ENSEMBLES OF RADIAL BASIS FUNCTION NETWORKS FOR SPECTROSCOPIC DETECTION OF CERVICAL PRECANCER

The mortality related to cervical cancer can be substantially reduced through early detection and treatment. However, current detection tech niques, such as Pap smear and colposcopy, fail to achieve a concurrent ly high sensitivity and specificity. lit vivo fluorescence spectroscop y is a technique which quickly, noninvasively and quantitatively probe s the biochemical and morphological changes that occur in precancerous tissue. A multivariate statistical algorithm was used to extract clin ically useful information from tissue spectra acquired from 361 cervic al sites from 95 patients at 337-, 380-, and 460-nm excitation wavelen gths, The multivariate statistical analysis was also employed to reduc e the number of fluorescence excitation-emission wave-length pairs req uired to discriminate healthy tissue samples from precancerous tissue samples. The use of connectionist methods such as multilayered percept rons, radial basis function (RBF) networks, and ensembles of such netw orks was investigated. RBF ensemble algorithms based on fluorescence s pectra potentially provide automated and near real-time implementation of precancer detection in the hands of nonexperts. The results are mo re reliable, direct, and accurate than those achieved by either human experts or multivariate statistical algorithms.