Conventional Scores Research Articles

Clinical and Genomic-Based Decision Support System to Define the Optimal Timing of Allogeneic Hematopoietic Stem Cell Transplantation in Patients with Myelodysplastic Syndromes (MDS)

Purpose. Allogeneic hematopoietic stem cell transplantation (HSCT) is the only potentially curative treatment for patients with myelodysplastic syndromes (MDS). Several issues must be considered when evaluating the benefits and risks of HSCT for patients with MDS, with the timing of transplantation during the disease course being a crucial question. Recently the integration of genomic screening (by Molecular International Prognostic Scoring System, IPSS-M) into patient's assessment has resulted into a significant improvement in predicting clinical outcomes with respect to the conventional prognostic score (Revised IPSS, IPSS-R), including better stratification of post-HSCT outcome. Here, we aimed to develop and validate a Decision Support System to define the optimal timing of HSCT in MDS patients based on clinical and genomic information as provided by IPSS-M vs conventional IPSS-R. Patients and methods. We studied a retrospective, international cohort of 8,326 patients with MDS for whom complete set of clinical and molecular information was available. Patients were stratified into a learning and validation cohorts (65%, n= 5411 vs. 35%, n=2915). The decision-strategy analysis was divided into two parts. First, we developed a multi-state model to estimate transition hazards between different disease strata. Age and IPSS-R/IPSS-M were used as explanatory variables. The time elapsed between the diagnosis of MDS and HSCT was also considered as covariate in the models, which were further adjusted for disease-modifying therapy (if any). Next, we developed a Semi-Markov multi-state decision model based on microsimulation to compare various transplantation strategies according to the timing of the procedure, conditionally on covariates of interest. The microsimulation strategy simulates a hypothetical randomized clinical trial where subjects are randomized to receive HSCT at different time points. Results were used to estimate the average survival time over an 8-year time horizon (Restricted Mean Survival Time, RMST) for each combination of covariates, and then determine the optimal transplantation policy by compared the different strategies. Results. When considering patient life expectancy for patients who received a HSCT (Restricted Mean Survival Time, RMST) in the learning cohort, in both scenarios of IPSS-R and IPSS-M based transplantation policies, early disease stage was associated with a better clinical outcome. Indeed, life expectancy after transplantation was higher for younger vs elderly patients ( Figure 1 and 2). Under an IPSS-R based policy, patients with low-risk disease clearly benefit from a delayed transplantation policy across all age groups, while in patients belonging to intermediate, high/very high risk categories immediate transplantation was associated with a prolonged life expectancy, regardless of age. ( Figure 1) Considering an IPSS-M based policy, patients with either low and moderate-low risk benefit from a delayed transplantation policy (across all age groups), while in those belonging to moderate-high, high and very high risk categories immediate transplantation was associated with a prolonged life expectancy (across all age groups, Figure 2). Importantly, all these results were confirmed in the validation cohort, thus providing evidence for their reliability and generalizability. Modelling decision analysis on IPSS-M vs. original IPSS-R changed transplantation policy in a significant proportion of patients, resulting in a gain-in-life expectancy under an IPSS-M based policy across all age groups (P<0.001). Specifically, 19% of candidates to be immediately transplanted under an IPSS-R based policy would benefit from a delayed strategy under an IPSS-M based policy, while 21% of candidates to delayed transplantation under an IPSS-R based policy, would benefit from immediate HSCT under an IPSS-M based policy. We have created a prototype web application for the MDS Transplantation Decision Support System, that allows clinicians to define the best timing for HSCT starting from individual patient demographics, IPSS-R, and IPSS-M information. Conclusion. We provided evidence for the clinical relevance of including genomic features into the transplantation decision making process, specifically regarding the optimal timing of HSCT, allowing personalizing of the hazards and effectiveness of HSCT in patients with MDS.

Going beyond carbon: An "Earth system impact" score to better capture corporate and investment impacts on the earth system

Corporations are responsible for a significant portion of observed impacts on the Earth system, including green-house gas (GHG) emissions, but also water extraction, landuse change and other pressures on nature. These nature-related impacts are essential to consider and capture because they have local impacts on a range of ecosystem functions on which companies and economies depend, but they also fundamentally affect our ability to mitigate and adapt to a changing climate. Furthermore, climate, land and water interact and affect each other in various ways, such that climate change can be exacerbated by degraded ecosystems, which in turn are dependent on water.This paper tests a novel metric developed to capture corporate Earth system impact (ESI) beyond merely direct GHG emissions and explores how such a tool could be used to improve assessments of corporate environmental impacts and support decisions on where to direct public and private investments. We use the mining sector as a test case to illustrate the applicability of the ESI score and examine the impact of the the five largest (by market cap) mining companies in the precious metal mining sector and the top five in the non-precious metal mining sector. We find that many of the mining assets have non-negligible impacts on land and water, and we show that the ESI metric identifies a different set of asset for targeted action than conventional carbon intensity scores would do.

