Machine Learning Approach with Harmonized Multinational Datasets for Enhanced Prediction of Hypothyroidism in Patients with Type 2 Diabetes.

Type 2 diabetes (T2D) is a global health concern with increasing prevalence. Comorbid hypothyroidism (HT) exacerbates kidney, cardiac, neurological and other complications of T2D; these risks can be mitigated pharmacologically upon detecting HT. The current HT standard of care (SOC) screening in T2D is infrequent, delaying HT diagnosis and treatment. We present a first-to-date machine learning algorithm (MLA) clinical decision tool to classify patients as low vs. high risk for developing HT comorbid with T2D; the MLA was developed using readily available patient data from harmonized multinational datasets. The MLA was trained on data from NIH All of US (AoU) and UK Biobank (UKBB) (Combined dataset) and achieved a high negative predictive value (NPV) of 0.989 and an AUROC of 0.762 in the Combined dataset, exceeding AUROCs for the models trained on AoU or UKBB alone (0.666 and 0.622, respectively), indicating that increasing dataset diversity for MLA training improves performance. This high-NPV automated tool can supplement SOC screening and rule out T2D patients with low HT risk, allowing for the prioritization of lab-based testing for at-risk patients. Conversely, an MLA output that designates a patient to be at risk of developing HT allows for tailored clinical management and thereby promotes improved patient outcomes.

OpenMedLM: prompt engineering can out-perform fine-tuning in medical question-answering with open-source large language models.

LLMs can accomplish specialized medical knowledge tasks, however, equitable access is hindered by the extensive fine-tuning, specialized medical data requirement, and limited access to proprietary models. Open-source (OS) medical LLMs show performance improvements and provide the transparency and compliance required in healthcare. We present OpenMedLM, a prompting platform delivering state-of-the-art (SOTA) performance for OS LLMs on medical benchmarks. We evaluated OS foundation LLMs (7B-70B) on medical benchmarks (MedQA, MedMCQA, PubMedQA, MMLU medical-subset) and selected Yi34B for developing OpenMedLM. Prompting strategies included zero-shot, few-shot, chain-of-thought, and ensemble/self-consistency voting. OpenMedLM delivered OS SOTA results on three medical LLM benchmarks, surpassing previous best-performing OS models that leveraged costly and extensive fine-tuning. OpenMedLM displays the first results to date demonstrating the ability of OS foundation models to optimize performance, absent specialized fine-tuning. The model achieved 72.6% accuracy on MedQA, outperforming the previous SOTA by 2.4%, and 81.7% accuracy on MMLU medical-subset, establishing itself as the first OS LLM to surpass 80% accuracy on this benchmark. Our results highlight medical-specific emergent properties in OS LLMs not documented elsewhere to date and validate the ability of OS models to accomplish healthcare tasks, highlighting the benefits of prompt engineering to improve performance of accessible LLMs for medical applications.

Family-Centric Applied Behavior Analysis Facilitates Improved Treatment Utilization and Outcomes.

Background/Objective: Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by lifelong impacts on functional social and daily living skills, and restricted, repetitive behaviors (RRBs). Applied behavior analysis (ABA), the gold-standard treatment for ASD, has been extensively validated. ABA access is hindered by limited availability of qualified professionals and logistical and financial barriers. Scientifically validated, parent-led ABA can fill the accessibility gap by overcoming treatment barriers. This retrospective cohort study examines how our ABA treatment model, utilizing parent behavior technicians (pBTs) to deliver ABA, impacts adaptive behaviors and interfering behaviors (IBs) in a cohort of children on the autism spectrum with varying ASD severity levels, and with or without clinically significant IBs. Methods: Clinical outcomes of 36 patients ages 3-15 years were assessed using longitudinal changes in Vineland-3 after 3+ months of pBT-delivered ABA treatment. Results: Within the pBT model, our patients demonstrated clinically significant improvements in Vineland-3 Composite, domain, and subdomain scores, and utilization was higher in severe ASD. pBTs utilized more prescribed ABA when children initiated treatment with clinically significant IBs, and these children also showed greater gains in their Composite scores. Study limitations include sample size, inter-rater reliability, potential assessment metric bias and schedule variability, and confounding intrinsic or extrinsic factors. Conclusion: Overall, our pBT model facilitated high treatment utilization and showed robust effectiveness, achieving improved adaptive behaviors and reduced IBs when compared to conventional ABA delivery. The pBT model is a strong contender to fill the widening treatment accessibility gap and represents a powerful tool for addressing systemic problems in ABA treatment delivery.

