Real-world Experience With Neurotrophic Tyrosine Receptor Kinase Fusion-positive Tumors and Tropomyosin Receptor Kinase Inhibitors in Veterans.

PURPOSE: Neurotrophic tyrosine receptor kinase 1-3 (NTRK1-3) gene fusions are found in a broad range of tumor types. Clinical trials demonstrated high response rates to tropomyosin receptor kinase (TRK) inhibitors in NTRK fusion-positive cancers, but few reports have described real-world experience with these targeted agents. We evaluated the prevalence of NTRK fusions and the outcomes with TRK inhibitor therapy in a real-world population of patients in the Veterans Health Administration. METHODS: Patients with NTRK fusions or rearrangements were identified from the Veterans Affairs (VA) National Precision Oncology Program (NPOP), and patients who were prescribed TRK inhibitors were identified from the Corporate Data Warehouse. Baseline data and clinical outcomes were obtained by retrospective review of medical records. RESULTS: A total of 33 patients with NTRK fusions or rearrangements were identified, including 25 patients comprising 0.12% of all patients with solid tumors sequenced through VA NPOP. Twelve patients with NTRK fusions or rearrangements were treated with TRK inhibitors, none of whom had objective responses. Eight patients experienced toxicities leading to drug interruption, dose reduction, or discontinuation. CONCLUSION: In this retrospective study of VA patients, NTRK fusions and rearrangements were less common than in previous studies, and objective responses to TRK inhibitors were not observed. Real-world experience with TRK inhibitors differs markedly from clinical trial findings, possibly due to differences in patient demographics, tumor types, and sequencing methods. Our findings highlight the need to study TRK inhibitors in the real-world setting and in populations underrepresented in clinical trials.

Molecular-Guided Off-Label Targeted Therapy in a Large-Scale Precision Oncology Program.

PURPOSE: Increasing utilization of comprehensive genomic profiling (CGP) and a growing number of targeted agents (TAs) have led to substantial improvements in outcomes among patients with cancer with actionable mutations. We sought to evaluate real-world experience with off-label TAs among Veterans who underwent CGP. METHODS: The National Precision Oncology Program database and VA Corporate Data Warehouse were queried to identify patients who underwent CGP between February 2019 and December 2021 and were prescribed 1 of 73 TAs for malignancy. OncoKB annotations were used to select patients who received off-label TAs based upon CGP results. Chart abstraction was performed to review response, toxicities, and time to progression. RESULTS: Of 18,686 patients who underwent CGP, 2,107 (11%) were prescribed a TA and 169 (0.9%) were prescribed a total of 183 regimens containing off-label TAs for variants in 31 genes. Median age was 68 years and 83% had prior systemic therapy, with 28% receiving three or more lines. Frequency of off-label TA prescriptions was highest for patients undergoing CGP for thyroid (8.6%) and breast (7.6%) cancers. Most patients harbored alterations in BRCA1/BRCA2/ATM (22.5%), ERBB2 (19.5%), and BRAF (19.5%). Among the 160 regimens prescribed > 4 weeks, 43 (27%) led to response. Median progression-free survival and overall survival were 5.3 (4.2-6.5) and 9.7 (7.5-11.9) months, respectively. Patients with OncoKB level 2/3A/3B annotations had longer median progression-free survival (5.8 [4.5-7] months v 3.7 [1.6-7.7] months; hazard ratio, 0.45; 95% CI, 0.24 to 0.82; P = .01) compared with those receiving level 4 treatments. CONCLUSION: Although administration of off-label TAs is infrequent after CGP, more than one quarter of treatment regimens led to response. TAs associated with level 4 annotations lead to worse outcomes than TAs bearing higher levels of evidence.

Effects of multi-resource addition on grassland plant productivity and biodiversity along a resource gradient.

