Rapid diagnosis of membranous nephropathy based on serum and urine Raman spectroscopy combined with deep learning methods

Abstract

Membranous nephropathy is the main cause of nephrotic syndrome, which has an insidious onset and may progress to end-stage renal disease with a high mortality rate, such as renal failure and uremia. At present, the diagnosis of membranous nephropathy mainly relies on the clinical manifestations of patients and pathological examination of kidney biopsy, which are expensive, time-consuming, and have certain chance and other disadvantages. Therefore, there is an urgent need to find a rapid, accurate and non-invasive diagnostic technique for the diagnosis of membranous nephropathy. In this study, Raman spectra of serum and urine were combined with deep learning methods to diagnose membranous nephropathy. After baseline correction and smoothing of the data, Gaussian white noise of different decibels was added to the training set for data amplification, and the amplified data were imported into ResNet, AlexNet and GoogleNet models to obtain the evaluation results of the models for membranous nephropathy. The experimental results showed that the three deep learning models achieved an accuracy of 1 for the classification of serum data of patients with membranous nephropathy and control group, and the discrimination of urine data was above 0.85, among which AlexNet was the best classification model for both samples. The above experimental results illustrate the great potential of serum- and urine-based Raman spectroscopy combined with deep learning methods for rapid and accurate identification of patients with membranous nephropathy.

RepDwNet: Lightweight Deep Learning Model for Special Biological Blood Raman Spectra Analysis

“…To assess the performance of Raman RepDwNet in Raman spectroscopy classification, we conducted a comparative analysis with other network classification models proposed by scholars in the last two years. These models include a one-dimensional VGG Raman spectroscopy classification network proposed by Sang [27], a model combining LSTM with convolutional neural networks proposed by Bratchenkoa [28], a one-dimensional AlexNet Raman spectroscopy classification network designed by Zhang [29], as well as a pure multi-head attention mechanism network adopted by Liu et al [30] and an adaptive multi-scale convolutional neural network designed by Deng et al [19]. In the process of experimental comparison, we utilized the same experimental dataset and employed a 10-fold cross-validation method.…”

RepDwNet: Lightweight Deep Learning Model for Spectral blood Raman spectra classification

“…To assess the Raman spectral classification potential exhibited by Raman RepDwNet, we conducted a comprehensive comparative analysis with network classification models proposed by other researchers in the past two years. These models encompass the one-dimensional VGG Raman spectral classification network introduced by Song et al [11], the model that combines LSTM with convolutional neural networks as devised by Bratchenko et al [22], the one-dimensional AlexNet Raman spectral classification network designed by Zhang et al [23], and the purely multi-head attention mechanism network adopted by Liu et al [9] .Throughout the experimental comparison process, we employed the same experimental dataset as the benchmark. We employed a 10-fold cross-validation strategy and fine-tuned each model to ensure the attainment of optimal classification predictive performance.…”

Toward Practical Augmentation of Raman Spectra for Deep Learning Classification of Contamination in HDD

“…The authors investigated and compared four neural networks: a multilayer perceptron, a simple recursive neural network, a simple CNN, and AlexNet. Zhang et al [22] proposed DLbased methods to identify patients with membranous nephropathy using the Raman spectra of serum, urine, and DL models. Among the investigated models (AlexNet, GoogL-Net, and ResNet), AlexNet yields the best accuracy.…”