An AI-based IHC quantification technique for assisting in the differentiation of MCL from CLL/SLL

Abstract

This study explores the application of artificial intelligence technology for the quantitative analysis of immunohistochemical markers to differentiate between mantle cell lymphoma and chronic lymphocytic leukaemia/small lymphocytic lymphoma. Utilizing an AI-based platform, the research analysed the expression of CD3, CD5, CD10, CD20, CD23, Cyclin D1, BCL-2, BCL-6 and Ki-67 in 91 samples from 84 patients. The findings demonstrate that, compared to manual interpretation, the AI system provides more objective, reproducible and accurate measurement results. Additionally, this study introduces a virtual dual immunohistochemical labelling technique for simultaneous antigen visualization. Although limited by its single-centre retrospective design, the research establishes a promising AI-assisted framework that enhances the accuracy, standardization and diagnostic efficiency in distinguishing between these two clinically distinct lymphomas.

Keywords: artificial intelligence; chronic lymphocytic leukaemia/small lymphocytic lymphoma; immunohistochemistry; mantle cell lymphoma.