The change of plant biodiversity caused by resource-enhancing global changes has greatly affected grassland productivity. However, it remains unclear how multi-resource enrichment induces the effects of multifaceted biodiversity on grassland productivity under different site resource constraints. We conducted a multiple resource addition (MRA) experiment of water and nutrients at three sites located along a resource gradient in northern China. This allowed us to assess the response of aboveground net primary productivity (ANPP), species (species richness and plant density), functional (functional richness and community-weighted mean of traits) and phylogenetic (phylogenetic richness) diversity to increasing number of MRA. We used structural equation model (SEM) to examine the direct and indirect effects of MRA and multifaceted biodiversity on ANPP. The combined addition of the four resources increased ANPP at all three sites. But with increasing number of MRA, biodiversity varied at the three sites. At the high resource constraint site, species richness, plant density and leaf nitrogen concentration (LNC) increased. At the medium resource constraint site, plant height and LNC increased, leaf dry matter content (LDMC) decreased. At the low resource constraint site, species, functional and phylogenetic richness decreased, and height increased. The SEM showed that MRA increased ANPP directly at all three sites, and indirectly by increasing plant density at the high constraint site and height at the medium constraint site. Independent of MRA, ANPP was affected by height at the high resource constraint site and LNC at the low resource constraint site. Our results illustrate that multi-resource addition positively affects productivity, while affects biodiversity depending on site resource constraint. The study highlights that site resource constraint conditions need to be taken into consideration to better predict grassland structure and function, particularly under the future multifaceted global change scenarios.

Electronic Health Record-Based Deep Learning Prediction of Death or Severe Decompensation in Heart Failure Patients.

BACKGROUND: Surgical mechanical ventricular assistance and cardiac replacement therapies, although life-saving in many heart failure (HF) patients, remain high-risk. Despite this, the difficulty in timely identification of medical therapy nonresponders and the dire consequences of nonresponse have fueled early, less selective surgical referral. Patients who would have ultimately responded to medical therapy are therefore subjected to the risk and life disruption of surgical therapy. OBJECTIVES: The purpose of this study was to develop deep learning models based upon commonly-available electronic health record (EHR) variables to assist clinicians in the timely and accurate identification of HF medical therapy nonresponders. METHODS: The study cohort consisted of all patients (age 18 to 90 years) admitted to a single tertiary care institution from January 2009 through December 2018, with International Classification of Disease HF diagnostic coding. Ensemble deep learning models employing time-series and densely-connected networks were developed from standard EHR data. The positive class included all observations resulting in severe progression (death from any cause or referral for HF surgical intervention) within 1 year. RESULTS: A total of 79,850 distinct admissions from 52,265 HF patients met observation criteria and contributed >350 million EHR datapoints for model training, validation, and testing. A total of 20% of model observations fit positive class criteria. The model C-statistic was 0.91. CONCLUSIONS: The demonstrated accuracy of EHR-based deep learning model prediction of 1-year all-cause death or referral for HF surgical therapy supports clinical relevance. EHR-based deep learning models have considerable potential to assist HF clinicians in improving the application of advanced HF surgical therapy in medical therapy nonresponders.

Shear-induced fabrication of SiO2 nano-meshes for efficient uranium capture.

The extraction of uranium from seawater and the safe treatment of wastewater containing uranium (VI) were important to ensure the sustainable development of nuclear-related energy sources. Two-dimensional silica nanomaterials have been extensively investigated in the field of uranium adsorption due to their high adsorption capacity, short equilibration times, and easily modified surface groups. However, the two-dimensional mesoporous silica nanomaterial preparation has become a challenge due to the lack of natural sheet templating agents. The reason will hinder the development of silica nanomaterials for uranium extraction. Here, the specific surface area silica nanomeshes (HSMSMs) uranium adsorbent was prepared by a high shear method to induce nanobubble formation. HSMSMs showed a high uranium adsorption capacity of 822 mg-U/g-abs in seawater with the uranium adsorption concentration was 50 mg/L, which was approximately 2 times higher than the conventional mesoporous silica nanomaterials. Compared to HSMSMs, the amidoxime-modified high specific surface area silica nanomesh (HSMSMs-AO) demonstrated good selectivity for U(VI), and the uranium ions uptake was 877 mg-U/g-abs in 50 mg/L uranium-spiked simulated seawater. Due to HSMSMs-AO's stable chemical properties and high mechanical strength, HSMSMs-AO also displayed long service life. Benefiting from the simple preparation method and high adsorption capacity of HSMSMs, HSMSMs could be a promising candidate for large-scale extraction of uranium from seawater.