Development of a plasma-free amino acid-based risk score for the incidence of cardiovascular diseases in a general population: The Nagahama study

Levels of circulating amino acids (AAs) have been suggested to be associated with cardiovascular diseases (CVDs). This study aimed to develop a plasma-free amino acid (PFAA)-based CVD risk-prediction model in a general population. The study participants consisted of 9220 community residents (mean age, 53.2 years; standard deviation, 13.3 years). Circulating levels of 19 PFAAs were measured via high-performance liquid chromatography/electrospray ionization mass spectrometry. The incidence of CVDs was determined by reviewing participants' clinical records. The prediction model was developed using the Cox proportional hazards model with the brute force variable selection and then cross-validated. During the 8.5-year follow-up, 220 CVD events were observed. Six AAs (alanine, citrulline, glycine, histidine, serine, and tyrosine) were identified as components of the prediction model, of which the C-index was 0.72. The association between the fourth quartile of the risk score calculated using the prediction model and the CVD events was independent of conventional risk factors (adjusted hazard ratio [HR], 1.9; 95% confidence interval, 1.1-3.3). When examining crude relationships between conventional risk factors and the PFAA-based risk score by subgroup analyses, the association was significant for most subpopulations, men [crude HR=6.4 (2.0-20.2)] and women [crude HR=4.9 (2.6-9.3)], and individuals with [crude HR=4.7 (2.5-8.9)] and without [crude HR=7.2 (2.7-18.9)] lifestyle-related diseases, but not for older (≥70 years) participants [crude HR=3.3 (0.8-13.5)]. The risk score successfully identified at-risk individuals [HR=2.1 (1.2-3.5)] from participants who were classified as low risk by a conventional CVD risk score. The PFAA-based risk score predicted CVD events independently of conventional risk factors.

The Price of Explainability in Machine Learning Models for 100-Day Readmission Prediction in Heart Failure: Retrospective, Comparative, Machine Learning Study.

Sensitive and interpretable machine learning (ML) models can provide valuable assistance to clinicians in managing patients with heart failure (HF) at discharge by identifying individual factors associated with a high risk of readmission. In this cohort study, we delve into the factors driving the potential utility of classification models as decision support tools for predicting readmissions in patients with HF. The primary objective of this study is to assess the trade-off between using deep learning (DL) and traditional ML models to identify the risk of 100-day readmissions in patients with HF. Additionally, the study aims to provide explanations for the model predictions by highlighting important features both on a global scale across the patient cohort and on a local level for individual patients. The retrospective data for this study were obtained from the Regional Health Care Information Platform in Region Halland, Sweden. The study cohort consisted of patients diagnosed with HF who were over 40 years old and had been hospitalized at least once between 2017 and 2019. Data analysis encompassed the period from January 1, 2017, to December 31, 2019. Two ML models were developed and validated to predict 100-day readmissions, with a focus on the explainability of the model's decisions. These models were built based on decision trees and recurrent neural architecture. Model explainability was obtained using an ML explainer. The predictive performance of these models was compared against 2 risk assessment tools using multiple performance metrics. The retrospective data set included a total of 15,612 admissions, and within these admissions, readmission occurred in 5597 cases, representing a readmission rate of 35.85%. It is noteworthy that a traditional and explainable model, informed by clinical knowledge, exhibited performance comparable to the DL model and surpassed conventional scoring methods in predicting readmission among patients with HF. The evaluation of predictive model performance was based on commonly used metrics, with an area under the precision-recall curve of 66% for the deep model and 68% for the traditional model on the holdout data set. Importantly, the explanations provided by the traditional model offer actionable insights that have the potential to enhance care planning. This study found that a widely used deep prediction model did not outperform an explainable ML model when predicting readmissions among patients with HF. The results suggest that model transparency does not necessarily compromise performance, which could facilitate the clinical adoption of such models.

