Radical radiotherapy for early glottic cancer: Results in a series of 1087 patients from two Italian radiation oncology centers. I. The case of T1N0 disease.

PURPOSE: To retrospectively evaluate local control rates, late damage incidence, functional results, and second tumor occurrence according to the different patient, tumor, and treatment features in a large bi-institutional series of T1 glottic cancer. METHODS AND MATERIALS: A total of 831 T1 glottic cancer cases treated consecutively with radical intent at the Florence University Radiation Oncology Department (FLO) and at the Radiation Oncology Department of the University of Brescia-Istituto del Radio "O. Alberti" (BS) were studied. Actuarial cumulative local control probability (LC), disease-specific (DSS), and overall survival (OS) rates have been calculated and compared in the different clinical and therapeutic subgroups with both univariate and multivariate analysis. Types of relapse and their surgical salvage have been evaluated, along with the functional results of treatment. Late damage incidence and second tumor cumulative probability (STP) have been also calculated. RESULTS: In the entire series, 3-, 5-, and 10-year OS was equal to 86%, 77%, and 57%, respectively. Corresponding values for LC were 86%, 84%, and 83% and for DSS 96%, 95%, and 93%, taking into account surgical salvage of relapsed cases. Eighty-seven percent of the patients were cured with function preserved. Main determinants of a worse LC at univariate analysis were: male gender, earlier treatment period, larger tumor extent, anterior commissure involvement, and the use of Cobalt 60. At multivariate analysis, only gender, tumor extent, anterior commissure involvement, and beam type retained statistical significance. Higher total doses and larger field sizes are significantly related (logistic regression) with a higher late damage incidence. Scatterplot analysis of various combinations of field dimensions and total dose showed that field dimensions >35 and <49 cm2, together with doses of >65 Gy, offer the best local control results together with an acceptably low late damage incidence. Twenty-year STP was equal to 23%, with second tumor deaths being more frequent than larynx cancer deaths (67 of 831 vs. 46/831). CONCLUSION: The results of this study support the opinion, suggested by some international guidelines, that radiotherapy is standard treatment for T1 glottic cancer. Better results are obtained in patients with less extended disease and with 4-6 MV photon beams. The use of doses in excess of 65 Gy and of field sizes of 36-49 cm2 is probably the best technical choice available. Late damage is infrequent, but careful follow-up is warranted to detect early not only relapses (because conservative salvage surgery is feasible), but also second malignant tumors, which constitute the main cause of death in these patients and are potentially curable.