Assessing variability of in vivo tissue raman spectra

Raman spectroscopy (RS) has received increasing attention as a potential tool for clinical diagnostics. However, the unknown comparability of multiple tissue RS systems remains a major issue for technique standardization and future multisystem trials. In this study, we evaluated potential factors affecting data collection and interpretation, utilizing the skin as an example tissue. The effects of contact pressure and probe angle were characterized as potential user-induced variability sources. Similarly, instrumentation-induced variability sources of system stability and system-dependent response were also analyzed on skin and a nonvolatile biological tissue analog. Physiologically induced variations were studied on multiple tissue locations and patients. The effect of variability sources on spectral line shape and dispersion was analyzed with analysis-ofvariance methods, and a new metric for comparing spectral dispersion was defined. In this study, in vivo measurements were made on multiple sites of skin from five healthy volunteers, with four stand-alone fiber optic probe-based tissue RS systems. System stability and controlled userinduced variables had no effects on obtained spectra. By contrast, instrumentation and anatomical location of measurement were significant sources of variability. These findings establish the comparability of tissue Raman spectra obtained by unique systems. Furthermore, we suggest steps for further procedural and instrumentation standardization prior to broad clinical applications of the technique.