Authors:
Adamson, BJS, Cohen, AB, Cheever, MA, Hooley, IJ, Williams, ER, Tymejczyk, O, Peeples, M, Hernandez, M, Meropol, NJ, Uldrick, TS
Background: Patients with HIV are commonly excluded from clinical trials of immunotherapy to treat cancer. Clinical adoption of checkpoint inhibitor therapy after US Food and Drug Administration (FDA) approval is unknown in this population. We aimed to 1) describe the real-world uptake of checkpoint inhibitors (anti-CTLA-4, anti-PD-1, or anti-PDL1) among cancer patients with HIV and 2) compare overall survival (OS) by HIV-infection status of lung cancer patients using these agents. Methods: The retrospective cohort study included patients diagnosed with advanced non-small cell lung cancer (NSCLC) or any of 13 other cancers who initiated systemic cancer therapy between January 1, 2013 and January 31, 2019. The data source was a de-identified nationwide electronic health record (EHR)-derived database from Flatiron Health. HIV status was obtained using a previously described algorithm (ISPOR 2019), with positive cases confirmed. The probability of any checkpoint inhibitor treatment (i.e., nivolumab, pembrolizumab, atezolizumab, avelumab, durvalumab, and ipilimumab) in any line of therapy was compared by HIV status using multivariable logistic regression adjusted for age, race, gender, practice type, and cancer type. Overall survival for advanced NSCLC patients administered checkpoint inhibitors was compared between HIV+ and HIV- patients using unadjusted, adjusted, and propensity score matched analyses. Results: Among 112,761 cancer patients, 304 were identified as HIV+. The HIV+ patients were more likely to be younger, male, and black than HIV- patients. The frequency of real-world treatment with checkpoint inhibitors was 14.8% among HIV+ patients and 20.4% among HIV- patients with adjusted odds ratio of 0.73 (p = 0.09). For patients with advanced NSCLC, in the unadjusted analysis, median OS of HIV- patients administered checkpoint inhibitors (n = 12,073) was 10 months (95% CI: 9.7 - 10.4); the median OS of HIV+ NSCLC patients administered checkpoint inhibitors (n = 19) was 13.5 months (95% CI: 6.9 - NA). After adjusting for age, practice type, race, therapy line number, and stage at diagnosis, the hazard ratio (HR) of OS for NSCLC patients using checkpoint inhibitors was 0.89 (95% CI: 0.48-1.66) for HIV+ vs. HIV-. In the propensity score matched analysis, the overall survival of NSCLC patients administered checkpoint inhibitors was similar comparing HIV+ (n = 19) to matched HIV- (n = 209) patients (HR for death for HIV+: 0.77; 95% CI: 0.44-1.36). Conclusions: Our findings indicate that the use of FDA-approved checkpoint inhibitors among HIV+ patients with cancer is comparable to the HIV- population, even in the context of limited clinical evidence for the HIV+ subgroup. There was no evidence that the effectiveness of checkpoint inhibitors differed by HIV status among patients with advanced NSCLC. Results align with recent initiatives to broaden cancer therapy clinical trial inclusion criteria to allow for more generalizable prospective data, as recommended by the National Cancer Institute Cancer Therapy Evaluation Program and American Society of Clinical Oncology-Friends of Cancer Research HIV Working Group, and support the recent FDA Draft Guidance for Industry regarding cancer clinical trial eligibility criteria for patients with HIV.
Sources:
International Conference on Malignancies in HIV/AIDS