Patient-reported outcomes in German vasculitis patients-Data from the KOBRA quality project

Current data on the care of patients with vasculitis in Germany are scarce. Patient-reported outcome (PRO) questionnaires can capture aspects of the disease that escape conventional scores for disease activity, remission, and damage. For this reason, the Association of Rheumatological Acute Care Clinics (VRA) initiated adata analysis as part of the KOBRA quality project, the results of which are presented here. Patients with vasculitis of vessels of any size or with polymyalgia rheumatica were included. The prospective survey included data on demographics, disease, pain, treatment, follow-up and satisfaction at the time of inpatient admission, discharge and follow-up after 2.5months. All patients completed the AAV-PRO and EQ-5D-3L questionnaires on admission and follow-up. In this study 420 patients were recruited and follow-up data were available from 302. On average, improvements were documented in all 5dimensions of the EQ-5D, with the strongest effects in self-care and coping with activities of daily living. In the AAV-PRO, highly significant differences were seen in the domains systemic symptoms and physical functioning. Satisfaction with medical and nursing treatment was very high and did not correlate with pain level or with the AAV-PRO measures. Under zreatment patient-reported outcomes improve at least partially in vasculitis patients. Satisfaction with medical treatment quality is independent of these outcomes.

A comparison of random survival forest and Cox regression for prediction of mortality in patients with hemorrhagic stroke

ObjectiveTo evaluate RSF and Cox models for mortality prediction of hemorrhagic stroke (HS) patients in intensive care unit (ICU).MethodsIn the training set, the optimal models were selected using five-fold cross-validation and grid search method. In the test set, the bootstrap method was used to validate. The area under the curve(AUC) was used for discrimination, Brier Score (BS) was used for calibration, positive predictive value(PPV), negative predictive value(NPV), and F1 score were combined to compare.ResultsA total of 2,990 HS patients were included. For predicting the 7-day mortality, the mean AUCs for RSF and Cox regression were 0.875 and 0.761, while the mean BS were 0.083 and 0.108. For predicting the 28-day mortality, the mean AUCs for RSF and Cox regression were 0.794 and 0.649, while the mean BS were 0.129 and 0.174. The mean AUCs of RSF and Cox versus conventional scores for predicting patients’ 7-day mortality were 0.875 (RSF), 0.761 (COX), 0.736 (SAPS II), 0.723 (OASIS), 0.632 (SIRS), and 0.596 (SOFA), respectively.ConclusionsRSF provided a better clinical reference than Cox. Creatine, temperature, anion gap and sodium were important variables in both models.

E-HEART score: A novel scoring system for undifferentiated chest pain in the emergency department.

Cardiovascular disease is the leading cause of death worldwide. As there is an increase in the global burden of ischemic heart disease, there are multiple scoring systems established in the emergency department (ED) to risk stratify and manage acute coronary syndrome (ACS) in patients with chest pain. The objective of this study was to integrate point-of-care echo into the existing history, electrocardiogram, age, risk factors, and troponin (HEART) score and evaluate a novel scoring system, the echo HEART (E-HEART) score in risk stratification of patients presenting with undifferentiated chest pain to the ED. The E-HEART Score was also compared with existing traditional scoring systems for risk-stratifying acute chest pain. A diagnostic accuracy study involving 250 patients with chest pain at the ED of a single tertiary care teaching hospital in India was conducted. The emergency physicians assessed the E-HEART score after integrating their point-of-care echo/focused echo findings into the conventional HEART score on presentation. The primary endpoint was the occurrence of major adverse cardiovascular events (MACE) within 4 weeks of initial presentation. The accuracy of the E-HEART score was compared with other conventional risk stratification scoring systems such as the thrombolysis in myocardial infarction (TIMI), history, electrocardiogram, age, and risk factors, Troponin Only Manchester ACS (T-MACS), and HEART scores. A total of 250 patients with a median age of 53 years (42.25-63.00) were part of the study. Low E-HEART scores (values 0-3) were calculated in 121 patients with no occurrence of MACE in this category. Eighty-one patients with moderate E-HEART scores (4-6) were found to have 30.9% MACE. In 48 patients with high E-HEART scores (values 7-11), MACE occurred in 97.9%. The area under receiver operating characteristics (AUROC) of E-HEART score is 0.992 (95% confidence interval: 0.98-0.99), which is significantly higher than AUROC values for HEART (0.978), TIMI (0.889), T-MACS (0.959), and HEAR (0.861), respectively (P < 0.0001). At a cutoff of E-HEART score >6, it accurately predicted ACS with a sensitivity of 92% and a specificity of 99% with a diagnostic accuracy of 97%. The E-HEART score gives the clinician a quick and accurate forecast of outcomes in undifferentiated chest pain presenting to the ED. Low E-HEART scores (0-3) have an extremely low probability for short-term MACE and may aid in faster disposition from the ED. The elevated risk of MACE in patients with high E-HEART scores (7-11) may facilitate more aggressive workup measures and avoid disposition errors. E-HEART is an easily adaptable scoring system with improved accuracy compared to conventional scoring systems.

