LOCAL RECURRENCES AND SECOND PRIMARY TUMORS FROM SQUAMOUS CELL CARCINOMA OF THE ORAL CAVITY: A RETROSPECTIVE ANALYTIC STUDY OF 500 PATIENTS

Background. The purpose of this study was to evaluate the incidence of local recurrences (LRs) and second primary tumors (SPTs) from squamous cell carcinoma (SCC) of the oral cavity primarily treated with surgery and to further study their relationship with several primary tumor clinical and pathological features. Methods. Five hundred of 522 patients with SCC of the oral cavity primarily treated with surgery were retrospectively analyzed for the appearance of LRs and SPTs within the oral cavity. All patients with SPTs fulfilled the Warren & Gates criteria. Several clinical features were analyzed. Histological study included TNM classification, tumor size, tumor thickness, surgical margins, perineural infiltration, peritumoral inflammation, and bone involvement. In the univariant analysis, the possible association between different clinical and pathological features and the presence of LRs or SPTs was analyzed by means of the chi-square test for categorical data and the Student's t test for parametric data. The appearance of LRs and SPTs was also studied by binary logistic regression as time-dependant phenomena, in the univariant analysis. Logistic regression was also used for the multivariant analysis between the selected variables. The Kaplan-Meier method was used to estimate the probability of SPT- or LR-free survival. Results. The mean duration of the follow-up period was 52.27 +/- 49.52 months. At the end of this time, 53.82% of the patients were alive without evidence of disease, whereas 31.48% had specifically died of disease. Twenty-eight (5.6%) patients developed an SPT within the oral cavity, whereas 95 (19%) patients developed an LR during the whole follow-up period. The 5-year disease-specific survival rate for the whole series was 67.2%, in contrast to 34.9% in the group of patients with SPT and/or LR. In relation to the univariant analysis, T classification, TNM staging, pT classification, surgical margins, bone involvement, and postoperative radiotherapy (RT) were found to be predictive for LR. In relation to the multivariant analysis, only postoperative RT and bone involvement were predictive for the development of LR. Conclusion. The identification of preoperative and postoperative clinical and pathological features that prelude a higher risk for the appearance of LRs and/or SPTs may be of potential interest in determining which patients should benefit of a closer regular follow-up. When considering together the whole clinical and pathological features, only postoperative RT and bone involvement were predictive for the development of LRs. Because of the poor survival rate of the affected patients, we strongly recommend aggressive surgical treatment following the appearance of an SPT or LR. (C) 2009 Wiley Periodicals, Inc. Head Neck 31: 1168-1180, 2009