Liat Ben Sira1, Lior Weizman2,
Leo Joskowicz2, Ronit Precel1, Shlomi Constantini3,4,
Dafna Ben Bashat5
1Department of Radiology , Tel Aviv
Sourasky Medical Center, Tel-Aviv, Israel; 2School of Eng and
Computer Science, The Hebrew Univeristy of Jerusalem, Jerusalem, Israel; 3The
Paediatric Neurosurgery Department, Tel Aviv Sourasky Medical Center,
Tel-Aviv, Israel; 4Sackler Faculty of Medicine, Tel Aviv
University, Tel-Aviv, Israel; 5The Wohl Institute for Advanced
Imaging, Brain Imaging Center, Tel Aviv Sourasky Medical Center , Tel-Aviv,
Israel
Accurate
and consistent volumetric measurements of optic pathway gliomas (OPG), the
most common tumor in the brain in patients with Neurofibromatosis, are
clinically crucial. In this study we present an automatic method for
segmentation of OPGs from multi-spectral MRI datasets. The method effectively
incorporates prior location of the OPG, its shape and intensity and
accurately identifies the boundaries in a consistent and repeatable manner.
The method was tested on 15 data sets, the optimal threshold was derived from
a receiver operating characteristic curve, and a significant correlation was obtained
between the volume calculated using this method compared to manual
measurements.