Mean Area Under Curve Research Articles

Predicting atrial fibrillation within 7 years follow-up using machine learning models

Abstract Background The prevalence of atrial fibrillation (AF), the most common type of cardiac arrhythmia, has increased globally. AF is also the leading cardiac factor for several cardiovascular and cerebrovascular events. Early prediction and treatment of AF prevents major cardiovascular and cerebrovascular events. Purpose Our objective is to predict AF risk within 7-years follow-up using only twelve basic factors and applying machine learning (ML) models. Methods Data from the German epidemiological trial on ankle brachial index (getABI) were utilized in this work. GetABI includes 6,457 patients with mean age of 72.53 years. 3,985 individuals were assessed for AF of which 393 patients suffered from AF at 7-years follow-up. The feature selection was initially done manually, ending up in 45 features. The 10 best features were extracted from this dataset according to the SHAP (SHapley Additive exPlanations) values. The final features were body mass index, sensitive CRP, cholesterol, LDL-cholesterol, troponin I, NT-pro-BNP, glomerular filtration rate, ratio of triglycerides and HDL-cholesterol, gamma-GT, heart rate, including age and sex as major factors. The handling of missing values was achieved by SimpleImputer, estimating the mean values. XGBoost and Random Forest predicted AF. Logistic Regression was also used as a baseline model. The data were split into training and test set and ML models were applied for 100 iterative runs, extracting results with mean values. Results XGBoost was the most accurate model, predicting the AF with 71.19 % mean accuracy, 66.37 % mean sensitivity and 75.61 % mean specificity. Figure 1 depicts the performance of all models. Additionally, Receiver Operating Characteristic Curve (ROC) and Area Under Curve (AUC) for each model are presented in Figure 2. XGBoost also outperformed all the others models, achieving the highest mean AUC value, equal to 70.99 %. Contrariwise, logistic regression had the lowest overall performance. Conclusion ML models, specifically XGBoost, achieved accurate AF prediction in 3,985 individuals utilizing only twelve clinical-routine data.Evaluation results of each used model.ROC curve and AUC value of each model.

Screening of lung cancer serum biomarkers based on Boruta-shap and RFC-RFECV algorithms

ObjectiveThis study aimed to identify a set of serum miRNAs as potential biomarkers for lung cancer diagnosis using algorithmic approaches. MethodsSerum miRNA expression data from lung cancer patients and non-tumor controls were obtained. The top six miRNAs were selected using Boruta-shap and RFC-RFECV algorithms. A Gaussian Naive Bayes (NB) classifier was trained and evaluated using cross-validation, ROC curve analysis, and evaluation metrics. ResultsSix miRNAs (hsa-miRNA-144, hsa-miRNA-107, hsa-miRNA-484, hsa-miRNA-103, hsa-miRNA-26b, and hsa-miRNA-641) were identified as feature genes. The NB classifier achieved an area under curve (AUC) of 0.8966 and a mean AUC of 0.88 in cross-validation. Accuracy, recall, and F1 scores exhibited promising results, with an accuracy of 82%. In the validation set, the AUC values for the NB and SVC classifiers were 0.9345 and 0.9423, respectively, with a mean AUC of 0.95 in cross-validation. The classifiers demonstrated an accuracy of 89% in diagnosing lung cancer. ConclusionThis study identified a panel of six serum miRNAs with potential as non-invasive biomarkers for lung cancer diagnosis. These miRNAs show promise in providing sensitive and specific tools for detecting lung cancer. SignificanceLung cancer is one of the top cancers worldwide, threatening the health and lives of tens of thousands of people. miRNA is a biomarker, which can be used as a potential clinical tool for diagnosis and prognosis of cancer patients. Therefore, the use of multiple miRNAs to construct diagnostic models may be one of the future methods of accurate diagnosis of lung cancer. In this study, we used the Boruta-shap and RFC-RFECV algorithms to automatically identify and extract characteristic miRNAs highly associated with lung cancer, thereby establishing an accurate classifier for the diagnosis of lung cancer with characteristic miRNAs.

Nomogram for predicting bleeding events in nonvalvular atrial fibrillation patients receiving rivaroxaban: A retrospective study.

To construct a bleeding events prediction model of nonvalvular atrial fibrillation (NVAF) patients receiving rivaroxaban. We conducted a retrospective cohort study in patients with NVAF who received rivaroxaban from June 2017 to March 2019. Demographic information and clinical characteristics were obtained from the electronic medical system. Univariate analysis was used to find the primary predictive factors of bleeding events in patients receiving rivaroxaban. Multiple analysis was conducted to screen the primary independent predictive factors selected from the univariate analysis. Finally, the independent influencing factors were applied to build a prediction model by using R software; then, a nomogram was established according to the selected variables visually, and the sensitivity and specificity of the model was evaluated. Twelve primary predictive factors were selected by univariate analysis from 46 variables, and multivariate analysis showed that older age, higher prothrombin time (PT) values, history of heart failure and stroke were independent risk factors of bleeding events. The area under curve (AUC) for this novel nomogram model was 0.828 (95% CI: 0.763-0.894). The mean AUC over 10-fold stratified cross-validation was 0.787, and subgroup analysis validation also showed a satisfied AUC. In addition, the decision curve analysis showed that the PT in combination with CHA2DS2-VASc and HASBLED was more practical and accurate for predicting bleeding events than using CHA2DS2-VASc and HASBLED alone. PT in combination with CHA2DS2-VASc and HASBLED could be considered as a more practical and accurate method for predicting bleeding events in patients taking rivaroxaban.