Ethanol extract of Eclipta prostrata induces multiple myeloma ferroptosis via Keap1/Nrf2/HO-1 axis.

BACKGROUND: Multiple myeloma (MM) is an incurable hematological malignancy with limited therapeutic efficacy. Eclipta prostrata is a traditional Chinese medicinal plant reported to possess antitumor properties. However, the effects of E. prostrata in MM have not been explored. PURPOSE: The aim of this study was to define the mechanism of the ethanol extract of E. prostrata (EEEP) in treating MM and identify its major components. METHODS: The pro-ferroptotic effects of EEEP on cell death, cell proliferation, iron accumulation, lipid peroxidation, and mitochondrial morphology were determined in RPMI-8226 and U266 cells. The expression levels of nuclear factor erythroid 2-related factor 2 (Nrf2), kelch-like ECH-associated protein 1 (Keap1), heme oxygenase-1 (HO-1), glutathione peroxidase 4 (GPX4), and 4-hydroxynonenal (4HNE) were detected using western blotting during EEEP-mediated ferroptosis regulation. The RPMI-8226 and U266 xenograft mouse models were used to explore the in vivo anticancer effects of EEEP. Finally, high performance liquid chromatography (HPLC) and ultra-high-performance liquid chromatography-quadrupole/time-of-flight mass spectrometry system (UPLC-Q/TOF-MS) were used to identify the major constituents of EEEP. RESULTS: EEEP inhibited MM cell growth and induced cell death in vitro and in vivo. By promoting malondialdehyde and Fe2+ accumulation, lipid peroxidation, and GSH suppression, EEEP triggers ferroptosis in MM. Mechanistically, EEEP regulates the Keap1/Nrf2/HO-1 axis and stimulates ferroptosis. EEEP-induced lipid peroxidation and malondialdehyde accumulation were blocked by the Nrf2 activator NK-252. In addition, HPLC and UPLC-Q/TOF-MS analysis elucidated the main components of EEEP, including demethylwedelolactone, wedelolactone, chlorogenic acid and apigenin, which may play important roles in the anti-tumor function of EEEP. CONCLUSION: In summary, EEEP exerts its anti-MM function by inducing MM cell death and inhibiting tumor growth in mice. We also showed that EEEP can induce lipid peroxidation and accumulation of ferrous irons in MM cells both in vivo and in vitro, leading to ferroptosis. In addition, this anti-tumor function may be achieved by the EEEP activation of Keap1/Nrf2/HO-1 axis. This is the first study to reveal that EEEP exerts anti-MM activity through the Keap1/Nrf2/HO-1-dependent ferroptosis regulatory axis, making it a promising candidate for MM treatment.

Sesamin attenuates UVA-induced keratinocyte injury via inhibiting ASK-1-JNK/p38 MAPK pathways.

BACKGROUND: Ultraviolet (UV) exposure-stimulated reactive oxygen species (ROS) formation in keratinocytes is a crucial factor in skin aging. Phytochemicals have become widely popular for protecting the skin from UV-induced cell injury. Sesamin (SSM) has been shown to play a role in extensive pharmacological activity and exhibit photoprotective effects. AIM: To assess the protective effect of SSM on UVA-irradiated keratinocytes and determine its potential antiphotoaging effect. METHODS: HaCaT keratinocytes pretreated with SSM were exposed to UVA radiation at 8 J/cm2 for 10 min. Cell viability and oxidative stress indicators were evaluated using a cell counting kit-8 and lactate dehydrogenase (LDH), malondialdehyde (MDA), glutathione (GSH), and superoxide dismutase (SOD) assay kits. Apoptosis and intracellular ROS levels were analyzed using annexin V-fluorescein isothiocyanate/propyridine iodide and dichlorodihydrofluorescein diacetate staining, respectively. Protein levels of matrix metalloprotein-1 (MMP-1), MMP-9, Bax/Bcl-2, and mitogen-activated protein kinase (MAPK) pathway proteins, phospho-apoptosis signal-regulating kinase-1 (p-ASK-1)/ASK-1, phospho-c-Jun N-terminal protein kinase (p-JNK)/JNK, and p-p38/p38 were determined using western blotting. RESULTS: Sesamin showed no cytotoxicity until 160 µmol/L on human keratinocytes. Sesamin pretreatment (20 and 40 µM) reversed the suppressed cell viability, increased LDH release and MDA content, decreased cellular antioxidants GSH and SOD, and elevated intracellular ROS levels, which were induced by UVA irradiation. Additionally, SSM inhibited the expression of Bax, MMP-1, and MMP-9 and stimulated Bcl-2 expression. In terms of the regulatory mechanisms, we demonstrated that SSM inhibits the phosphorylation of ASK-1, JNK, and p38. CONCLUSION: The results suggest that SSM attenuates UVA-induced keratinocyte injury by inhibiting the ASK-1-JNK/p38 MAPK pathways.

