Rapid and accurate identification of colon cancer by Raman spectroscopy coupled with convolutional neural networks

Colonoscopy is regarded as the gold standard in colorectal tumor diagnosis, but it is costly and time-consuming. Raman spectroscopy has shown promise for differentiating cancerous from non-cancerous tissue and is expected to be a new tool for oncological diagnosis. However, traditional Raman spectroscopy analysis requires tedious preprocessing, and the classification accuracy needs to be improved. In this work, a novel Raman spectral qualitative classification method based on convolutional neural network (CNN) is proposed for the identification of three different colon tissue samples, including adenomatous polyp, adenocarcinoma and normal tissues. Experimental results show that this CNN model has superior feature extraction ability. For the spectral data of new individuals, the trained CNN model presents much better classification performance than traditional machine learning methods, such as the k-nearest neighbor, random forest, and support vector machine. Raman spectroscopy combined with CNN can be used as an effective auxiliary tool for the early diagnosis of colon cancer.