Harnessing the power of artificial intelligence in predicting all-cause mortality in transcatheter aortic valve replacement: a systematic review and meta-analysis.

In recent years, the use of artificial intelligence (AI) models to generate individualised risk assessments and predict patient outcomes post-Transcatheter Aortic Valve Implantation (TAVI) has been a topic of increasing relevance in literature. This study aims to evaluate the predictive accuracy of AI algorithms in forecasting post-TAVI mortality as compared to traditional risk scores. Following the Preferred Reporting Items for Systematic Reviews and Meta-analyses for Systematic Reviews (PRISMA) standard, a systematic review was carried out. We searched four databases in total-PubMed, Medline, Embase, and Cochrane-from 19 June 2023-24 June, 2023. From 2,239 identified records, 1,504 duplicates were removed, 735 manuscripts were screened, and 10 studies were included in our review. Our pooled analysis of 5 studies and 9,398 patients revealed a significantly higher mean area under curve (AUC) associated with AI mortality predictions than traditional score predictions (MD: -0.16, CI: -0.22 to -0.10, p < 0.00001). Subgroup analyses of 30-day mortality (MD: -0.08, CI: -0.13 to -0.03, p = 0.001) and 1-year mortality (MD: -0.18, CI: -0.27 to -0.10, p < 0.0001) also showed significantly higher mean AUC with AI predictions than traditional score predictions. Pooled mean AUC of all 10 studies and 22,933 patients was 0.79 [0.73, 0.85]. AI models have a higher predictive accuracy as compared to traditional risk scores in predicting post-TAVI mortality. Overall, this review demonstrates the potential of AI in achieving personalised risk assessment in TAVI patients. This systematic review and meta-analysis was registered under the International Prospective Register of Systematic Reviews (PROSPERO), under the registration name "All-Cause Mortality in Transcatheter Aortic Valve Replacement Assessed by Artificial Intelligence" and registration number CRD42023437705. A review protocol was not prepared. There were no amendments to the information provided at registration. https://www.crd.york.ac.uk/, PROSPERO (CRD42023437705).

UJI AKTIVITAS EKSTRAK ETANOL DAUN BIDARA LAUT (Strychnos ligustrina Biume) TERHADAP PENYEMBUHAN LUKA BAKAR PADA KELINCI (Oryctolagus cuniculus)

Burns are damage or loss of skin tissue caused by contact with heat sources such as hot water, fire, chemicals, electricity, friction and radiation. Sea lotus (Strychnos ligustrina Blume) is known to contain compounds that have the potential to heal burns such as flavonoids, saponins and tannins. This study aims to determine the activity of ethanol extract of bidara de mer leaves on burn healing as a function of reduction in burn diameter, wound discoloration, edema and crust formation. This study used 5 Oryctolagus cuniculus that received 4 types of treatment after being induced by hot plate, namely wound I (bioplacenton® gel positive control), wound 2 (HPMC 5% negative control), wound 3 (extract concentration 10%) and wound 4 (concentration of extract 10%) 25%). Apply the test sample to the burns twice a day and observe the condition of the wound daily. Burn healing as a function of wound diameter was determined based on the AUC (Area Under Curve) value and statistically analyzed using SPSS 21.0. The test results showed that the mean AUC value at 10% and 25% concentrations was 0.495 cm/day and 0.532 cm/day, respectively. The results of the statistical analysis showed that the ethanol extract of bidara sea leaves gave significant results compared to the negative control. treatment (HPMC). The conclusion obtained is that the ethanolic extract of sea bidara leaves accelerates the healing process of burns in Oryctolagus cuniculus.

Co prescription of anti-acid therapy reduces the bioavailability of mycophenolate mofetil in systemic sclerosis patients: A crossover trial

ObjectiveMycophenolate mofetil (MMF) is an effective treatment option for interstitial lung disease (ILD) in systemic sclerosis (SSc). Many patients require co-administration of proton pump inhibitors (PPI) or H2 receptor blockers (HRB) because of various gastrointestinal (GI) manifestations in SSc. Co-treatment with PPI or HRB have shown to reduce serum drug levels in post-transplant patients. We wanted to see if there is a similar phenomenon for Mycophenolate in SSc. MethodsTwenty SSc patients, who were on a stable dose of MMF (1.5-3 g) underwent a sequential cross over study with MMF alone in the first month, followed by co-treatment with Ranitidine and then Esomeprazole in the second and third month respectively. Estimation of 12-hour area under curve (AUC) of Mycophenolic Acid (MPA) levels and total GI score were calculated at the end of each month and compared between the treatment arms. [Trial registration: CTRI/2020/06/025,939] ResultsCo-administration of esomeprazole was associated with 32.7% (mean difference = 22.28 μg h ml–1) reduction in mean AUC MPA, whereas ranitidine caused a reduction of 21.97% (mean difference = 14.93 μg h ml–1) in MPA AUC when compared to MMF without anti-acid therapies. The addition of ranitidine or esomeprazole resulted in significant reduction in the total GI score. ConclusionCo-administration of PPI or HRB can significantly reduce the bioavailability of MMF in patients with SSc. To avoid therapeutic failure of MMF drug level monitoring is essential when these agents are co-prescribed with MMF.

