Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning

Abstract

Visual inspection of histopathology slides is one of the main methods used by pathologists to assess the stage, type and subtype of lung tumors. Adenocarcinoma (LUAD) and squamous cell carcinoma (LUSC) are the most prevalent subtypes of lung cancer, and their distinction requires visual inspection by an experienced pathologist. In this study, we trained a deep convolutional neural network (inception v3) on whole-slide images obtained from The Cancer Genome Atlas to accurately and automatically classify them into LUAD, LUSC or normal lung tissue. The performance of our method is comparable to that of pathologists, with an average area under the curve (AUC) of 0.97. Our model was validated on independent datasets of frozen tissues, formalin-fixed paraffin-embedded tissues and biopsies. Furthermore, we trained the network to predict the ten most commonly mutated genes in LUAD. We found that six of them-STK11, EGFR, FAT1, SETBP1, KRAS and TP53-can be predicted from pathology images, with AUCs from 0.733 to 0.856 as measured on a held-out population. These findings suggest that deep-learning models can assist pathologists in the detection of cancer subtype or gene mutations. Our approach can be applied to any cancer type, and the code is available at https://github.com/ncoudray/DeepPATH .

MicroRNAs associated with lung squamous cell carcinoma: New prognostic biomarkers and therapeutic targets

“…Squamous cell carcinoma of the lung (LUSC) is a common type of NSCLC with poor clinical outcomes, with only 15% of patients surviving for 5 or more years . Due to a lack of clear histological features for LUSC, although some treatments are effective there is no clear personalized treatment strategy as to whether conventional radiotherapy, chemotherapy, or targeted therapy will be most effective . Based on the current state of onset and treatment of patients with tumor, the importance of finding relevant tumor markers has become increasingly prominent .…”

A deep learning algorithm for simulating autonomous driving considering prior knowledge and temporal information

“…This network has been shown to attain greater than 78.1% accuracy on the ImageNet dataset and achieves 21.2% top 1 and 5.6% top 5 error rates (i.e., the percentage of test examples for which the correct class was not in the top 1 and top 5 predicted classes, respectively). In one of the most recent applications of the Inception v3 network structure, a research team from New York University trained the network to classify different types of lung cancer (Coudray et al., ).…”

The coming 15 years in gynaecological pathology: digitisation, artificial intelligence, and new technologies

“…A machine learning algorithm that can aid in histological classification, combined with the ability to provide an exact measurement of surface area, would not only simplify this process, but also make it more accurate and meaningful to clinicians and patients. Going one step further, there has also been some early success in using machine learning algorithms to predict the mutational status of tumours . This has the potential to dramatically change the practice of surgical pathology.…”