More Like this

1. INPREM: An Interpretable and Trustworthy Predictive Model for

   Zhang, X; Qian, B; Cao, S; Li, Y; Chen, H; Zheng, Y; Davidson, I (August 2020, KDD)

   null (Ed.)

   Building a predictive model based on historical Electronic Health Records (EHRs) for personalized healthcare has become an active research area. Benefiting from the powerful ability of feature ex- traction, deep learning (DL) approaches have achieved promising performance in many clinical prediction tasks. However, due to the lack of interpretability and trustworthiness, it is difficult to apply DL in real clinical cases of decision making. To address this, in this paper, we propose an interpretable and trustworthy predictive model (INPREM) for healthcare. Firstly, INPREM is designed as a linear model for interpretability while encoding non-linear rela- tionships into the learning weights for modeling the dependencies between and within each visit. This enables us to obtain the contri- bution matrix of the input variables, which is served as the evidence of the prediction result(s), and help physicians understand why the model gives such a prediction, thereby making the model more in- terpretable. Secondly, for trustworthiness, we place a random gate (which follows a Bernoulli distribution to turn on or off) over each weight of the model, as well as an additional branch to estimate data noises. With the help of the Monto Carlo sampling and an ob- jective function accounting for data noises, the model can capture the uncertainty of each prediction. The captured uncertainty, in turn, allows physicians to know how confident the model is, thus making the model more trustworthy. We empirically demonstrate that the proposed INPREM outperforms existing approaches with a significant margin. A case study is also presented to show how the contribution matrix and the captured uncertainty are used to assist physicians in making robust decisions. 

   more » « less
2. Uncertainty quantification in the radiogenomics modeling of EGFR amplification in glioblastoma

   https://doi.org/10.1038/s41598-021-83141-z  

   Hu, Leland S.; Wang, Lujia; Hawkins-Daarud, Andrea; Eschbacher, Jennifer M.; Singleton, Kyle W.; Jackson, Pamela R.; Clark-Swanson, Kamala; Sereduk, Christopher P.; Peng, Sen; Wang, Panwen; et al (December 2021, Scientific Reports)

   null (Ed.)

   Abstract Radiogenomics uses machine-learning (ML) to directly connect the morphologic and physiological appearance of tumors on clinical imaging with underlying genomic features. Despite extensive growth in the area of radiogenomics across many cancers, and its potential role in advancing clinical decision making, no published studies have directly addressed uncertainty in these model predictions. We developed a radiogenomics ML model to quantify uncertainty using transductive Gaussian Processes (GP) and a unique dataset of 95 image-localized biopsies with spatially matched MRI from 25 untreated Glioblastoma (GBM) patients. The model generated predictions for regional EGFR amplification status (a common and important target in GBM) to resolve the intratumoral genetic heterogeneity across each individual tumor—a key factor for future personalized therapeutic paradigms. The model used probability distributions for each sample prediction to quantify uncertainty, and used transductive learning to reduce the overall uncertainty. We compared predictive accuracy and uncertainty of the transductive learning GP model against a standard GP model using leave-one-patient-out cross validation. Additionally, we used a separate dataset containing 24 image-localized biopsies from 7 high-grade glioma patients to validate the model. Predictive uncertainty informed the likelihood of achieving an accurate sample prediction. When stratifying predictions based on uncertainty, we observed substantially higher performance in the group cohort (75% accuracy, n = 95) and amongst sample predictions with the lowest uncertainty (83% accuracy, n = 72) compared to predictions with higher uncertainty (48% accuracy, n = 23), due largely to data interpolation (rather than extrapolation). On the separate validation set, our model achieved 78% accuracy amongst the sample predictions with lowest uncertainty. We present a novel approach to quantify radiogenomics uncertainty to enhance model performance and clinical interpretability. This should help integrate more reliable radiogenomics models for improved medical decision-making. 

   more » « less
3. A Bayesian Network Approach for Modeling Resident Physician Vigilance

   https://doi.org/10.1177/10711813251369394  

   Prendez, David; Kettel, Veronika; Higgins, Elizabeth; Li, Jiaxin; Kwon, Younghoon; Roach, Victoria; Kim, Ji-Eun (September 2025, Proceedings of the Human Factors and Ergonomics Society Annual Meeting)

