Histopathological Meningioma classification based on Nuclear Phenotypes

What is it about?

In this paper, a nuclear phenotypes based evolutionary framework is proposed for the histopathological classification of benign meningioma into four subtypes. This framework investigates the imperative role of RGB color channels for discrimination of tumor subtypes on account of nuclear phenotypes. The nuclear regions are scanned rather than the image as a whole in order to capture the nuclear variations among the histological patterns belonging to different subtypes. A large set of features including structural, statistical and spectral features is extracted from RGB color channels of segmented nuclei to characterize nuclear morphology and texture. A classifier-based evolutionary framework is developed to tune classifier parameters and to select the best possible combination of extracted phenotypes that improved the classification accuracy (94.88%) on meningioma histology dataset. These statistics show that computational framework can robustly discriminate four subtypes of benign meningioma and may aid pathologists in the diagnosis and classification of these lesions.

Featured Image

Why is it important?

In contrast to most of the histopathological classification studies, we take a more unbiased automatic data-driven approach, where we develop a nuclear phenotypes based classification framework to extract a relatively high dimensional nuclei feature set from both spatial and spectral domains in three different color channels. Our framework established that combining nuclear phenotypes from different image representations (spatial and spectral) and exploiting evolutionary algorithms for optimal selection of nuclear features and classifier's parameters can achieve high accuracy for meningioma subtypes classification. Building on this study, we plan to extend the computational framework for the classification and grading of additional challenging brain lesions including pre-invasive and invasive cancer, which represents a major challenge in diagnostic pathology.