LCK and HOMER1 gene expression-based classifier distinguishes early mycosis fungoides from eczema and psoriasis

Abstract Mycosis fungoides (MF) is a rare cutaneous T cell lymphoma that, in its early stages, can closely mimic eczema and psoriasis in both clinical appearance and histopathologic features, leading to frequent misdiagnosis, inappropriate treatment, and delayed care. Reliable adjunctive biomarkers are lacking, underscoring the need for improved diagnostic strategies. We developed a biomarker discovery framework based on bulk RNA sequencing of skin biopsies (19 MF, 112 psoriasis, 105 eczema), which identified seven candidate diagnostic genes. RT-PCR analysis in FFPE tissue specimens from 65 MF and 101 eczema/psoriasis samples verified LCK and HOMER1 as robust discriminator genes. A logistic regression model based on LCK and HOMER1 gene expression differentiated MF from psoriasis and eczema with 91% sensitivity and 94% specificity. Independent validation on 7 additional international cohorts (MF n=58, eczema/psoriasis n=55) confirmed robust performance. Spatial and single-cell transcriptomic analyses revealed biological underpinnings of classifier accuracy: LCK was enriched in malignant and specific T cell subsets in MF, whereas HOMER1 was confined to keratinocytes in eczema and psoriasis but nearly absent in MF. Case studies demonstrated that the classifier identified MF in routine biopsies earlier than histopathology. This molecular diagnostic approach enables earlier and more reliable distinction of MF from common inflammatory dermatoses, offering a clinically applicable tool to reduce diagnostic uncertainty, accelerate appropriate treatment, and might improve patient outcomes. One Sentence Summary A two-gene classifier reliably distinguishes mycosis fungoides from eczema and psoriasis, enabling early and accurate diagnosis.