Stromal tumor-infiltrating lymphocytes (sTIL) in laryngeal squamous cell carcinoma: association with pharyngocutaneous fistula, recurrence, and mortality.

Abstract

Purpose: Previous studies suggest a favorable prognostic role for tumor-infiltrating lymphocytes in laryngeal squamous cell carcinoma, although findings vary by immune-cell phenotype, assessment method, tumor compartment, and cohort characteristics. This study evaluated associations between sTIL percentage and pharyngocutaneous fistula, recurrence, and mortality after surgery.

Methods: This single-center retrospective cohort included 81 patients who underwent primary or salvage total laryngectomy between January 2019 and March 2024. sTIL was assessed on H&E sections according to International TILs Working Group guidelines. Because sTIL was non-normally distributed (Shapiro-Wilk W = 0.908; p < 0.001), non-parametric tests were used. Associations were assessed using Mann-Whitney U, Kruskal-Wallis, chi-square/Fisher's exact tests, and Spearman correlation. Survival was analyzed using Kaplan-Meier and Cox regression.

Results: Median sTIL was 30% (IQR: 10-50%). sTIL was not significantly associated with fistula, recurrence, mortality, surgical approach, lymphovascular or perineural invasion, T stage, or differentiation grade (all p > 0.05). Mortality was associated with older age (68 vs. 60 years; p = 0.004). Deceased patients had longer observed follow-up, reflecting time-to-event structure rather than a prognostic association. Kaplan-Meier analysis showed no survival difference by sTIL group (log-rank p = 0.914). In the reduced multivariable Cox model, age (HR: 1.05; p = 0.043) and lymphovascular invasion (HR: 4.79; p < 0.001) were independently associated with mortality, whereas sTIL was not (HR: 1.01; p = 0.600).

Conclusions: H&E-based sTIL scoring was not significantly associated with pharyngocutaneous fistula, recurrence, or overall survival. Given the limited number of fistula events, the fistula-related findings should be considered exploratory and hypothesis-generating. Age and lymphovascular invasion were independently associated with mortality. These non-confirmatory findings require validation in larger studies using spatial and phenotypic immune profiling.