Clinical Validation of Imaging Biomarkers in Mycosis Fungoides

Abstract

The composite index lesion severity (CAILS) score is used to monitor disease and therapeutic response in mycosis fungoides (MF), but is limited by interobserver variability and low sensitivity. Emerging imaging techniques, such as multispectral imaging (MSI), colourimetry and laser speckle contrast imaging (LSCI), offer objective alternatives for quantifying CAILS parameters. The aim of this study was to evaluate non-invasive imaging modalities for objective and reliable quantification of disease extent in MF. Sixty-six participants were enrolled in two prospective studies: a cross-sectional discovery cohort to assess baseline characteristics of 35 MF patients (IA–IVB) and 10 healthy controls using CAILS and MSI, and a longitudinal confirmation cohort including 21 early-stage MF patients (IA–IIA) treated with chlormethine gel 0.016% for 16 weeks, in whom lesional and non-lesional skin were assessed using CAILS, MSI, colourimetry and LSCI at multiple time points. Candidate biomarkers were required to meet five clinical validation criteria: disease discrimination, repeatability, treatment responsiveness, correlation with CAILS and patient acceptability. In the discovery cohort, MSI detected significant differences in erythema, pigmentation, elevation and desquamation between healthy, non-lesional and lesional skin. In the confirmation cohort, four candidate biomarkers met all validation criteria: MSI CIELAB a*, MSI average haemoglobin, and colourimetry CIELAB a* (DSMIII) for quantifying erythema, and MSI individual typology angle (ITA) for pigmentation. These biomarkers reliably discriminated lesional from non-lesional skin (p ≤ 0.001), showed strong test–retest reliability (CV < 10%, ICC > 0.84), detected treatment effects, showed moderate concordance with CAILS, and were associated with low patient burden (mean 3.4/100). These findings show that MSI- and colourimetry-derived biomarkers can objectively monitor disease extent in MF and complement existing clinical assessments.