Developing a machine-learning model for real-time prediction of successful extubation in mechanically ventilated patients using time-series ventilator-derived parameters.

Background: Successful weaning from mechanical ventilation is important for patients admitted to intensive care units. However, models for predicting real-time weaning outcomes remain inadequate. Therefore, this study aimed to develop a machine-learning model for predicting successful extubation only using time-series ventilator-derived parameters with good accuracy. Methods: Patients with mechanical ventilation admitted to the Yuanlin Christian Hospital in Taiwan between August 2015 and November 2020 were retrospectively included. A dataset with ventilator-derived parameters was obtained before extubation. Recursive feature elimination was applied to select the most important features. Machine-learning models of logistic regression, random forest (RF), and support vector machine were adopted to predict extubation outcomes. In addition, the synthetic minority oversampling technique (SMOTE) was employed to address the data imbalance problem. The area under the receiver operating characteristic (AUC), F1 score, and accuracy, along with the 10-fold cross-validation, were used to evaluate prediction performance. Results: In this study, 233 patients were included, of whom 28 (12.0%) failed extubation. The six ventilatory variables per 180 s dataset had optimal feature importance. RF exhibited better performance than the others, with an AUC value of 0.976 (95% confidence interval [CI], 0.975-0.976), accuracy of 94.0% (95% CI, 93.8-94.3%), and an F1 score of 95.8% (95% CI, 95.7-96.0%). The difference in performance between the RF and the original and SMOTE datasets was small. Conclusion: The RF model demonstrated a good performance in predicting successful extubation in mechanically ventilated patients. This algorithm made a precise real-time extubation outcome prediction for patients at different time points.

The role of extracranial carotid duplex in predicting functional outcome in first time lacunae strokes.

BACKGROUND: Stroke is the third largest cause of death both worldwide and in Taiwan. Among the various stroke subtypes, lacunae strokes account for 20 to 30% of the total stroke population. Through vigorous risk control factors, the effective prevention and the long-term functional outcome remains are yet to be investigated. Carotid duplex is a non-invasive neuroimaging modality that is routinely applied to stroke patients. In the current research, we correlate baseline carotid duplex examination parameters with functional outcome assessment in an attempt to ascertain important variables for early outcome prediction. METHOD: We retrospectively recruited lacunae ischemic stroke patients from the Department of Neurology, Changhua Christian Hospital. Patient baseline demographics, biochemistry, neuroimaging, and outcome assessments were thoroughly evaluated via a modified Rankin Scale (mRS), NIHSS, and Barthel index scores, both prior and subsequent to patient discharge for a period of up to one year. Patients were divided into two groups based on their functional recovery status. This current study utilizes the Kaplan-Meier method to draw the survival curve and adopts the Cox proportional hazards model to estimate the hazard ratio between the two groups. The risk modeling method is also applied to formulate the best fit in seeking and identifying the variables, which has the most impact on determining theoutcome predictor. All statistical analyses were performed using R for Windows (version 3.6.3), with p < 0.05 meaning statistical significance and 0.05 < p < 0.1 denoting marginal statistical significance. RESULTS: One hundred forty-one ischemic lacunae stroke patients are included in the study, of which the follow-ups of 69 patients showed no physical functionality improvement, whereas 72 patients demonstrated favorable functional outcomes. In terms of underlying diseases, more than 80% of the patients suffered from hyperlipidemia upon hospitalization and more than 40% of the patients suffered from diabetes mellitus and heart disease. There was a significant difference in pulmonary disease between improvement and non-improvement in disability (1.4% vs 10.1%, p = 0.034). The average age of the improved group was lower than that of the non-improved group (65.5 vs. 72, p = 0.014), and the WBC was higher (8 vs 6.95, p = 0.005); furthermore, higher NIHSS scores (6.5 vs 3, p = 0.001) and lower Barthel index scores (47.5 vs 80, p < 0.001) were also noted. The group with improved mRS required a longer hospital stay (with a median of 10 days), while their un-improved counterpart required 8 days (p < 0.05). The majority of patients were discharged within 30 days, and there was a slight difference between the two survival curves (marginal statistical significance, 0.05 < p < 0.1). The hazard ratio of the improved group was 0.7188. There was a difference in the end diastolic velocity (EDV) of common carotid artery (CCA) between the lesion and non-lesion sides of the patients (p < 0.05). Risk-stratifying models indicate that diabetes mellitus, peak systolic velocity (PSV) of the internal carotid artery (ICA) on the lesion side, the diameter of the ICA on the non-lesion side, and the pulsatility index (PI) of the ICA and external carotid artery (ECA) were selected as the most important factors affecting the end diastolic velocity (EDV) of the common carotid artery (CCA) in all four models. CONCLUSION: In the current study, the most suitable length of stay in hospital for lacunae stroke patients to fully recuperate is ten days. Additionally, CCA EDV might be the key determinant, with baseline diabetes mellitus acting in a medium role, capable of altering long-term outcome functionality and recovery.

Functional outcome prediction of ischemic stroke patients with atrial fibrillation accepting post-acute care training.