Refining decomposition stage estimation: A study of seasonal insect composition and carcass weight change

AbstractIn this study, we explore the interplay of seasonal insect assemblages and carcass weight change to refine the estimation of minimum post‐mortem interval (PMImin), a key variable in forensic investigations. By integrating these parameters, we shed light on their combined influence on decomposition stages, thus paving the way for a more robust method of PMImin estimation. Our approach relies on rigorous field studies examining pig carcasses across spring, summer and autumn, with comprehensive measurements of carcass weight and detailed documentation of insect succession. We observed distinct insect assemblages and weight‐change trends characteristic of each season, underlining the influence of seasonal variability on decomposition dynamics. To further enhance the reliability of our model, we combined the novel parameter of carcass weight change with the conventional total body score (TBS) method, ensuring a more objective assessment of decomposition stages. Although our study indicates promising advancements, it also acknowledges the limitations of using a single pig carcass per season, suggesting future research should incorporate larger sample sizes.

A Cautionary Note on Using Propensity Score Calibration to Control for Unmeasured Confounding Bias When the Surrogacy Assumption Is Absent.

Conventional propensity score methods encounter challenges when unmeasured confounding is present, as it becomes impossible to accurately estimate the gold-standard propensity score when data on certain confounders are unavailable. Propensity score calibration (PSC) addresses this issue by constructing a surrogate for the gold-standard propensity score under the surrogacy assumption. This assumption posits that the error-prone propensity score, based on observed confounders, is independent of the outcome when conditioned on the gold-standard propensity score and the exposure. However, this assumption implies that confounders cannot directly impact the outcome and that their effects on the outcome are solely mediated through the propensity score. This raises concerns regarding the applicability of PSC in practical settings where confounders can directly affect the outcome. While PSC aims to target a conditional treatment effect by conditioning on a subject's unobservable propensity score, the causal interest in the latter case lies in a conditional treatment effect conditioned on a subject's baseline characteristics. Our analysis reveals that PSC is generally biased unless the effects of confounders on the outcome and treatment are proportional to each other. Furthermore, we identify 2 sources of bias: 1) the noncollapsibility of effect measures, such as the odds ratio or hazard ratio and 2) residual confounding, as the calibrated propensity score may not possess the properties of a valid propensity score.

Early warning system enables accurate mortality risk prediction for acute gastrointestinal bleeding admitted to intensive care unit.

Acute gastrointestinal (GI) bleeding are potentially life-threatening conditions. Early risk stratification is important for triaging patients to the appropriate level of medical care and intervention. Patients admitted to intensive care unit (ICU) has a high mortality, but risk tool is scarce for these patients. This study aimed to develop and validate a risk score to improve the prognostication of death at the time of patient admission to ICU. We developed and internally validated a nomogram for mortality in patients with acute GI bleeding from the eICU Collaborative Research Database (eICU-CRD), and externally validated it in patients from the Medical Information Mart for Intensive Care III database (MIMIC-III) and Wuhan Tongji Hospital. The performance of the model was assessed by examining discrimination (C-index), calibration (calibration curves) and usefulness (decision curves). 4750 patients were included in the development cohort, with 1184 patients in the internal validation cohort, 1406 patients in the MIMIC-III validation cohort, and 342 patients in the Tongji validation cohort. The nomogram, which incorporated ten variables, showed good calibration and discrimination in the training and validation cohorts, yielded C-index ranged from 0.832 (95%CI 0.811-0.853) to 0.926 (95CI% 0.905-0.947). The nomogram-defined high-risk group had a higher mortality than the low-risk group (44.8% vs. 3.5%, P < 0.001; 41.4% vs 3.1%, P < 0.001;53.6% vs 7.5%, P < 0.001; 38.2% vs 4.2%, P < 0.001). The model performed better than the conventional Glasgow-Blatchford score, AIMS65 and the newer Oakland and Sengupta scores for mortality prediction in both the derivation and validation cohorts concerning discrimination and usefulness. Our nomogram is a reliable prognostic tool that might be useful to identify high-risk acute GI bleeding patients admitted to ICU.

