SBRT in 3 Fractions for T1 Glottic Cancer

Abstract

The purpose of this study was to report the results of a mature phase II study. This was a prospective study at a single institution that were enrolling patients with T1 glottic cancer. The true vocal cords (TVCs) were divided into thirds and the third(s) containing disease prescribed 36 Gy in 3 fractions. The portions of the TVCs next to the involved one were planned to receive 30 Gy in 3 fxs. The remaining parts of the TVCs as well the rest of the larynx was tentatively spared. SBRT was delivered by a LINAC-based approach using multiple arcs (VMAT). Toxicity was scored by CTCAE and late events were considered those occurring 3 months after SBRT. Voice quality was investigated by the voice handicap index (VHI) at regular intervals. A generalized linear mixed model was used to analyze longitudinal functional data and the actuarial approach was used to report late toxicity prevalence rates. The planned sample size, 75 patients, was calculated to reject the hypothesis of a 2-yr local control (LC) of 85% against the alternative hypothesis of 95%. The median follow-up time is 40 months (IQR: 31.5-47.3 months). Accrual was discontinued after 33 patients due to concerns on late toxicity rates on both the mucosa and the cartilages. Median age was 66 yrs (IQR: 60-71 yrs). T stage was as follows: T1a: 23 pts (69.7%); T1b: 10 pts (30.3%). Only two patients never smoked, while of the remaining ones, 13 (41.9%) kept smoking after SBRT despite appropriate counseling. At last, follow up, all patients are alive and without evidence of disease but one patient who died 21.3 months after SBRT for intercurrent causes. The observed local control rate of 100% allows to reject the null hypothesis because, according to the binomial distribution, the probability of not observing at least one recurrence out of the accrued 33 patients is <0.5% if a 2-yr LR ≤85% would be true. All but 5 patients developed transient confluent mucositis of the treated volume with a prevalence peak at 19 days after SBRT. Six patients (18.2%) developed soft tissue necrosis (N = 4) or cartilage necrosis (N = 2) after a median time of 14.9 months from SBRT (range: 9.1-24.5 months). The actuarial prevalence of any necrosis is 3.0%, 1.3% and 0.8% at 12, 24 and 36 months after SBRT, respectively, with a peak at 6.1% between 9.9 and 11.9 months. All 4 soft tissue events healed with conservative therapy after a median time of 2.5 months (range: 1-12.2 months). One patient with chondronecrosis refused total laryngectomy and is alive with persistent toxicity 22.9 months after its onset; the other patient underwent endoscopic removal of the necrotic arytenoid and he is alive without evidence of disease or further complications more than 2 years afterwards. Patient reported outcomes (VHI) will be presented as well. SBRT to 36/30 Gy in 3 fractions is highly effective in controlling T1 true vocal cord carcinoma, but necrosis, though mostly transient, is a concern. Based on the present results, a reduction in total doses is justified and a novel phase II study is ongoing.