Prediction of postoperative survival in patients with head and neck acinic cell carcinoma: a population-based study

Surgical intervention constitutes a pivotal therapeutic strategy for patients with head and neck acinic cell carcinoma (HNACC). However, precise long-term postoperative outcome prediction for individual patients remains suboptimal. This study aimed to develop and validate a nomogram model for "prognosis prediction and risk stratification" to evaluate postoperative outcomes in patients with HNACC.

Patients with HNACC diagnosed between 2004 and 2015 were extracted from the Surveillance, Epidemiology, and End Results (SEER) database and randomly allocated to training and validation cohorts in a 7:3 ratio. Cancer-specific survival (CSS) determinants were identified through Cox regression analysis. The nomogram's predictive accuracy and calibration were assessed using the concordance index (C-index), calibration plots, decision curve analysis (DCA), and Kaplan-Meier survival curves. Comparative analysis with previous models was conducted via the receiver operating characteristic (ROC) curve and the C-index.

A total of 931 surgically treated patients with HNACC were identified, of whom 651 comprised the training cohort for nomogram construction, while 280 constituted the validation cohort. Multivariate Cox proportional hazards regression analysis revealed tumor size, histological grade, American Joint Committee on Cancer (AJCC) stage, and M stage as independent prognostic factors for CSS. To enhance clinical decision-making, both clinically pertinent and statistically significant variables, including N stage and radiotherapy, were incorporated into the nomogram. The C-index for CSS prediction reached 0.898 in the training cohort and 0.840 in the validation cohort. The area under the curve (AUC) for 8- and 10-year survival in the training cohort was 0.905 [95% confidence interval (CI): 0.856-0.954] and 0.902 (95% CI: 0.856-0.948), respectively, while in the validation cohort, it was 0.824 (95% CI: 0.726-0.922) and 0.799 (95% CI: 0.685-0.899), respectively. Throughout the study, the nomogram demonstrated superior time-varying C-index and ROC curve performance compared to tumor-node-metastasis (TNM) staging and pathological grading.

The proposed nomogram model for "prognosis prediction and risk stratification" demonstrated enhanced prognostic accuracy for patients with HNACC following surgical treatment.