PNPLA3 Genotype and Diabetes Identify Patients With Nonalcoholic Fatty Liver Disease at High Risk of Incident Cirrhosis

BACKGROUND & AIMS

Non-alcoholic fatty liver disease (NAFLD) can progress to cirrhosis and hepatic decompensation, but whether genetic variants influence the rate of progression to cirrhosis or are useful in risk stratification among patients with NAFLD is uncertain.

METHODS

We included participants from 2 independent cohorts, they Michigan Genomics Initiative (MGI) and UK Biobank (UKBB), who had NAFLD defined by elevated alanine aminotransferase (ALT) levels in the absence of alternative chronic liver disease. The primary predictors were genetic variants and metabolic comorbidities associated with cirrhosis. We conducted time-to-event analyses using Fine-Gray competing risk models.

RESULTS

We included 7893 and 46,880 participants from MGI and UKBB, respectively. In univariable analysis, PNPLA3-rs738409-GG genotype, diabetes, obesity, and ALT of ≥2× upper limit of normal were associated with higher incidence rate of cirrhosis in both MGI and UKBB. PNPLA3-rs738409-GG had additive effects with clinical risk factors including diabetes, obesity, and ALT elevations. Among patients with indeterminate fibrosis-4 (FIB4) scores (1.3-2.67), those with diabetes and PNPLA3-rs738409-GG genotype had an incidence rate of cirrhosis comparable to that of patients with high-risk FIB4 scores (>2.67) and 2.9-4.8 times that of patients with diabetes but CC/CG genotypes. In contrast, FIB4 <1.3 was associated with an incidence rate of cirrhosis significantly lower than that of FIB4 of >2.67, even in the presence of clinical risk factors and high-risk PNPLA3 genotype.

CONCLUSIONS

PNPLA3-rs738409 genotype and diabetes identified patients with NAFLD currently considered indeterminate risk (FIB4 1.3-2.67) who had a similar risk of cirrhosis as those considered high-risk (FIB4 >2.67). PNPLA3 genotyping may improve prognostication and allow for prioritization of intensive intervention.

Validation of the Hepatocellular Carcinoma Early Detection Screening (HES) Algorithm in a Cohort of Veterans With Cirrhosis

BACKGROUND & AIMS

Early detection of hepatocellular carcinoma (HCC) through surveillance reduces mortality associated with this cancer. Guidelines recommend HCC surveillance every 6 months for patients with cirrhosis, via ultrasonography, with or without measurement of serum level of alpha fetoprotein (AFP).

METHODS

We previously developed and internally validated an HCC early detection screening (HES) algorithm that included patient's current level of AFP, rate of AFP change, age, level of alanine aminotransferase, and platelet count in a department of Veterans affairs (VA) cohort with active hepatitis C virus-related cirrhosis. HES score was associated with 3.84% absolute improvement in sensitivity of detection of HCC compared with AFP alone, at 90% specificity, within 6 months prior to diagnosis of this cancer. We externally validated the HES algorithm in a cohort of 38,431 patients with cirrhosis of any etiology evaluated at a VA medical center from 2010 through 2015.

RESULTS

A total of 4804 cases of HCC developed during a median follow-up time of 3.12 years. At 90% specificity, the HES algorithm identified patients with HCC with 52.56% sensitivity, compared to 48.13% sensitivity for the AFP assay alone, within 6 months prior to diagnosis; this was an absolute improvement of 4.43% (P < .0005). In HCC screening, a positive result leads to follow-up evaluation by computed tomography or magnetic resonance imaging. We estimated that the number of HCC cases detected per 1000 imaging analyses was 198.57 for the HES algorithm vs 185.52 for the AFP assay alone, or detection of 13 additional cases of HCC (P < .0005).

CONCLUSION

We validated the HES algorithm in detection of HCC in patients with cirrhosis of any etiology evaluated at VA medical centers. The algorithm offers a modest but useful advantage over AFP alone in HCC surveillance.