Analysis of actionable genetic alterations in lung carcinoma from the VA National Precision Oncology Program.

Lung cancer is the leading cause of cancer mortality in men and women. Genomic sequencing of non-small cell lung cancer (NSCLC) is critical for the optimal treatment of NSCLC. In this study we sought to describe the frequencies of highly actionable driver mutations in lung adenocarcinoma (LUAD), squamous cell (LUSQ) and other NSCLC histologies (LUOT) in Veterans tested through the VA's National Precision Oncology Program (NPOP) and compare these frequencies to other published datasets from highly specialized academic cancer centers. The NPOP cohort included 3,376 unique Veterans with a diagnosis of lung carcinoma tested between February 2019 and January 2021 including 1892 with LUAD, 940 with LUSQ, and 549 with LUOT. Among patients with LUAD, 27.5% had highly actionable genetic variants. The frequency of targetable mutations was as follows: ALK rearrangement 0.8%, BRAF V600E 2.1%, EGFR exon 20 insertion mutation 0.48%, EGFR sensitizing mutations 6.6%, ERBB2 small variants 1.2%, KRAS G12C 14.0%, MET exon 14 skipping mutation 1.5%, NTRK rearrangement 0.1%, RET rearrangement 0.4%, and ROS1 rearrangement 0.3%. The frequency of EGFR mutations, RET rearrangement, MET exon 14 and ERBB2 small variants frequencies were significantly lower in NPOP compared to other published reports while MET amplification was more common in NPOP. Combined rates of highly actionable genetic variants were 2.7% and 13.4% in LUSQ and LUOT, respectively. In this study, 27.5% of Veterans with lung adenocarcinoma have actionable genetic alterations eligible for FDA approved targeted therapies, a frequency only slightly lower than other published datasets despite higher smoking rates in Veterans. Genomic sequencing should be performed in all Veterans with advanced LUAD and LUOT.

Two Dominant Herbaceous Species Have Different Plastic Responses to N Addition in a Desert Steppe.

Nitrogen (N) deposition rates are increasing in the temperate steppe due to human activities. Understanding the plastic responses of plant dominant species to increased N deposition through the lens of multiple traits is crucial for species selection in the process of vegetation restoration. Here, we measured leaf morphological, physiological, and anatomical traits of two dominant species (Stipa glareosa and Peganum harmala) after 3-year N addition (0, 1, 3, and 6 g N m-2 year-1, designated N0, N1, N3, and N6, respectively) in desert steppe of Inner Mongolia. We separately calculated the phenotypic plasticity index (PI) of each trait under different N treatments and the mean phenotypic plasticity index (MPI) of per species. The results showed that N addition increased the leaf N content (LNC) in both species. N6 increased the contents of soluble protein and proline, and decreased the superoxide dismutase (SOD) and the peroxidase (POD) activities of S. glareosa, while increased POD and catalase (CAT) activities of P. harmala. N6 increased the palisade tissue thickness (PT), leaf thickness (LT), and palisade-spongy tissue ratio (PT/ST) and decreased the spongy tissue-leaf thickness ratio (ST/LT) of S. glareosa. Furthermore, we found higher physiological plasticity but lower morphological and anatomical plasticity in both species, with greater anatomical plasticity and MPI in S. glareosa than P. harmala. Overall, multi-traits comparison reveals that two dominant desert-steppe species differ in their plastic responses to N addition. The higher plasticity of S. glareosa provides some insight into why S. glareosa has a broad distribution in a desert steppe.

Development and Structure of an Accurate Machine Learning Algorithm to Predict Inpatient Mortality and Hospice Outcomes in the Coronavirus Disease 2019 Era.

