Histological Classification of Thymoma

Evaluation by Computer-Assisted Morphometry

Abstract

Background: Many thymoma classifications have been followed and have been updated by newer or alternative schemes. Many classifications were based on the morphology and histogenesis of normal thymus as the backbone, while other classifications have followed a more simplified scheme, whereby thymomas were grouped based on biological behavior. The WHO classification is currently the advocated one, which is based on “organotypical” features (i.e. histological characteristics mimicking those observed in the normal thymus) including cytoarchitecture (encapsulation and a “lobular architecture”) and the cellular composition, mostly the nuclear morphology is generally appreciated.

Objectives: This study aims to re-classify thymomas by establishing certain morphometric parameters to evaluate the epithelial cells nuclei. An appraisal of thymoma classification as cortical/ lymphocytic/ type B1 and B2 or medullary/ spindle cells/ type A will be attempted as objective re-evaluation of thymoma.

Patients: This study is a retrospective evaluation of 50 cases of thymoma, 20 cases of thymic hyperplasia and 10 cases of normal thymus (control group). Using a 5 µm formalin-fixed paraffin embeded tissue sections, stained with hematoxylin-eosin stain, these cases were previously classified histologically into lymphocytic (6 cases), lymphoepithelial (mixed) (28 cases) and epithelial (16 cases), including 2 cases of spindle cell thymoma.

Methods: Computer-assisted morphometry was performed for 80 cases. This involves digitization of the histopathological features and application of morphometric analysis through software. The morphometric parameters used are nuclear area, maximum nuclear diameter and Form factor of epithelial cells nuclei. Ten normal thymus glands from the archived cases were also evaluated as a control groups.

Results: The results showed that epithelial thymomas possess significantly different nuclear areas from that of a normal thymus, the maximum nuclear diameter (D-Max) follows the same pattern and adds no further outcomes. The numerical morphometric analysis showed no significant differences between lymphocytic predominant thymoma and those classified as cortical thymoma (Type B2). Thus it does not support such a re-classification. Form factor is an indication of pleomorphism, but it should be cautiously used when spindle cells are present since it may give a false indication of pleomorphism.

Conclusion: Computer-Assisted morphometric analysis provides an objective, reproducible and comparable results for thymoma histological classification.

Keywords: Thymoma, Histological classification, Grading, Morphometry, Computer-assistedmorphometry.