Predictive Modeling for Metabolomics Data

In recent years, mass spectrometry (MS)-based metabolomics has been extensively applied to characterize biochemical mechanisms, and study physiological processes and phenotypic changes associated with disease. Metabolomics has also been important for identifying biomarkers of interest suitable for clinical diagnosis. For the purpose of predictive modeling, in this chapter, we will review various supervised learning algorithms such as random forest (RF), support vector machine (SVM), and partial least squares-discriminant analysis (PLS-DA). In addition, we will also review feature selection methods for identifying the best combination of metabolites for an accurate predictive model. We conclude with best practices for reproducibility by including internal and external replication, reporting metrics to assess performance, and providing guidelines to avoid overfitting and to deal with imbalanced classes. An analysis of an example data will illustrate the use of different machine learning methods and performance metrics.

Trajectories of the heart rate characteristics index, a physiomarker of sepsis in premature infants, predict Neonatal ICU mortality

OBJECTIVE

Trajectories of physiomarkers over time can be useful to define phenotypes of disease progression and as predictors of clinical outcomes. The aim of this study was to identify phenotypes of the time course of late-onset sepsis in premature infants in Neonatal Intensive Care Units.

METHODS

We examined the trajectories of a validated continuous physiomarker, abnormal heart rate characteristics, using functional data analysis and clustering techniques.

PARTICIPANTS

We analyzed continuous heart rate characteristics data from 2989 very low birth weight infants (<1500 grams) from nine NICUs from 2004-2010.

RESULT

Despite the relative homogeneity of the patients, we found extreme variability in the physiomarker trajectories. We identified phenotypes that were indicative of seven and 30 day mortality beyond that predicted by individual heart rate characteristics values or baseline demographic information.

CONCLUSION

Time courses of a heart rate characteristics physiomarker reveal snapshots of illness patterns, some of which were more deadly than others.

Detecting model misconducts in decentralized healthcare federated learning

BACKGROUND

To accelerate healthcare/genomic medicine research and facilitate quality improvement, researchers have started cross-institutional collaborations to use artificial intelligence on clinical/genomic data. However, there are real-world risks of incorrect models being submitted to the learning process, due to either unforeseen accidents or malicious intent. This may reduce the incentives for institutions to participate in the federated modeling consortium. Existing methods to deal with this "model misconduct" issue mainly focus on modifying the learning methods, and therefore are more specifically tied with the algorithm.

BASIC PROCEDURES

In this paper, we aim at solving the problem in an algorithm-agnostic way by (1) designing a simulator to generate various types of model misconduct, (2) developing a framework to detect the model misconducts, and (3) providing a generalizable approach to identify model misconducts for federated learning. We considered the following three categories: Plagiarism, Fabrication, and Falsification, and then developed a detection framework with three components: Auditing, Coefficient, and Performance detectors, with greedy parameter tuning.

MAIN FINDINGS

We generated 10 types of misconducts from models learned on three datasets to evaluate our detection method. Our experiments showed high recall with low added computational cost. Our proposed detection method can best identify the misconduct on specific sites from any learning iteration, whereas it is more challenging to precisely detect misconducts for a specific site and at a specific iteration.

PRINCIPAL CONCLUSIONS

We anticipate our study can support the enhancement of the integrity and reliability of federated machine learning on genomic/healthcare data.

Exploring supervised machine learning approaches to predicting Veterans Health Administration chiropractic service utilization

BACKGROUND

Chronic spinal pain conditions affect millions of US adults and carry a high healthcare cost burden, both direct and indirect. Conservative interventions for spinal pain conditions, including chiropractic care, have been associated with lower healthcare costs and improvements in pain status in different clinical populations, including veterans. Little is currently known about predicting healthcare service utilization in the domain of conservative interventions for spinal pain conditions, including the frequency of use of chiropractic services. The purpose of this retrospective cohort study was to explore the use of supervised machine learning approaches to predicting one-year chiropractic service utilization by veterans receiving VA chiropractic care.

METHODS

We included 19,946 veterans who entered the Musculoskeletal Diagnosis Cohort between October 1, 2003 and September 30, 2013 and utilized VA chiropractic services within one year of cohort entry. The primary outcome was one-year chiropractic service utilization following index chiropractic visit, split into quartiles represented by the following classes: 1 visit, 2 to 3 visits, 4 to 6 visits, and 7 or greater visits. We compared the performance of four multiclass classification algorithms (gradient boosted classifier, stochastic gradient descent classifier, support vector classifier, and artificial neural network) in predicting visit quartile using 158 sociodemographic and clinical features.