BACKGROUND: The coronavirus disease 2019 (COVID-19) pandemic has challenged the accuracy and racial biases present in traditional mortality scores. An accurate prognostic model that can be applied to hospitalized patients irrespective of race or COVID-19 status may benefit patient care. RESEARCH DESIGN: This cohort study utilized historical and ongoing electronic health record features to develop and validate a deep-learning model applied on the second day of admission predicting a composite outcome of in-hospital mortality, discharge to hospice, or death within 30 days of admission. Model features included patient demographics, diagnoses, procedures, inpatient medications, laboratory values, vital signs, and substance use history. Conventional performance metrics were assessed, and subgroup analysis was performed based on race, COVID-19 status, and intensive care unit admission. SUBJECTS: A total of 35,521 patients hospitalized between April 2020 and October 2020 at a single health care system including a tertiary academic referral center and 9 community hospitals. RESULTS: Of 35,521 patients, including 9831 non-White patients and 2020 COVID-19 patients, 2838 (8.0%) met the composite outcome. Patients who experienced the composite outcome were older (73 vs. 61 y old) with similar sex and race distributions between groups. The model achieved an area under the receiver operating characteristic curve of 0.89 (95% confidence interval: 0.88, 0.91) and an average positive predictive value of 0.46 (0.40, 0.52). Model performance did not differ significantly in White (0.89) and non-White (0.90) subgroups or when grouping by COVID-19 status and intensive care unit admission. CONCLUSION: A deep-learning model using large-volume, structured electronic health record data can effectively predict short-term mortality or hospice outcomes on the second day of admission in the general inpatient population without significant racial bias.

Contrasting effects of plant inter- and intraspecific variation on community trait responses to nitrogen addition and drought in typical and meadow steppes.

BACKGROUND: Inter- and intraspecific variation in plant traits play an important role in grassland community assembly under global change scenarios. However, explorations of how these variations contribute to the responses of community traits to nitrogen (N) addition and drought in different grassland types are lacking. We measured the plant height, leaf area (LA), specific leaf area (SLA), leaf dry matter content (LDMC), leaf N content (LNC) and the ratio of leaf carbon (C) to leaf N (C:N) in a typical and a meadow steppe after three years of N addition, drought and their interaction. We determined the community-weighted means (CWMs) of the six traits to quantify the relative contribution of inter- and intraspecific variation to the responses of community traits to N addition and drought in the two steppes. RESULTS: The communities in the two steppes responded to N addition and the interaction by increasing the CWM of LNC and decreasing C:N. The community in the meadow steppe responded to drought through increased CWM of LNC and reduced C:N. Significant differences were observed in SLA, LDMC, LNC and C:N between the two steppes under different treatments. The SLA and LNC of the community in the meadow steppe were greater than those of the typical steppe, and the LDMC and C:N exhibited the opposite results. Moreover, variation in community traits in the typical steppe in response to N addition and drought was caused by intraspecific variation. In contrast, the shifts in community traits in the meadow steppe in response to N addition and drought were influenced by both inter- and intraspecific variation. CONCLUSIONS: The results demonstrate that intraspecific variation contributed more to community functional shifts in the typical steppe than in the meadow steppe. Intraspecific variation should be considered to understand better and predict the response of typical steppe communities to global changes. The minor effects of interspecific variation on meadow steppe communities in response to environmental changes also should not be neglected.

Association of STS database variables with repair durability in ischemic mitral regurgitation using machine learning.

