Analysis of recurrences in 322 Tis, T1, or T2 glottic carcinomas treated by carbon dioxide laser.

An endoscopic approach to early glottic carcinoma is considered a sound treatment for both previously untreated lesions and selected recurrent lesions. Between January 1988 and December 2000, we treated 322 patients by CO2 laser at a single institution; 37 had Tis, 191 T1a, 55 T1b, and 39 T2 lesions (mean follow-up, 77 months; range, 6 to 180 months). Kaplan-Meier curves showed a 5-year overall survival rate of 88%, a determinate survival rate of 99%, a disease-free survival rate of 81%, a rate of ultimate local control with laser alone of 91%, and a laryngeal preservation rate of 97%. Univariate and multivariate analysis showed that the only factor that statistically affected endoscopic control was lateral extension of the tumor with involvement of the bottom of the ventricle (hazard risk ratio, 4.0; 95% confidence interval, 1.71 to 9.35). The 58 recurrences were classified according to their location compared with the site of the primary tumor as follows: 14 in the same area (group A), 27 in adjacent subsites with superficial spreading or multifocal distribution (group B), and 17 in adjacent sites by submucosal diffusion to the visceral spaces, cartilaginous framework, or extralaryngeal tissues (group C). For each group, we analyzed the rate of patients who underwent salvage by endoscopic or open neck procedures and the rate of laryngeal preservation. Recurrences in groups A and B were endoscopically treated in 86% and 74% of cases, respectively. By contrast, in group C no patient was endoscopically cured, and there was a low laryngeal preservation rate (47%). The pathways of spread in recurrent carcinoma are therefore the single most important factor in predicting its endoscopic curability.