RESULTS

The selected algorithms demonstrated poor prediction capabilities. Subset accuracy was 42.1% for the gradient boosted classifier, 38.6% for the stochastic gradient descent classifier, 41.4% for the support vector classifier, and 40.3% for the artificial neural network. The micro-averaged area under the precision-recall curve for each one-versus-rest classifier was 0.43 for the gradient boosted classifier, 0.38 for the stochastic gradient descent classifier, 0.43 for the support vector classifier, and 0.42 for the artificial neural network. Performance of each model yielded only a small positive shift in prediction probability (approximately 15%) compared to naïve classification.

CONCLUSIONS

Using supervised machine learning to predict chiropractic service utilization remains challenging, with only a small shift in predictive probability over naïve classification and limited clinical utility. Future work should examine mechanisms to improve model performance.

Prediction of Acute Kidney Injury in the Intensive Care Unit: Preliminary Findings in a European Open Access Database

Acute kidney injury (AKI) is a common complication in critically ill patients and is associated with long-term complications and an increased mortality. This work presents preliminary findings from the first freely available European intensive care database released by Amsterdam UMC. A machine learning (ML) model was developed to predict AKI in the intensive care unit 12 hours before the actual event. Main features of the model included medications and hemodynamic parameters. Our models perform with an accuracy of 81.8% on moderate to severe AKI and 79.8% on all AKI patients. Those results can compete with models reported in the literature and introduce an ML model for AKI based on European patient data.

Assessing artificial intelligence ability in predicting hospitalization duration for pleural empyema patients managed with uniportal video-assisted thoracoscopic surgery: a retrospective observational study

BACKGROUND

This retrospective observational research evaluates the potential applicability of artificial intelligence models to predict the length of hospital stay for patients with pleural empyema who underwent uniportal video-assisted thoracoscopic surgery.

METHODS

Data from 56 patients were analyzed using two artificial intelligence models. A Random Forest Regressor, the initial model, was trained using clinical data unique to each patient. Weighted factors from evidence-based research were incorporated into the second model, which was created using a prediction approach informed by the literature.

RESULTS

The two models tested showed poor prediction accuracy. The first one had a mean absolute error of 4.56 days and a negative R2 value. The literature-informed model performed similarly, with a mean absolute error of 4.53 days and an R2 below zero.

CONCLUSIONS

While artificial intelligence holds promise in supporting clinical decision-making, this study demonstrates the challenges of predicting length of stay in pleural empyema patients due to significant clinical variability and the current limitations of AI-based models. Future research should focus on integrating larger, multi-center datasets and more advanced machine learning approaches to enhance predictive accuracy.

Time in targeted blood glucose range as an independent predictor of 28-Day mortality in ICU Patients: A retrospective study

OBJECTIVE

This study aimed to evaluate the relationship between time in targeted blood glucose range (TIR) and 28-day mortality in critically ill patients.

METHODS

A retrospective cohort analysis was conducted using data from the MIMIC-IV database. Patients (n = 18,905) were stratified into four quartiles based on TIR values. The association between TIR and mortality was assessed using multivariable logistic regression models with adjustments for potential confounders.

RESULTS

In the fully adjusted model, each percentage point increase in TIR was associated with a 1 % reduction in 28-day mortality risk (OR = 0.99, 95 % CI: 0.98-0.99, P < 0.001). Patients in the highest TIR quartile showed a 60 % lower mortality risk compared to those in the lowest quartile (OR = 0.40, 95 % CI: 0.22-0.74, P = 0.003). The predictive model demonstrated good discriminative ability (AUC = 0.7543).

CONCLUSION

Time in targeted blood glucose range is independently associated with 28-day mortality in ICU patients, suggesting its potential value as a metric for risk stratification and glycemic management optimization.

Computational Approaches for Predicting Drug Interactions with Human Organic Anion Transporter 4 (OAT4)

Human Organic Anion Transporter 4 (OAT4) is predominantly expressed in the kidneys, particularly in the apical membrane of the proximal tubule cells. This transporter is involved in the renal handling of endogenous and exogenous organic anions (OAs), making it an important transporter for drug-drug interactions (DDIs). To better understand OAT4-compound interactions, we generated single concentration (25 μM) in vitro inhibition data for over 1400 small molecules against the uptake of the fluorescent OA 6-carboxyfluorescein (6-CF) in Chinese hamster ovary (CHO) cells. Several drugs exhibiting higher than 50% inhibition in this initial screen were selected to determine IC50 values against three structurally distinct OAT4 substrates: estrone sulfate (ES), ochratoxin A (OTA), and 6-CF. These IC50 values were then compared to the drug plasma concentration as per the 2020 FDA drug-drug interaction (DDI) guidance. Several screened compounds, including some not previously reported, emerged as novel inhibitors of OAT4. These data were also used to build machine learning classification models to predict the activity of potential OAT4 inhibitors. We compared multiple machine learning algorithms and data cleaning techniques to model these screening data and investigated the utility of conformal predictors to predict OAT4 inhibition of a leave-out set. These experimental and computational approaches allowed us to model diverse and unbalanced data to enable predictions for DDIs mediated by this transporter.