   Vigilance, the ability to sustain attention, is critical in healthcare, yet resident physicians face significant sleep deprivation, increasing their risk of vigilance decrement and medical errors. This study aimed to develop a predictive model of vigilance in this population using contextual factors, physiological measures, and eye-tracking data. Fifteen resident physicians participated in psychomotor vigilance tests (PVT) under sleep-deprived and non-sleep-deprived conditions, and completed questionnaires assessing sleep, anxiety, and workload. Bayesian Networks (BN) were employed to model vigilance, featuring layers for contextual factors (sleep, anxiety), performance (PVT reaction time), and observable features (eye movement, physiological responses). The three-layered BN integrating both contextual and multi-sensor (eye-tracking and physiological) data demonstrated the best prediction accuracy, compared to BNs with fewer layers and/or only one sensor type. This demonstrates that combining continuous physiological and eye-tracking data with contextual information enhances the prediction of vigilance decrement in resident physicians. This study contributes to the development of predictive tools for mitigating vigilance decrement and the future design of intervention strategies in demanding clinical settings. 

   more » « less
4. Predicting Cardiac Complications of Myocardial Infarction Patients Using Machine Learning

   https://doi.org/10.1109/BigData62323.2024.10826091  

   Baidya, Shriyansh; Gupta, Vibhuti (January 2025, IEEE)

   In the United States, heart disease is the leading cause of death, killing about 695,000 people each year. Myocardial infarction (MI) is a cardiac complication which occurs when blood flow to a portion of the heart decreases or halts, leading to damage in the heart muscle. Heart failure and Atrial fibrillation (AF) are closely associated with MI. Heart failure is a common complication of MI and a risk factor for AF. Machine learning (ML) and deep learning techniques have shown potential in predicting cardiovascular conditions. However, developing a sim- plified predictive model, along with a thorough feature analysis, is challenging due to various factors, including lifestyle, age, family history, medical conditions, and clinical variables for cardiac complications prediction. This paper aims to develop simplified models with comprehensive feature analysis and data preprocessing for predicting cardiac complications, such as heart failure and atrial fibrillation linked with MI, using a publicly available dataset of myocardial infarction patients. This will help the students and health care professionals understand various factors responsible for cardiac complications through a simplified workflow. By prioritizing interpretability, this paper illustrates how simpler models, like decision trees and logistic regression, can provide transparent decision-making processes while still maintaining a balance with accuracy. Additionally, this paper examines how age-specific factors affect heart failure and atrial fibrillation conditions. Overall this research focuses on making machine learning accessible and interpretable. Its goal is to equip students and non-experts with practical tools to understand how ML can be applied in healthcare, particularly for the cardiac complications prediction for patients having MI. 

   more » « less
5. Failure Detection in Deep Neural Networks for Medical Imaging

   https://doi.org/10.3389/fmedt.2022.919046  

   Ahmed, Sabeen; Dera, Dimah; Hassan, Saud Ul; Bouaynaya, Nidhal; Rasool, Ghulam (July 2022, Frontiers in Medical Technology)

   Deep neural networks (DNNs) have started to find their role in the modern healthcare system. DNNs are being developed for diagnosis, prognosis, treatment planning, and outcome prediction for various diseases. With the increasing number of applications of DNNs in modern healthcare, their trustworthiness and reliability are becoming increasingly important. An essential aspect of trustworthiness is detecting the performance degradation and failure of deployed DNNs in medical settings. The softmax output values produced by DNNs are not a calibrated measure of model confidence. Softmax probability numbers are generally higher than the actual model confidence. The model confidence-accuracy gap further increases for wrong predictions and noisy inputs. We employ recently proposed Bayesian deep neural networks (BDNNs) to learn uncertainty in the model parameters. These models simultaneously output the predictions and a measure of confidence in the predictions. By testing these models under various noisy conditions, we show that the (learned) predictive confidence is well calibrated. We use these reliable confidence values for monitoring performance degradation and failure detection in DNNs. We propose two different failure detection methods. In the first method, we define a fixed threshold value based on the behavior of the predictive confidence with changing signal-to-noise ratio (SNR) of the test dataset. The second method learns the threshold value with a neural network. The proposed failure detection mechanisms seamlessly abstain from making decisions when the confidence of the BDNN is below the defined threshold and hold the decision for manual review. Resultantly, the accuracy of the models improves on the unseen test samples. We tested our proposed approach on three medical imaging datasets: PathMNIST, DermaMNIST, and OrganAMNIST, under different levels and types of noise. An increase in the noise of the test images increases the number of abstained samples. BDNNs are inherently robust and show more than 10% accuracy improvement with the proposed failure detection methods. The increased number of abstained samples or an abrupt increase in the predictive variance indicates model performance degradation or possible failure. Our work has the potential to improve the trustworthiness of DNNs and enhance user confidence in the model predictions. 

   more » « less