scholarly journals Natural Language Processing Enhances Prediction of Functional Outcome After Acute Ischemic Stroke

2021 ◽  
Author(s):  
Sheng‐Feng Sung ◽  
Chih‐Hao Chen ◽  
Ru‐Chiou Pan ◽  
Ya‐Han Hu ◽  
Jiann‐Shing Jeng
Keyword(s):  
Ischemic Stroke ◽  
Natural Language Processing ◽  
Natural Language ◽  
Acute Ischemic Stroke ◽  
Receiver Operating Characteristic ◽  
Functional Outcome ◽  
Language Processing ◽  
Operating Characteristic ◽  
Receiver Operating Characteristic Curves ◽  
Receiver Operating

Background Conventional prognostic scores usually require predefined clinical variables to predict outcome. The advancement of natural language processing has made it feasible to derive meaning from unstructured data. We aimed to test whether using unstructured text in electronic health records can improve the prediction of functional outcome after acute ischemic stroke. Methods and Results Patients hospitalized for acute ischemic stroke were identified from 2 hospital stroke registries (3847 and 2668 patients, respectively). Prediction models developed using the first cohort were externally validated using the second cohort, and vice versa. Free text in the history of present illness and computed tomography reports was used to build machine learning models using natural language processing to predict poor functional outcome at 90 days poststroke. Four conventional prognostic models were used as baseline models. The area under the receiver operating characteristic curves of the model using history of present illness in the internal and external validation sets were 0.820 and 0.792, respectively, which were comparable to the National Institutes of Health Stroke Scale score (0.811 and 0.807). The model using computed tomography reports achieved area under the receiver operating characteristic curves of 0.758 and 0.658. Adding information from clinical text significantly improved the predictive performance of each baseline model in terms of area under the receiver operating characteristic curves, net reclassification improvement, and integrated discrimination improvement indices (all P <0.001). Swapping the study cohorts led to similar results. Conclusions By using natural language processing, unstructured text in electronic health records can provide an alternative tool for stroke prognostication, and even enhance the performance of existing prognostic scores.

scholarly journals Coarse-Grained Sentiment Analysis Berbasis Natural Language Processing – Ulasan Hotel

2021 ◽  
Vol 10 (1) ◽  
pp. 41-48
Author(s):  
Warnia Nengsih ◽  
M. Mahrus Zein ◽  
Nazifa Hayati
Keyword(s):  
Natural Language Processing ◽  
Random Forest ◽  
Natural Language ◽  
Receiver Operating Characteristic ◽  
Sentiment Analysis ◽  
Language Processing ◽  
Operating Characteristic ◽  
Coarse Grained ◽  
Receiver Operating

Sentiment analysis adalah metode untuk memperoleh data dari berbagai platform yang tersedia di internet. Kemajuan teknologi memungkinkan mesin untuk mengenali suatu istilah yang dianggap sebagai opini positif maupun sebaliknya. Data-data dan opini tersebut berperan penting sebagai umpan balik produk, layanan, dan topik lainnya. Tanpa perlu memperoleh opini secara langsung dari masyarakat, pihak penyedia telah mendapatkan evaluasi yang penting guna mengembangkan diri. Bisnis perhotelan merupakan bidang yang terkait dengan jasa memberikan layanan pada pelanggan. Indikator keberlangsungan bisnis ini juga bergantung pada umpan balik pelanggannya dan dijadikan sebagai acuan untuk pengambilan kebijakan strategis. Teknik sentiment analysis berbasis Natural Language Processing dapat mengatasi permasalahan tersebut. Pada makalah ini prediksi dilakukan menggunakan classifier Random Forest (RF), sementara untuk merangkum kualitas classifier, digunakan kurva Receiver Operating Characteristic (ROC). Kurva ROC berupa grafik yang baik untuk merangkum kualitas classifier. Semakin tinggi kurva berada di atas garis diagonal, semakin baik prediksinya, dengan nilai kurva ROC yang diperoleh sebesar 0,90. Terlihat hasil ulasan terhadap opini pelanggan terhadap jasa dan pelayanan yang diberikan oleh hotel untuk kategori positif lebih banyak daripada kategori negatif. Polaritas dari ulasan diperoleh 68% ulasan pelanggan berada pada area positif dan 32% berada pada area negatif.

scholarly journals Association of Serum Fatty Acids at Admission with the Age of Onset of Acute Ischemic Stroke