Blood virome of patients with traumatic sepsis.

Sepsis is one of the possible outcomes of severe trauma, and it poses a dire threat to human life, particularly in immunocompromised people. The most prevalent pathogens are bacteria and fungi, but viruses should not be overlooked. For viral metagenomic analysis, we collected blood samples from eight patients with post-traumatic sepsis before and seven days after treatment. The results demonstrated that Anellovirus predominated the viral community, followed by Siphoviridae and Myoviridae, and that the variations in viral community and viral load before and after treatment were not statistically significant. This study allows us to investigate methods for establishing NGS-based viral diagnostic instruments for detecting viral infections in the blood of sepsis patients so that antiviral therapy can be administered quickly.

Erratum: Corrigendum: Water extract of cacumen platycladi promotes hair growth through the Akt/GSK3ß/ß-catenin signaling pathway.

[This corrects the article DOI: 10.3389/fphar.2023.1038039.].

Three New Benzophenone Derivatives from Selaginella tamariscina.

Six compounds including three new benzophenones, selagibenzophenones D-F (1-3), two known selaginellins (4-5) and one known flavonoid (6), were isolated from Selaginella tamariscina. The structures of new compounds were established by 1D-, 2D-NMR and HR-ESI-MS spectral analyses. Compound 1 represents the second example of diarylbenzophenone from natural sources. Compound 2 possesses an unusual biphenyl-bisbenzophenone structure. Their cytotoxicity against human hepatocellular carcinoma HepG2 and SMCC-7721 cells and inhibitory activities on lipopolysaccharide-induced nitric oxide (NO) production in RAW264.7 cells were evaluated. Compound 2 showed moderate inhibitory activity against HepG2 and SMCC-7721 cells, and compounds 4 and 5 showed moderate inhibitory activity to HepG2 cells. Compounds 2 and 5 also exhibited inhibitory activities on lipopolysaccharide-induced nitric oxide (NO) production.

Machine learning determination of applied behavioral analysis treatment plan type.

BACKGROUND: Applied behavioral analysis (ABA) is regarded as the gold standard treatment for autism spectrum disorder (ASD) and has the potential to improve outcomes for patients with ASD. It can be delivered at different intensities, which are classified as comprehensive or focused treatment approaches. Comprehensive ABA targets multiple developmental domains and involves 20-40 h/week of treatment. Focused ABA targets individual behaviors and typically involves 10-20 h/week of treatment. Determining the appropriate treatment intensity involves patient assessment by trained therapists, however, the final determination is highly subjective and lacks a standardized approach. In our study, we examined the ability of a machine learning (ML) prediction model to classify which treatment intensity would be most suited individually for patients with ASD who are undergoing ABA treatment. METHODS: Retrospective data from 359 patients diagnosed with ASD were analyzed and included in the training and testing of an ML model for predicting comprehensive or focused treatment for individuals undergoing ABA treatment. Data inputs included demographics, schooling, behavior, skills, and patient goals. A gradient-boosted tree ensemble method, XGBoost, was used to develop the prediction model, which was then compared against a standard of care comparator encompassing features specified by the Behavior Analyst Certification Board treatment guidelines. Prediction model performance was assessed via area under the receiver-operating characteristic curve (AUROC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). RESULTS: The prediction model achieved excellent performance for classifying patients in the comprehensive versus focused treatment groups (AUROC: 0.895; 95% CI 0.811-0.962) and outperformed the standard of care comparator (AUROC 0.767; 95% CI 0.629-0.891). The prediction model also achieved sensitivity of 0.789, specificity of 0.808, PPV of 0.6, and NPV of 0.913. Out of 71 patients whose data were employed to test the prediction model, only 14 misclassifications occurred. A majority of misclassifications (n = 10) indicated comprehensive ABA treatment for patients that had focused ABA treatment as the ground truth, therefore still providing a therapeutic benefit. The three most important features contributing to the model's predictions were bathing ability, age, and hours per week of past ABA treatment. CONCLUSION: This research demonstrates that the ML prediction model performs well to classify appropriate ABA treatment plan intensity using readily available patient data. This may aid with standardizing the process for determining appropriate ABA treatments, which can facilitate initiation of the most appropriate treatment intensity for patients with ASD and improve resource allocation.

