Sensitivity of Time-Resolved Fluorescence Analysis Methods for Disease Detection

Abstract

Time-resolved fluorescence (TRF) measurements of biological tissue provide chemical and structural information useful for detecting subtle disease processes, including cancers originating in mucosal tissues. A number of techniques for analyzing such TRF measurements exist, but they have not yet been compared to determine which of them can provide the greatest sensitivity to differences between a TRF measurement from nondiseased tissue and one from diseased tissue. We have evaluated four TRF analysis methods in this study: biexponential curve fitting, monoexponential curve fitting, Laguerre function representation, and computing the area under the decay curve (AUDC). We performed this study on a large dataset of computer-generated TRF decay curves based on colonic mucosa and typical measurement instrumentation. We statistically determined the minimum detectable change (MDC) in the relative contribution of each mucosal layer to the total TRF signal with each analysis method. We also determined the MDC in fluorescence lifetime of the upper mucosal layer. These two types of changes are due to the structural and biochemical changes expected in mucosa with onset of neoplasia. Under a wide range of baseline conditions, the monoexponential, AUDC, and Laguerre analysis methods all yield dramatically superior sensitivity and robustness over the standard biexponential method, with several caveats.