Using deep neural networks and interpretability methods to identify gene expression patterns that predict radiomic features and histology in non-small cell lung cancer

Abstract

Purpose: Integrative analysis combining diagnostic imaging and genomic information can uncover biological insights into lesions that are visible on radiologic images. We investigate techniques for interrogating a deep neural network trained to predict quantitative image (radiomic) features and histology from gene expression in non-small cell lung cancer (NSCLC).Approach: Using 262 training and 89 testing cases from two public datasets, deep feedforward neural networks were trained to predict the values of 101 computed tomography (CT) radiomic features and histology. A model interrogation method called gene masking was used to derive the learned associations between subsets of genes and a radiomic feature or histology class [adenocarcinoma (ADC), squamous cell, and other].Results: Overall, neural networks outperformed other classifiers. In testing, neural networks classified histology with area under the receiver operating characteristic curves (AUCs) of 0.86 (ADC), 0.91 (squamous cell), and 0.71 (other). Classification performance of radiomics features ranged from 0.42 to 0.89 AUC. Gene masking analysis revealed new and previously reported associations. For example, hypoxia genes predicted histology (>0.90 AUC). Previously published gene signatures for classifying histology were also predictive in our model (>0.80 AUC). Gene sets related to the immune or cardiac systems and cell development processes were predictive (>0.70 AUC) of several different radiomic features. AKT signaling, tumor necrosis factor, and Rho gene sets were each predictive of tumor textures.Conclusions: This work demonstrates neural networks' ability to map gene expressions to radiomic features and histology types in NSCLC and to interpret the models to identify predictive genes associated with each feature or type.

Construction of a multiple-class classifier based on mRNAs and lncRNA FAM66A and PSORS1C3 for predicting distant metastasis in lung adenocarcinoma

“…Wei et al also used a CNN to classify NSCLC histopathology types and transcriptomic subtypes ( 37 ), and Le Page et al used CNN to classify NSCLC based on diagnostic histopathology HES images ( 38 ). Smedley et al used deep neural networks and interpretability methods to identify gene expression patterns that predict radiomic features and histology in NSCLC ( 39 ).…”

Structured Reporting of Computed Tomography and Magnetic Resonance in the Staging of Pancreatic Adenocarcinoma: A Delphi Consensus Proposal

“…These data could create the basis of a large database, allowing not only for the carrying out of epidemiological statistical analysis, but they could be used to build a radiomics model by combining radiological features and clinical data [33][34][35][36][37][38][39]. In this context, the added value of genomic data could be used to develop a model of radiogenomics, which was helpful regarding the highest level of personalized risk stratification and the advanced precision medicine process [40][41][42][43][44][45].…”

Structured Reporting of Lung Cancer Staging: A Consensus Proposal

“…Natural language processing applications can help extract the data from the reports with variable terminology, allowing the compilation of standardized data, which could then be used to develop multi-institutional data registries, as well as in clinical and research analyses. Moreover, the possibility of combining genomic data and radiological features allows for developing models of radiogenomics—models that today represent the highest level of advanced-precision medicine processes [ 38 , 39 , 40 , 41 ]. The fact that the present SR can be included in the picture archiving and communication system (PACS) is an added value; therefore, it is only necessary to enter these data once upon first entry into the radiology department.…”