Infections and their impact on patients on pembrolizumab-based therapies for head and neck cancer

• Published in: Journal of clinical oncology Vol. 40; no. 16_suppl; p. 6035
• Main Authors: Zhang, Yuqi, Xu, Jiaqiong, Muhsen, Ibrahim, Burns, Ethan, Shah, Shivan M, Umoru, Godsfavour, Mylavarapu, Charisma, Sun, Kai, Crenshaw, Aubrey, Esmail, Abdullah, Kieser, Ryan Blair, Guerrero, Carlo, Gong, Zimu, Gee, Kelly, Heyne, Kirk, Singh, Monisha, Zhang, Jun, Abdelrahim, Maen
• Format: Journal Article
• Language: English
• Published: 01.06.2022
• ISSN: 0732-183X; 1527-7755
• DOI: 10.1200/JCO.2022.40.16_suppl.6035

6035 Background: Immune checkpoint inhibitors (ICIs) stimulate the antitumor activity of the adaptive immune response and are used in combination with chemotherapy or as a monotherapy in metastatic head and neck squamous cell carcinoma (mHNSCC). Infections are known to impact outcomes in patients receiving chemotherapy, but the incidence and impact in patients receiving ICI is poorly defined. This analysis aims to identify the infectious burden and associated impact on clinical outcome in patients receiving pembrolizumab for mHNSCC. Methods: This is a retrospective study across 7 hospitals analyzing patients who received pembrolizumab from 1/1/2017-8/1/2021 and followed until 12/1/2021. Patients with mHNSCC that received pembrolizumab were dichotomized into infected and uninfected cohorts. Comparative covariates included gender, race, ECOG, comorbidities, anti-infectives use at ICI initiation, chronic infections, line of therapy (first or 1L, second or 2L, or > 2L). Risk factors for infections were assessed by univariable and multivariable analysis with reported odds ratio (OR) and 95% confidence intervals (CI). Morbidity and mortality was assessed by all-cause emergency department (ED), inpatient, and intensive care unit (ICU) admissions, median cycles of ICI-based therapy, and overall survival (OS). OS was assessed via Kaplan-Meier analysis. P-value < 0.05 was considered statistically significant. Results: Of the 45 evaluable patients with mHNSCC, 22 (48.9%) had at least one infection. Baseline characteristics and comorbidities were similar across cohorts. Risk factors for developing infection included anti-infective use at ICI initiation (OR 13.27, 95% CI: 1.13-156.53, p = 0.04). Fewer infections were observed with ICI as > 2L compared to 1L (OR 0.1, 95% CI: 0.01-0.91, p = 0.041). Infection was the primary cause of death in 2 (9.5%) patients. Compared to the uninfected, infected patients received fewer cycles of therapy (4 vs 7, p = 0.026), and were more frequently admitted to the hospital [20 (90.9%) vs. 8 (34.8%), p < 0.001)] and ICU [5 (11%) vs. 0, p = 0.022]. Number of ED visits were similar at 8 (36.4%) in the infected vs. 5 (21.7%) in the uninfected (p = 0.28). At median follow-up of 8 months, 13 (59.1%) in the infected cohort and 8 (34.8%) in the uninfected had died (p = 0.1). Median OS was significantly lower in the infected (9.2; 95% CI 4.9-12.8 months) compared to the uninfected (28.1; 95% CI 14.3-unreached months) (p = 0.009). Conclusions: For mHNSCC patients on pembrolizumab, infections are more common in the 1L setting and is associated with fewer treatment cycles, more hospitalizations, and significantly shorter OS. Risk factors for infection include anti-infective use at ICI initiation.