Abstract GS5-04: Identification of symptoms that are associated with irAEs in the I-SPY clinical trial

Abstract Background. Immunotherapy has emerged as an important component of neoadjuvant therapy for some patients with breast cancer (BC). As a result, immune-related adverse events (irAEs) are increasing and have effects on both short and long term symptoms significantly impacting patient quality of life. BC patients may develop new conditions including arthralgias, gastrointestinal issues, endocrinopathies, and fatigue during or after cancer therapy that may be acute or long-lasting in nature. Monitoring for early onset and severity of symptoms, and adjusting treatment and symptom management could optimize therapy for a particular patient, maximizing potential efficacy while mitigating toxicity. We sought to identify patient demographic characteristics and symptom patterns associated with risk for development of irAEs in the context of a randomized trial for patients with early-stage high-risk breast cancer. Methods. I-SPY2 is a multi-center, phase 2 trial using response-adaptive randomization for high-risk early-stage women with BC. The study population for this analysis includes enrolled patients receiving combinations of experimental immunotherapy and chemotherapy. Groups considered for statistical comparisons included those that developed an irAE versus those that did not develop an irAE up until the surgery timepoint. In I-SPY adverse events are documented through the Common Terminology Criteria for Adverse Events (CTCAEv5.0). Hypothyroidism, adrenal insufficiency, and pneumonitis were the irAEs considered in this study. A chi-square test was used to assess associations between race and ethnicity (White, Asian, Black, non-Hispanic) and irAEs. One-way ANOVA was used to evaluate the association between age (&amp;gt;50 vs &amp;lt; 50) and irAEs. 33 symptoms reported at CTCAE grade 2 or higher were included in the analyses and a symptom burden score was calculated using area under curve (AUC) which combined the duration of each symptom between baseline and week 6 of treatment, and grade of adverse event. Regularized regression using leave-one out cross validation was used to evaluate early symptoms (as quantified by the symptom burden score) as predictors, and irAEs as surrogate responses. Results. Out of 461 patients, percentages of patients wth irAEs of interest included hypothyroidism (13%), adrenal insufficiency (9%), and pneumonitis (4%). Demographic information was available for 333 patients, of which 270 (81%) were White, 23 (7%) were Asian, 37 (11%) were African American (AA) and 278 (17%) were non-Hispanic. There were proportionately higher number of white patients that developed hypothyroidism than non-white patients (35 of 265 (13%) vs 2 of 63 (3%), P &amp;lt; 0.04). Pneumonitis was more common in patients over 50 years old than under 50 years old (P &amp;lt; 0.02). Symptoms that were most commonly reported up to week 6 of treatment among patients who developed an irAE included: diarrhea (36%), fatigue (15%), dizziness (12%) and shortness of breath (SOB) (11%). Symptoms associated with the development of hypothyroidism included fatigue (15%, mean AUC=11.8 vs 5.8 for those that did not develop irAE), SOB (11%, 4.3 vs 2.8), and blurry vision (1%, 1.0 vs 0.12). Development of adrenal insufficiency was associated with early reports of diarrhea (36%, 19.0 vs 10.5), SOB (11%, 7.8 vs 2.6), joint pain (3%, 2.29 vs 0.58), decreased appetite (3%, 3.55 vs 0.91), and constipation (1%, 3.6 vs 0.02). No significant early symptoms emerged for pneumonitis due to a limited number of events. Conclusion. Our study utilizes an analysis framework that is aimed to determine symptom clusters that predict the development of irAEs. We describe specific symptoms presenting early with the development of hypothyroidism and adrenal insufficiency, in recognition of allowing physicians to be more diligent in active and post treatment monitoring. Citation Format: Amrita Basu, Saumya Umashankar, Michelle Melisko, Ruixiao Lu, Hongmei Yu, Christina Yau, Smita Asare, Maria Pitsouni, Rebecca A. Shatsky, Claudine Isaacs, Angela DeMichele, Dawn Hershman, Rita Nanda, Mi-Ok Kim, Laura J. Esserman, Hope Rugo. Identification of symptoms that are associated with irAEs in the I-SPY clinical trial [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr GS5-04.

Impact on renal function of daily and on-demand HIV pre-exposure prophylaxis in the ANRS-PREVENIR study.

