Survival to oral cancer. A study of clinical risk markers with independent prognostic value.

BACKGROUND: The aims of this study have been to describe survival to oral cancer and to identify clinical variables with independent influence on its prognosis before treatment. METHODS: 94 oral cancer patients treated during 1991-99 entered the study. The variables considered were: age, sex, location of the lesion, clinical presentation, symptoms, TNM classification, months elapsed since treatment and ploidy pattern. A descriptive study of the data was performed, along with a survival analysis using Kaplan-Meier curves (log rank test for comparison among curves) and single and multivariate Cox regression. RESULTS: Multivariate analysis recognized a prognostic value for the age of the patient (OR = 1.06; CI95%: 1.02-1.09) and also for tumour size. Tumour stage resulted also selected, but its predictive value was lower than size's, so it was excluded from the predictive model. No statistically significant differences in terms of survival were identified on the rest of variables considered in the study. CONCLUSIONS: This study suggests the need for considering age and tumour size as the most relevant clinical variables for predicting survival to oral cancer at the time of diagnosis.