Authors: Jemma G Kelly Plamen P Angelov Júlio Trevisan Anastasia Vlachopoulou Evangelos Paraskevaidis Pierre L MartinHirsch Francis L Martin
Publish Date: 2010/09/21
Volume: 398, Issue: 5, Pages: 2191-2201
Abstract
Although the UK cervical screening programme has reduced mortality associated with invasive disease advancement from a highthroughput predictive methodology that is costeffective and robust could greatly support the current system We combined analysis by attenuated total reflection Fouriertransform infrared spectroscopy of cervical cytology with selflearning classifier eClass This predictive algorithm can cope with vast amounts of multidimensional data with variable characteristics Using a characterised dataset set A consisting of UK cervical specimens designated as normal n = 60 lowgrade n = 60 or highgrade n = 60 and one further dataset set B consisting of n = 30 lowgrade samples we set out to determine whether this approach could be robustly predictive Variously extending the training set consisting of set A with set B data produced good classification rates with three twoclass cascade classifiers However a single threeclass classifier was equally efficient producing a userfriendly applicable methodology with improved interpretability ie better classification with only one set of fuzzy rules As data from set B were added incrementally to the training set the model learned and evolved Additionally monitoring of results of the set B lowgrade specimens known to be lowgrade cervical cytology specimens provided the opportunity to explore the possibility of distinguishing patients likely to progress towards invasive disease eClass exhibited a remarkably robust predictive power in a userfriendly fashion ie high throughput ease of use compared to other classifiers knearest neighbours support vector machines artificial neural networks Development of eClass to classify such datasets for applications such as screening exhibits robustness in identifying a dichotomous marker of invasive disease progression
Keywords: