Impact of ECOG performance status on recurrent/metastatic head and neck squamous cell carcinomas treated with anti-PD1 inhibitors.

Abstract

e18004 Background: Anti-PD1 checkpoint inhibitors (ICI) represent an established standard of care for patients with recurrent/metastatic head & neck squamous cell carcinoma (RMHNSCC). Landmark studies excluded patients with ECOG performance status (PS) ≥ 2; the benefit of ICI in this population is therefore unknown. Methods: We retrospectively reviewed RMHNSCC patients who received at least 1 dose of ICI at our institution. Demographic data and clinical outcomes were obtained; the latter included objective response to ICI (ORR), physician-documented CTCAE grade 2+ toxicity (irAE), and any unplanned hospitalization within 100-days of last ICI dose (UH). Associations between demographic data and clinical outcomes were explored using both uni- and multivariate analysis. Overall survival (OS) was estimated using a Cox proportional hazards model; ORR, irAE, and UH were evaluated with logistic regression. This project was approved by our institutional IRB. Results: We identified 152 RMHNSCC patients who were treated with ICI between 1/2013 and 1/2019. ECOG PS was 0 in 42 (27%), 1 in 75 (50%), 2 in 27 (18%), 3 in 2 (1%), and unknown in 6 (4%) patients. The median age was 61 (range: 25 - 90). 124 (82%) were male, 124 (82%) were white, and 69 (45%) were never-smokers. The most common primary sites were the oropharynx (n = 59, 40%), oral cavity (n = 39, 26%), nasopharynx (n = 11, 7%), and larynx (n = 10, 6%). 54 (36%) were p16+ oropharynx cancers. CPS score was available in 10 (6.6%). Single agent ICI was received by 118 (77%) patients. 66 (44%) had a documented irAE and 54 (36%) had an UH. A multivariate model for OS containing PS, smoking status and HPV status showed a strong association between inferior OS and ECOG 2/3 compared to 0/1 (p < 0.001; HR = 3.30, CI = 2.01-5.41), as well as former (vs. never) smoking status (p < 0.001; HR = 2.17, CI = 1.41-3.35). Current smoking (p = 0.25) did not reach statistical significance. On univariate analysis, poor PS was associated with inferior ORR (p = 0.03; OR = 0.25, CI = 0.06-0.77) and increased UH (p = 0.04; OR = 2.43, CI = 1.05—5.71). There was no significant association between irAE and any patient characteristic. Conclusions: We observed inferior overall survival among ICI-treated RMHNSCC patients with ECOG 2/3 in our single-institution, retrospective series. Our findings help frame discussion of therapeutic options in this poor-risk population. Further study must be done to determine which interventions are of greatest benefit for RMHNSCC patients with declining performance status.