Predicting HIV Diagnosis Among Emerging Adults Using Electronic Health Records and Health Survey Data in All of Us Research Program

The global decline in HIV incidence has not been mirrored in the United States, where young adults (ages 18-29) continue to account for a significant portion of new infections. In this study, we leverage the All of Us (AoU) Research Program's extensive electronic health records (EHRs) and health survey data to develop machine learning models capable of predicting HIV diagnoses at least three months before clinical identification. Among various models tested, the Support Vector Machine (SVM) model demonstrated a balanced performance, integrating clinically relevant features with robust predictive accuracy (AUC = 0.91). Risky drinking behaviors emerged as consistent top predictors across models, highlighting the importance of targeted interventions in this age group. Our findings underscore the potential of predictive analytics in enhancing HIV prevention strategies and informing public health efforts aimed at reducing HIV transmission among emerging adults.

Overcoming gaps: regional collaborative to optimize capacity management and predict length of stay of patients admitted with COVID-19

Objective

Ensuring an efficient response to COVID-19 requires a degree of inter-system coordination and capacity management coupled with an accurate assessment of hospital utilization including length of stay (LOS). We aimed to establish optimal practices in inter-system data sharing and LOS modeling to support patient care and regional hospital operations.

Materials and Methods

We completed a retrospective observational study of patients admitted with COVID-19 followed by 12-week prospective validation, involving 36 hospitals covering the upper Midwest. We developed a method for sharing de-identified patient data across systems for analysis. From this, we compared 3 approaches, generalized linear model (GLM) and random forest (RF), and aggregated system level averages to identify features associated with LOS. We compared model performance by area under the ROC curve (AUROC).

Results

A total of 2068 patients were included and used for model derivation and 597 patients for validation. LOS overall had a median of 5.0 days and mean of 8.2 days. Consistent predictors of LOS included age, critical illness, oxygen requirement, weight loss, and nursing home admission. In the validation cohort, the RF model (AUROC 0.890) and GLM model (AUROC 0.864) achieved good to excellent prediction of LOS, but only marginally better than system averages in practice.

Conclusion

Regional sharing of patient data allowed for effective prediction of LOS across systems; however, this only provided marginal improvement over hospital averages at the aggregate level. A federated approach of sharing aggregated system capacity and average LOS will likely allow for effective capacity management at the regional level.

Using artificial intelligence and predictive modelling to enable learning healthcare systems (LHS) for pandemic preparedness

In anticipation of potential future pandemics, we examined the challenges and opportunities presented by the COVID-19 outbreak. This analysis highlights how artificial intelligence (AI) and predictive models can support both patients and clinicians in managing subsequent infectious diseases, and how legislators and policymakers could support these efforts, to bring learning healthcare system (LHS) from guidelines to real-world implementation. This report chronicles the trajectory of the COVID-19 pandemic, emphasizing the diverse data sets generated throughout its course. We propose strategies for harnessing this data via AI and predictive modelling to enhance the functioning of LHS. The challenges faced by patients and healthcare systems around the world during this unprecedented crisis could have been mitigated with an informed and timely adoption of the three pillars of the LHS: Knowledge, Data and Practice. By harnessing AI and predictive analytics, we can develop tools that not only detect potential pandemic-prone diseases early on but also assist in patient management, provide decision support, offer treatment recommendations, deliver patient outcome triage, predict post-recovery long-term disease impacts, monitor viral mutations and variant emergence, and assess vaccine and treatment efficacy in real-time. A patient-centric approach remains paramount, ensuring patients are both informed and actively involved in disease mitigation strategies.

