Major salivary gland carcinomas: A 21-center cohort study of 892 patients with external validation of the UICC 9th edition n-classification

OBJECTIVES: Major salivary gland carcinomas are rare malignancies with limited prognostic evidence. This study assessed survival outcomes, independent prognostic factors, intraparotid lymph node involvement, and stromal tumor-infiltrating lymphocyte (sTIL) density. MATERIALS AND METHODS: Multicenter retrospective cohort study of 892 patients, classified per the 2022 WHO Classification, treated surgically at 21 Italian referral centers (2010-2023). RESULTS: Median follow-up 61.4 months. The cohort included 61.1% males, median age 59.0 years; parotid gland was the primary site in 90.4%. AdCC (28.4%) and MEC (27.2%) were most frequent; 63.8% at Stage I-II. Five-year OS, DSS, and RFS were 87.9% (95% CI 85.4-90.1%), 93.1% (91.0-94.8%), and 77.7% (74.6-80.4%). Histological facial nerve infiltration was the most robust independent prognostic factor across all endpoints (OS: HR 1.78, 95% CI 1.12-2.81; DSS: HR 2.49, 1.37-4.50; RFS: HR 2.22, 1.54-3.21). Age, lymphovascular invasion, and atypical mitoses were additional independent predictors. Among 688 parotid patients with intraparotid nodal data available, 29.2% had intraparotid metastasis; 60.7% lacked cervical involvement, with LN+/pN0 patients showing worse RFS than LN-negative patients (71.2% vs 85.3%; p = 0.002). In 285 patients with available sTIL assessment (32.0%), higher sTIL density predicted improved RFS (HR 0.51 per 10% increment, 95% CI 0.35-0.73; p < 0.001). UICC 9th edition restaging reclassified 27.3% of pN0 patients as node-positive, with significant survival stratification (p < 0.001). CONCLUSIONS: Histological facial nerve infiltration was the most consistently independent prognostic factor, supporting systematic pathological reporting. Current TNM N-staging may inadequately capture intraparotid lymph node involvement, and sTILs showed a continuous prognostic effect on RFS, supporting prospective biomarker validation.