To assess the impact of the radiation therapy (RT) regimen and image guidance (image guided radiation therapy [IGRT]) protocol on local control (LC) for T2N0 glottic cancer treated with partial laryngeal intensity modulated radiation therapy (IMRT).

All patients with T2N0 glottic cancer treated with IMRT from 2006 to 2013 at a single institution were retrospectively reviewed. The gross tumor volume (GTV), delineated from endoscopic and/or radiologic findings, was expanded 0.5 cm for the high-dose clinical target volume and an additional 0.5 cm for the lower-dose clinical target volume (total of 1.0 cm from GTV). The planning target volume margin was 0.5 cm radially and 1 cm superiorly and inferiorly. RT regimens evolved from hypofractionated IMRT (RT-hypo, 60 Gy in 25 fractions over a period of 5 weeks) to moderately accelerated IMRT (RT-acc, 66-70 Gy in 33-35 fractions over a period of 5.5-6 weeks) since 2010. The IGRT matching surrogate changed from cervical vertebral bone (IGRT-bone) to laryngeal soft tissue (IGRT-larynx) in 2008. LC was compared between 3 sequential cohorts: RT-hypo/IGRT-bone, RT-hypo/IGRT-larynx, and RT-acc/IGRT-larynx. Multivariable analysis assessed the relative impact of RT regimen and IGRT technique on local failure separately.

Among 139 eligible patients (median follow-up period, 5.03 years [range, 0.8-10.5 years]), we identified 28 local, 6 regional, and 2 distant failures. A higher 3-year LC rate was observed for RT-acc/IGRT-larynx (89% [95% CI: 78%-95%]) versus RT-hypo/IGRT-larynx (80% [95% CI: 54%-91%]) and RT-hypo/IGRT-bone (70% [95% CI: 53%-80%]) (P = .02). Multivariable analysis adjusted for GTV (in cubic centimeters) and smoking status confirmed that IGRT-larynx versus IGRT-bone (hazard ratio, 0.40; P = .019) and RT-acc versus RT-hypo (hazard ratio, 0.34; P = .012) both reduced the risk of local failure.

This single-institution cohort study shows a high LC rate (89%) for T2N0 glottic cancer following moderately accelerated partial laryngeal IMRT with daily laryngeal soft tissue matching IGRT. These results appear to represent an improvement attributable to changes in both IGRT matching and dose delivered, but their independent significance is unknown and further confirmation in a larger cohort is warranted.