BACKGROUND: Machine learning (ML) can identify nonintuitive clinical variable combinations that predict clinical outcomes. To assess the potential predictive contribution of standardized Society of Thoracic Surgeons (STS) Database clinical variables, we used ML to detect their association with repair durability in ischemic mitral regurgitation (IMR) patients in a single institution study. METHODS: STS Database variables (n = 53) served as predictors of repair durability in ML modeling of 224 patients who underwent surgical revascularization and mitral valve repair for IMR. Follow-up mortality and echocardiography data allowed 1-year outcome analysis in 173 patients. Supervised ML analyses were performed using recurrence (≥3+ IMR) or death versus nonrecurrence (<3+ IMR) as the binary outcome classification. RESULTS: We tested standard ML and deep learning algorithms, including support vector machines, logistic regression, and deep neural networks. Following training, final models were utilized to predict class labels for the patients in the test set, producing receiver operating characteristic (ROC) curves. The three models produced similar area under the curve (AUC), and predicted class labels with promising accuracy (AUC = 0.72-0.75). CONCLUSIONS: Readily-available STS Database variables have potential to play a significant role in the development of ML models to direct durable surgical therapy in IMR patients.

Women and ethnoracial minorities with poor cardiovascular health measures associated with a higher risk of developing mood disorder.

BACKGROUND: Mood disorders (MDS) are a type of mental health illness that effects millions of people in the United States. Early prediction of MDS can give providers greater opportunity to treat these disorders. We hypothesized that longitudinal cardiovascular health (CVH) measurements would be informative for MDS prediction. METHODS: To test this hypothesis, the American Heart Association's Guideline Advantage (TGA) dataset was used, which contained longitudinal EHR from 70 outpatient clinics. The statistical analysis and machine learning models were employed to identify the associations of the MDS and the longitudinal CVH metrics and other confounding factors. RESULTS: Patients diagnosed with MDS consistently had a higher proportion of poor CVH compared to patients without MDS, with the largest difference between groups for Body mass index (BMI) and Smoking. Race and gender were associated with status of CVH metrics. Approximate 46% female patients with MDS had a poor hemoglobin A1C compared to 44% of those without MDS; 62% of those with MDS had poor BMI compared to 47% of those without MDS; 59% of those with MDS had poor blood pressure (BP) compared to 43% of those without MDS; and 43% of those with MDS were current smokers compared to 17% of those without MDS. CONCLUSIONS: Women and ethnoracial minorities with poor cardiovascular health measures were associated with a higher risk of development of MDS, which indicated the high utility for using routine medical records data collected in care to improve detection and treatment for MDS among patients with poor CVH.

Discovering disease-disease associations using electronic health records in The Guideline Advantage (TGA) dataset.

Certain diseases have strong comorbidity and co-occurrence with others. Understanding disease-disease associations can potentially increase awareness among healthcare providers of co-occurring conditions and facilitate earlier diagnosis, prevention and treatment of patients. In this study, we utilized the valuable and large The Guideline Advantage (TGA) longitudinal electronic health record dataset from 70 outpatient clinics across the United States to investigate potential disease-disease associations. Specifically, the most prevalent 50 disease diagnoses were manually identified from 165,732 unique patients. To investigate the co-occurrence or dependency associations among the 50 diseases, the categorical disease terms were first mapped into numerical vectors based on disease co-occurrence frequency in individual patients using the Word2Vec approach. Then the novel and interesting disease association clusters were identified using correlation and clustering analyses in the numerical space. Moreover, the distribution of time delay (Δt) between pair-wise strongly associated diseases (correlation coefficients ≥ 0.5) were calculated to show the dependency among the diseases. The results can indicate the risk of disease comorbidity and complications, and facilitate disease prevention and optimal treatment decision-making.

The National COVID Cohort Collaborative: Analyses of Original and Computationally Derived Electronic Health Record Data.