Spatial and temporal changes of air quality in Shandong Province from 2016 to 2022 and model prediction

This study investigates the spatial and temporal dynamics of air quality in Shandong Province from 2016 to 2022. The Air Quality Index (AQI) showed a seasonal pattern, with higher values in winter due to temperature inversions and heating emissions, and lower values in summer aided by favorable dispersion conditions. The AQI improved significantly, decreasing by approximately 39.4 % from 6.44 to 3.90. Coastal cities exhibited better air quality than inland areas, influenced by industrial activities and geographical features. For instance, Zibo's geography restricts pollutant dispersion, resulting in poor air quality. CO levels remained stable, while O3 increased seasonally due to photochemical reactions in summer, with correlation coefficients indicating a strong positive correlation with temperature (r = 0.65). Winter saw elevated NO2 levels linked to heating and vehicular emissions, with an observed increase in correlation with AQI (r = 0.78). PM2.5 and PM10 concentrations were higher in colder months due to heating and atmospheric dust, showing a significant decrease of 45 % and 40 %, respectively, over the study period. Predictive modeling forecasts continued air quality improvements, contingent on sustained policy enforcement and technological advancements. This approach provides a comprehensive framework for future air quality management and improvement.

Precision Opioid Prescription in ICU Surgery: Insights from an Interpretable Deep Learning Framework

Purpose

Appropriate opioid management is crucial to reduce opioid overdose risk for ICU surgical patients, which can lead to severe complications. Accurately predicting postoperative opioid needs and understanding the associated factors can effectively guide appropriate opioid use, significantly enhancing patient safety and recovery outcomes. Although machine learning models can accurately predict postoperative opioid needs, lacking interpretability hinders their adoption in clinical practice.

Methods

We developed an interpretable deep learning framework to evaluate individual feature's impact on postoperative opioid use and identify important factors. A Permutation Feature Importance Test (PermFIT) was employed to assess the impact with a rigorous statistical inference for machine learning models including Support Vector Machines, eXtreme Gradient Boosting, Random Forest, and Deep Neural Networks (DNN). The Mean Squared Error (MSE) and Pearson Correlation Coefficient (PCC) were used to evaluate the performance of these models.

Results

We conducted analysis utilizing the electronic health records of 4,912 surgical patients from the Medical Information Mart for Intensive Care database. In a 10-fold cross-validation, the DNN outperformed other machine learning models, achieving the lowest MSE (7889.2 mcg) and highest PCC (0.283). Among 25 features, 13-including age, surgery type, and others-were identified as significant predictors of postoperative opioid use (p < 0.05).

Conclusion

The DNN proved to be an effective model for predicting postoperative opioid consumption and identifying significant features through the PermFIT framework. This approach offers a valuable tool for precise opioid prescription tailored to the individual needs of ICU surgical patients, improving patient outcomes and enhancing safety.

An Ensemble Deep Learning Algorithm for Structural Heart Disease Screening Using Electrocardiographic Images: PRESENT SHD

Background

Identifying structural heart diseases (SHDs) early can change the course of the disease, but their diagnosis requires cardiac imaging, which is limited in accessibility.

Objective

To leverage images of 12-lead ECGs for automated detection and prediction of multiple SHDs using an ensemble deep learning approach.

Methods

We developed a series of convolutional neural network models for detecting a range of individual SHDs from images of ECGs with SHDs defined by transthoracic echocardiograms (TTEs) performed within 30 days of the ECG at the Yale New Haven Hospital (YNHH). SHDs were defined as LV ejection fraction <40%, moderate-to-severe left-sided valvular disease (aortic/mitral stenosis or regurgitation), or severe left ventricular hypertrophy (IVSd > 1.5cm and diastolic dysfunction). We developed an ensemble XGBoost model, PRESENT-SHD, as a composite screen across all SHDs. We validated PRESENT-SHD at 4 US hospitals and the prospective, population-based Brazilian Longitudinal Study of Adult Health (ELSA-Brasil), with concurrent protocolized ECGs and TTEs. We also used PRESENT-SHD for risk stratification of new-onset SHD or heart failure (HF) in clinical cohorts and the population-based UK Biobank (UKB).