Clinical Outcomes of a Hybrid Model Approach to Applied Behavioral Analysis Treatment.

Objective This study examines the implementation of a hybrid applied behavioral analysis (ABA) treatment model to determine its impact on autism spectrum disorder (ASD) patient outcomes.  Methods Retrospective data were collected for 25 pediatric patients to measure progress before and after the implementation of a hybrid ABA treatment model under which therapists consistently captured session notes electronically regarding goals and patient progress. ABA treatment was streamlined for consistent delivery, with improved software utilization for tracking scheduling and progress. Eleven goals within three domains (behavioral, social, and communication) were examined.  Results After the implementation of the hybrid model, the goal success rate improved by 9.7% compared to the baseline; 41.8% of goals showed improvement, 38.4% showed a flat trend, and 19.8% showed deterioration. Multiple goals trended upwards in 76% of the patients.  Conclusion This pilot study demonstrated that enhancing the consistency with which ABA treatment is monitored/delivered can improve patient outcomes as seen through improved attainment of goals.

Water extract of cacumen platycladi promotes hair growth through the Akt/GSK3ß/ß-catenin signaling pathway.

Cacumen Platycladi (CP) consists of the dried needles of Platycladus orientalis L.) Franco. It was clinically demonstrated that it effectively regenerates hair, but the underlying mechanism remains unknown. Thus, we employed shaved mice to verify the hair growth-promoting capability of the water extract of Cacumen Platycladi (WECP). The morphological and histological analyses revealed that WECP application could significantly promote hair growth and hair follicles (HFs) construction, in comparison to that of control group. Additionally, the skin thickness and hair bulb diameter were significantly increased by the application of WECP in a dose-dependent manner. Besides, the high dose of WECP also showed an effect similar to that of finasteride. In an in vitro assay, WECP stimulated dermal papilla cells (DPCs) proliferation and migration. Moreover, the upregulation of cyclins (cyclin D1, cyclin-dependent kinase 2 (CDK2), and cyclin-dependent kinase 4 (CDK4)) and downregulation of P21 in WECP-treated cell assays have been evaluated. We identified the ingredients of WECP using ultra-high-performance liquid chromatography-quadrupole time-of-flight mass spectrometry (UPLC-Q/TOF-MS) and endeavored to predict their relevant molecular mechanisms by network analysis. We found that the Akt (serine/threonine protein kinase) signaling pathway might be a crucial target of WECP. It has been demonstrated that WECP treatment activated the phosphorylation of Akt and glycogen synthase kinase-3-beta (GSK3ß), promoted ß-Catenin and Wnt10b accumulation, and upregulated the expression of lymphoid enhancer-binding factor 1 (LEF1), vascular endothelial growth factor (VEGF), and insulin-like growth factor 1 (IGF1). We also found that WECP significantly altered the expression levels of apoptosis-related genes in mouse dorsal skin. The enhancement capability of WECP on DPCs proliferation and migration could be abrogated by the Akt-specific inhibitor MK-2206 2HCl. These results suggested that WECP might promote hair growth by modulating DPCs proliferation and migration through the regulation of the Akt/GSK3ß/ß-Catenin signaling pathway.

Stimulation of Surface Polysaccharide Production under Aerobic Conditions Confers Aerotolerance in Campylobacter jejuni.

