Abstract:Background: Our previous studies have demonstrated that Raman spectroscopy could be used for skin cancer detection with good sensitivity and specificity. The objective of this study is to determine if skin cancer detection can be further improved by combining deep neural networks and Raman spectroscopy. Patients and methods: Raman spectra of 731 skin lesions were included in this study, containing 340 cancerous and precancerous lesions (melanoma, basal cell carcinoma, squamous cell carcinoma and actinic keratosis) and 391 benign lesions (melanocytic nevus and seborrheic keratosis). One-dimensional convolutional neural networks (1D-CNN) were developed for Raman spectral classification. The stratified samples were divided randomly into training (70%), validation (10%) and test set (20%), and were repeated 56 times using parallel computing. Different data augmentation strategies were implemented for the training dataset, including added random noise, spectral shift, spectral combination and artificially synthesized Raman spectra using one-dimensional generative adversarial networks (1D-GAN). The area under the receiver operating characteristic curve (ROC AUC) was used as a measure of the diagnostic performance. Conventional machine learning approaches, including partial least squares for discriminant analysis (PLS-DA), principal component and linear discriminant analysis (PC-LDA), support vector machine (SVM), and logistic regression (LR) were evaluated for comparison with the same data splitting scheme as the 1D-CNN. Results: The ROC AUC of the test dataset based on the original training spectra were 0.886±0.022 (1D-CNN), 0.870±0.028 (PLS-DA), 0.875±0.033 (PC-LDA), 0.864±0.027 (SVM), and 0.525±0.045 (LR), which were improved to 0.909±0.021 (1D-CNN), 0.899±0.022 (PLS-DA), 0.895±0.022 (PC-LDA), 0.901±0.020 (SVM), and 0.897±0.021 (LR) respectively after augmentation of the training dataset (p<0.0001, Wilcoxon test). Paired analyses of 1D-CNN with conventional machine learning approaches showed that 1D-CNN had a 1–3% improvement (p<0.001, Wilcoxon test). Conclusions: Data augmentation not only improved the performance of both deep neural networks and conventional machine learning techniques by 2–4%, but also improved the performance of the models on spectra with higher noise or spectral shifting. Convolutional neural networks slightly outperformed conventional machine learning approaches for skin cancer detection by Raman spectroscopy.

Serum Raman spectroscopy combined with Gaussian—convolutional neural network models to quickly detect liver cancer patients

Rapid diagnosis of lung cancer and glioma based on serum Raman spectroscopy combined with deep learning

Classification of Focal Liver Lesions Using Deep Learning with Fine-Tuning

Deep Learning-Assisted Diagnosis of Liver Tumors Using Non-Contrast Magnetic Resonance Imaging: A Multi-Center Study

Deep Learning for Accurate Diagnosis of Liver Tumor Based on Magnetic Resonance Imaging and Clinical Data

Laser tweezer Raman spectroscopy combined with deep neural networks for identification of liver cancer cells

High-Precision Intelligent Cancer Diagnosis Method: 2D Raman Figures Combined with Deep Learning

Serum Raman spectroscopy combined with multiple classification models for rapid diagnosis of breast cancer

Serum analysis based on SERS combined with 2D convolutional neural network and Gramian angular field for breast cancer screening

Rapid and precise detection of cancers via label-free SERS and deep learning

A Novel and Rapid Serum Detection Technology for Non-Invasive Screening of Gastric Cancer Based on Raman Spectroscopy Combined With Different Machine Learning Methods

Multi-Stream Scale-Insensitive Convolutional And Recurrent Neural Networks For Liver Tumor Detection In Dynamic Ct Images

Improving skin cancer detection by Raman spectroscopy using convolutional neural networks and data augmentation

Noninvasive Cholangitis and Cholangiocarcinoma Screening Based on Serum Raman Spectroscopy and Support Vector Machine

Screening Ovarian Cancers with Raman Spectroscopy of Blood Plasma Coupled with Machine Learning Data Processing.

Fully connected neural network-based serum surface-enhanced Raman spectroscopy accurately identifies non-alcoholic steatohepatitis

Machine-learning-assisted spontaneous Raman spectroscopy classification and feature extraction for the diagnosis of human laryngeal cancer

[Using ANN and Serum Protein Pattern Models in Liver Cancer Diagnosis].

Deep Convolutional Neural Networks Combine Raman Spectral Signature of Serum for Prostate Cancer Bone Metastases Screening.

Application of Serum Mid-Infrared Spectroscopy Combined with Machine Learning in Rapid Screening of Breast Cancer and Lung Cancer

Human serum mid-infrared spectroscopy combined with machine learning algorithms for rapid detection of gliomas