Classifying breast cancer tissue by Raman spectroscopy with one-dimensional convolutional neural network.

As the most common cancer in women, breast cancer is becoming lethal worldwide. However, the current breast diagnosis technologies are not enough to meet the requirements in clinic due to some shortages of early-stage insensitiveness, time consumption and relying on the doctor's experience, etc. It's necessary to develop a creative method for the automatical diagnosis of breast cancer. Therefore, Raman spectroscopy and one-dimensional convolutional neural network (1D-CNN) algorithm were combined together for the first time to classify the healthy and cancerous breast tissues in this study. First, a number of Raman spectra were collected from breast samples of 20 patients for spectral analysis. Then, a 1D-CNN model was developed and trained for classification. In addition, the Fisher Discrimination Analysis (FDA) and Support Vector Machine (SVM) classifiers were trained and tested with the same spectral data for comparison. The best classification performance, namely the overall diagnostic accuracy of 92%, the sensitivity of 98% and the specificity of 86%, has been achieved by using 1D-CNN model. This study proves that 1D-CNN combined with Raman spectroscopy can classify breast tissues effectively and automatically and lay the foundation for automated cancer diagnosis in the future.