SEGMENTATION OF ACOUSTIC NEUROMAS WITH MAGNETIC-RESONANCE-IMAGING ANDEIGEN IMAGE FILTERING

Objective: This study aimed to determine whether magnetic resonance im agine with Eigen image filtering can segment acoustic neuromas as a pr eliminary requirement to the development and validation of a volumetri c method of measuring tumor size and growth using Eigen image filterin g. Study Design: This was an observational study. Setting: The study w as performed in an academic, comprehensive multi-specialty group pract ice. Patients: Patients were a convenience sample of adults of both se xes who had acoustic neuromas identified by magnetic resonance imaging . Tumors ranged widely in size. Intervention: Magnetic resonance imagi ng with digital image analysis using Eigen image filtering was the int ervention. Main Outcome Measure: Observation and analysis of magnetic resonance images was the main outcome measure. Results: The ability of magnetic resonance imaging with Eigen image filtering to segment acou stic neuromas of various sizes and shapes is illustrated. Conclusions: Magnetic resonance imaging with digital image analysis using Eigen im age filtering is a promising method for volumetric measurement of acou stic neuroma size and growth. The potential for acoustic neuromas to g row may be underestimated by linear methods because of their relative lack of precision and of the geometric error that occurs in rep resent ing the volumetric growth of a three-dimensional tumor as a linear dia meter.