Nutrients ◽  
2020 ◽  
Vol 12 (8) ◽  
pp. 2411 ◽  
Author(s):  
Takahisa Mori ◽  
Kazuhiro Yoshioka ◽  
Yuhei Tanno ◽  
Shigen Kasakura
Keyword(s):  
Ischemic Stroke ◽  
Acute Ischemic Stroke ◽  
Receiver Operating Characteristic ◽  
Operating Characteristic ◽  
Age Of Onset ◽  
Linear Regression Analysis ◽  
Multiple Linear Regression Analysis ◽  
Characteristic Curves ◽  
Receiver Operating Characteristic Curves ◽  
Receiver Operating

Dietary triglycerides influence fatty acid (FA) serum concentrations and weight percentages (wt %), which may be associated with the age of onset of acute ischemic stroke (AIS). We investigated the correlations between serum FA levels and proportions at admission and the age of onset of AIS. We evaluated patients with AIS admitted between 2016 and 2019 within 24 h of AIS onset and calculated the correlation coefficients between their ages, serum FA concentrations, and FA wt % values. Multiple linear regression analysis was performed to identify independent FAs indicating AIS age of onset. Furthermore, we estimated the threshold values of independent FAs for age of onset <60 years using receiver operating characteristic curves by logistic regression. A total of 525 patients (median age: 75 years) met the inclusion criteria. The concentration of dihomo-gamma-linolenic acid (DGLA) and wt % of docosahexaenoic acid (DHA) were significant independent variables for age of onset of AIS, and receiver operating characteristic curves for age of onset <60 years showed thresholds of ≥117.7 µmol/L for DGLA and ≤3.7% for DHA. An increased DGLA concentration and decreased DHA wt % were significantly associated with onset of AIS at a younger age.

scholarly journals Predicting high-cost care in a mental health setting

BJPsych Open ◽  
2020 ◽  
Vol 6 (1) ◽  
Author(s):  
Craig Colling ◽  
Mizanur Khondoker ◽  
Rashmi Patel ◽  
Marcella Fok ◽  
Robert Harland ◽  
...  
Keyword(s):  
Natural Language Processing ◽  
Natural Language ◽  
Receiver Operating Characteristic ◽  
Language Processing ◽  
Service Use ◽  
Operating Characteristic ◽  
Characteristic Curves ◽  
Receiver Operating Characteristic Curves ◽  
High Service ◽  
Derived Data

Background The density of information in digital health records offers new potential opportunities for automated prediction of cost-relevant outcomes. Aims We investigated the extent to which routinely recorded data held in the electronic health record (EHR) predict priority service outcomes and whether natural language processing tools enhance the predictions. We evaluated three high priority outcomes: in-patient duration, readmission following in-patient care and high service cost after first presentation. Method We used data obtained from a clinical database derived from the EHR of a large mental healthcare provider within the UK. We combined structured data with text-derived data relating to diagnosis statements, medication and psychiatric symptomatology. Predictors of the three different clinical outcomes were modelled using logistic regression with performance evaluated against a validation set to derive areas under receiver operating characteristic curves. Results In validation samples, the full models (using all available data) achieved areas under receiver operating characteristic curves between 0.59 and 0.85 (in-patient duration 0.63, readmission 0.59, high service use 0.85). Adding natural language processing-derived data to the models increased the variance explained across all clinical scenarios (observed increase in r2 = 12–46%). Conclusions EHR data offer the potential to improve routine clinical predictions by utilising previously inaccessible data. Of our scenarios, prediction of high service use after initial presentation achieved the highest performance.

Bispectral Index in predicting in-hospital mortality in patients with ischemic stroke: A methodological study

2020 ◽  
pp. 102490792090867
Author(s):  
Sultan Tuna Akgol Gur ◽  
Ilker Akbas ◽  
Muhammed Zubeyir Kose ◽  
Abdullah Osman Kocak ◽  
Alper Eren ◽  
...  
Keyword(s):  
Ischemic Stroke ◽  
Glasgow Coma Scale ◽  
Receiver Operating Characteristic ◽  
Bispectral Index ◽  
Charlson Comorbidity Index ◽  
Operating Characteristic ◽  
Coma Scale ◽  
Receiver Operating Characteristic Curves ◽  
Comorbidity Index ◽  
Receiver Operating

