IMAGE ANALYSIS FOR AUTOMATED ASSESSMENT OF GRADE OF NEUROBLASTIC DIFFERENTIATION

Peripheral neuroblastic tumors (pNTs) make the most commonly encountered tumor groups in children. Neuroblastoma, one of the categories in pNTs, is known to have unique biological behaviors with variable clinical prognoses of the patients. Part of the neuroblastoma prognosis is closely related with grade of neuroblastic differentiation. In this work, we present an automatic classification system that includes a novel segmentation method using the EM algorithm with the Fisher-Rao criterion as its kernel. This is followed by a classification stage with classifiers applied to the actual neuroblastoma images. The good classification accuracy suggests that the developed method is promising in automating this pathological assessment