Implementing deep convolutional neural networks on adenocarcinoma images for classification and genomic association analysis
Antonio Victor Andrew Asuncion · Institutional Repositories DataBase (IRDB) · 2018
A number of advances involving convolutional neural network (CNN) architectures have demonstrated that extracting morphological features from histological images can be effective for classification of subtypes of various diseases, especially cancer.On the other hand, varying types of gene expression or mutation data have been a rich and ubiquitous resource for studies involving cancer prognosis, survival, and others.In this study, we present a method for classifying transcriptome subtypes of lung adenocarcinoma from slices of pathological images whose features come from convolutional autoencoders pretrained on smaller images.We also attempted to provide a stepping stone for whole slide image analysis by performing classification and correlation analysis.Whole slide images were processed based on features extracted from Google's Inception architecture.A classification was then implemented on the processed images.Furthermore, a correlation between the image features and their corresponding gene expression data were investigated.Variants of autoencoders as building blocks of pretrained convolutional layers of neural networks were implemented on histological images gathered from the Cancer Genome Atlas database.From here, a sparse deep autoencoder was proposed and applied to images of size 2048x2048.We applied this model for feature extraction from pathological images of lung adenocarcinoma, which is comprised