Background: Ischemic stroke is a leading cause of death and functional disability worldwide. Several clinical scores or stroke scales, biological test or markers, clinical signs, and radiological imaging have been performed to predict both worse neurologic outcome and mortality for ischemic stroke. Objectives: The aim of our study was to investigate the association between early Bispectral Index scores and in-hospital mortality in patients with ischemic stroke. Methods: This is a comparative prospective methodological study, in which we evaluated the predictive accuracies of Bispectral Index, Glasgow Coma Scale, and Charlson Comorbidity Index for in-hospital mortality of patients with ischemic stroke. Receiver operating characteristic analysis was used for comparing the accuracy of the scoring systems, areas under receiver operating characteristic curves were calculated, and Youden J index was used for estimating associated cut-off values. Results: Among the 80 ischemic stroke patients, in-hospital mortality rate was 38.8% (n = 31). The areas under receiver operating characteristic curves were 0.984, 0.960, and 0.863 for Bispectral Index, Glasgow Coma Scale, and Charlson Comorbidity Index, respectively. The difference between areas under receiver operating characteristic curves for Bispectral Index and Glasgow Coma Scale was statistically similar. Besides, the difference between areas under receiver operating characteristic curves for Bispectral Index and Charlson Comorbidity Index, and the difference between areas under receiver operating characteristic curves for Glasgow Coma Scale and Charlson Comorbidity Index were statistically significant. The associated cut-off values were ⩽74, ⩽12, and >4 for Bispectral Index, Glasgow Coma Scale, and Charlson Comorbidity Index, respectively. For these cut-off points, sensitivity and specificity of Bispectral Index were 93.6% and 95.9%, sensitivity and specificity of Glasgow Coma Scale were 100.0% and 83.7%, and sensitivity and specificity of Charlson Comorbidity Index were 83.9% and 69.4%, respectively. However, accuracy of Bispectral Index was 95.0%, accuracy of Glasgow Coma Scale was 90.0%, and accuracy of Charlson Comorbidity Index was 75.0. Conclusion: Knowledge of the risk factors for mortality in patients with ischemic stroke can help to identify which patients have a higher risk of fatal outcome. The Bispectral Index score improved discrimination and classified patients with higher mortality better than both Glasgow Coma Scale and Charlson Comorbidity Index.

scholarly journals Infarct-core CT perfusion parameters in predicting post-thrombolysis hemorrhagic transformation of acute ischemic stroke

2018 ◽  
Vol 53 (1) ◽  
pp. 25-30 ◽  
Author(s):  
Crt Langel ◽  
Katarina Surlan Popovic
Keyword(s):  
Ischemic Stroke ◽  
Acute Ischemic Stroke ◽  
Receiver Operating Characteristic ◽  
Operating Characteristic ◽  
Receiver Operating Characteristic Analysis ◽  
Ct Perfusion ◽  
Hemorrhagic Transformation ◽  
Characteristic Analysis ◽  
Infarct Core ◽  
Receiver Operating

Abstract Background Intravenous thrombolysis (IVT) is the method of choice in reperfusion treatment of patients with signs and symptoms of acute ischemic stroke (AIS) lasting less than 4.5 hours. Hemorrhagic transformation (HT) of acute ischemic stroke is a serious complication of IVT and occurs in 4.5–68.0% of clinical cases. The aim of our study was to determine the infarct core CT perfusion parameter (CTPP) most predictive of HT. Patients and methods Seventy-five patients with AIS who had undergone CT perfusion (CTP) imaging and were treated with IVT were enrolled in this retrospective study. Patients with and without HT after IVT were defined as cases and controls, respectively. Controls were found by matching for time from AIS symptom onset to IVT ± 0.5 h. The following CTPPs were measured: cerebral blood flow (CBF), cerebral blood volume (CBV), mean transit time (MTT), relative CBF (rCBF) and relative CBV (rCBV). Receiver operating characteristic analysis curves of significant CTPPs determined cut-off values that best predict HT. Results There was a significant difference between cases and controls for CBF (p = 0.004), CBV (p = 0.009), rCBF (p < 0.001) and rCBV (p = 0.001). Receiver operating characteristic analysis revealed that rCBF < 4.5% of the contralateral mean (area under the curve = 0.736) allowed prediction of HT with a sensitivity of 71.0% and specificity of 52.5%. Conclusions CTP imaging has a considerable role in HT prediction, assisting in selection of patients that are likely to benefit from IVT. rCBF proved to have the highest HT predictive value.

