The role of computed tomography in the T classification of laryngeal carcinoma

BACKGROUND. The objectives of this study were 1) to describe patterns of us e of computed tomography (CT) in laryngeal carcinoma, and 2) to characteriz e the contribution of CT to the T classification of laryngeal carcinoma. METHODS. The study population comprised 1195 patients with laryngeal carcin oma diagnosed from 1982 through 1995 chosen randomly from the Ontario provi ncial cancer registry. A chart review was conducted to obtain data on each case. Patient-related, tumor-related, and health-system-related factors wer e analyzed to identify factors associated with the use of CT. Descriptions of clinical exams and CT reports were reviewed to see holy CT information m odified T classification. Actuarial local control and cause specific surviv al curves were plotted by clinical T classification without and with CT to evaluate stage migration. The percentage of the variance in outcome explain ed by T classification in a Cox analysis was used to evaluate whether the p rognostic accuracy of T classification was improved with the use of informa tion from CT. RESULTS. Patients with glottic (20.1%) and supraglottic (41.7%) carcinoma u nderwent CT. The use of CT increased over time in glottic and supraglottic carcinoma combined from 17.2% in 1982-5 to 33.9% in 1991-5. Computed tomogr aphy was used less often in older patients with a 16% (95% confidence inter val, 5-27%) decrease in the odds of having CT with each 10-year age increme nt. Computed tomography use varied considerably across the cancer center re gions in Ontario. Computed tomography altered the T classification in 20.2% of those patients who had CT, with most being "upstages." Stage migration due to CT was demonstrated. Using information from CT in the assignment of T classification for 27.8% of this study population did not make a signific ant contribution to the ability of T classification to predict outcome over the entire group. CONCLUSIONS. There is large variation in the use of CT among different age groups and regions. The ability to compare outcomes by stage across geograp hic areas is compromised when the use of CT varies. (C) 2001 American Cance r Society.