From Morphology to Gene Expression Profiling in Mycosis Fungoides: Is It Still a Diagnostic Challenge?

TL;DR

This paper reviews the challenges in diagnosing mycosis fungoides and explores how gene expression profiling and artificial intelligence might improve diagnosis and understanding of the disease.

Contribution

The paper highlights the potential of gene expression profiling and AI in advancing the diagnosis and understanding of mycosis fungoides.

Findings

Mycosis fungoides diagnosis remains challenging despite decades of research.

Genetic alterations are mainly observed in late stages, with unclear roles in disease initiation.

Artificial intelligence may help identify biomarkers for disease progression.

Abstract

Background: We herein review the most important clinico-pathological features of mycosis fungoides (MF). These evolving clinico-pathological aspects are paired with innovative therapeutic schemes. Moreover, we indicate cutaneous lymphomas as a new frontier of artificial intelligence application. Methods: We encompass new diagnostic and prognostic data derived from the recent medical literature describing the possible histological features which could be the targets of deep learning in conjunction with available clinical data. Results: In spite of decades of research, MF diagnosis still represents the most challenging debate from a dermatopathologist’s point of view. Genetic alterations have been identified mainly in late stages of the disease, and their importance for disease initiation is still unclear. The exploration of the genome-wide expression of individual genes in skin samples may be useful in elucidating MF pathogenesis and improving early diagnosis, while artificial intelligence could offer the possibility of searching for biomarkers of disease progression. Conclusions: MF still deserves the name of the ‘great imitator’, both clinically and histopathologically. The goal of summing up all the clinico-pathological information before reaching a final diagnosis is the approach needed to reach diagnostic accuracy, especially in early MF cases. It is advisable to think of the most common clinical presentations, to be aware of the most common histopathological features, and to interpret the results of ancillary studies only in the right clinico-pathological context.

Figures