The ability of a foodborne pathogen to tolerate environmental stress critically affects food safety by increasing the risk of pathogen survival and transmission in the food supply chain. Campylobacter jejuni, a leading bacterial cause of foodborne illnesses, is an obligate microaerophile and is sensitive to atmospheric levels of oxygen. Currently, the molecular mechanisms of how C. jejuni withstands oxygen toxicity under aerobic conditions have not yet been fully elucidated. Here, we show that when exposed to aerobic conditions, C. jejuni develops a thick layer of bacterial capsules, which in turn protect C. jejuni under aerobic conditions. The presence of both capsular polysaccharides and lipooligosaccharides is required to protect C. jejuni from excess oxygen in oxygen-rich environments by alleviating oxidative stress. Under aerobic conditions, C. jejuni undergoes substantial transcriptomic changes, particularly in the genes of carbon metabolisms involved in amino acid uptake, the tricarboxylic acid (TCA) cycle, and the Embden-Meyerhof-Parnas (EMP) pathway despite the inability of C. jejuni to grow aerobically. Moreover, the stimulation of carbon metabolism by aerobiosis increases the level of glucose-6-phosphate, the EMP pathway intermediate required for the synthesis of surface polysaccharides. The disruption of the TCA cycle eliminates aerobiosis-mediated stimulation of surface polysaccharide production and markedly compromises aerotolerance in C. jejuni. These results in this study provide novel insights into how an oxygen-sensitive microaerophilic pathogen survives in oxygen-rich environments by adapting its metabolism and physiology. IMPORTANCE Oxygen-sensitive foodborne pathogens must withstand oxygen toxicity in aerobic environments during transmission to humans. C. jejuni is a major cause of gastroenteritis, accounting for 400 million to 500 million infection cases worldwide per year. As an obligate microaerophile, C. jejuni is sensitive to air-level oxygen. However, it has not been fully explained how this oxygen-sensitive zoonotic pathogen survives in aerobic environments and is transmitted to humans. Here, we show that under aerobic conditions, C. jejuni boosts its carbon metabolism to produce a thick layer of bacterial capsules, which in turn act as a protective barrier conferring aerotolerance. The new findings in this study improve our understanding of how oxygen-sensitive C. jejuni can survive in aerobic environments.

Early prediction of prostate cancer risk in younger men using polygenic risk scores and electronic health records.

BACKGROUND: Prostate cancer (PCa) screening is not routinely conducted in men aged 55 and younger, although this age group accounts for more than 10% of cases. Polygenic risk scores (PRSs) and patient data applied toward early prediction of PCa may lead to earlier interventions and increased survival. We have developed machine learning (ML) models to predict PCa risk in men 55 and under using PRSs combined with patient data. METHODS: We conducted a retrospective study on 91,106 male patients aged 35-55 using the UK Biobank database. Five gradient boosting models were developed and validated utilizing routine screening data, PRSs, additional clinical data, or combinations of the three. RESULTS: Combinations of PRSs and patient data outperformed models that utilized PRS or patient data only, and the highest performing models achieved an area under the receiver operating characteristic curve of 0.788. Our models demonstrated a substantially lower false positive rate (35.4%) in comparison to standard screening using prostate-specific antigen (60%-67%). CONCLUSION: This study provides the first preliminary evidence for the use of PRSs with patient data in a ML algorithm for PCa risk prediction in men aged 55 and under for whom screening is not standard practice.

Machine Learning Approach for Improved Longitudinal Prediction of Progression from Mild Cognitive Impairment to Alzheimer's Disease.

Mild cognitive impairment (MCI) is cognitive decline that can indicate future risk of Alzheimer's disease (AD). We developed and validated a machine learning algorithm (MLA), based on a gradient-boosted tree ensemble method, to analyze phenotypic data for individuals 55-88 years old (n = 493) diagnosed with MCI. Data were analyzed within multiple prediction windows and averaged to predict progression to AD within 24-48 months. The MLA outperformed the mini-mental state examination (MMSE) and three comparison models at all prediction windows on most metrics. Exceptions include sensitivity at 18 months (MLA and MMSE each achieved 0.600); and sensitivity at 30 and 42 months (MMSE marginally better). For all prediction windows, the MLA achieved AUROC ≥ 0.857 and NPV ≥ 0.800. With averaged data for the 24-48-month lookahead timeframe, the MLA outperformed MMSE on all metrics. This study demonstrates that machine learning may provide a more accurate risk assessment than the standard of care. This may facilitate care coordination, decrease healthcare expenditures, and maintain quality of life for patients at risk of progressing from MCI to AD.

Expanding known viral diversity in plants: virome of 161 species alongside an ancient canal.

