A novel surgical T staging system for recurrent nasopharyngeal carcinoma: advancing prognostic accuracy and clinical applicability.

Abstract

The aim of this study was to develop and validate a novel surgical T (sT) staging system for recurrent nasopharyngeal carcinoma (rNPC) that better reflects anatomical barriers to resection and improves prognostic accuracy compared with the American Joint Committee on Cancer (AJCC) staging system. Cadaveric dissections and a retrospective analysis of 211 patients with rNPC undergoing endoscopic nasopharyngectomy were conducted. The proposed sT staging system incorporates 3 key anatomical barriers-the pharyngobasilar fascia (PBF), interpterygoid fascia (IPF), and dura mater-to stratify rNPC into 4 stages (sT1-sT4). In a retrospective cohort of 211 patients who underwent endoscopic nasopharyngectomy, the 5-year overall survival (OS) rates based on AJCC rT staging were 92.9% (rT1), 72.9% (rT2), 69.3% (rT3), and 37.1% (rT4). In contrast, the 5-year OS rates under the sT staging system were 93.4% (sT1), 80.3% (sT2), 60.0% (sT3), and 34.0% (sT4), demonstrating improved separation of survival outcomes across stages (p < 0.001). In the held-out validation cohort under a hash-locked 80/20 split, the sT model achieved superior discrimination compared with AJCC rT (C-index, 0.624 vs 0.546; 2-year time area under the curve [AUC] 0.709 [simple/IPCW 0.709/0.740] vs 0.632 [0.632/0.667]) with acceptable calibration at 2-4 years. The 5-year AUC was not estimable because no patients remained event free beyond 5 years. By encoding barrier-guided spread aligned with endoscopic resectability (PBF, IPF, and dura), the sT system improves external discrimination and stage separation, thereby enhancing preoperative risk stratification and surgical planning. External, multi-institutional validation with longer follow-up is warranted.