BACKGROUND: Computationally derived ("synthetic") data can enable the creation and analysis of clinical, laboratory, and diagnostic data as if they were the original electronic health record data. Synthetic data can support data sharing to answer critical research questions to address the COVID-19 pandemic. OBJECTIVE: We aim to compare the results from analyses of synthetic data to those from original data and assess the strengths and limitations of leveraging computationally derived data for research purposes. METHODS: We used the National COVID Cohort Collaborative's instance of MDClone, a big data platform with data-synthesizing capabilities (MDClone Ltd). We downloaded electronic health record data from 34 National COVID Cohort Collaborative institutional partners and tested three use cases, including (1) exploring the distributions of key features of the COVID-19-positive cohort; (2) training and testing predictive models for assessing the risk of admission among these patients; and (3) determining geospatial and temporal COVID-19-related measures and outcomes, and constructing their epidemic curves. We compared the results from synthetic data to those from original data using traditional statistics, machine learning approaches, and temporal and spatial representations of the data. RESULTS: For each use case, the results of the synthetic data analyses successfully mimicked those of the original data such that the distributions of the data were similar and the predictive models demonstrated comparable performance. Although the synthetic and original data yielded overall nearly the same results, there were exceptions that included an odds ratio on either side of the null in multivariable analyses (0.97 vs 1.01) and differences in the magnitude of epidemic curves constructed for zip codes with low population counts. CONCLUSIONS: This paper presents the results of each use case and outlines key considerations for the use of synthetic data, examining their role in collaborative research for faster insights.

Application of a time-series deep learning model to predict cardiac dysrhythmias in electronic health records.

BACKGROUND: Cardiac dysrhythmias (CD) affect millions of Americans in the United States (US), and are associated with considerable morbidity and mortality. New strategies to combat this growing problem are urgently needed. OBJECTIVES: Predicting CD using electronic health record (EHR) data would allow for earlier diagnosis and treatment of the condition, thus improving overall cardiovascular outcomes. The Guideline Advantage (TGA) is an American Heart Association ambulatory quality clinical data registry of EHR data representing 70 clinics distributed throughout the US, and has been used to monitor outpatient prevention and disease management outcome measures across populations and for longitudinal research on the impact of preventative care. METHODS: For this study, we represented all time-series cardiovascular health (CVH) measures and the corresponding data collection time points for each patient by numerical embedding vectors. We then employed a deep learning technique-long-short term memory (LSTM) model-to predict CD from the vector of time-series CVH measures by 5-fold cross validation and compared the performance of this model to the results of deep neural networks, logistic regression, random forest, and Naïve Bayes models. RESULTS: We demonstrated that the LSTM model outperformed other traditional machine learning models and achieved the best prediction performance as measured by the average area under the receiver operator curve (AUROC): 0.76 for LSTM, 0.71 for deep neural networks, 0.66 for logistic regression, 0.67 for random forest, and 0.59 for Naïve Bayes. The most influential feature from the LSTM model were blood pressure. CONCLUSIONS: These findings may be used to prevent CD in the outpatient setting by encouraging appropriate surveillance and management of CVH.

Predicting mortality among patients with liver cirrhosis in electronic health records with machine learning.

OBJECTIVE: Liver cirrhosis is a leading cause of death and effects millions of people in the United States. Early mortality prediction among patients with cirrhosis might give healthcare providers more opportunity to effectively treat the condition. We hypothesized that laboratory test results and other related diagnoses would be associated with mortality in this population. Our another assumption was that a deep learning model could outperform the current Model for End Stage Liver disease (MELD) score in predicting mortality. MATERIALS AND METHODS: We utilized electronic health record data from 34,575 patients with a diagnosis of cirrhosis from a large medical center to study associations with mortality. Three time-windows of mortality (365 days, 180 days and 90 days) and two cases with different number of variables (all 41 available variables and 4 variables in MELD-NA) were studied. Missing values were imputed using multiple imputation for continuous variables and mode for categorical variables. Deep learning and machine learning algorithms, i.e., deep neural networks (DNN), random forest (RF) and logistic regression (LR) were employed to study the associations between baseline features such as laboratory measurements and diagnoses for each time window by 5-fold cross validation method. Metrics such as area under the receiver operating curve (AUC), overall accuracy, sensitivity, and specificity were used to evaluate models. RESULTS: Performance of models comprising all variables outperformed those with 4 MELD-NA variables for all prediction cases and the DNN model outperformed the LR and RF models. For example, the DNN model achieved an AUC of 0.88, 0.86, and 0.85 for 90, 180, and 365-day mortality respectively as compared to the MELD score, which resulted in corresponding AUCs of 0.81, 0.79, and 0.76 for the same instances. The DNN and LR models had a significantly better f1 score compared to MELD at all time points examined. CONCLUSION: Other variables such as alkaline phosphatase, alanine aminotransferase, and hemoglobin were also top informative features besides the 4 MELD-Na variables. Machine learning and deep learning models outperformed the current standard of risk prediction among patients with cirrhosis. Advanced informatics techniques showed promise for risk prediction in patients with cirrhosis.

