This paper compares results for T2-T3 carcinoma of the supraglottic larynx treated with radiotherapy (RT), supraglottic laryngectomy, or total laryngectomy. Two hundred thirty patients with T2-T3 carcinoma of the supraglottic larynx were treated at the University of Florida between 1964 and 1989. All had a minimum 2-year follow-up. Patients were treated with RT alone (160), supraglottic laryngectomy (25), or total laryngectomy (45). The probabilities of local control, local-regional control, distant metastasis, and survival were calculated by the Kaplan-Meier product-limit method and comparisons were made between the three treatment methods. American Joint Committee on Cancer (AJCC) stage, T stage, anatomic suitability for a supraglottic laryngectomy, treatment group, and presence or absence of a pretreatment tracheostomy were evaluated in multivariate analyses for the various end points. The 2-year local control rates for patients treated with RT alone, supraglottic laryngectomy, and total laryngectomy for T 2 lesions were 87%, 82%, and loo%, respectively, and rates forT3 lesions were 70%, 71%, and 81%. In the multivariate analyses, T stage was of independent prognostic significance for the end point of local control, whereas AJCC stage was a significant prognostic factor for control above the clavicles and cause-specific survival. The incidence of severe complications was as follows: RT (6%), supraglottic laryngectomy (20%), and total laryngectomy (16%). In this analysis of T2-T3 supraglottic larynx carcinomas, treatment with RT offered equivalent local control rates, local-regional control rates, and survival rates with fewer severe complications compared with supraglottic laryngectomy. Although total laryngectomy was associated with a modest improvement in local-regional control, the differences were not statistically significant, and this procedure would rarely be indicated today for T 2 lesions and uncommonly for T 3 lesions. Radiut Oncol Invest 1995;2:237-244. o 1995 Wiey-Liss, hc.