MAFLD fibrosis score: Using routine measures to identify advanced fibrosis in metabolic-associated fatty liver disease.

Early screening may prevent fibrosis progression in metabolic-associated fatty liver disease (MAFLD). We developed and validated MAFLD fibrosis score (MFS) for identifying advanced fibrosis (≥F3) among MAFLD patients. This cross-sectional, multicentre study consecutively recruited MAFLD patients receiving tertiary care (Malaysia as training cohort [n = 276] and Hong Kong and Wenzhou as validation cohort [n = 431]). Patients completed liver biopsy, vibration-controlled transient elastography (VCTE), and clinical and laboratory assessment within 1 week. We used machine learning to select 'highly important' predictors of advanced fibrosis, followed by backward stepwise regression to construct MFS formula. MFS was composed of seven variables: age, body mass index, international normalised ratio, aspartate aminotransferase, gamma-glutamyl transpeptidase, platelet count, and history of type 2 diabetes. MFS demonstrated an area under the receiver-operating characteristic curve of 0.848 [95% CI 0.800-898] and 0.823 [0.760-0.886] in training and validation cohorts, significantly higher than aminotransferase-to-platelet ratio index (0.684 [0.603-0.765], 0.663 [0.588-0.738]), Fibrosis-4 index (0.793 [0.735-0.854], 0.737 [0.660-0.814]), and non-alcoholic fatty liver disease fibrosis score (0.785 [0.731-0.844], 0.750 [0.674-0.827]) (DeLong's test p < 0.05). MFS could include 92.3% of patients using dual cut-offs of 14 and 15, with a correct prediction rate of 90.4%, resulting in a larger number of patients with correct diagnosis compared to other scores. A two-step MFS-VCTE screening algorithm demonstrated positive and negative predictive values and overall diagnostic accuracy of 93.4%, 89.5%, and 93.2%, respectively, with only 4.0% of patients classified into grey zone. MFS outperforms conventional non-invasive scores in predicting advanced fibrosis, contributing to screening in MAFLD patients.

Quantifying acute kidney injury in an Ischaemia-Reperfusion Injury mouse model using deep-learning-based semantic segmentation in histology.

This study focuses on ischaemia-reperfusion injury (IRI) in kidneys, a cause of acute kidney injury (AKI) and end-stage kidney disease (ESKD). Traditional kidney damage assessment methods are semi-quantitative and subjective. This study aims to use a convolutional neural network (CNN) to segment murine kidney structures after IRI, quantify damage via CNN-generated pathological measurements, and compare this to conventional scoring. The CNN was able to accurately segment the different pathological classes, such as Intratubular casts and Tubular necrosis, with an F1 score of over 0.75. Some classes, such as Glomeruli and Proximal tubules, had even higher statistical values with F1 scores over 0.90. The scoring generated based on the segmentation approach statistically correlated with the semiquantitative assessment (Spearman's rank correlation coefficient=0.94). The heatmap approach localised the intratubular necrosis mainly in the outer stripe of the outer medulla, while the tubular casts were also present in more superficial or deeper portions of the cortex and medullary areas. This study presents a CNN model capable of segmenting multiple classes of interest, including acute IRI-specific pathological changes, in a whole mouse kidney section and can provide insights into the distribution of pathological classes within the whole mouse kidney section.

Development of Electronic Health Record-Based Machine Learning Models to Predict Barrett's Esophagus and Esophageal Adenocarcinoma Risk.

