Off-label use of checkpoint inhibitor (CPI) monotherapy in PD-L1–negative or unknown recurrent/metastatic head and neck cancer (R/M HNSCC).

person Margaret Stalker Hospital of the University of Pennsylvania, Philadelphia, PA info_outline Margaret Stalker, Kewen Qu, Wei-Ting Hwang, Roger B. Cohen, Ronac Mamtani, Lova Sun

Authors person Margaret Stalker Hospital of the University of Pennsylvania, Philadelphia, PA info_outline Margaret Stalker, Kewen Qu, Wei-Ting Hwang, Roger B. Cohen, Ronac Mamtani, Lova Sun Organizations Hospital of the University of Pennsylvania, Philadelphia, PA, University of Pennsylvania, Philadelphia, PA, Department of Biostatistics, Epidemiology, and Informatics, University of Pennsylvania, Philadelphia, PA, Penn Medicine Abramson Cancer Center, Philadelphia, PA Abstract Disclosures Research Funding No funding sources reported Background: The KN-048 study established that the CPI pembrolizumab with chemotherapy should be used for PD-L1 negative R/M HNSCC, while CPI monotherapy has limited efficacy (ORR 4%) in PD-L1 negative disease and should be reserved for PD-L1 positive disease. National patterns of PD-L1 testing and PD-L1 guided treatment selection are unknown. We examined PD-L1 testing rates, the use of CPI overall, and the use of CPI monotherapy in PD-L1 negative or unknown R/M HNSCC (“off-label”), and associated factors. Methods: This retrospective analysis included adult patients starting treatment for R/M HNSCC from 2019-2023 in the Flatiron Health electronic health record (EHR)-derived de-identified national database. Demographics, treatment type, and PD-L1 test results were summarized using descriptive statistics. Specifically, first-line therapy was categorized as CPI monotherapy, CPI with chemotherapy, or chemotherapy and/or cetuximab without CPI. “Off-label” use of CPI was defined as single-agent use without concurrent chemotherapy in patients with negative or unknown PD-L1. Factors associated with “off-label” use were identified using multivariable logistic regression analysis. Results: Our cohort included 3,395 patients with median age 66 (IQR 59-73), 65% White, 77% treated in a community setting, 76% with smoking history, 37% HPV positive, and 72% ECOG PS 0-1. Almost half of patients (44%) did not have a recorded PD-L1 test result; of those with known PD-L1 status (n=1886), distribution of combined positive score (CPS) 0, 1-19, and 20 was 19%, 41%, and 40%, respectively. The most common frontline treatment was CPI monotherapy (43%), followed by chemotherapy/cetuximab (33%) and CPI with chemotherapy (25%). CPI monotherapy use was highly prevalent in patients aged ≥75 (54%) and with ECOG PS ≥ 2 (52%). Among the subgroup of PD-L1 negative or unknown patients (n=1831), 37% (678) received CPI monotherapy (“off-label”). Factors associated with “off-label” CPI monotherapy use included ECOG PS ≥ 2 (OR 1.3), age ≥ 75 (OR 1.3), community practice (OR 1.7), and earlier year (HR 1.2) (all p<0.05). Conclusions: Most US patients with R/M HNSCC are now receiving CPI-based therapy in the frontline setting, but PD-L1 testing rates remain suboptimal. Use of CPI monotherapy in PD-L1 negative or unknown HNSCC is common, particularly in elderly patients and those with poor performance status. Treatment for frontline R/M HNSCC by PD-L1 levels, 2019-2023. PD-L1/Treatment PD-L1 CPS ≥ 20 PD-L1 CPS 1-19 PD-L1 CPS <1 PD-L1 unknown CPI monotherapy 398 (53%) 371 (47%) 116 (33%) 562 (38%) CPI + chemotherapy 181 (24%) 232 (30%) 121 (34%) 289 (20%) Chemotherapy +/- cetuximab 173 (23%) 179 (23%) 115 (33%) 628 (42%) Total 752 782 352 1479