Background: Ischemic stroke poses a major threat to human health and represents the third leading cause of death worldwide and in Taiwan. Post-acute care (PAC) training has been reported to be beneficial for post-index stroke events. However, knowledge is still lacking on the outcome of stroke events with cardiac origin. The focus of the current study is to investigate the effectiveness of PAC in this subgroup of patients as well as identify key baseline pointers that are capable of early prediction of patients' physical recovery. In addition, the authors hypothesize that the routinely arranged non-invasive carotid duplex that evaluates the characteristics of the carotid lumen could play a significant role in providing an early outcome prediction. Methods: For the current research, 142 ischemic stroke patients with underlying cardiac arrhythmia (atrial fibrillation) were retrospectively recruited. The patients' basic demographics, neuroimaging, carotid duplex, and basic biochemistry datasets were accurately documented. The pre and post-admission National Institutes of Health Stroke Scale (NIHSS) (6-month follow-ups), Barthel Index, and mRS score (12-month follow-ups) were also recorded. All statistical analyses were performed using R for Windows (version 3.6.3). Barthel Index, NIHSS, and mRS scores obtained before and after hospitalization were compared to determine the patients' outcomes and were classified as improved or unimproved. A multivariate logistic analysis was designed and applied to assess the significance of risk factors and to obtain the odds ratios (ORs). The receiver operating characteristic (ROC) curve and the Youden Index was used to find the important cut-off point information, and the area under the curve (AUC) was calculated to provide accuracy. Results: The average age of the 142 ischemic stroke patients enrolled in the current study was about 66 years, of which 88 patients were male and 54, female. Many of them had other comorbidities: 86 patients had mixed hyperlipidemia (60.56%), 115 had hypertension (80.99%), and 49 suffered from diabetes mellitus (34.51%). The mRS showed an improvement in the condition of only 40 patients (28.175%), whereas the Barthel Index showed improvement in 71 patients (50%), and 68 patients (47.89%) showed recovery on the NIHSS. The Barthel Index and NIHSS were selected because they already had an almost equal number of samples among the improved and unimproved groups (50%), rather than mRS, which had a lower number (28.17%) of improved cases. While conducting the EuroQol-5 Dimension (EQ-5D) assessment, anxiety/depression stood out as the most prominent issue, affecting 44 patients (30.99%). Self-care was another factor that was involved in the ongoing improvement of 36 patients (25.35%). Multivariate logistic analysis of both NIHSS and Barthel Index showed improvement with a contralateral plaque index statistical significance (P<0.05), whereas NIHSS showed a relevant significance in anxiety/depression and Barthel Index registered usual activity in the data analysis (P<0.05). ROC curve and Youden index analysis showed similar results in both NIHSS and Barthel Index of contralateral plaque index of 4.5, this being the cutoff point value for this group of patients. Conclusion: In the current study, nearly half of the enrolled patients showed favorable functional recovery. The outcome assessments seem to correlate well with NIHSS and Barthel Index scores, rather than mRS. The anxiety/depression and usual activities domains of the EQ-5D results are associated with and have a great impact after the patients undertake the PAC rehabilitative strategy. Moreover, the variables obtained through carotid duplex and plaque index might also play a significant role in determining the patient's functional outcome.

Long-Term Functionality Prediction for First Time Ischemic Middle Cerebral Artery Stroke Patients Receiving Conventional Medical Treatment.

BACKGROUND: Middle cerebral artery (MCA) ischemic stroke poses a major threat to human beings and prompts intravenous thrombolytic and/or thrombectomy management remains the gold standard treatment. However, not all MCA stroke patients fit in the inclusion and exclusion criteria that many patients only receive conventional medical therapy. We attempt to seek the baseline parameters that can effectively predict patients' long-term functionality, as well as hypothesizing that the carotid duplex derived resistance/pulsatility index might be capable of fulfilling this purpose. METHODS: The 741 MCA ischemic stroke patients have been retrospectively recruited for the project. Under the initial screening, matching the inclusion and exclusion criteria, there are 471 participants to be enrolled in the study. The patients' basic information, along with outcome assessments, pre-admission Barthel index and NIHSS score, and pre- and post-treatment mRS are recorded. All statistical analyses were performed using R for Windows (version 3.6.3). The significance level was set at P < 0.05 for all analyses. RESULTS: Of the 471 patients, 239 participants show a net mRS improvement, whereas the other 232 show deterioration. Hyperlipidaemia, chronic kidney disease, and dementia are related to long-term functionality improvement. The multivariate logistic regression analysis shows that right common carotid artery (CCA) resistance index (RI) and ischemic heart disease play a significant role in favourable outcome functionality. The ROC and Youden Index models are formulated, and it shows that Barthel Index and the NIHSS are most significant in the outcome measurement group (AUC: 0.675, 0.653; cut-off point: 57.5, 3.5, respectively). The right-side CCA RI is the solely important outcome predictor for the baseline carotid duplex study (AUC: 0.5; cut-off point: 0.785). CONCLUSION: The favourable long-term functionality of MCA ischemic stroke patients receiving conventional medical treatment seems to correlate fairly with pre-admission NIHSS and Barthel index scores. Underlying hyperlipidaemia, chronic kidney disease, and dementia are conversely associated with favourable long-term capability. Moreover, the value of CCA RI appears to significantly alter the long-term outcomes in this group of patients.