Screening for Barrett's esophagus (BE) is suggested in those with risk factors, but remains underutilized. BE/esophageal adenocarcinoma (EAC) risk prediction tools integrating multiple risk factors have been described. However, accuracy remains modest (area under the receiver-operating curve [AUROC] ≤0.7), and clinical implementation has been challenging. We aimed to develop machine learning (ML) BE/EAC risk prediction models from an electronic health record (EHR) database. The Clinical Data Analytics Platform, a deidentified EHR database of 6 million Mayo Clinic patients, was used to predict BE and EAC risk. BE and EAC cases and controls were identified using International Classification of Diseases codes and augmented curation (natural language processing) techniques applied to clinical, endoscopy, laboratory, and pathology notes. Cases were propensity score matched to 5 independent randomly selected control groups. An ensemble transformer-based ML model architecture was used to develop predictive models. We identified 8,476 BE cases, 1,539 EAC cases, and 252,276 controls. The BE ML transformer model had an overall sensitivity, specificity, and AUROC of 76%, 76%, and 0.84, respectively. The EAC ML transformer model had an overall sensitivity, specificity, and AUROC of 84%, 70%, and 0.84, respectively. Predictors of BE and EAC included conventional risk factors and additional novel factors, such as coronary artery disease, serum triglycerides, and electrolytes. ML models developed on an EHR database can predict incident BE and EAC risk with improved accuracy compared with conventional risk factor-based risk scores. Such a model may enable effective implementation of a minimally invasive screening technology.

Voxel Extraction and Multiclass Classification of Identified Brain Regions across Various Stages of Alzheimer's Disease Using Machine Learning Approaches.

This study sought to investigate how different brain regions are affected by Alzheimer's disease (AD) at various phases of the disease, using independent component analysis (ICA). The study examines six regions in the mild cognitive impairment (MCI) stage, four in the early stage of Alzheimer's disease (AD), six in the moderate stage, and six in the severe stage. The precuneus, cuneus, middle frontal gyri, calcarine cortex, superior medial frontal gyri, and superior frontal gyri were the areas impacted at all phases. A general linear model (GLM) is used to extract the voxels of the previously mentioned regions. The resting fMRI data for 18 AD patients who had advanced from MCI to stage 3 of the disease were obtained from the ADNI public source database. The subjects include eight women and ten men. The voxel dataset is used to train and test ten machine learning algorithms to categorize the MCI, mild, moderate, and severe stages of Alzheimer's disease. The accuracy, recall, precision, and F1 score were used as conventional scoring measures to evaluate the classification outcomes. AdaBoost fared better than the other algorithms and obtained a phenomenal accuracy of 98.61%, precision of 99.00%, and recall and F1 scores of 98.00% each.

Acute Vascular Events: Cellular and Molecular Mechanisms

Cardiovascular diseases (CVDs) are the leading cause of death worldwide. An estimated 17.9 million individuals died from CVDs in 2019, representing 32% of all global deaths. Of these deaths, 85% were due to heart attack and stroke. Cardiometabolic risks, such as hypertension, excess weight, obesity, type 2 diabetes, and vascular diseases, contribute significantly to the progression of coronary artery disease. Known sequelae of events that lead to these cardiometabolic diseases include oxidative stress, inflammation, development of dysfunction of vascular adipose tissue, altered blood pressure and blood lipids, altered glucose metabolism, hardening of the arteries, endothelial dysfunction, development of atherosclerotic plaques, and activation of platelet and coagulation pathways. The Framingham Heart Study Group has developed a Risk Score that estimates the risk of developing heart disease in a 10-year period. This group of experts has developed mathematical functions for predicting clinical coronary disease events. These prediction capabilities are derived by assigning weights to major CVD risk factors such as sex, age, blood pressure, total cholesterol, low-density lipoprotein, high-density lipoprotein cholesterol, smoking behavior, and diabetes status. Currently, there is a growing interest in the use of artificial intelligence and machine learning applications. AI-based mimetic pattern-based algorithms seem to be better than the conventional Framingham Risk Score, in predicting clinical events related to CVDs. However, there are limitations to these applications as they do not have access to data on the specific factors that trigger acute vascular events, such as heart attack and stroke. This overview briefly discusses some salient cellular and molecular mechanisms involved in precipitating thrombotic conditions. Further improvements in emerging technologies will provide greater opportunities for patient selection and treatment options. Several clinical studies have demonstrated that most CVDs can be prevented by addressing behavioral risk factors such as tobacco use, unhealthy diet and obesity, physical activity, and harmful use of alcohol. Early detection and better management of the modifiable risks seem to be the only way to reduce, reverse, or prevent these diseases.