To assess the impact on the estimated glomerular filtration rate (eGFR) of different tenofovir disoproxil/emtricitabine dosing regimens for HIV pre-exposure prophylaxis (PrEP). We included in the study individuals with baseline eGFR > 50 mL/min/1.73 m2 who initiated PrEP in the ongoing ANRS-PREVENIR PrEP cohort. We retrospectively classified PrEP users in three groups: 'on-demand' (reported at ≥75% of study visits), 'daily' (≥75% of study visits) or 'switches'. We compared the area under curve (AUC) of the eGFR variation from baseline (ΔeGFR) between groups using analysis of covariance, and assessed factors associated with a negative AUC of ΔeGFR. From May 2017 to October 2020, 1253 PrEP-naïve participants (98% of MSM) were included in the study with a median follow-up of 22 months. 499 (40%), 494 (39%) and 260 (21%) users were in the group daily, on-demand and switches, respectively, for a median number of pills taken per week of 6, 1.7 and 4. The mean AUC of the ΔeGFR was -1.09 mL/min/1.73 m2 in the daily PrEP group, -0.69 mL/min/1.73 m2 in the switches group and +0.18 mL/min/1.73 m2 with on-demand PrEP. In a model adjusted on baseline age and eGFR, the AUC of the ΔeGFR was significantly higher with on-demand PrEP compared to daily PrEP (P = 0.037). Independent factors associated with a negative AUC of ΔeGFR were a daily PrEP regimen, a switches regimen, an age > 40 years and a baseline eGFR≥90 mL/min/1.73 m². On-demand PrEP dosing had a smaller impact on eGFR evolution than daily PrEP, but the difference was not clinically relevant.

A prediction model for dosimetric-based lung adaptive radiotherapy.

Anatomical changes occurred during the treatment course of radiation therapy for lung cancer patients may introduce clinically unacceptable dosimetric deviations from the planned dose. Adaptive radiotherapy (ART) can compensate these dosimetric deviations in subsequent treatments via plan adaption. Determining whether and when to trigger plan adaption during the treatment course is essential to the effectiveness and efficiency of ART. In this study, we aimed to develop a prediction model as an auxiliary decision-making tool for lung ART to identify the patients with intrathoracic anatomical changes that would potentially benefit from the plan adaptions during the treatment course. Seventy-one pairs of weekly cone-beam computer tomography (CBCT) and planning CT (pCT) from 17 advanced non-small cell lung cancer patients were enrolled in this study. To assess the dosimetric impacts brought by anatomical changes observed on each CBCT, dose distribution of the original treatment plan on the CBCT anatomy was calculated on a virtual CT generated by deforming the corresponding pCT to the CBCT and compared to that of the original plan. A replan was deemed needed for the CBCT anatomy once the recalculated dose distribution violated our dosimetric-based trigger criteria. A three-dimensional region of significant anatomical changes (region of interest, ROI) between each CBCT and the corresponding pCT was identified, and 16 morphological features of the ROI were extracted. Additionally, eight features from the overlapped volume histograms (OVHs) of patient anatomy were extracted for each patient to characterize the patient-specific anatomy. Based on the 24 extracted features and the evaluated replanning needs of the pCT-CBCT pairs, a nonlinear supporting vector machine was used to build a prediction model to identify the anatomical changes on CBCTs that would trigger plan adaptions. The most relevant features were selected using the sequential backward selection (SBS) algorithm and a shuffling-and-splitting validation scheme was used for model evaluation. Fifty-five CBCT-pCT pairs were identified of having an ROI, among which 21 CBCT anatomies required plan adaptions. For these 21 positive cases, statistically significant improvements in the sparing of lung, esophagus and spinal cord were achieved by plan adaptions. A high model performance of 0.929 AUC (area under curve) and 0.851 accuracy was achieved with six selected features, including five ROI shape features and one OVH feature. Without involving the OVH features in the feature selection process, the mean AUC and accuracy of the model significantly decreased to 0.826 and 0.779, respectively. Further investigation showed that poor prediction performance with AUC of 0.76 was achieved by the univariate model in solving this binary classification task. We built a prediction model based on the features of patient anatomy and the anatomical changes captured by on-treatment CBCT imaging to trigger plan adaption for lung cancer patients. This model effectively associated the anatomical changes with the dosimetric impacts for lung ART. This model can be a promising tool to assist the clinicians in making decisions for plan adaptions during the treatment courses.

Contrast between traditional and machine learning algorithms based on a urine culture predictive model: a multicenter retrospective study in patients with urinary calculi.

BackgroundQuick and accurate identification of urinary calculi patients with positive urinary cultures is critical to the choice of the treatment strategy. Predictive models based on machine learning algorithms provide a new way to solve this problem. This study aims to determine the predictive value of machine learning algorithms using a urine culture predictive model based on patients with urinary calculi.MethodsData were collected from four clinical centers in the period of June 2016, to May 2019. 2,054 cases were included in the study. The dataset was randomly split into ratios of 5:5, 6:4, and 7:3 for model construction and validation. Predictive models of urine culture outcomes were constructed and validated by logistic regression, random forest, adaboost, and gradient boosting decision tree (GBDT) models. Each ratio’s construction and verification were repeated five times independently for cross-validation. The Matthews correlation coefficient (MMC), F1-score, receiver operating characteristic (ROC) curve with the area under curve (AUC) was used to evaluate the performance of each prediction model. The additive net reclassification index (NRI) and absolute NRI were used to assess the predictive capabilities of the models.ResultsFour prediction models of urinary culture results in patients with urinary calculi were constructed. The mean AUCs of the logistic regression, random forest, adaboost, and GBDT models were 0.761 (95% CI: 0.753–0.770), 0.790 (95% CI: 0.782–0.798), 0.779 (95% CI: 0.766–0.791), and 0.831 (95% CI: 0.823–0.840), respectively. Moreover, the average MMC and F1-score of GBDT model was 0.460 and 0.588, which was improved compared to logistic regression model of 0.335 and 0.501. The additive NRI and absolute NRI of the GBDT and logistic regression models were 0.124 (95% CI: 0.106–0.142) and 0.065 (95% CI: 0.060–0.069), respectively.ConclusionsOur results indicate that machine learning algorithms may be useful tools for urine culture outcome prediction in patients with urinary calculi because they exhibit superior performance compared with the logistic regression model.

