Classification and Segmentation of Hyperspectral Data of Hepatocellular Carcinoma Samples Using 1‐D Convolutional Neural Network

Abstract

Pathological diagnosis plays an important role in the diagnosis and treatment of hepatocellular carcinoma (HCC). The traditional method of pathological diagnosis of most cancers requires freezing, slicing, hematoxylin and eosin staining, and manual analysis, limiting the speed of the diagnosis process. In this study, we designed a one‐dimensional convolutional neural network to classify the hyperspectral data of HCC sample slices acquired by our hyperspectral imaging system. A weighted loss function was employed to promote the performance of the model. The proposed method allows us to accelerate the diagnosis process of identifying tumor tissues. Our deep learning model achieved good performance on our data set with sensitivity, specificity, and area under receiver operating characteristic curve of 0.871, 0.888, and 0.950, respectively. Meanwhile, our deep learning model outperformed the other machine learning methods assessed on our data set. The proposed method is a potential tool for the label‐free and real‐time pathologic diagnosis. © 2019 International Society for Advancement of Cytometry

Hyperspectral Imaging during Normothermic Machine Perfusion—A Functional Classification of Ex Vivo Kidneys Based on Convolutional Neural Networks

“…The combination of CNN and HSI has been successfully applied to assist in the diagnosis of various tissue types. Examples of histologic tissue examination include classification or segmentation of head and neck cancer [ 39 , 50 ], breast cancer [ 37 , 51 ], gastric cancer [ 38 ], oral cancer [ 52 ], esophagus [ 53 ], hepatocellular carcinoma [ 54 ], glioblastoma tumor cells [ 40 ], and blood cells [ 55 ]. Its use in neurosurgery has been investigated for in vivo real-time segmentation of the tumor margin [ 40 ].…”

“…ROIs can be determined by visual inspection confirmed by a physician [ 38 , 52 , 54 ] or by appropriate labeling of acquired images by a surgeon using a semi-automated tool based on the spectral angle mapper algorithm [ 40 ]. Masks are also used to allow ROI selection by avoiding specular reflections [ 39 , 50 ].…”

“…In the literature, the application of different training parameters to optimize CNN models in HSI analysis was described. The following methods were mainly implemented: dropout [ 37 , 39 , 50 , 54 , 57 ], learning rate optimization [ 37 , 38 , 40 , 51 , 53 , 55 , 57 ], weight decay [ 51 , 55 , 57 ], and momentum factor [ 38 , 51 ]. The three most commonly used training parameters were investigated in this work for their suitability to improve the KidneyResNet model performance (see Table 2 ).…”

Artificial intelligence in liver diseases: Improving diagnostics, prognostics and response prediction

“…developed a DL model which accurately identified tumour tissue in hyperspectral data of unstained HCC samples. 34 Sun et al. used the DL technique of multiple instance learning to distinguish between HCC and normal liver tissue in WSIs, reporting AUCs of nearly 1.00.…”

Classification of hyperspectral images for detection of hepatic carcinoma cells based on spectral–spatial features of nucleus

“…Moreover, the identi¯cation of hepatic carcinoma cells with our method was better than that with the convolutional neural network algorithm. 25 However, there are some limitations of this classi¯cation method. The threshold for the correlation coe±cient was set based on the results for 500 samples.…”