Worst pattern of invasion (WPOI) has been evaluated in many single-institute cohorts. Our goal was to perform a large multicentre evaluation of WPOI as a prognostic marker in oral squamous cell carcinoma (OSCC). Retrospective pathology data was collated from 14 institutions and compared with clinical outcome in 1374 OSCC patients with upfront curative resection. Most cases were of oral tongue (n = 645, 47%); T2 (33%) and N0 (59%). WPOI 1-3 frequency was 29.4%, WPOI 4 47% and WPOI 5 22%. On univariable analysis, the 3-year disease free survival (DFS) was 54.2% for WPOI 5 vs. 69.7% for WPOI 1-4 (p < 0.001). The locoregional control (LRC) was 68.9% vs 79.2% (p = 0.001), and overall survival (OS) 68.4% vs 83.8% (p < 0.001). On multivariable Cox-regression in the entire cohort, WPOI 4 or 5 was strongly correlated with other known poor prognostic factors and not an independent predictor of OS (HR 1.10, 95% CI 0.92-1.52), LRC or DFS. However, in early-stage (pT1-2 N0) patients treated with surgery alone without adjuvant radiotherapy, WPOI 5 was a robust independent predictor of DFS (HR 4.36, 95% CI 1.54-12.32, p = 0.006), OS (HR 3.69, 95% CI 1.23-11.1, p = 0.020) and LRC (HR 3.52, 95% CI 2.13-5.82, p <0.001) after applying inverse probability weighting to correct for selection bias. Furthermore, in the entire cohort of early-stage patients, interaction modeling showed that adjuvant radiotherapy significantly reduces the risk for both DFS and LRC for those with WPOI-5 (Interaction p = 0.002). Therefore, it may act as a predictive biomarker for the benefit of adjuvant radiotherapy. The prognostic and predictive role of WPOI-5 should be validated in prospective trials.
Keywords: Adjuvant radiation; Oral cavity squamous carcinoma; WPOI 4; WPOI 5; Worst pattern on invasion.
Copyright © 2026 Elsevier Ltd. All rights reserved.