Abstract Objective: The objective of this study is to evaluate the efficacy of
laser-induced fluorescence (LIF) data obtained at 325-nm
pulsed laser excitation for the discrimination of normal,
benign, and
malignant ovarian tissues. Background Data: Several studies have reported that the
autofluorescence technique has a high specificity and sensitivity for discrimination between diseased and non-diseased tissues of various cancers, and also has the advantages of being non-invasive and producing a real-time diagnosis. When using this technique on
ovarian tissues in most of the previously reported studies,
multivariate statistical tools were used and classification analyses were carried out. Materials and Methods:
Autofluorescence spectra of normal,
benign, and
malignant ovarian tissues were recorded with 325-nm
pulsed laser excitation in the spectral region from 350-600 nm in vitro. The
spectral analysis for discrimination between the different types of tissues was carried out using
principal component analysis (PCA)-based
non-parametric k-nearest neighbor (k-NN) analysis. Results: A total of 97 (34 normal, 33
benign, and 30
malignant) spectra were obtained from 22 subjects with normal,
benign, and
malignant tissues. The discrimination analysis of data using a PCA-based k-NN algorithm showed very good discrimination. The performance of the analysis was evaluated by calculating
statistical parameters, specificity, sensitivity, and
accuracy and were found to be 100%, 90.90%, and 94.2%, respectively. Conclusion: The results show that the discrimination of normal,
benign, and
malignant ovarian conditions can be achieved quite successfully using LIF.
