A novel in vivo fluorescence spectroscopic diagnostic method has been developed in an animal model to make a quantified precancer diagnosis. In the study, 40 golden hamsters were randomly divided into four groups (groups A, B, C, and D), with group A being the control group and the other three groups being inducted at different precancer stages. A 1% Rose Bengal (RB) solution was used for the fluorescence spectroscopic diagnosis. A parameter K defined as K = IRB/Iauto was introduced to reflect the amount of RB in the tissue, where IRB and Iauto represent the fluorescence peak intensity of the RB in the tissue and the autofluorescence intensity of tissue at 580 nm, respectively. The average K values of the four groups were calculated and statistically analyzed by analysis of variance (ANOVA), which revealed statistically significant differences within each group as well as between groups (p < 0.001). After analysis by Clementine 11.1 C&R Tree modeling (CART), the following diagnostic criteria were set: normal, K ≤8.91; simple hyperplasia, 8.91 < K ≤41.92; mild dysplasia, 41.92 < K ≤ 70.79; moderate and severe dysplasia, K >70.79. The sensitivity and specificity to detect precancerous lesions compared with scalpel biopsy were calculated. The results of this study showed that the spectrofluorometric method mediated by RB could accurately discriminate different precancer stages.
All Science Journal Classification (ASJC) codes