BACKGROUND: Since viral metagenomic approach was applied to discover plant viruses for the first time in 2006, many plant viruses had been identified from cultivated and non-cultivated plants. These previous researches exposed that the viral communities (virome) of plants have still largely uncharacterized. Here, we investigated the virome in 161 species belonging to 38 plant orders found in a riverside ecosystem. RESULTS: We identified 245 distinct plant-associated virus genomes (88 DNA and 157 RNA viruses) belonging to 27 known viral families, orders, or unclassified virus groups. Some viral genomes were sufficiently divergent to comprise new species, genera, families, or even orders. Some groups of viruses were detected that currently are only known to infect organisms other than plants. It indicates a wider host range for members of these clades than previously recognized theoretically. We cannot rule out that some viruses could be from plant contaminating organisms, although some methods were taken to get rid of them as much as possible. The same viral species could be found in different plants and co-infections were common. CONCLUSIONS: Our data describe a complex viral community within a single plant ecosystem and expand our understanding of plant-associated viral diversity and their possible host ranges.

Viral metagenomics of pharyngeal secretions from children with acute respiratory diseases with unknown etiology revealed diverse viruses.

Acute respiratory tract infections are a major public health problem and the leading cause of morbidity in children younger than 5 years old. This study investigated the potential reasons of unexplained acute respiratory infections in children in Xuzhou and its environs during 2018-2019.We collected pharyngeal swab samples from 411 children under the age of five who presented with symptoms of unexplained acute respiratory infection and were negative for bacteria, mycoplasma, and influenza viruses. Using viral metagenomic techniques, viral nucleic acids were extracted, enriched, and sequenced from the samples. Results indicated that Picornaviridae, Parvoviridae, Paramyxoviridae, Coronaviridae, and Anelloviridae were the five virus families with the highest relative content of sequence reads. And we detected 35 HBoV-positive and 12 HEV-positive samples out of 411 samples by the polymerase chain reaction (PCR). Partial or nearly complete genome sequences of viruses belonging to the families Picornaviridae, Parvoviridae, and Anelloviridae were characterized, and phylogenetic trees were constructed based on the nucleic acid or amino acid sequences of the predicted viral open reading frames (ORFs), as well as genotyping of the viruses. In addition, we observed recombination events in the Saffold virus and Coxsackievirus A9 by analyzing the genetic characteristics of the viruses revealed in this study. This study provides vital information for the prevention and treatment of acute respiratory infections in children younger than five years old.

Machine learning early prediction of respiratory syncytial virus in pediatric hospitalized patients.

Respiratory syncytial virus (RSV) causes millions of infections among children in the US each year and can cause severe disease or death. Infections that are not promptly detected can cause outbreaks that put other hospitalized patients at risk. No tools besides diagnostic testing are available to rapidly and reliably predict RSV infections among hospitalized patients. We conducted a retrospective study from pediatric electronic health record (EHR) data and built a machine learning model to predict whether a patient will test positive to RSV by nucleic acid amplification test during their stay. Our model demonstrated excellent discrimination with an area under the receiver-operating curve of 0.919, a sensitivity of 0.802, and specificity of 0.876. Our model can help clinicians identify patients who may have RSV infections rapidly and cost-effectively. Successfully integrating this model into routine pediatric inpatient care may assist efforts in patient care and infection control.

Massive external validation of a machine learning algorithm to predict pulmonary embolism in hospitalized patients.

