Computer aided discrimination between primary and secondary brain tumors on MRI: From 2D to 3D texture analysis

Abstract

Three dimensional texture analysis of volumetric brain MR images have been identified as an important indicator for discriminating among different brain pathologies. The aim of the present study was to evaluate the efficiency of three dimensional textural features using a pattern recognition system in the task of discriminating primary from metastatic brain tissues on T1 post-contrast MRI series. The dataset consisted of sixty seven brain MRI series obtained from patients with verified and untreated intracranial tumors. The pattern recognition system was designed employing a probabilistic neural network classifier, specially modified in order to integrate the non-linear least squares feature transformation logic in its discriminant function. The latter, in conjunction with using three dimensional textural features, enabled boosting up the performance of the system in discriminating primary from metastatic with accuracy of 95.52%. The proposed system might be used as an assisting tool for brain tumor characterization on volumetric MRI series.