Dataset Information

Shedding Light on the Black Box: Explaining Deep Neural Network Prediction of Clinical Outcomes.

ABSTRACT: Deep neural network models are emerging as an important method in healthcare delivery, following the recent success in other domains such as image recognition. Due to the multiple non-linear inner transformations, deep neural networks are viewed by many as black boxes. For practical use, deep learning models require explanations that are intuitive to clinicians. In this study, we developed a deep neural network model to predict outcomes following major cardiovascular procedures, using temporal image representation of past medical history as input. We created a novel explanation for the prediction of the model by defining impact scores that associate clinical observations with the outcome. For comparison, a logistic regression model was fitted to the same dataset. We compared the impact scores and log odds ratios by calculating three types of correlations, which provided a partial validation of the impact scores. The deep neural network model achieved an area under the receiver operating characteristics curve (AUC) of 0.787, compared to 0.746 for the logistic regression model. Moderate correlations were found between the impact scores and the log odds ratios. Impact scores generated by the explanation algorithm has the potential to shed light on the "black box" deep neural network model and could facilitate its adoption by clinicians.

SUBMITTER: Shao Y

PROVIDER: S-EPMC7983057 | biostudies-literature | 2021 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Shedding Light on the Black Box: Explaining Deep Neural Network Prediction of Clinical Outcomes.

Shao Yijun Y Cheng Yan Y Shah Rashmee U RU Weir Charlene R CR Bray Bruce E BE Zeng-Treitler Qing Q

Journal of medical systems 20210104 1

Deep neural network models are emerging as an important method in healthcare delivery, following the recent success in other domains such as image recognition. Due to the multiple non-linear inner transformations, deep neural networks are viewed by many as black boxes. For practical use, deep learning models require explanations that are intuitive to clinicians. In this study, we developed a deep neural network model to predict outcomes following major cardiovascular procedures, using temporal i ...[more]

PMID: 33404886

Similar Datasets

Project description:We propose a BlackBox Counterfactual Explainer, designed to explain image classification models for medical applications. Classical approaches (e.g., , saliency maps) that assess feature importance do not explain how imaging features in important anatomical regions are relevant to the classification decision. Such reasoning is crucial for transparent decision-making in healthcare applications. Our framework explains the decision for a target class by gradually exaggerating the semantic effect of the class in a query image. We adopted a Generative Adversarial Network (GAN) to generate a progressive set of perturbations to a query image, such that the classification decision changes from its original class to its negation. Our proposed loss function preserves essential details (e.g., support devices) in the generated images. We used counterfactual explanations from our framework to audit a classifier trained on a chest X-ray dataset with multiple labels. Clinical evaluation of model explanations is a challenging task. We proposed clinically-relevant quantitative metrics such as cardiothoracic ratio and the score of a healthy costophrenic recess to evaluate our explanations. We used these metrics to quantify the counterfactual changes between the populations with negative and positive decisions for a diagnosis by the given classifier. We conducted a human-grounded experiment with diagnostic radiology residents to compare different styles of explanations (no explanation, saliency map, cycleGAN explanation, and our counterfactual explanation) by evaluating different aspects of explanations: (1) understandability, (2) classifier's decision justification, (3) visual quality, (d) identity preservation, and (5) overall helpfulness of an explanation to the users. Our results show that our counterfactual explanation was the only explanation method that significantly improved the users' understanding of the classifier's decision compared to the no-explanation baseline. Our metrics established a benchmark for evaluating model explanation methods in medical images. Our explanations revealed that the classifier relied on clinically relevant radiographic features for its diagnostic decisions, thus making its decision-making process more transparent to the end-user.

Project description:MotivationA majority of known genetic variants associated with human-inherited diseases lie in non-coding regions that lack adequate interpretation, making it indispensable to systematically discover functional sites at the whole genome level and precisely decipher their implications in a comprehensive manner. Although computational approaches have been complementing high-throughput biological experiments towards the annotation of the human genome, it still remains a big challenge to accurately annotate regulatory elements in the context of a specific cell type via automatic learning of the DNA sequence code from large-scale sequencing data. Indeed, the development of an accurate and interpretable model to learn the DNA sequence signature and further enable the identification of causative genetic variants has become essential in both genomic and genetic studies.ResultsWe proposed Deopen, a hybrid framework mainly based on a deep convolutional neural network, to automatically learn the regulatory code of DNA sequences and predict chromatin accessibility. In a series of comparison with existing methods, we show the superior performance of our model in not only the classification of accessible regions against background sequences sampled at random, but also the regression of DNase-seq signals. Besides, we further visualize the convolutional kernels and show the match of identified sequence signatures and known motifs. We finally demonstrate the sensitivity of our model in finding causative noncoding variants in the analysis of a breast cancer dataset. We expect to see wide applications of Deopen with either public or in-house chromatin accessibility data in the annotation of the human genome and the identification of non-coding variants associated with diseases.Availability and implementationDeopen is freely available at https://github.com/kimmo1019/Deopen.Contactruijiang@tsinghua.edu.cn.Supplementary informationSupplementary data are available at Bioinformatics online.

Dataset Information

Shedding Light on the Black Box: Explaining Deep Neural Network Prediction of Clinical Outcomes.

Publications

Shedding Light on the Black Box: Explaining Deep Neural Network Prediction of Clinical Outcomes.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets