Early and fair COVID-19 outcome risk assessment using robust feature selection.

Personalized medicine plays an important role in treatment optimization for COVID-19 patient management. Early treatment in patients at high risk of severe complications is vital to prevent death and ventilator use. Predicting COVID-19 clinical outcomes using machine learning may provide a fast and data-driven solution for optimizing patient care by estimating the need for early treatment. In addition, it is essential to accurately predict risk across demographic groups, particularly those underrepresented in existing models. Unfortunately, there is a lack of studies demonstrating the equitable performance of machine learning models across patient demographics. To overcome this existing limitation, we generate a robust machine learning model to predict patient-specific risk of death or ventilator use in COVID-19 positive patients using features available at the time of diagnosis. We establish the value of our solution across patient demographics, including gender and race. In addition, we improve clinical trust in our automated predictions by generating interpretable patient clustering, patient-level clinical feature importance, and global clinical feature importance within our large real-world COVID-19 positive patient dataset. We achieved 89.38% area under receiver operating curve (AUROC) performance for severe outcomes prediction and our robust feature ranking approach identified the presence of dementia as a key indicator for worse patient outcomes. We also demonstrated that our deep-learning clustering approach outperforms traditional clustering in separating patients by severity of outcome based on mutual information performance. Finally, we developed an application for automated and fair patient risk assessment with minimal manual data entry using existing data exchange standards.

Use of Natural Language Processing of Patient-Initiated Electronic Health Record Messages to Identify Patients With COVID-19 Infection.

Importance: Natural language processing (NLP) has the potential to enable faster treatment access by reducing clinician response time and improving electronic health record (EHR) efficiency. Objective: To develop an NLP model that can accurately classify patient-initiated EHR messages and triage COVID-19 cases to reduce clinician response time and improve access to antiviral treatment. Design, Setting, and Participants: This retrospective cohort study assessed development of a novel NLP framework to classify patient-initiated EHR messages and subsequently evaluate the model's accuracy. Included patients sent messages via the EHR patient portal from 5 Atlanta, Georgia, hospitals between March 30 and September 1, 2022. Assessment of the model's accuracy consisted of manual review of message contents to confirm the classification label by a team of physicians, nurses, and medical students, followed by retrospective propensity score-matched clinical outcomes analysis. Exposure: Prescription of antiviral treatment for COVID-19. Main Outcomes and Measures: The 2 primary outcomes were (1) physician-validated evaluation of the NLP model's message classification accuracy and (2) analysis of the model's potential clinical effect via increased patient access to treatment. The model classified messages into COVID-19-other (pertaining to COVID-19 but not reporting a positive test), COVID-19-positive (reporting a positive at-home COVID-19 test result), and non-COVID-19 (not pertaining to COVID-19). Results: Among 10 172 patients whose messages were included in analyses, the mean (SD) age was 58 (17) years; 6509 patients (64.0%) were women and 3663 (36.0%) were men. In terms of race and ethnicity, 2544 patients (25.0%) were African American or Black, 20 (0.2%) were American Indian or Alaska Native, 1508 (14.8%) were Asian, 28 (0.3%) were Native Hawaiian or other Pacific Islander, 5980 (58.8%) were White, 91 (0.9%) were more than 1 race or ethnicity, and 1 (0.01%) chose not to answer. The NLP model had high accuracy and sensitivity, with a macro F1 score of 94% and sensitivity of 85% for COVID-19-other, 96% for COVID-19-positive, and 100% for non-COVID-19 messages. Among the 3048 patient-generated messages reporting positive SARS-CoV-2 test results, 2982 (97.8%) were not documented in structured EHR data. Mean (SD) message response time for COVID-19-positive patients who received treatment (364.10 [784.47] minutes) was faster than for those who did not (490.38 [1132.14] minutes; P = .03). Likelihood of antiviral prescription was inversely correlated with message response time (odds ratio, 0.99 [95% CI, 0.98-1.00]; P = .003). Conclusions and Relevance: In this cohort study of 2982 COVID-19-positive patients, a novel NLP model classified patient-initiated EHR messages reporting positive COVID-19 test results with high sensitivity. Furthermore, when responses to patient messages occurred faster, patients were more likely to receive antiviral medical prescription within the 5-day treatment window. Although additional analysis on the effect on clinical outcomes is needed, these findings represent a possible use case for integration of NLP algorithms into clinical care.

Explainable synthetic image generation to improve risk assessment of rare pediatric heart transplant rejection.

Expert microscopic analysis of cells obtained from frequent heart biopsies is vital for early detection of pediatric heart transplant rejection to prevent heart failure. Detection of this rare condition is prone to low levels of expert agreement due to the difficulty of identifying subtle rejection signs within biopsy samples. The rarity of pediatric heart transplant rejection also means that very few gold-standard images are available for developing machine learning models. To solve this urgent clinical challenge, we developed a deep learning model to automatically quantify rejection risk within digital images of biopsied tissue using an explainable synthetic data augmentation approach. We developed this explainable AI framework to illustrate how our progressive and inspirational generative adversarial network models distinguish between normal tissue images and those containing cellular rejection signs. To quantify biopsy-level rejection risk, we first detect local rejection features using a binary image classifier trained with expert-annotated and synthetic examples. We converted these local predictions into a biopsy-wide rejection score via an interpretable histogram-based approach. Our model significantly improves upon prior works with the same dataset with an area under the receiver operating curve (AUROC) of 98.84% for the local rejection detection task and 95.56% for the biopsy-rejection prediction task. A biopsy-level sensitivity of 83.33% makes our approach suitable for early screening of biopsies to prioritize expert analysis. Our framework provides a solution to rare medical imaging challenges currently limited by small datasets.

Explainable Artificial Intelligence Methods in Combating Pandemics: A Systematic Review.

Despite the myriad peer-reviewed papers demonstrating novel Artificial Intelligence (AI)-based solutions to COVID-19 challenges during the pandemic, few have made a significant clinical impact, especially in diagnosis and disease precision staging. One major cause for such low impact is the lack of model transparency, significantly limiting the AI adoption in real clinical practice. To solve this problem, AI models need to be explained to users. Thus, we have conducted a comprehensive study of Explainable Artificial Intelligence (XAI) using PRISMA technology. Our findings suggest that XAI can improve model performance, instill trust in the users, and assist users in decision-making. In this systematic review, we introduce common XAI techniques and their utility with specific examples of their application. We discuss the evaluation of XAI results because it is an important step for maximizing the value of AI-based clinical decision support systems. Additionally, we present the traditional, modern, and advanced XAI models to demonstrate the evolution of novel techniques. Finally, we provide a best practice guideline that developers can refer to during the model experimentation. We also offer potential solutions with specific examples for common challenges in AI model experimentation. This comprehensive review, hopefully, can promote AI adoption in biomedicine and healthcare.

Accelerating Multi-site Health Informatics with Streamlined Data Infrastructure using OMOP-on-FHIR.

Shriners Children's (SHC) is a hospital system whose mission is to advance the treatment and research of pediatric diseases. SHC success has generated a wealth of clinical data. Unfortunately, barriers to healthcare data access often limit data-driven clinical research. We decreased this burden by allowing access to clinical data via the standardized data access standard called FHIR (Fast Healthcare Interoperability Resources). Specifically, we converted existing data in the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) standard into FHIR data elements using a technology called OMOP-on-FHIR. In addition, we developed two applications leveraging the FHIR data elements to facilitate patient cohort curation to advance research into pediatric musculoskeletal diseases. Our work enables clinicians and clinical researchers to use hundreds of currently available open-sourced FHIR applications. Our successful implementation of OMOP-on-FHIR within a large hospital system will accelerate advancements in pediatric disease treatment and research.

Domain Adaptation Using Convolutional Autoencoder and Gradient Boosting for Adverse Events Prediction in the Intensive Care Unit.

More than 5 million patients have admitted annually to intensive care units (ICUs) in the United States. The leading causes of mortality are cardiovascular failures, multi-organ failures, and sepsis. Data-driven techniques have been used in the analysis of patient data to predict adverse events, such as ICU mortality and ICU readmission. These models often make use of temporal or static features from a single ICU database to make predictions on subsequent adverse events. To explore the potential of domain adaptation, we propose a method of data analysis using gradient boosting and convolutional autoencoder (CAE) to predict significant adverse events in the ICU, such as ICU mortality and ICU readmission. We demonstrate our results from a retrospective data analysis using patient records from a publicly available database called Multi-parameter Intelligent Monitoring in Intensive Care-II (MIMIC-II) and a local database from Children's Healthcare of Atlanta (CHOA). We demonstrate that after adopting novel data imputation on patient ICU data, gradient boosting is effective in both the mortality prediction task and the ICU readmission prediction task. In addition, we use gradient boosting to identify top-ranking temporal and non-temporal features in both prediction tasks. We discuss the relationship between these features and the specific prediction task. Lastly, we indicate that CAE might not be effective in feature extraction on one dataset, but domain adaptation with CAE feature extraction across two datasets shows promising results.

Proposing Causal Sequence of Death by Neural Machine Translation in Public Health Informatics.

Each year there are nearly 57 million deaths worldwide, with over 2.7 million in the United States. Timely, accurate and complete death reporting is critical for public health, especially during the COVID-19 pandemic, as institutions and government agencies rely on death reports to formulate responses to communicable diseases. Unfortunately, determining the causes of death is challenging even for experienced physicians. The novel coronavirus and its variants may further complicate the task, as physicians and experts are still investigating COVID-related complications. To assist physicians in accurately reporting causes of death, an advanced Artificial Intelligence (AI) approach is presented to determine a chronically ordered sequence of conditions that lead to death (named as the causal sequence of death), based on decedent's last hospital discharge record. The key design is to learn the causal relationship among clinical codes and to identify death-related conditions. There exist three challenges: different clinical coding systems, medical domain knowledge constraint, and data interoperability. First, we apply neural machine translation models with various attention mechanisms to generate sequences of causes of death. We use the BLEU (BiLingual Evaluation Understudy) score with three accuracy metrics to evaluate the quality of generated sequences. Second, we incorporate expert-verified medical domain knowledge as constraints when generating the causal sequences of death. Lastly, we develop a Fast Healthcare Interoperability Resources (FHIR) interface that demonstrates the usability of this work in clinical practice. Our results match the state-of-art reporting and can assist physicians and experts in public health crisis such as the COVID-19 pandemic.

COVID-19 Automatic Diagnosis With Radiographic Imaging: Explainable Attention Transfer Deep Neural Networks.

Researchers seek help from deep learning methods to alleviate the enormous burden of reading radiological images by clinicians during the COVID-19 pandemic. However, clinicians are often reluctant to trust deep models due to their black-box characteristics. To automatically differentiate COVID-19 and community-acquired pneumonia from healthy lungs in radiographic imaging, we propose an explainable attention-transfer classification model based on the knowledge distillation network structure. The attention transfer direction always goes from the teacher network to the student network. Firstly, the teacher network extracts global features and concentrates on the infection regions to generate attention maps. It uses a deformable attention module to strengthen the response of infection regions and to suppress noise in irrelevant regions with an expanded reception field. Secondly, an image fusion module combines attention knowledge transferred from teacher network to student network with the essential information in original input. While the teacher network focuses on global features, the student branch focuses on irregularly shaped lesion regions to learn discriminative features. Lastly, we conduct extensive experiments on public chest X-ray and CT datasets to demonstrate the explainability of the proposed architecture in diagnosing COVID-19.

Regularization of Deep Neural Networks for EEG Seizure Detection to Mitigate Overfitting.

Seizure detection is a major goal for simplifying the workflow of clinicians working on EEG records. Current algorithms can only detect seizures effectively for patients already presented to the classifier. These algorithms are hard to generalize outside the initial training set without proper regularization and fail to capture seizures from the larger population. We proposed a data processing pipeline for seizure detection on an intra-patient dataset from the world's largest public EEG seizure corpus. We created spatially and session invariant features by forcing our networks to rely less on exact combinations of channels and signal amplitudes, but instead to learn dependencies towards seizure detection. For comparison, the baseline results without any additional regularization on a deep learning model achieved an F1 score of 0.544. By using random rearrangements of channels on each minibatch to force the network to generalize to other combinations of channels, we increased the F1 score to 0.629. By using random rescale of the data within a small range, we further increased the F1 score to 0.651 for our best model. Additionally, we applied adversarial multi-task learning and achieved similar results. We observed that session and patient specific dependencies were causing overfitting of deep neural networks, and the most overfitting models learnt features specific only to the EEG data presented. Thus, we created networks with regularization that the deep learning did not learn patient and session-specific features. We are the first to use random rearrangement, random rescale, and adversarial multitask learning to regularize intra-patient seizure detection and have increased sensitivity to 0.86 comparing to baseline study.

Advancing Medical Imaging Informatics by Deep Learning-Based Domain Adaptation.

INTRODUCTION: There has been a rapid development of deep learning (DL) models for medical imaging. However, DL requires a large labeled dataset for training the models. Getting large-scale labeled data remains a challenge, and multi-center datasets suffer from heterogeneity due to patient diversity and varying imaging protocols. Domain adaptation (DA) has been developed to transfer the knowledge from a labeled data domain to a related but unlabeled domain in either image space or feature space. DA is a type of transfer learning (TL) that can improve the performance of models when applied to multiple different datasets. OBJECTIVE: In this survey, we review the state-of-the-art DL-based DA methods for medical imaging. We aim to summarize recent advances, highlighting the motivation, challenges, and opportunities, and to discuss promising directions for future work in DA for medical imaging. METHODS: We surveyed peer-reviewed publications from leading biomedical journals and conferences between 2017-2020, that reported the use of DA in medical imaging applications, grouping them by methodology, image modality, and learning scenarios. RESULTS: We mainly focused on pathology and radiology as application areas. Among various DA approaches, we discussed domain transformation (DT) and latent feature-space transformation (LFST). We highlighted the role of unsupervised DA in image segmentation and described opportunities for future development. CONCLUSION: DA has emerged as a promising solution to deal with the lack of annotated training data. Using adversarial techniques, unsupervised DA has achieved good performance, especially for segmentation tasks. Opportunities include domain transferability, multi-modal DA, and applications that benefit from synthetic data.

Response and remediation actions following the detection of Naegleria fowleri in two treated drinking water distribution systems, Louisiana, 2013-2014.

Naegleria fowleri causes the usually fatal disease primary amebic meningoencephalitis (PAM), typically in people who have been swimming in warm, untreated freshwater. Recently, some cases in the United States were associated with exposure to treated drinking water. In 2013, a case of PAM was reported for the first time in association with the exposure to water from a US treated drinking water system colonized with culturable N. fowleri. This system and another were found to have multiple areas with undetectable disinfectant residual levels. In response, the water distribution systems were temporarily converted from chloramine disinfection to chlorine to inactivate N. fowleri and reduced biofilm in the distribution systems. Once >1.0 mg/L free chlorine residual was attained in all systems for 60 days, water testing was performed; N. fowleri was not detected in water samples after the chlorine conversion. This investigation highlights the importance of maintaining adequate residual disinfectant levels in drinking water distribution systems. Water distribution system managers should be knowledgeable about the ecology of their systems, understand potential water quality changes when water temperatures increase, and work to eliminate areas in which biofilm growth may be problematic and affect water quality.

Improved Prediction on Heart Transplant Rejection Using Convolutional Autoencoder and Multiple Instance Learning on Whole-Slide Imaging.

Heart transplant rejection is one major threat for the survival of patients with a heart transplant. Endomyocardial biopsies are effective in showing signs of heart transplant rejection even before patients have any symptoms. Manually examining the tissue samples is costly, time-consuming and error-prone. With recent advances in deep learning (DL) based image processing methods, automatic training and prediction on heart transplant rejection using whole-slide images expect to be promising. This paper develops an advanced pipeline for quality control, feature extraction, clustering and classification. We first implement a stacked convolutional autoencoder to extract feature maps for each tile; we then incorporate multiple instance learning (MIL) with dimensionality reduction and unsupervised clustering prior to classification. Our results show that utilizing unsupervised clustering after feature extraction can achieve higher classification results while preserving the capability for multi-class classification.

Feature Exploration and Causal Inference on Mortality of Epilepsy Patients Using Insurance Claims Data.

Approximately 0.5-1% of the global population is afflicted with epilepsy, a neurological disorder characterized by repeated seizures. Sudden Unexpected Death in Epilepsy (SUDEP) is a poorly understood complication that claims the lives of nearly 1-in-1000 epilepsy patients every year. This paper aims to explore diagnosis codes, demographic and payment features on mortality of epilepsy patients. We design a mortality prediction model with diagnosis codes and non-diagnosis features extracted from US commercial insurance claims data. We present classification accuracy of 0.91 and 0.85 by using different feature vectors. After analyzing the aforementioned features in prediction model, we extend the work to causal inference between modified diagnosis codes and selected non-diagnosis features. The uplift test of causal inference using three algorithms indicates that a patient is more likely to survive if upgrading from a low-coverage healthcare plan into a high-coverage plan.

Assessing spatial variability of soil petroleum contamination using visible near-infrared diffuse reflectance spectroscopy.

Visible near-infrared (VisNIR) diffuse reflectance spectroscopy (DRS) is a rapid, non-destructive method for sensing the presence and amount of total petroleum hydrocarbon (TPH) contamination in soil. This study demonstrates the feasibility of VisNIR DRS to be used in the field to proximally sense and then map the areal extent of TPH contamination in soil. More specifically, we evaluated whether a combination of two methods, penalized spline regression and geostatistics could provide an efficient approach to assess spatial variability of soil TPH using VisNIR DRS data from soils collected from an 80 ha crude oil spill in central Louisiana, USA. Initially, a penalized spline model was calibrated to predict TPH contamination in soil by combining lab TPH values of 46 contaminated and uncontaminated soil samples and the first-derivative of VisNIR reflectance spectra of these samples. The r(2), RMSE, and bias of the calibrated penalized spline model were 0.81, 0.289 log(10) mg kg(-1), and 0.010 log(10) mg kg(-1), respectively. Subsequently, the penalized spline model was used to predict soil TPH content for 128 soil samples collected over the 80 ha study site. When assessed with a randomly chosen validation subset (n = 10) from the 128 samples, the penalized spline model performed satisfactorily (r(2) = 0.70; residual prediction deviation = 2.0). The same validation subset was used to assess point kriging interpolation after the remaining 118 predictions were used to produce an experimental semivariogram and map. The experimental semivariogram was fitted with an exponential model which revealed strong spatial dependence among soil TPH [r(2) = 0.76, nugget = 0.001 (log(10) mg kg(-1))(2), and sill 1.044 (log(10) mg kg(-1))(2)]. Kriging interpolation adequately interpolated TPH with r(2) and RMSE values of 0.88 and 0.312 log(10) mg kg(-1), respectively. Furthermore, in the kriged map, TPH distribution matched with the expected TPH variability of the study site. Since the combined use of VisNIR prediction and geostatistics was promising to identify the spatial patterns of TPH contamination in soils, future research is warranted to evaluate the approach for mapping spatial variability of petroleum contaminated soils.

Use of portable X-ray fluorescence spectrometry for environmental quality assessment of peri-urban agriculture.

Urban expansion into traditional agricultural lands has augmented the potential for heavy metal contamination of soils. This study examined the utility of field portable X-ray fluorescence (PXRF) spectrometry for evaluating the environmental quality of sugarcane fields near two industrial complexes in Louisiana, USA. Results indicated that PXRF provided quality results of heavy metal levels comparable to traditional laboratory analysis. When coupled with global positioning system technology, the use of PXRF allows for on-site interpolation of heavy metal levels in a matter of minutes. Field portable XRF was shown to be an effective tool for rapid assessment of heavy metals in soils of peri-urban agricultural areas.