Clinical outcomes, prognosis, and predictors of salvage surgery in oral cavity squamous cell carcinoma.

e18040 Background: Salvage surgery (SS) is one of the curative options for oral cavity squamous cell carcinoma (OCSCC) patients with locoregional recurrence (LRR) or secondary primary, but the role of SS should be reevaluated between life expectancy, morbidity, and quality of life. Selecting suitable OCSCC patients receiving SS is important. Methods: From 2010 to 2018, newly diagnosed OCSCC patients who progressed to LRR or secondary primary were recorded. Clinical outcomes, prognostic factors, and predictors were analyzed for the patients receiving SS. Cox regression analyses were performed for PSS, defined from the date of SS to the date of death or last follow-up. Survival was estimated using the Kaplan–Meier method and log-rank tests. Results: A total of 263 newly diagnosed OCSCC patients progressing to LRR or secondary primary were recorded. Half (55.1%, 145/263) of them received SS, and one-third (29.7%, 43/145) of the SS group received twice and more times SS. Median survivals after disease progression were 65.6 and 10.6 months for patients with or without SS, respectively (P < 0.001). A total of 214 SS events were enrolled for analysis. Nearly twenty percentage (20.1%, 39/194) of SS events would progress to death within 1 year after surgery (PSS < 1 year). PSSs of the first, secondary, third, and fourth or more times SS were 64.2, 47.6, 40.9, and 18.9 months, respectively (P = 0.217). Surgical features of the last surgery (perineural invasion and depth of invasion), the interval between the last and current surgery, and clinical N staging of the current surgery were the four independent factors for PSS. To predict the patients with PSS < 1 year, a scoring system was established that each of the independent factors was scored one point. The area under the curve of the scoring system was 0.755, and sensitivity, specificity, positive predictive value, and negative predictive value were 66.7%, 76.0%, 39.0%, and 90.8%, respectively (Table). Conclusions: A scoring system with a high negative predictive value was established to predict PSS < 1 year.[Table: see text]