23. ZOE-50 and ZOE-70 Placebo Groups Data Shows that Burden of Pain Associated with Herpes Zoster Interferes with Activities of Daily Living

Abstract Background Adults ≥ 50 years of age (YOA) are at increased risk of herpes zoster (HZ), a condition which can cause long-term pain and discomfort. In this analysis, we assessed the burden of pain associated with HZ and its interference with activities of daily living (ADL) in patients ≥ 50 YOA. Methods ZOE-50 (NCT01165177) and ZOE-70 (NCT01165229) were phase III, observer-blind, placebo-controlled, 1:1 randomized studies in adults ≥ 50 YOA (ZOE-50) and ≥ 70 YOA (ZOE-70) who received 2 doses of the adjuvanted recombinant zoster vaccine or placebo, 2 months apart. To correctly evaluate HZ pain, the analysis was performed only on the placebo groups data of ZOE-50 (≥ 50 YOA) and pooled ZOE-50/70 (≥ 70 YOA pooled data) in the modified total vaccinated cohort (mTVC, primary population for efficacy analysis) HZ-confirmed cases. HZ pain and interference with ADL was assessed by the Zoster Brief Pain Inventory (ZBPI) instrument completed daily by patients for the first 28 days and then weekly until resolution. Time to resolution of clinically significant pain was analyzed using Kaplan Meier methods. We estimated the cumulative area under curve (AUC) of the ZBPI worst pain score and the ZBPI ADL score up to 182 days post-HZ rash onset. A high AUC reflects a higher severity/longer duration of pain. Results Overall, 254 patients ≥ 50 YOA and 284 patients ≥ 70 YOA were included in the mTVC HZ-confirmed cases for the ZOE-50 and ZOE-70 pooled analysis, respectively. In HZ patients ≥ 50 YOA, 94.6% reported any pain (ZBPI pain score &amp;gt; 0), 87.6% clinically significant pain (ZBPI pain score ≥ 3) and 65.1% severe pain (ZBPI pain score ≥ 7). Similarly, in HZ patients ≥ 70 YOA, 93.2% reported any pain, 90.9% clinically significant pain and 68.4% severe pain. It was estimated that 11.6% and 18.3% of patients aged ≥ 50 and ≥ 70 YOA, respectively, had clinically significant pain 3 months after the onset of HZ. The mean AUC at 182 days was 137.24 (≥ 50 YOA) and 190.68 (≥ 70 YOA), for the ZBPI worst pain score and 92.75 (≥ 50 YOA) and 130.89 (≥ 70 YOA), for the ZBPI ADL score. Conclusion Analysis of data provided by patients with confirmed HZ shows that the burden of HZ pain is high and is associated with interference on patients’ ADL. Funding GlaxoSmithKline Biologicals SA Acknowledgment M Maior/S Hulsmans provided medical writing/editorial support (Modis c/o GSK) Disclosures Eliazar Sabater Cabrera, MSc, GSK (Employee, Shareholder) Desmond Curran, PhD, The GSK group of companies (Employee, Shareholder) Sean Matthews, MSc, GSK (Independent Contractor) Céline Boutry, PhD, GSK (Employee) Nicolas Lecrenier, Ing, PhD, The GSK group of companies (Employee, Shareholder) Anthony L. Cunningham, F.A.H.M.S., MD, M.B.B.S., B. Med. Sci. (Hons), F.R.A.C.P., F.R.C.P.A., F.A.S.M., GSK group of companies (Grant/Research Support, Advisor or Review Panel member, Speaker’s Bureau)

Epidemiological study of snakebite cases in Sikkim: Risk modeling with regard to the habitat suitability of common venomous snakes.

