Radiotherapy outcome and prognostic factors in early glottic carcinoma

Objective: To evaluate radiotherapy outcome and prognostic factors in early glottic carcinoma. Methods: The medical records of 76 patients with glottic T1N0 or T2N0 squamous cell carcinoma diagnosed at Tampere University Hospital from 1970 to 1991 and treated with primary radiotherapy were retrospectively reviewed. Except for three patients treated during the last years of the study period, radiotherapy was delivered by split-course technique with a pause of 1–3 weeks in the middle of the treatment. Primary locoregional control and disease-specific survival were analysed by the Kaplan–Meier method, and the log rank test was applied to compare the survival functions. Prognostic factors were evaluated by uni- and multivariate Cox regression analysis. Results: The 10-year locoregional control rate after radiotherapy was 83 and 48% for patients with T1 and T2 tumours, respectively (P=0.0005). The 10-year disease-specific survival was 91% for T1 and 69% for T2 disease (P=0.0018). The larynx could be preserved in 87% of T1 and 44% of T2 cases. Tumour extent expressed by the number of vocal cord thirds involved was the only factor with significant prognostic value in the multivariate analysis of primary locoregional control (hazard ratio (HR) 3.2, 95%CI 1.8–5.8, P=0.0001). Involvement of the posterior vocal cord third (HR 8.4, 95%CI 1.0–69.5, P=0.04899) and T-category (HR 3.0, 95% CI 0.9–10.2, P=0.0790) were connected with poorer prognosis in the multivariate analysis of disease-specific survival. In the multivariate analysis of T1 cases only, higher number of vocal cord thirds involved (HR 5.9, 95%CI 2.2–16.2, P=0.0005) and longer treatment duration (HR 1.1, 95%CI 1.0–1.3, P=0.0188) indicated worse locoregional control. Treatment duration (HR 1.2, 95%CI 1.0–1.3, P=0.0384) together with posterior cord involvement (HR 9.9, 95% CI 1.1–92.7, P=0.0437) signified poorer survival. Conclusion: Our findings indicate that the extent of the tumour is the most important predictor of radiotherapy outcome in early glottic carcinoma. This suggests that a classification based on the actual size of the tumour could be a better prognostic indicator than the conventional T-grouping. Although treatment duration was significant only in separate analysis of T1 cases, the split-course regimen resulting in long treatment times may be considered a major contributor to our relatively low local control rate also in T2 disease.