2533 - A nomogram based prognostic score to predict overall survival (OS) in recurrent-metastatic (R/M) head and neck squamous cell carcinoma (HNSCC) patients (pts) treated with immune checkpoint inhibitors (ICI).

Background

ICIs are pivotal therapies in the treatment of pts with R/M HNSCC. The dismal OS of some pts underscores the need to detect pre-treatment prognostic scores.

Methods

Pts with R/M HNSCC receiving pembrolizumab (P) or nivolumab (N) were included in this retrospective analysis. We analyzed the impact of the following pre-treatment variables for OS using a Cox proportional hazards regression model: drug type, age, sex, tumor site, line of therapy, p-16, albumin (alb), hemoglobin (Hb), lactic dehydrogenase (LDH), lymphocyte count (LC), platelets (PLT), and neutrophil counts (NC).

Results

66 pts (55%) treated with N and 55 (45%) with P were included. Characteristics: Sex: Male 100 pts (83%), female 21 pts (17%); Mean Age: 62 (SD 11); Performance Status 0-1: 63 pts (52%), 2-3: 58 pts (48%); Line of therapy, 1: 43 pts (36%), 2: 58 pts (48%), 3-4: 20 pts (16%); P-16: positive 44 pts (36%), negative 77 pts (64%); Tumor site: oropharynx 54 pts (45%), oral cavity 23 pts (19%), larynx 16 pts (13%), other 28 pts (23%); Response: complete 11 pts (9%), partial 3 pts (2%), stable disease 47 pts (39%); and disease progression: 60 pts (50%). The OS at 1 and 2 years were 52% (95% confidence interval (CI), 43%-62%) and 27% (CI 17%-42%), respectively. Age, sex, PLT, NC, LC, p-16, alb, Hb, and LDH were included in the multivariable Cox model. Adjusted hazard ratio estimates are: P-16: 0.53 (CI 0.3-0.95, P = 0.03); LC: 0.67 (CI 0.46-0.97, P = 0.03); NC: 1.25 (CI 1.01-1.56, P = 0.04); LDH: 1.67 (CI 1.01-2.78, P = 0.04); Hb: 1.69 (CI 1.00-2.87, P = 0.05); alb: 1.69 (CI 0.94-3.02, P = 0.08); PLT: 1.06 (CI 0.82, 1.37, P = 0.7); Age: 1.02 (CI 0.78-1.34, P = 0.9); and Sex: 0.98 (CI 0.5-1.91, P = 0.9). This model was used to create a nomogram based prognostic score for 1 and 2 year OS probability and median OS time. Internal model validation using bootstrapped based bias corrected estimates showed Nagelkerke’s R2 = 0.18, calibration slope of 0.67, and C-index of 0.7.

Conclusions

Pre-treatment LC, NC as well as LDH, Hb, alb, and p-16 had the largest impact on OS based on a prognostic nomogram. The nomogram may help treatment decisions regarding ICI use in this population.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The Ohio State University Comprehensive Cancer Center – Arthur G. James Cancer Hospital and Richard J. Solove Research Institute.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.