BackgroundSnakebite envenoming is listed as category ‘A’ Neglected Tropical Disease. To achieve the target of WHO (World Health Organization) 2019, it becomes necessary to understand various attributes associated with snakebite including community awareness, improvisation of medical facilities and to map the potential distribution of venomous snakes responsible for the bite. Hence this study is conducted in Sikkim, India to understand the epidemiology of snakebite in Sikkim. The potential distribution and risk mapping of five common venomous snakes are done for effective management of snakebite cases.Methods and findingsThe snakebite cases registered in six district hospitals and four PHCs (Primary Health Centers) of Sikkim were collected from the year 2011 to 2018. Community survey was also conducted to supplement the data. Ecological Niche Modeling (ENM) was performed to predict the potential habitat of five common venomous snakes of Sikkim. The risk modeling of snakebite cases was done at the level of Gram Panchayat Unit (GPU) using Geographically Weighted Regression (GWR) and Ordinary Linear Square (OLS) model. We found higher number of male victims inflicted with snakebite envenomation. The potential distribution of the five venomous snakes showed satisfactory mean AUC (Area under Curve) value. Both the models showed significant positive association of snakebite cases with habitat suitability of the venomous snakes. Hospital data revealed no death cases whereas community data reported 24 deaths.ConclusionsDeath from snakebite reflected in community data but not in hospital data strongly indicates the people’s belief in traditional medicine. Though people of Sikkim have rich traditional knowledge, in case of snakebite traditional practices may be ineffective leading to loss of life. Sensitizing people and improving medical facilities along with proper transport facilities in rural areas might significantly reduce the snakebite casualties in the state.

Role of Internal Ca2+ Stores in KCa Channel Feedback Control of Human Uterine Contractions

Background Minute-to-minute regulation of intracellular calcium concentrations is essential to allow a laboring uterus to phasically contract and relax in a manner required for delivery. Cytosolic calcium levels increase through influx of calcium through L-type calcium channels and are amplified by SR calcium release through inositol 1, 4, 5-triphosphate receptor (IP3R) channels and calcium induced calcium release. A specific class of small conductance calcium-activated potassium (KCa2.3) channels has been strongly implicated in the negative feedback control of membrane potential and intracellular calcium levels. Uterine contractions can be suppressed by amplifying this calcium-dependent feedback system. CyPPA, a positive modulator of KCa2.3, reduces contraction amplitudes in human and rodent tissue. Previous studies of endothelial and neuronal cells suggest focused Ca2+ release from internal stores in close proximity to the plasma membrane, may play an important role in KCa channel control. Indeed, close proximity and interaction of endoplasmic reticulum Ca2+ release channels (e.g. IP3Rs) and KCa channels may regulate intracellular calcium in a variety of tissues. This study aims to assess whether internal stores contribute to KCa2.3 dependent feedback regulation of contraction in the human myometrium. Methods Myograph studies were performed on human uterine tissues collected from non-pregnant patients undergoing hysterectomy. Uterine strips were sutured end-to-end to form rings, mounted onto myograph pins, and allowed to develop spontaneous phasic contractions. Concentration-dependent suppression of contractions with the positive KCa2.3 modulator CyPPA (3 μM, 10 μM, 30 μM) was assessed following depletion of internal calcium stores (using the SERCA inhibitor cyclopiazonic acid, CPA) or following treatment with the vehicle DMSO (controls). Results Utilizing myographical recordings and LabChart software, contraction area under curve (AUC) was calculated and normalized to time. Addition of 3 μM CyPPA resulted in a control group mean AUC of 1.287 mN/min (± 0.417) and 2.261 mN/min (± 1.338) for the treatment group [n=5]. The mean AUC with 10 μM CyPPA was 1.129 mN/min (± 0.571) for the control group and 2.164 mN/min (± 1.387) for the treatment group. After 30 μM CyPPA the mean AUC for the control group was 0.734 mN/min (± 0.471) and 1.863 mN/min (± 1.364) for the treatment group. Our group found that CPA pretreated uterine tissue had a significantly higher AUC than the control group (p<0.5), after addition of 30 μM CyPPA. Further analysis demonstrated no significant difference with CyPPA concentrations of 3 and 10 μM, but did reveal a trend of reduced AUC in the control group compared to the treatment group. These findings suggest blocking SERCA and inhibiting SR reuptake of calcium may decrease CyPPAs ability to diminish contractions, especially at higher concentrations. Conclusions Phasic contractions of isolated human uterine strips vary considerably with respect to amplitude and frequency. Overall, the findings suggest internal calcium store release may play an important role in moderating uterine contractions and augmenting KCa2.3 channel suppression of uterine contractility. Extended studies will assess whether this mechanism can be exploited as an effective tocolytic strategy.

Non-invasive wearable seizure detection using long–short-term memory networks with transfer learning

Objective. The detection of seizures using wearable devices would improve epilepsy management, but reliable detection of seizures in an ambulatory environment remains challenging, and current studies lack concurrent validation of seizures using electroencephalography (EEG) data. Approach. An adaptively trained long–short-term memory deep neural network was developed and trained using a modest number of seizure data sets from wrist-worn devices. Transfer learning was used to adapt a classifier that was initially trained on intracranial electroencephalography (iEEG) signals to facilitate classification of non-EEG physiological datasets comprising accelerometry, blood volume pulse, skin electrodermal activity, heart rate, and temperature signals. The algorithm’s performance was assessed with and without pre-training on iEEG signals and transfer learning. To assess the performance of the seizure detection classifier using long-term ambulatory data, wearable devices were used for multiple months with an implanted neurostimulator capable of recording iEEG signals, which provided independent electrographic seizure detections that were reviewed by a board-certified epileptologist. Main results. For 19 motor seizures from 10 in-hospital patients, the algorithm yielded a mean area under curve (AUC), a sensitivity, and an false alarm rate per day (FAR/day) of 0.98, 0.93, and 2.3, respectively. Additionally, for eight seizures with probable motor semiology from two ambulatory patients, the classifier achieved a mean AUC of 0.97 and an FAR of 2.45 events/day at a sensitivity of 0.9. For all seizure types in the ambulatory setting, the classifier had a mean AUC of 0.82 with a sensitivity of 0.47 and an FAR of 7.2 events/day. Significance. The performance of the algorithm was evaluated using motor and non-motor seizures during in-hospital and ambulatory use. The classifier was able to detect multiple types of motor and non-motor seizures, but performed significantly better on motor seizures.

