Deep Learning Applied to Raman Spectroscopy for the Detection of Microsatellite Instability/MMR Deficient Colorectal Cancer

Abstract

Defective DNA mismatch repair is one pathogenic pathway to colorectal cancer. It is characterised by microsatellite instability which provides a molecular biomarker for its detection. Clinical guidelines for universal testing of this biomarker are not met due to resource limitations; thus, there is interest in developing novel methods for its detection. Raman spectroscopy (RS) is an analytical tool able to interrogate the molecular vibrations of a sample to provide a unique biochemical fingerprint. The resulting datasets are complex and high-dimensional, making them an ideal candidate for deep learning, though this may be limited by small sample sizes. This study investigates the potential of using RS to distinguish between normal, microsatellite stable (MSS) and microsatellite unstable (MSI-H) adenocarcinoma in human colorectal samples and whether deep learning provides any benefit to this end over traditional machine learning models. A 1D convolutional neural network (CNN) was developed to discriminate between healthy, MSI-H and MSS in human tissue and compared to a principal component analysis–linear discriminant analysis (PCA–LDA) and a support vector machine (SVM) model. A nested cross-validation strategy was used to train 30 samples, 10 from each group, with a total of 1490 Raman spectra. The CNN achieved a sensitivity and specificity of 83% and 45% compared to PCA–LDA, which achieved a sensitivity and specificity of 82% and 51%, respectively. These are competitive with existing guidelines, despite the low sample size, speaking to the molecular discriminative power of RS combined with deep learning. A number of biochemical antecedents responsible for this discrimination are also explored, with Raman peaks associated with nucleic acids and collagen being implicated.

Classifying Raman Spectra of Colon Cells Based on Machine Learning Algorithms

“…The eXtreme Gradient Boosting trees method performed better than the other two methods, although these authors also used the whole dataset to build the classification models, which were tested with a leave-one-out cross-validation approach. As a third example, N. Blake et al [40] investigated the potential of using Raman spectroscopy to distinguish between normal cells and adenocarcinoma in human colorectal tissue samples. In particular, they obtained discrimination accuracy values between 71% and 75% by means of PCA-LDA, SVM, and CNN.…”

Comparative Study on Handheld, Modular, and Laboratory Raman Instruments for the Analysis of Colon Tissues and Colorectal Polyps

“…This goal can be achieved by comparing individual characteristic Raman bands of biochemical markers of carcinogenesis or through the discriminatory analysis of Raman spectra using multivariate statistical methods. Previous works have used various approaches to separate and classify normal and pathological tissue and biofluid samples based on Raman spectra, such as HCA [6,43], PCA [40], partial least squares discriminant analysis (PLS-DA) [51,52], SVM and other machine learning methods [43], PCA in combination with linear discriminant analysis (LDA), SVM, SIMCA [53] and other discriminating parameters like Mahalanobis distance or spectral residuals [45,52,[54][55][56][57], and convolutional neural networks (CNN) [58][59][60][61]. Of particular interest for early cancer diagnosis is the immediate analysis of intracellular Raman spectra using a biomolecular component analysis (BCA) algorithm [62].…”