Growth of untreated vestibular schwannoma: a prospective study

Abstract

Object Small vestibular schwannomas (VSs) are often conservatively managed and treated only upon growth. Growth is usually reported in mm/year, but describing the growth of a 3D structure by a single diameter has been questioned. As a result, VS growth dynamics should be further investigated. In addition, baseline clinical parameters that could predict growth would be helpful. In this prospective study the authors aimed to describe growth dynamics in a cohort of conservatively managed VSs. They also compared different growth models and evaluated the ability of baseline parameters to predict future growth. Methods Between 2000 and 2006, 178 consecutive patients with unilateral de novo small-sized VSs identified among the Norwegian population of 4.8 million persons were referred to a tertiary care center and were included in a study protocol of conservative management. Tumor size was defined by MR imaging–based volume estimates and was recorded along with clinical data at regular visits. Mixed-effects models were used to analyze the relationships between observations. Three growth models were compared using statistical diagnostic tests: a mm/year–based model, a cm3/year–based model, and a volume doubling time (VDT)-based model. A receiver operating characteristic curve analysis was used to determine a cutoff for the VDT-based model for distinguishing growing and nongrowing tumors. Results A mean growth rate corresponding to a VDT of 4.40 years (95% CI 3.49–5.95) was found. Other growth models in this study revealed mean growth rates of 0.66 mm/year (95% CI 0.47–0.86) and 0.19 cm3/year (95% CI 0.12–0.26). Volume doubling time was found to be the most realistic growth model. All baseline variables had p values > 0.09 for predicting growth. Conclusions Based on the actual measurements, VDT was the most correct way to describe VS growth. The authors found that a cutoff of 5.22 years provided the best value to distinguish growing from nongrowing tumors. None of the investigated baseline predictors were usable as predictors of growth.