Noninvasive label-free monitoring of skin AGE autofluorescence using highly robust biomedical spectroscopy

Noninvasive spectral sensing with molecular specificity is an active area for clinical applications, however, the lack of monitoring stability hinders these spectral sensing methods from becoming medical gold standards. Here, we introduce a highly robust biomedical spectral monitoring system, termed the dual-mode pressure and autofluorescence sensor (DuPAFs), which enables synchronous real-time monitoring of contact pressure and autofluorescence spectra. By increasing the detection efficiency, DuPAFs achieves rapid autofluorescence spectral acquisition in as little as 0.1 s per spectrum. We developed an Adaptive Spectral Correction Algorithm (ASCA) that reduces perturbations from in vivo tissue parameters and autofluorescence crosstalk. Validated against human urine advanced glycation end-products (AGEs) concentrations, the application of DuPAFs and ASCA in measurements on palm skin greatly improved the correlation coefficient between skin autofluorescence and urine AGEs concentration from 0.64 to 0.8979. Notably, the correction algorithm is based on the physical principle of differing quantum efficiencies among various fluorophores (endogenous fluorophores such as collagen, NADH, flavins, and AGEs), giving it universal application potential. Our work characterizes the ten-day trajectory of AGEs and their diurnal rhythms, and further documents the changes in skin AGEs autofluorescence under a sustained high-sugar, high-fat dietary intervention.