Effects of habitat types on the dynamic changes in allocation in carbon and nitrogen storage of vegetation-soil system in sandy grasslands: How habitat types affect C and N allocation?

The progressively restoration of degraded vegetation in semiarid and arid desertified areas undoubtedly formed different habitat types. The most plants regulate their growth by fixing carbon with their energy deriving from photosynthesis; carbon (C) and nitrogen (N) play the crucial role in regulating plant growth, community structure, and function in the vegetation restoration progress. However, it is still unclear how habitat types affect the dynamic changes in allocation in C and N storage of vegetation-soil system in sandy grasslands. Here, we investigated plant community characteristics and soil properties across three successional stages of habitat types: semi-fixed dunes (SFD), fixed dunes (FD), and grasslands (G) in 2011, 2013, and 2015. We also examined the C and N concentrations of vegetation-soil system and estimated their C and N storage. The C and N storage of vegetation system, soil, and vegetation-soil system remarkably increased from SFD to G. The litter C and N storage in SFD, N storage of vegetation system in SFD, and N storage of soil and vegetation-soil system in FD increased from 2011 to 2015, while aboveground plant C and N storage of FD were higher in 2011 than in 2013 and 2015. Most of C and N were sequestered in soil in the vegetation restoration progress. These results suggest that the dynamic changes in allocation in C and N storage in vegetation-soil systems varied with habitat types. Our study highlights that SFD has higher N sequestration rate in vegetation, while FD has the considerably N sequestration rate in the soil.

Drought of early time in growing season decreases community aboveground biomass, but increases belowground biomass in a desert steppe.

BACKGROUND: Increasing drought induced by global climate changes is altering the structure and function of grassland ecosystems. However, there is a lack of understanding of how drought affects the trade-off of above- and belowground biomass in desert steppe. We conducted a four-year (2015-2018) drought experiment to examine the responses of community above-and belowground biomass (AGB and BGB) to manipulated drought and natural drought in the early period of growing season (from March to June) in a desert steppe. We compared the associations of drought with species diversity (species richness and density), community-weighted means (CWM) of five traits, and soil factors (soil Water, soil carbon content, and soil nitrogen content) for grass communities. Meanwhile, we used the structural equation modeling (SEM) to elucidate whether drought affects AGB and BGB by altering species diversity, functional traits, or soil factors. RESULTS: We found that manipulated drought affected soil water content, but not on soil carbon and nitrogen content. Experimental drought reduced the species richness, and species modified the CWM of traits to cope with a natural drought of an early time in the growing season. We also found that the experimental and natural drought decreased AGB, while natural drought increased BGB. AGB was positively correlated with species richness, density, CWM of plant height, and soil water. BGB was negatively correlated with CWM of plant height, CWM of leaf dry matter content, and soil nitrogen content, while was positively correlated with CWM of specific leaf area, CWM of leaf nitrogen content, soil water, and soil carbon content. The SEM results indicated that the experimental and natural drought indirectly decreased AGB by reducing species richness and plant height, while natural drought and soil nitrogen content directly affected BGB. CONCLUSIONS: These results suggest that species richness and functional traits can modulate the effects of drought on AGB, however natural drought and soil nitrogen determine BGB. Our findings demonstrate that the long-term observation and experiment are necessary to understand the underlying mechanism of the allocation and trade-off of community above-and belowground biomass.

Using electronic health records and claims data to identify high-risk patients likely to benefit from palliative care.