Results

The models were developed using 261,228 ECGs from 93,693 YNHH patients and evaluated on a single ECG from 11,023 individuals at YNHH (19% with SHD), 44,591 across external hospitals (20-27% with SHD), and 3,014 in the ELSA-Brasil (3% with SHD). In the held-out test set, PRESENT-SHD demonstrated an AUROC of 0.886 (0.877-894), 90% sensitivity, and 66% specificity. At hospital-based sites, PRESENT-SHD had AUROCs ranging from 0.854-0.900, with sensitivities and specificities of 93-96% and 51-56%, respectively. The model generalized well to ELSA-Brasil (AUROC, 0.853 [0.811-0.897], 88% sensitivity, 62% specificity). PRESENT-SHD demonstrated consistent performance across demographic subgroups, novel ECG formats, and smartphone photographs of ECGs from monitors and printouts. A positive PRESENT-SHD screen portended a 2- to 4-fold higher risk of new-onset SHD/HF, independent of demographics, comorbidities, and the competing risk of death across clinical sites and UKB, with high predictive discrimination.

Conclusion

We developed and validated PRESENT-SHD, an AI-ECG tool identifying a range of SHD using images of 12-lead ECGs, representing a robust, scalable, and accessible modality for automated SHD screening and risk stratification.

Psychosocial problems and high amount of opioid administration are associated with opioid dependence and abuse after first exposure for chronic pain patients

Controversy surrounding the use of opioids for the treatment and the unique characteristics of chronic pain heighten the risks for abuse and dependence; however, it's unclear if higher doses of opioids and first exposure are associated with dependence and abuse. This study aimed to identify patients who developed dependence or opioid abuse after exposed to opioids for the first time and what were the risks factors associated with the outcome. A retrospective observational cohort study analyzed 2,411 patients between 2011 and 2017 who had a diagnosis of chronic pain and received opioids for the first time. A logistic regression model was used to estimate the likelihood of opioid dependence/abuse after the first exposure based on their mental health conditions, prior substance abuse disorders, demographics, and the amount of MME per day patients received. From 2,411 patients, 5.5 % of the patients had a diagnosis of dependence or abuse after the first exposure. Patients who were depressed (OR = 2.09), previous non-opioid substance dependence or abuse (OR = 1.59) or received greater than 50 MME per day (OR = 1.03) showed statistically significant relationship with developing opioid dependence or abuse, while age (OR = -1.03) showed to be a protective factor. Further studies should stratify chronic pain patients into groups who is in higher risk in developing opioid dependence or abuse and develop alternative strategies for pain management and treatments beyond opioids. This study reinforces the psychosocial problems as determinants of opioid dependence or abuse and risk factors, and the need for safer opioid prescribing practices.

An exploration of ontology-based EMR data abstraction for diabetic kidney disease prediction

Diabetic Kidney Disease (DKD) is a critical and morbid complication of diabetes and the leading cause of chronic kidney disease in the developed world. Electronic medical records (EMRs) hold promise for supporting clinical decision-making with its nationwide adoption as well as rich information characterizing patients' health care experience. However, few retrospective studies have fully utilized the EMR data to model DKD risk. This study examines the effectiveness of an unbiased data driven approach in identifying potential DKD patients in 6 months prior to onset by utilizing EMR on a broader spectrum. Meanwhile, we evaluate how different levels of data granularity of Medications and Diagnoses observations would affect prediction performance and knowledge discovery. The experimental results suggest that different data granularity may not necessarily influence the prediction accuracy, but it would dramatically change the internal structure of the predictive models.

Exploring the power of MRI and clinical measures in predicting AD neuropathology

Predicting Alzheimer's disease (AD) pathology prior to clinical diagnosis is important for identifying individuals at high risk of developing AD dementia. However, there remains a gap in leveraging MRI and clinical data to predict AD pathology. This study examines a novel machine learning approach that integrates the combined vascular (white matter hyperintensities, WMHs) and structural brain changes (gray matter, GM) with clinical factors (cognitive scores) to predict post-mortem neuropathology. Participants from the Alzheimer's Disease Neuroimaging Initiative dataset (ADNI) and National Alzheimer's Coordinating Center (NACC) with both post-mortem neuropathology data and antemortem MRI and clinical data were included. Machine learning models were applied towards feature selection of the top seven MRI, clinical, and demographic data to identify the best performing set of variables that could predict postmortem neuropathology outcomes (i.e., neurofibrillary tangles, neuritic plaques, diffuse plaques, senile/amyloid plaques, and amyloid angiopathy). The best-performing neuropathology predictors from ADNI were then validated in NACC to compare results and ensure that the feature selection process did not lead to overfitting. In ADNI, the best-performing model included total and temporal lobe WMHs and achieved r = 0.87(RMSE = 0.62) during cross-validation for neuritic plaques. Overall, post-mortem neuropathology outcomes were predicted up to 14 years before death with high accuracies (~ 90%). Similar results were observed in the NACC dataset. These findings highlight that MRI features are critical to successfully predict AD-related pathology years in advance.