Nighttime melatonin secretion and sleep architecture: different associations in perimenopausal and postmenopausal women

BackgroundSleep quality typically decreases after menopause, but the underlying mechanisms are poorly understood. Concentrations of melatonin are lower and its secretion profiles different before and after menopause. However, whether and how melatonin and sleep architecture are associated in women of different reproductive states have not been examined to date. MethodsOvernight serum melatonin samples were taken from 17 perimenopausal and 18 postmenopausal healthy women. Sleep quality was measured with all-night polysomnography recordings. ResultsMelatonin concentrations tended to be the lowest during NREM sleep, and were associated with higher odds of transitions from wake to NREM sleep. The curves of predicted overnight melatonin values from linear mixed models varied according to sleep phases (NREM, REM, Wake) in perimenopausal, but not in postmenopausal women. In perimenopause higher melatonin area under curve (AUC) correlated with higher slow-wave activity (p = 0.043), and higher minimum concentrations with shorter slow-wave sleep (SWS) latency (p = 0.029). In postmenopause higher mean and maximum melatonin concentrations and AUC correlated with lower SWS percentage (p = 0.044, p = 0.029, p = 0.032), and higher mean (p = 0.032), maximum (p = 0.032) and minimum (p = 0.037) concentrations with more awakenings from REM sleep. In the age- and BMI- adjusted regression models, the association between higher maximum (p = 0.046) melatonin concentration and lower SWS percentage remained. ConclusionsThe relationship between melatonin and sleep architecture differed in perimenopausal and postmenopausal women. After menopause, high melatonin concentrations were associated with worse sleep. Whether these different patterns are related to aging of the reproductive system, and to decrease in menopausal sleep quality, remains to be elucidated.

Machine learning is a valid method for predicting prehospital delay after acute ischemic stroke.

ObjectivesThis study aimed to identify the influencing factors associated with long onset‐to‐door time and establish predictive models that could help to assess the probability of prehospital delay in populations with a high risk for stroke.Materials and MethodsPatients who were diagnosed with acute ischemic stroke (AIS) and hospitalized between 1 November 2018 and 31 July 2019 were interviewed, and their medical records were extracted for data analysis. Two machine learning algorithms (support vector machine and Bayesian network) were applied in this study, and their predictive performance was compared with that of the classical logistic regression models after using several variable selection methods. Timely admission (onset‐to‐door time < 3 hr) and prehospital delay (onset‐to‐door time ≥ 3 hr) were the outcome variables. We computed the area under curve (AUC) and the difference in the mean AUC values between the models.ResultsA total of 450 patients with AIS were enrolled; 57 (12.7%) with timely admission and 393 (87.3%) patients with prehospital delay. All models, both those constructed by logistic regression and those by machine learning, performed well in predicting prehospital delay (range mean AUC: 0.800–0.846). The difference in the mean AUC values between the best performing machine learning model and the best performing logistic regression model was negligible (0.014; 95% CI: 0.013–0.015).ConclusionsMachine learning algorithms were not inferior to logistic regression models for prediction of prehospital delay after stroke. All models provided good discrimination, thereby creating valuable diagnostic programs for prehospital delay prediction.

Using Sensor Data to Detect Lameness and Mastitis Treatment Events in Dairy Cows: A Comparison of Classification Models.

The aim of this study was to develop classification models for mastitis and lameness treatments in Holstein dairy cows as the target variables based on continuous data from herd management software with modern machine learning methods. Data was collected over a period of 40 months from a total of 167 different cows with daily individual sensor information containing milking parameters, pedometer activity, feed and water intake, and body weight (in the form of differently aggregated data) as well as the entered treatment data. To identify the most important predictors for mastitis and lameness treatments, respectively, Random Forest feature importance, Pearson’s correlation and sequential forward feature selection were applied. With the selected predictors, various machine learning models such as Logistic Regression (LR), Support Vector Machine (SVM), K-nearest neighbors (KNN), Gaussian Naïve Bayes (GNB), Extra Trees Classifier (ET) and different ensemble methods such as Random Forest (RF) were trained. Their performance was compared using the receiver operator characteristic (ROC) area-under-curve (AUC), as well as sensitivity, block sensitivity and specificity. In addition, sampling methods were compared: Over- and undersampling as compensation for the expected unbalanced training data had a high impact on the ratio of sensitivity and specificity in the classification of the test data, but with regard to AUC, random oversampling and SMOTE (Synthetic Minority Over-sampling) even showed significantly lower values than with non-sampled data. The best model, ET, obtained a mean AUC of 0.79 for mastitis and 0.71 for lameness, respectively, based on testing data from practical conditions and is recommended by us for this type of data, but GNB, LR and RF were only marginally worse, and random oversampling and SMOTE even showed significantly lower values than without sampling. We recommend the use of these models as a benchmark for similar self-learning classification tasks. The classification models presented here retain their interpretability with the ability to present feature importances to the farmer in contrast to the “black box” models of Deep Learning methods.