OBJECTIVES: Palliative care has been demonstrated to have positive effects for patients, families, health care providers, and health systems. Early identification of patients who are likely to benefit from palliative care would increase opportunities to provide these services to those most in need. This study predicted all-cause mortality of patients as a surrogate for patients who could benefit from palliative care. STUDY DESIGN: Claims and electronic health record (EHR) data for 59,639 patients from a large integrated health care system were utilized. METHODS: A deep learning algorithm-a long short-term memory (LSTM) model-was compared with other machine learning models: deep neural networks, random forest, and logistic regression. We conducted prediction analyses using combined claims data and EHR data, only claims data, and only EHR data, respectively. In each case, the data were randomly split into training (80%), validation (10%), and testing (10%) data sets. The models with different hyperparameters were trained using the training data, and the model with the best performance on the validation data was selected as the final model. The testing data were used to provide an unbiased performance evaluation of the final model. RESULTS: In all modeling scenarios, LSTM models outperformed the other 3 models, and using combined claims and EHR data yielded the best performance. CONCLUSIONS: LSTM models can effectively predict mortality by using a combination of EHR data and administrative claims data. The model could be used as a promising clinical tool to aid clinicians in early identification of appropriate patients for palliative care consultations.

Predicting cardiovascular health trajectories in time-series electronic health records with LSTM models.

BACKGROUND: Cardiovascular disease (CVD) is the leading cause of death in the United States (US). Better cardiovascular health (CVH) is associated with CVD prevention. Predicting future CVH levels may help providers better manage patients' CVH. We hypothesized that CVH measures can be predicted based on previous measurements from longitudinal electronic health record (EHR) data. METHODS: The Guideline Advantage (TGA) dataset was used and contained EHR data from 70 outpatient clinics across the United States (US). We studied predictions of 5 CVH submetrics: smoking status (SMK), body mass index (BMI), blood pressure (BP), hemoglobin A1c (A1C), and low-density lipoprotein (LDL). We applied embedding techniques and long short-term memory (LSTM) networks - to predict future CVH category levels from all the previous CVH measurements of 216,445 unique patients for each CVH submetric. RESULTS: The LSTM model performance was evaluated by the area under the receiver operator curve (AUROC): the micro-average AUROC was 0.99 for SMK prediction; 0.97 for BMI; 0.84 for BP; 0.91 for A1C; and 0.93 for LDL prediction. Model performance was not improved by using all 5 submetric measures compared with using single submetric measures. CONCLUSIONS: We suggest that future CVH levels can be predicted using previous CVH measurements for each submetric, which has implications for population cardiovascular health management. Predicting patients' future CVH levels might directly increase patient CVH health and thus quality of life, while also indirectly decreasing the burden and cost for clinical health system caused by CVD and cancers.

Machine Learning Outcome Prediction in Dilated Cardiomyopathy Using Regional Left Ventricular Multiparametric Strain.

The clinical presentation of idiopathic dilated cardiomyopathy (IDCM) heart failure (HF) patients who will respond to medical therapy (responders) and those who will not (non-responders) is often similar. A machine learning (ML)-based clinical tool to identify responders would prevent unnecessary surgery, while targeting non-responders for early intervention. We used regional left ventricular (LV) contractile injury patterns in ML models to identify IDCM HF non-responders. MRI-based multiparametric strain analysis was performed in 178 test subjects (140 normal subjects and 38 IDCM patients), calculating longitudinal, circumferential, and radial strain over 18 LV sub-regions for inclusion in ML analyses. Patients were identified as responders based upon symptomatic and contractile improvement on medical therapy. We tested the predictive accuracy of support vector machines (SVM), logistic regression (LR), random forest (RF), and deep neural networks (DNN). The DNN model outperformed other models, predicting response to medical therapy with an area under the receiver operating characteristic curve (AUC) of 0.94. The top features were longitudinal strain in (1) basal: anterior, posterolateral and (2) mid: posterior, anterolateral, and anteroseptal sub-regions. Regional contractile injury patterns predict response to medical therapy in IDCM HF patients, and have potential application in ML-based HF patient care.