Prediction of gastrointestinal bleeding hospitalization risk in hemodialysis using machine learning

BACKGROUND

Gastrointestinal bleeding (GIB) is a clinical challenge in kidney failure. INSPIRE group assessed if machine learning could determine a hemodialysis (HD) patient's 180-day GIB hospitalization risk.

METHODS

An eXtreme Gradient Boosting (XGBoost) and logistic regression model were developed using an HD dataset in United States (2017-2020). Patient data was randomly split (50% training, 30% validation, and 20% testing). HD treatments ≤ 180 days before GIB hospitalization were classified as positive observations; others were negative. Models considered 1,303 exposures/covariates. Performance was measured using unseen testing data.

RESULTS

Incidence of 180-day GIB hospitalization was 1.18% in HD population (n = 451,579), and 1.12% in testing dataset (n = 38,853). XGBoost showed area under the receiver operating curve (AUROC) = 0.74 (95% confidence interval (CI) 0.72, 0.76) versus logistic regression showed AUROC = 0.68 (95% CI 0.66, 0.71). Sensitivity and specificity were 65.3% (60.9, 69.7) and 68.0% (67.6, 68.5) for XGBoost versus 68.9% (64.7, 73.0) and 57.0% (56.5, 57.5) for logistic regression, respectively. Associations in exposures were consistent for many factors. Both models showed GIB hospitalization risk was associated with older age, disturbances in anemia/iron indices, recent all-cause hospitalizations, and bone mineral metabolism markers. XGBoost showed high importance on outcome prediction for serum 25 hydroxy (25OH) vitamin D levels, while logistic regression showed high importance for parathyroid hormone (PTH) levels.

CONCLUSIONS

Machine learning can be considered for early detection of GIB event risk in HD. XGBoost outperforms logistic regression, yet both appear suitable. External and prospective validation of these models is needed. Association between bone mineral metabolism markers and GIB events was unexpected and warrants investigation.

TRIAL REGISTRATION

This retrospective analysis of real-world data was not a prospective clinical trial and registration is not applicable.

Pre-Operative Anemia is an Unsuspecting Driver of Machine Learning Prediction of Adverse Outcomes after Lumbar Spinal Fusion

BACKGROUND CONTEXT

Pre-operative risk assessment remains a challenge in spinal fusion operations. Predictive modeling provides data-driven estimates of post-surgical outcomes, guiding clinical decisions and improving patient care. Moreover, automated machine learning models are both effective and user-friendly, allowing healthcare professionals with minimal technical expertise to identify high-risk patients who may need additional pre-operative support.

PURPOSE

This study investigated the use of automated machine learning models to predict discharge disposition, length of hospital stay, and readmission post-surgery by analyzing pre-operative patient electronic medical record data and identifying key factors influencing adverse outcomes.

STUDY DESIGN/SETTING

Retrospective cohort study.

PATIENT SAMPLE

The sample includes electronic medical records of 3,006 unique surgical events from 2,855 patients who underwent lumbar spinal fusion surgeries at a single institution.

OUTCOME MEASURES

The adverse outcomes assessed were discharge disposition (non-home facility), length of hospital stay (extended stay), and readmission within 90 days post-surgery.

METHODS

We employed several inferential and predictive approaches, including the automated machine learning tool TPOT2 (Tree-based Pipeline Optimization Tool-2). TPOT2, which uses genetic programming to select optimal machine learning pipelines in a process inspired by molecular evolution, constructed, optimized and identified robust predictive models for all outcomes. Feature importance values were derived to identify major pre-operative predictive features driving optimal models.