Comparison of Various Abdominal Obesity Measures for Predicting Metabolic Syndrome, Diabetes, Nephropathy, and Dyslipidemia.

To evaluate abdominal volume index (AVI), body roundness index (BRI), body adiposity index (BAI), a body shape index (ABSI) and conicity index (C-Index) for differences in subjects with or without metabolic syndrome, diabetes, nephropathy, and dyslipidemia; and secondly, to evaluate the diagnostic performance through measuring area under curve (AUC) by ROC curve analysis for new and conventional obesity measures in diagnosing metabolic syndrome. Cross-sectional analytical study. PNS Hafeez Hospital, Islamabad, from January 2016 to December 2018. Baseline anthropometric measures including BMI, WHpR, WHtR, AVI, BRI, BAI, ABSI and C-Index were measured for 232 subjects along with measurement of various biochemical parameters. Differences among subjects with and without metabolic syndrome, diabetes, nephropathy, and groups based upon insulin resistance were noted. ROC curve analysis was utilised to measure AUC for all anthropometric measures for diagnosing metabolic syndrome. Pearson's correlation between obesity measures and lipid indices suggested highest correlation for AVI for most lipid indices followed by WHpR and WHtR. Mean AUC for obesity measures were greater than 0.80 for WHtR and AVI, followed by other parameters. The least AUC i.e. 0.320, was observed for ABSI. The differences between various anthropometric measures for groups based upon metabolic syndrome, diabetes, nephropathy, and insulin resistance remain variable indicating that each anthropometric index may depict a different aspect of the metabolic risk. WHtR and AVI showed the highest AUC to diagnose metabolic syndrome and were better associated with metabolic diseases.

A Machine Learning Approach to Classifying Self-Reported Health Status in a Cohort of Patients With Heart Disease Using Activity Tracker Data.

Constructing statistical models using personal sensor data could allow for tracking health status over time, thereby enabling the possibility of early intervention. The goal of this study was to use machine learning algorithms to classify patient-reported outcomes (PROs) using activity tracker data in a cohort of patients with stable ischemic heart disease (SIHD). A population of 182 patients with SIHD were monitored over a period of 12 weeks. Each subject received a Fitbit Charge 2 device to record daily activity data, and each subject completed eight Patient-Reported Outcomes Measurement Information Systems short form at the end of each week as a self-assessment of their health status. Two models were built to classify PRO scores using activity tracker data. The first model treated each week independently, whereas the second used a hidden Markov model (HMM) to take advantage of correlations between successive weeks. Retrospective analysis compared the classification accuracy of the two models and the importance of each feature. In the independent model, a random forest classifier achieved a mean area under curve (AUC) of 0.76 for classifying the physical function PRO. The HMM model achieved significantly better AUCs for all PROs (p < 0.05) other than Fatigue and Sleep Disturbance, with a highest mean AUC of 0.79 for the physical function-short form 10a. Our study demonstrates the ability of activity tracker data to classify health status over time. These results suggest that patient outcomes can be monitored in real time using activity trackers.

A review of the Rowland Universal Dementia Assessment Scale.

Objectives To explore the application, diagnostic accuracy and predictors effecting the performance of the Rowland Universal Dementia Assessment Scale (RUDAS) in older populations living outside Australia. Design A literature review was conducted using a simplified approach of thematic analysis. A comprehensive computerized search of peer reviewed published papers (2004–2017) was conducted in Academic Search Complete (EBSCO), Medline with Full-text, Pubmed and Google Scholar using the following search terms: Rowland Universal Dementia Assessment Scale, sensitivity and specificity. Inclusion criteria were papers in which the RUDAS was used with older populations outside Australia and written in English. Results Database and manual searching identified a total of 289 papers and after reviewing the abstracts, titles and full-text papers, 12 papers were included in this review. Outside the target populations of culturally and linguistically diverse communities in Australia, the RUDAS is applied internationally in Asia, Europe and Africa. The psychometric properties of the RUDAS were measured across these non-target population groups: sensitivity and specificity means, respectively were 80.9 (SD ±6.24; 95% CI) and 76.1 (SD ±12.2; 95% CI); RUDAS-MMSE mean correlation was 0.70 (SD ±0.14; 95% CI) and mean AUC (area under curve) for RUDAS was 0.83 (SD ±0.08; 95% CI). Conclusions Findings from this literature review showed that the RUDAS has strong psychometric properties across many population groups who are culturally and linguistically diverse in several countries. These findings provide support to undertake further research to evaluate the use of the RUDAS within culturally and linguistically diverse communities in other countries.