Advancing Education, Research, and Quality of Care for the Head and Neck oncology patient.
Background: We previously showed that in head and neck squamous cell carcinoma (HNSCC) treated with immune checkpoint inhibitors (ICIs), pretreatment higher lactate dehydrogenase (LDH) and absolute (abx) neutrophils as well as lower percent (%) lymphocytes correlated with worse overall survival (OS). Using these peripheral blood biomarkers (PBBMs), we developed a prognostic signature for OS risk stratification in HNSCC treated with ICIs. We then validated our signature in an independent cohort and compared the signature to the systemic inflammatory response index (SIRI), the most robust PBBM-based score in HNSCC currently.
Methods: From a cohort of recurrent/metastatic HNSCCs treated with ICIs at our institution from 08/2012-03/2021 (n=151), an OS prognostic score was developed using Cox proportional hazards regression. This score incorporated three previously-identified pretreatment PBBMs: lactate dehydrogenase (LDH), percent (%) lymphocytes, and absolute (abx) neutrophils. A grid search identified optimal cutpoints trichotomizing (low, medium, and high) risk groups. We validated the trichotomized signature by Cox regression in an independent cohort (n=54) and generated Kaplan-Meier curves. Cox regression was used to evaluate correlations between SIRI and OS. Receiver-operating characteristic (ROC) curves were created to compare the signature to SIRI.
Results: We previously developed an OS signature for HNSCC treated with ICIs: Risk score = 1.24*log10(LDH) - 1.95*log10(% lymphocytes) + 0.47*log10(abx neutrophils). Optimal risk score cutpoints for the trichotomized signature were identified, and Kaplan-Meier curves showed good separation among risk groups. The external validation dataset demonstrated no significant clinicodemographic differences from the dataset used to develop the signature, except that ICIs were more frequently first-line therapy in the validation dataset. Trichotomizing risk groups in the validation dataset using the predetermined cutpoints was borderline significant (p=0.068), with Kaplan-Meier curves showing low-risk group with best OS and some overlap between intermediate and high-risk groups.
SIRI was evaluated in our ICI treatment cohort. Higher SIRI scores correlated with worse OS (HR 2.6, [95% CI 1.0-4.2], p<0.001), with an even stronger association observed when the model was adjusted for ECOG performance status, p16 status, and smoking (HR 2.9, [95% CI 1.2-4.7], p<0.001). ROC curves showed similar performance between our prognostic model (AUC=0.641) and SIRI (AUC=0.623).
Conclusions: We previously developed a prognostic signature for OS based on previously-identified PBBMs for HNSCC treated with ICIs. Validation of the signature was promising, although limited by sample size. SIRI was shown for the first time to correlate with OS in the immunotherapy treatment setting and performs similarly to our prognostic signature. Our prognostic model may improve patient selection for ICIs and warrants validation in a larger cohort receiving first-line ICIs as well as correlation with CPS biomarker.