RESULTS

Adverse outcome rates were 25.9% for discharge to non-home facilities, 23.9% for extended hospital stay, and 24.7% for readmission within 90 days post-surgery. TPOT2 delivered the best-performing predictive models, achieving balanced accuracies ((Sensitivity [true positive rate] + Specificity [true negative rate)]) / 2) of 0.72 for discharge disposition, 0.72 for length of stay, and 0.67 for readmission. Notably, preoperative hemoglobin emerged as a consistently strong predictor in best-performing models across outcomes. Patients with severe anemia (hemoglobin <80g/dL) demonstrated higher associations with all adverse outcomes and common comorbidities associated with frailty (e.g., hypertension, type II diabetes, and chronic pain). Additional patient variables and comorbidities, including body mass index, age, and mental health status, influencing post-surgical outcomes were also highly predictive.

CONCLUSIONS

This study demonstrates the effectiveness of automated machine learning in predicting post-surgical adverse outcomes and identifying key pre-operative predictors associated with such outcomes. While factors like age, BMI, insurance type, and specific comorbidities showed notable effects on outcomes, preoperative hemoglobin consistently emerged as a significant predictor across outcomes, suggesting its critical role in pre-surgical assessment. These findings underscore the potential of enhancing patient care and preoperative assessment through advanced predictive modeling.

Radiomic-based Textural Analysis of Intraluminal Thrombus in Aortic Abdominal Aneurysms: A Demonstration of Automated Workflow

Our main objective is to investigate how the structural information of intraluminal thrombus (ILT) can be used to predict abdominal aortic aneurysms (AAA) growth status through an automated workflow. Fifty-four human subjects with ILT in their AAAs were identified from our database; those AAAs were categorized as slowly- (< 5 mm/year) or fast-growing (≥ 5 mm/year) AAAs. In-house deep-learning image segmentation models were used to generate 3D geometrical AAA models, followed by automated analysis. All features were fed into a support vector machine classifier to predict AAA's growth status.The most accurate prediction model was achieved through four geometrical parameters measuring the extent of ILT, two parameters quantifying the constitution of ILT, antihypertensive medication, and the presence of co-existing coronary artery disease. The predictive model achieved an AUROC of 0.89 and a total accuracy of 83%. When ILT was not considered, our prediction's AUROC decreased to 0.75 (P-value < 0.001).

A STRUCTURAL EQUATION MODEL PREDICTS CHRONIC WOUND HEALING TIME USING PATIENT CHARACTERISTICS AND WOUND MICROBIOME COMPOSITION

Wound etiology, host characteristics, and the wound microbiome contribute to chronic wound development. Yet, there is little accounting for the relative importance of these factors to predict wound healing. Here, a structural equation model was developed to provide such an explanatory and predictive framework. Chronic wounds from 565 patients treated at a clinic practicing biofilm-based wound care were included. Patient information included DNA sequencing-based wound microbiome clinical reports corresponding to initial clinical visit. Wound microbiome data was integrated into the SEM as a latent variable using a pre-modeling parcel optimization routine presented herein for the first time (available as R library parcelR). A microbiome latent construct associated with improved healing was validated, and the final SEM included this latent construct plus three species associated with diminished healing (Anaerococcus vaginalis, Finegoldia magna, Pseudomonas aeruginosa), as well as smoking, wound volume, slough, exudate, edema, percent granulation, and wound etiology This model explained 46% of variation in healing time with the microbiome contributing the largest proportion of variance explained. Model validity was confirmed with an independent cohort (n = 79) through which ~60% of variation in healing time was predicted. This model can serve as foundation for development of a predictive tool that may have clinical utility.

Machine Learning-Based Factor Analysis of Carbohydrate Metabolism Compensation for TDM2 Patients

Type 2 diabetes is one of the most common chronic diseases in the world. World Diabetes Federation experts predict that the diabetes patients' number by 2035 will increase by 205 million to reach 592 million. For health care, this diabetes type is one of the highest priority problems. This disease is associated with many concomitant diseases leading to early disability and high cardiovascular risk. A severity disease indicator is the degree of carbohydrate metabolism compensation. Decompensated and subcompensated carbohydrate metabolism patients have increased cardiovascular risks. Therefore, it is important to be able to select the right therapy to control carbohydrate metabolism. In this study, we propose a new method for selecting the optimal therapy automatically. The method includes creating personal optimal therapies. This kind of therapy has the highest probability of compensating carbohydrate metabolism for a patient within a six-month. The method includes models for predicting the results of different therapies. It is based on data from the previous medical history and current medical indicators of patients. This method provides high-quality predictions and medical recommendations. Therefore, medical professionals can use this method as part of the Support and Decision-Making Systems for working with T2DM patients.