BACKGROUND: Pulmonary embolism (PE) is a life-threatening condition associated with ~10% of deaths of hospitalized patients. Machine learning algorithms (MLAs) which predict the onset of pulmonary embolism (PE) could enable earlier treatment and improve patient outcomes. However, the extent to which they generalize to broader patient populations impacts their clinical utility. OBJECTIVE: To conduct the first large-scale external validation of a machine learning-based PE prediction model which uses EHR data from the first three hours of a patient's hospital stay to predict the occurrence of PE within the next 10 days of the inpatient stay. METHODS: This retrospective study included approximately two million adult hospital admissions across 44 medical institutions in the US from 2011 to 2017. Demographics, vital signs, and lab tests from adult inpatients at 12 institutions (n = 331,268; 3.3% PE positive) were used for training an XGBoost model. External validation of the model was conducted on patient populations from each of 32 medical institutions (total n = 1,660,715; 3.7% PE positive) without retraining. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC). Backward elimination regression was used to identify correlations between characteristics of the external validation sets and AUROC. RESULTS: The model performed well (AUROC = 0.87) on the 20% hold-out subset of the training set. Despite demographic differences between the 32 external validation populations (percent PE positive: min = 1.54%, max = 6.47%), without retraining, the model had excellent discrimination, with a mean AUROC of 0.88 (min = 0.79, max = 0.93). Fixing sensitivity at 0.80, the model had a mean specificity of 0.85 (min = 0.64, max = 0.93). Backward elimination regression identified a negative association (ß = -0.015, p < 0.001) between the percentage of PE positive encounters and AUROC. CONCLUSIONS: A PE prediction model performed remarkably well across 32 different external patient populations without retraining and despite significant differences in demographic characteristics, demonstrating its generalizability and potential as a clinical decision support tool to aid PE detection and improve patient outcomes in a clinical setting.

Multitask Learning With Recurrent Neural Networks for Acute Respiratory Distress Syndrome Prediction Using Only Electronic Health Record Data: Model Development and Validation Study.

BACKGROUND: Acute respiratory distress syndrome (ARDS) is a condition that is often considered to have broad and subjective diagnostic criteria and is associated with significant mortality and morbidity. Early and accurate prediction of ARDS and related conditions such as hypoxemia and sepsis could allow timely administration of therapies, leading to improved patient outcomes. OBJECTIVE: The aim of this study is to perform an exploration of how multilabel classification in the clinical setting can take advantage of the underlying dependencies between ARDS and related conditions to improve early prediction of ARDS in patients. METHODS: The electronic health record data set included 40,703 patient encounters from 7 hospitals from April 20, 2018, to March 17, 2021. A recurrent neural network (RNN) was trained using data from 5 hospitals, and external validation was conducted on data from 2 hospitals. In addition to ARDS, 12 target labels for related conditions such as sepsis, hypoxemia, and COVID-19 were used to train the model to classify a total of 13 outputs. As a comparator, XGBoost models were developed for each of the 13 target labels. Model performance was assessed using the area under the receiver operating characteristic curve. Heat maps to visualize attention scores were generated to provide interpretability to the neural networks. Finally, cluster analysis was performed to identify potential phenotypic subgroups of patients with ARDS. RESULTS: The single RNN model trained to classify 13 outputs outperformed the individual XGBoost models for ARDS prediction, achieving an area under the receiver operating characteristic curve of 0.842 on the external test sets. Models trained on an increasing number of tasks resulted in improved performance. Earlier prediction of ARDS nearly doubled the rate of in-hospital survival. Cluster analysis revealed distinct ARDS subgroups, some of which had similar mortality rates but different clinical presentations. CONCLUSIONS: The RNN model presented in this paper can be used as an early warning system to stratify patients who are at risk of developing one of the multiple risk outcomes, hence providing practitioners with the means to take early action.

Predicting pulmonary embolism among hospitalized patients with machine learning algorithms.

Background: Pulmonary embolisms (PE) are life-threatening medical events, and early identification of patients experiencing a PE is essential to optimizing patient outcomes. Current tools for risk stratification of PE patients are limited and unable to predict PE events before their occurrence. Objective: We developed a machine learning algorithm (MLA) designed to identify patients at risk of PE before the clinical detection of onset in an inpatient population. Materials and Methods: Three machine learning (ML) models were developed on electronic health record data from 63,798 medical and surgical inpatients in a large US medical center. These models included logistic regression, neural network, and gradient boosted tree (XGBoost) models. All models used only routinely collected demographic, clinical, and laboratory information as inputs. All were evaluated for their ability to predict PE at the first time patient vital signs and lab measures required for the MLA to run were available. Performance was assessed with regard to the area under the receiver operating characteristic (AUROC), sensitivity, and specificity. Results: The model trained using XGBoost demonstrated the strongest performance for predicting PEs. The XGBoost model obtained an AUROC of 0.85, a sensitivity of 81%, and a specificity of 70%. The neural network and logistic regression models obtained AUROCs of 0.74 and 0.67, sensitivity of 81% and 81%, and specificity of 44% and 35%, respectively. Conclusions: This algorithm may improve patient outcomes through earlier recognition and prediction of PE, enabling earlier diagnosis and treatment of PE.