Comparing Classification Methods for Diffuse Reflectance Spectra to Improve Tissue Specific Laser Surgery (BMC Medical Research Methodology)

Background: In the field of oral and maxillofacial surgery, newly developed laser scalpels have multiple advantagesover traditional metal scalpels. However, they lack haptic feedback. This is dangerous near e.g. nervetissue, which has to be preserved during surgery. One solution to this problem is to train an algorithmthat analyzes the reflected light spectra during surgery and can classify these spectra into differenttissue types, in order to ultimately send a warning or temporarily switch off the laser when criticaltissue is about to be ablated. Various machine learning algorithms are available for this task, but adetailed analysis is needed to assess the most appropriate algorithm. Methods: In this study, a small data set is used to simulate many larger data sets according to a multivariateGaussian distribution. Various machine learning algorithms are then trained and evaluated onthese data sets. The algorithms¿ performance is subsequently evaluated and compared by averagedconfusion matrices and ultimately by boxplots of misclassification rates. The results are validated onthe smaller, experimental data set. Results: Most classifiers have a median misclassification rate below 0.25 in the simulated data. The mostnotable performance was observed for the Penalized Discriminant Analysis, with a misclassifiactionrate of 0.00 in the simulated data, and an average misclassification rate of 0.02 in a 10-fold crossvalidation on the original data. Conclusion: The results suggest a Penalized Discriminant Analysis is the most promising approach, most probablybecause it considers the functional, correlated nature of the reflectance spectra.The results of this study improve the accuracy of real-time tissue discrimination and are an essentialstep towards improving the safety of oral laser surgery.