Application of Raman spectroscopy technology based on deep learning algorithm in the rapid diagnosis of glioma

Abstract

Medical diagnosis technology based on convolutional neural networks (CNNs) has achieved good performance. In this study, we collected serum samples from 38 glioma patients and 45 healthy controls and used partial least squares (PLS) analysis to reduce the dimension of the data. Different levels of noise were added to the reduced data onto data augmentation, and the AlexNet, ResNet, and GoogLeNet fine‐tuning models were applied for classification. To evaluate the performance of the models, we used five‐fold cross‐validation. The accuracy rates of the AlexNet, ResNet, and GoogLeNet fine‐tuning models were 98.50%, 98.24%, and 99.50%, respectively. The model with the best classification effect was GoogLeNet. The specificity and sensitivity of this model were 98.98% and 98.48%, respectively. In addition, the area under the receiver operating characteristic (ROC) curve (AUC) of the established diagnostic model was 0.9949. The results showed that the combination of serum Raman spectroscopy and the PLS‐Gaussian‐GoogLeNet model achieved a good diagnostic effect for glioma. This method has high clinical application value and is worthy of further popularization.

Rapid identification of human muscle disease with fibre optic Raman spectroscopy

“…Such a study would also permit testing of additional and more sophisticated machine learning algorithms, which may improve diagnostic performance. 36 A larger trial would also be able to explore if Raman spectra can be used to further classify specimens, for example, into different genetic subtypes of mitochondrial myopathy (as well as mitochondrial disease without myopathy), or into dystrophic or inflammatory myopathies. If these distinctions were possible, Raman analyses could potentially help direct further tests such as immunohistochemical and/or genetic analyses.…”

A new method for Raman spectral analysis: Decision fusion‐based transfer learning model

“…As a supervised dimensional reduction method, PLS enable the extracted characteristic variables to not only well summarize the information of the original variables, but also have a strong ability to explain the dependent variables. [50] After dummy processing of the dependent variable, we call this technique PLS-DA. [51][52][53] Based on the cumulative variance contribution rate of PCs and the number of PCs commonly used in Vieira et al, 54 we retained the results of PLS-DA processing in the first 50 dimensions of all data sets as input to ensure that the cumulative contribution rate of features and labels could reach more than 97%, as shown in Table S4.…”

Simple machine learning methods work surprisingly well for Ramanomics

“…Classification of Raman spectra by deep learning methods has been a topic of interest for over three decades 1 . With the recent success of deep learning, in scientific domains such as protein folding, 2 there also seems to be a growing wave of interest in deep learning for Ramanomics 3,4 . However, simpler machine learning methods could potentially outperform deep learning methods 5 .…”

“…1 With the recent success of deep learning, in scientific domains such as protein folding, 2 there also seems to be a growing wave of interest in deep learning for Ramanomics. 3,4 However, simpler machine learning methods could potentially outperform deep learning methods. 5 I personally found this to be true, on a dataset in the wild, when I could gain a 0.5% in accuracy for the classification of pathogenic bacteria by their Raman spectra, 6 by replacing the originally proposed 26-layer deep convolutional neural network with a sample-standardized softmax regression classifier.…”