Hybridizing machine learning in survival analysis of cardiac PET/CT imaging

BackgroundMachine Learning (ML) allows integration of the numerous variables delivered by cardiac PET/CT, while traditional survival analysis can provide explainable prognostic estimates from a restricted number of input variables. We implemented a hybrid ML-and-survival analysis of multimodal PET/CT data to identify patients who developed myocardial infarction (MI) or death in long-term follow up.MethodsData from 739 intermediate risk patients who underwent coronary CT and selectively stress 15O-water-PET perfusion were analyzed for the occurrence of MI and all-cause mortality. Images were evaluated segmentally for atherosclerosis and absolute myocardial perfusion through 75 variables that were integrated through ML into an ML-CCTA and an ML-PET score. These scores were then modeled along with clinical variables through Cox regression. This hybridized model was compared against an expert interpretation-based and a calcium score-based model.ResultsCompared with expert- and calcium score-based models, the hybridized ML-survival model showed the highest performance (CI .81 vs .71 and .64). The strongest predictor for outcomes was the ML-CCTA score.ConclusionPrognostic modeling of PET/CT data for the long-term occurrence of adverse events may be improved through ML imaging score integration and subsequent traditional survival analysis with clinical variables. This hybridization of methods offers an alternative to traditional survival modeling of conventional expert image scoring and interpretation.

Role of MRI Diffusion Tensor Imaging in Early Diagnosis of Cervical Spondylotic Myelopathy (CSM)

Abstract Background and Purpose Cervical spondylotic myelopathy (CSM) is a common degenerative disorder of the cervical spine and is considered as the most common cause of spinal cord dysfunction in older individuals worldwide. Assessing the severity of the disease objectively remains a challenge and there is still remaining controversy in terms of the optimal timing and indications for surgical intervention. The conventional MRI is the gold standard for radiographic evaluation of the spinal cord; however, it cannot assess the underlying microstructural deficits of the spinal cord and it has a limited application in determining prognosis and recovery of the disease [1]. We sought to assess the utility of MR diffusion tensor imaging as an imaging technique in clinically suggested CSM patients by obtaining the microstructural parameters (Fractional anisotropy; FA and apparent diffusion coefficient; ADC) in the cervical spinal cord segments. Patient &amp; methods our study was a prospective cross-sectional study including 40 patients with clinical manifestations of cervical spondylotic myelopathy. The patients were sorted according to the European myelopathy score into three categories as follows: grade 1 (with mild symptoms), grade 2 (with moderate symptoms) and grade 3 (with severe symptoms). All the patients were investigated with a 1.5 T MRI unit acquiring conventional MRI and DTI sequences. FA and ADC values at each spinal segment from C2/C3 to C6/C7 were generated and measured using manual drawing of multiple ROIs within the cervical cord in the sagittal and axial combined anatomical and color-coded images. For each of these cases, the patient’s age, gender, DTI parameters at each spinal segment level, maximum compression level, conventional T2WI characteristics and EMS score were recorded, analyzed and compared. Results The study included 23 females (57.5%) and 17 males (42.5%). Their age ranged from 30 to 72 years with a mean age of 47.3 ± 9.79 years. The study revealed a significant difference in the mean FA and ADC values between stenotic and non-stenotic segments of the cervical spinal cord. There was a significant correlation between the European myelopathy score and DTI parameters (FA and ADC). FA and ADC values performed better in recognizing myelopathic changes in patients who were classified as grade 1 according to EMS compared to conventional MRI T2WI which performed very poorly in recognizing myelopathy changes in the same patients, therefore DTI indices were more sensitive in the detection of CSM patients compared to conventional MRI especially in early myelopathic stages. T2 weighted images were highly significant in recognizing myelopathy changes in patients who were classified as grade 3 according to EMS. The CSM patients had significant FA decrease and ADC increase at the most compressed part of the cervical spinal cord and there was a significant correlation between the number of patients who had myelopathic changes by DTI parameters and the severity of disease. The FA sensitivity and specificity in recognition of myelopathic changes were 73.5% and 83.3% respectively and the ADC sensitivity and specificity were 100% and 70.6% respectively. Conclusion In conclusion, diffusion tensor imaging enhances the efficacy and accuracy of MRI in the diagnosis of cervical spondylotic myelopathy. Therefore, it can be used as a non-invasive modality for detection of the spondylotic myelopathy changes in the early stages helping to decide the appropriate timing of decompression surgery before attaining the irreversible chronic changes. Abbreviations ADC: Apparent diffusion coefficient; CSM: Cervical spondylotic myelopathy; DTI: Diffusion tensor imaging; FA: Fractional Anisotropy; EMS: European myelopathy score

The Potential of Gut Microbiota in Prediction of Stroke-Associated Pneumonia.

