BACKGROUND: We developed the first-of-its-kind handheld confocal Raman spectroscopy (CRS) system to quantify the concentration of natural moisturizing factors in the skin. OBJECTIVE: To evaluate the feasibility of our handheld CRS system and propose a novel quantitative index to measure skin barrier function. METHODS: This prospective study included 30 atopic dermatitis (AD) patients and 14 healthy volunteers. All AD participants were assessed using the Scoring Atopic Dermatitis (SCORAD) severity instrument, a vapometer for trans-epidermal water loss and a moisture meter for skin surface moisture. A handheld CRS operating at 785 nm laser was used to measure the biochemical constituents of the skin up to a depth of ∼100 μm. We trained a linear kernel-based support vector machine (SVM) model for eczema classification based on the water, ceramide and urocanic acid content. A novel Eczema Biochemical Index (EBI) was then formulated using the skin constituents measured from the AD participants to stage disease severity. RESULTS: The SVM model used to classify healthy participants and AD patients obtained high cross-validated area under the curve of 0.857 and accuracy of 0.841, with high sensitivity and specificity values of 0.857 and 0.833 respectively. EBI can be used to stratify AD patients of varying severity, based on the biochemical constituents in the skin. CONCLUSION: As compared to the standard CRS system, the handheld CRS offers higher portability and provides Raman measurements at various body regions with similar sensitivity. This suggests that a handheld CRS device could be a valuable point-of-care resource in both research and clinical use.
BACKGROUND: We developed the first-of-its-kind handheld confocal Raman spectroscopy (CRS) system to quantify the concentration of natural moisturizing factors in the skin. OBJECTIVE: To evaluate the feasibility of our handheld CRS system and propose a novel quantitative index to measure skin barrier function. METHODS: This prospective study included 30 atopic dermatitis (AD) patients and 14 healthy volunteers. All ADparticipants were assessed using the Scoring Atopic Dermatitis (SCORAD) severity instrument, a vapometer for trans-epidermal water loss and a moisture meter for skin surface moisture. A handheld CRS operating at 785 nm laser was used to measure the biochemical constituents of the skin up to a depth of ∼100 μm. We trained a linear kernel-based support vector machine (SVM) model for eczema classification based on the water, ceramide and urocanic acid content. A novel Eczema Biochemical Index (EBI) was then formulated using the skin constituents measured from the ADparticipants to stage disease severity. RESULTS: The SVM model used to classify healthy participants and ADpatients obtained high cross-validated area under the curve of 0.857 and accuracy of 0.841, with high sensitivity and specificity values of 0.857 and 0.833 respectively. EBI can be used to stratify ADpatients of varying severity, based on the biochemical constituents in the skin. CONCLUSION: As compared to the standard CRS system, the handheld CRS offers higher portability and provides Raman measurements at various body regions with similar sensitivity. This suggests that a handheld CRS device could be a valuable point-of-care resource in both research and clinical use.