remained nearly identical from treatment initiation through long-term follow-up, indicating that HIV status did not significantly impact survival outcomes in patients with lung cancer treated with ICIs.

Impact of irAEs on Survival
In both cohorts, patients who experienced irAEs exhibited significantly improved OS compared with those without irAEs. Among PLWH, those with irAEs had significantly better survival outcomes than those without irAEs (p < 0.0001) (Fig. 2B and C). This survival benefit associated with irAE development was similarly observed in PLWoH (p < 0.0001), suggesting that irAE occurrence may serve as a positive prognostic indicator regardless of HIV status.

Incidence of irAEs
The incidence of irAEs was similar between PLWH and PLWoH (59.1% versus 58.8%, respectively). Among patients who experienced irAEs, the majority in both groups experienced toxicities in only one organ system (54.84% of PLWH versus 53.50% of PLWoH). PLWH experienced irAEs affecting up to three organ systems, whereas PLWoH experienced irAEs in up to eight organ systems (Supplementary Table 2).
The pattern of organ system involvement differed somewhat between groups. In PLWH, the most common irAEs were neurologic, endocrine, renal, and cardiac. In contrast, PLWoH most typically experienced endocrine, renal, cutaneous, and cardiac irAEs. Incidence was statistically significantly different in neurologic and endocrine irAEs. (Supplementary Table 3).

Discussion
This real-world analysis of more than 21,000 patients with lung cancer treated with ICIs provides important insights into the safety and efficacy of immunotherapy in PLWH, a population historically excluded from clinical trials.
The median OS of approximately 11 to 12 months in both groups is consistent with real-world outcomes reported in the CATCH-IT study,15 which found no significant difference in survival between PWH and PLWoH with metastatic NSCLC treated with ICIs. These findings have important clinical implications, as they revealed that HIV status should not be a barrier to ICI treatment decisions in appropriate patients with NSCLC. The overlapping survival curves throughout the follow-up period indicate that any potential immunosuppression associated with HIV infection did not meaningfully impact ICI effectiveness, challenging historical concerns about reduced efficacy in this population. However, the relatively short median survival in both groups underscores the continued need for improved therapeutic strategies in advanced lung cancer, regardless of HIV status.
In addition to similar efficacy outcomes, the safety profile of ICIs in PLWH exhibited overall similarities to PLWoH, similar to previously reported in CATCH-IT,15 but with some notable differences in toxicity patterns. The different patterns of irAE distribution between PLWH and PLWoH contrast with the generally similar overall irAE rates, suggesting that HIV status may influence which organ systems are most susceptible to immune-related toxicity. The higher rate of neurologic irAEs in PLWH has important clinical implications, as neurologic toxicities can be among the most serious and difficult to manage irAEs, potentially requiring closer monitoring in this population. These distinct patterns of organ-specific irAEs raise important questions on whether HIV infection alters the presentation of immune-related toxicities, which could inform monitoring strategies and toxicity management protocols. However, these differences may also reflect variations in clinical monitoring practices or health care utilization between the two populations, highlighting the importance of standardized toxicity assessment in future prospective studies.
Several limitations should be acknowledged. The claims-based nature of this analysis limits the granularity of clinical information available, including HIV viral loads, CD4 counts, cancer staging, and detailed treatment histories. The relatively small number of PLWH limits statistical power for subgroup analyses. In addition, irAE identification through diagnosis codes may underestimate mild toxicities that do not result in coded health care encounters and overestimate attribution of toxicity to ICIs, particularly neurologic and cardiac irAEs.
Despite these limitations, to our knowledge, this study represents one of the largest real-world analyses of ICI outcomes in PLWH with lung cancer to date. These findings support current efforts to revise HIV-related exclusion criteria from clinical trials and ensure that PLWH have equitable access to standard-of-care immunotherapy.

Conclusions
This population-based analysis of 21,259 patients with lung cancer suggest comparable ICI outcomes between PLWH and PLWoH. The findings provide real-world evidence that HIV status does not affect survival or overall toxicity rates, supporting the use of ICIs as standard-of-care treatment for eligible PLWH. Notably, the differing patterns of organ-specific irAE highlight the need for further studies into underlying mechanisms and predictors of toxicity. Inclusion of PLWH in clinical trials remains important for optimizing care in this population.

CRediT Authorship Contribution Statement
Melinda Hsu: Conceptualization, Methodology, Validation, Resources, Writing – Original draft, Visualization, Funding acquisition.
Fangzhou Liu: Methodology, Software, Validation, Formal analysis, Data Curation, Writing – review & editing, Visualization.
Ravi Kyasaram: Software, Validation, Formal analysis, Investigation, Data Curation, Writing – review & editing.
Jeffrey Zhong: Data Curation, Writing – review & editing.
Afshin Dowlati: Methodology, Resources, Writing – review & editing, Supervision.

Disclosure
Dr. Hsu reports receiving consulting fees from Johnson and Johnson and Amgen; and participates in advisory board meetings for Takeda, AstraZeneca, and BMS. Dr. Dowlati reports participating in advisory board meetings for AstraZeneca, Jazz Pharmaceuticals, Seattle Genetics, BMS, Prelude Therapeutics, and Puma Biotechnology. The remaining authors declare no conflict of interest.