Stroke-associated pneumonia (SAP) is a common stroke complication, and the changes in the gut microbiota composition may play a role. Our study aimed to evaluate the predictive ability of gut microbiota for SAP. Acute ischemic stroke patients were prospectively enrolled and divided into two groups based on the presence or absence of SAP. The composition of gut microbiota was characterized by the 16S RNA Miseq sequencing. The gut microbiota that differed significantly between groups were incorporated into the conventional risk scores, the Acute Ischemic Stroke-Associated Pneumonia Score (AIS-APS), and the Age, Atrial fibrillation, Dysphagia, Sex, Stroke Severity Score (A2DS2). The predictive performances were assessed in terms of the area under the curve (AUC), the Net Reclassification Improvement (NRI), and the Integrated Discrimination Improvement (IDI) indices. A total of 135 patients were enrolled, of whom 43 had SAP (31%). The short-chain fatty acids (SCFAs)-producing bacteria, such as Bacteroides, Fusicatenibacter, and Butyricicoccus, were decreased in the SAP group. The integrated models showed better predictive ability for SAP (AUC = 0.813, NRI = 0.333, p = 0.052, IDI = 0.038, p = 0.018, for AIS-APS; AUC = 0.816, NRI = 0.575, p < 0.001, IDI = 0.043, p = 0.007, for A2DS2) in comparison to the differential genera (AUC = 0.699) and each predictive score (AUCAISAPS = 0.777; AUCA2DS2 = 0.777). The lower abundance of SCFAs-producing gut microbiota after acute ischemic stroke was associated with SAP and may play a role in SAP prediction.

A comparison of diagnostic performance of word-list and story recall tests for biomarker-determined Alzheimer’s disease

ABSTRACT Background Wordlist and story recall tests are routinely employed in clinical practice for dementia diagnosis. In this study, our aim was to establish how well-standard clinical metrics compared to process scores derived from wordlist and story recall tests in predicting biomarker determined Alzheimer’s disease, as defined by CSF ptau/Aβ42 ratio. Methods Data from 295 participants (mean age = 65 ± 9.) were drawn from the University of Wisconsin – Madison Alzheimer’s Disease Research Center (ADRC) and Wisconsin Registry for Alzheimer’s Prevention (WRAP). Rey’s Auditory Verbal Learning Test (AVLT; wordlist) and Logical Memory Test (LMT; story) data were used. Bayesian linear regression analyses were carried out with CSF ptau/Aβ42 ratio as outcome. Sensitivity analyses were carried out with logistic regressions to assess diagnosticity. Results LMT generally outperformed AVLT. Notably, the best predictors were primacy ratio, a process score indexing loss of information learned early during test administration, and recency ratio, which tracks loss of recently learned information. Sensitivity analyses confirmed this conclusion. Conclusions Our study shows that story recall tests may be better than wordlist tests for detection of dementia, especially when employing process scores alongside conventional clinical scores.

Development and validation of a novel nomogram for predicting recurrent atrial fibrillation after cryoballoon ablation.

Few studies have explored the use of machine learning models to predict the recurrence of atrial fibrillation (AF) in patients who have undergone cryoballoon ablation (CBA). We aimed to explore the risk factors for the recurrence of AF after CBA in order to construct a nomogram that could predict this risk. Data of 498 patients who had undergone CBA at Ruijin Hospital, Shanghai Jiaotong University School of Medicine, were retrospectively collected. Factors such as clinical characteristics and biophysical parameters during the CBA procedure were collected for the selection of variables. Scores for all the biophysical factors-such as time to pulmonary vein isolation (TTI) and balloon temperature-were calculated to enable construction of the model, which was then calibrated and compared with the risk scores. A 36-month follow-up showed that 177 (35.5%) of the 489 patients experienced AF recurrence. The left atrial volume, TTI, nadir cryoballoon temperature, and number of unsuccessful freezes were related to the recurrence of AF (P < .05). The area under the curve (AUC) of the nomogram's time-dependent receiver operating characteristic curve was 77.6%, 71.6%, and 71.0%, respectively, for the 1-, 2-, and 3-year prediction of recurrence in the training cohort and 77.4%, 74.7%, and 68.7%, respectively, for the same characteristics in the validation cohort. Calibration and data on the nomogram's clinical effectiveness showed it to be accurate for the prediction of recurrence in both the training and validation cohorts as compared with established risk scores. Biophysical parameters such as TTI and cryoballoon temperature have a great impact on AF recurrence. The predictive accuracy for recurrence of our nomogram was superior to that of conventional risk scores.