The World Health Organization classification of pituitary tumors provides a framework for pathologists and researchers to classify pituitary adenomas. From the perspective of a practicing pathologist, this classification can be improved by pooling immunohistochemical data in a more standardized way, and by deliberately distinguishing features that assist in classification from those that do not. This article illustrates one general workflow to examine classification features consisting of immunohistochemical stains for anterior pituitary tumors, in order to promote debate and advance an evidence-based framework for classification.
Keywords: Classification; Machine learning; Pituitary; Statistical learning.