Abstract 4030: Infarct Volume Thresholds for Prediction of Independent Functional Outcomes after Acute Ischemic Stroke

Stroke ◽  
2012 ◽  
Vol 43 (suppl_1) ◽  
Author(s):  
Raul G Nogueira ◽  
Tudor Jovin ◽  
David Liebeskind ◽  
Leticia M Souza ◽  
Qing Hao ◽  
...  
Keyword(s):  
Ischemic Stroke ◽  
Stroke Volume ◽  
Acute Ischemic Stroke ◽  
Receiver Operating Characteristic ◽  
Functional Outcome ◽  
Functional Outcomes ◽  
Operating Characteristic ◽  
Infarct Volume ◽  
Roc Curves ◽  
The Mean

Background and Purpose: The main premise of reperfusion therapy in acute ischemic stroke (AIS) is to prevent the conversion of the salvable penumbra into irreversible infarct core thereby reducing the final stoke size. Indeed, previous studies have demonstrated a strong correlation between final infarct volumes and functional outcomes. We sought to establish the optimal final infarct thresholds that best correlate with independent functional outcomes. Methods: Multicenter retrospective analysis across five large academic centers. Consecutive patients meeting the following criteria were included: (1) anterior circulation stroke; (2) available final stroke imaging volumetric analysis; (3) available modified Rankin scale (mRS) score at 90 days. Receiver operating characteristic (ROC) curves were created to help defining the optimal final stroke volume points that discriminate a 90-day mRS ≤2. Results: A total of 484 consecutive patients were identified. The mean age was 65.6±14.6 years. The mean baseline NIHSS was 14.2±7.1. The mean final stroke volume was 77.7±88.5cc (median, 40.5cc). A total of 201 out of the 484 (41.5%) of the patients achieved functional independence at 90 days. The ROC analysis demonstrated that final infarct volume (FIV) was a strong discriminator of independent outcomes with an area under the curve [AUC] of 0.778. The best cut-off point for discriminating 90-day mRS ≤2 was 35cc of FIV (70% specificity; 70% sensitivity). The AUC could be improved by excluding older patients. In patient <65 years, the AUC was 0.844 with an optimal discriminating point at 53cc (75% specificity; 75% sensitivity). The exclusion of patients >80 years yielded an AUC of 0.797 with an optimal FIV discriminating point at 40cc (72% specificity; 72% sensitivity). In this population, FIV of 29cc had 80% specificity and 62% sensitivity whereas FIV of 15cc had 90% specificity but only 40% sensitivity for an independent functional outcome. Conclusions: Final infarct volume is a strong surrogate for good outcomes after AIS. Our data suggest that the exclusion of patients with infarct volumes >35-40cc on baseline imaging may enhance the chances of a positive confirmatory clinical trial of reperfusion in AIS. Figure: Receiver operating characteristic (ROC) curves of Final Stroke Volume (cc) and Independent Functional Outcome (mRS ≤2) at 90 days.

A conditional approach for the receiver operating characteristic curve construction to evaluate diagnostic test performance in a family-matched case–control design

2021 ◽  
pp. 096228022199595
Author(s):  
Yalda Zarnegarnia ◽  
Shari Messinger
Keyword(s):  
Receiver Operating Characteristic Curve ◽  
Receiver Operating Characteristic ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Case Control ◽  
Paired Data ◽  
Characteristic Curves ◽  
Receiver Operating Characteristic Curves ◽  
Operating Characteristic Curve ◽  
Receiver Operating

Receiver operating characteristic curves are widely used in medical research to illustrate biomarker performance in binary classification, particularly with respect to disease or health status. Study designs that include related subjects, such as siblings, usually have common environmental or genetic factors giving rise to correlated biomarker data. The design could be used to improve detection of biomarkers informative of increased risk, allowing initiation of treatment to stop or slow disease progression. Available methods for receiver operating characteristic construction do not take advantage of correlation inherent in this design to improve biomarker performance. This paper will briefly review some developed methods for receiver operating characteristic curve estimation in settings with correlated data from case–control designs and will discuss the limitations of current methods for analyzing correlated familial paired data. An alternative approach using conditional receiver operating characteristic curves will be demonstrated. The proposed approach will use information about correlation among biomarker values, producing conditional receiver operating characteristic curves that evaluate the ability of a biomarker to discriminate between affected and unaffected subjects in a familial paired design.

Sign in / Sign up

Export Citation Format

Share Document