Classification of Healthy and Cancer States of Colon Epithelial Tissues Using Opto-magnetic Imaging Spectroscopy

Colorectal carcinoma (CRC) as a major health problem in industrialized countries is highly preventable and can be successfully treated in the early stages. However, incidence and mortality of CRC has increased over the last two decades. The reason could be that the current recommended options for screening are costly, unpleasant for patients, have low sensitivity and poor accessibility for screening. These reasons provide a strong rationale for the development of a new method. Opto-magnetic imaging spectroscopy (OMIS) as a new imaging method for the characterisation of various materials, including human tissues, is based on light-matter interaction, using a Poincare sphere for light properties and a Bloch sphere for electron properties, and allows the detection of biophysical characteristics within human tissue samples. Compared with histopathology examination, the OMIS method achieved an accuracy of 92.59% using Multilayer Perceptron Neural Network as a classifier, and 89.87% using Naive-Bayes, respectively. The obtained results, based on the investigation of 316 samples, both tumour and normal mucosa (162 cancer cases), strongly suggest that the new non-invasive OMIS method might be used for tissue characterization ex vivo to discriminate between the healthy and carcinoma state of the colon. However, it opens up the possibility of using the same method in in vivo studies to assist physicians in targeting biopsies of colorectal tissue.