Analysis of agricultural exports based on deep learning and text mining.

Agricultural exports are an important source of economic profit for many countries. Accurate predictions of a country's agricultural exports month on month are key to understanding a country's domestic use and export figures and facilitate advance planning of export, import, and domestic use figures and the resulting necessary adjustments of production and marketing. This study proposes a novel method for predicting the rise and fall of agricultural exports, called agricultural exports time series-long short-term memory (AETS-LSTM). The method applies Jieba word segmentation and Word2Vec to train word vectors and uses TF-IDF and word cloud to learn news-related keywords and finally obtain keyword vectors. This research explores whether the purchasing managers' index (PMI) of each industry can effectively use the AETS-LSTM model to predict the rise and fall of agricultural exports. Research results show that the inclusion of keyword vectors in the PMI values of the finance and insurance industries has a relative impact on the prediction of the rise and fall of agricultural exports, which can improve the prediction accuracy for the rise and fall of agricultural exports by 82.61%. The proposed method achieves improved prediction ability for the chemical/biological/medical, transportation equipment, wholesale, finance and insurance, food and textiles, basic materials, education/professional, science/technical, information/communications/broadcasting, transportation and storage, retail, and electrical and machinery equipment categories, while its performance for the electrical and optical categories shows improved prediction by combining keyword vectors, and its accuracy for the accommodation and food service, and construction and real estate industries remained unchanged. Therefore, the proposed method offers improved prediction capacity for agricultural exports month on month, allowing agribusiness operators and policy makers to evaluate and adjust domestic and foreign production and sales.

Hand-to-Hand Model for Bioelectrical Impedance Analysis to Estimate Fat Free Mass in a Healthy Population.

This study aimed to establish a hand-to-hand (HH) model for bioelectrical impedance analysis (BIA) fat free mass (FFM) estimation by comparing with a standing position hand-to-foot (HF) BIA model and dual energy X-ray absorptiometry (DXA); we also verified the reliability of the newly developed model. A total of 704 healthy Chinese individuals (403 men and 301 women) participated. FFM (FFMDXA) reference variables were measured using DXA and segmental BIA. Further, regression analysis, Bland-Altman plots, and cross-validation (2/3 participants as the modeling group, 1/3 as the validation group; three turns were repeated for validation grouping) were conducted to compare tests of agreement with FFMDXA reference variables. In male participants, the hand-to-hand BIA model estimation equation was calculated as follows: FFMmHH = 0.537 h²/ZHH - 0.126 year + 0.217 weight + 18.235 (r² = 0.919, standard estimate of error (SEE) = 2.164 kg, n = 269). The mean validated correlation coefficients and limits of agreement (LOAs) of the Bland-Altman analysis of the calculated values for FFMmHH and FFMDXA were 0.958 and -4.369-4.343 kg, respectively, for hand-to-foot BIA model measurements for men; the FFM (FFMmHF) and FFMDXA were 0.958 and -4.356-4.375 kg, respectively. The hand-to-hand BIA model estimating equation for female participants was FFMFHH = 0.615 h²/ZHH - 0.144 year + 0.132 weight + 16.507 (r² = 0.870, SEE = 1.884 kg, n = 201); the three mean validated correlation coefficient and LOA for the hand-to-foot BIA model measurements for female participants (FFMFHH and FFMDXA) were 0.929 and -3.880-3.886 kg, respectively. The FFMHF and FFMDXA were 0.942 and -3.511-3.489 kg, respectively. The results of both hand-to-hand and hand-to-foot BIA models demonstrated similar reliability, and the hand-to-hand BIA models are practical for assessing FFM.

Predicting body composition using foot-to-foot bioelectrical impedance analysis in healthy Asian individuals.

BACKGROUND: The objectives of this study were to develop a regression model for predicting fat-free mass (FFM) in a population of healthy Taiwanese individuals using standing foot-to-foot bioelectrical impedance analysis (BIA) and to test the model's performance in predicting FFM with different body fat percentages (BF%). METHODS: We used dual-energy X-ray absorptiometry (DXA) to measure the FFM of 554 healthy Asian subjects (age, 16-75 y; body mass index, 15.8-43.1 kg/m(2)). We also evaluated the validity of the developed multivariate model using a double cross-validation technique and assessed the accuracy of the model in an all-subjects sample and subgroup samples with different body fat levels. RESULTS: Predictors in the all-subjects multivariate model included height(2)/impedance, weight, year, and sex (FFM = 13.055 + 0.204 weight + 0.394 height(2)/Impedance - 0.136 age + 8.125 sex (sex: Female = 0, Male = 1), r(2) = 0.92, standard error of the estimate = 3.17 kg). The correlation coefficients between predictive FFM by BIA (FFMBIA) and DXA-measured FFM (FFMDXA) in female subjects with a total-subjects BF%DXA of <20 %, 20 %-30 %, 30 %-40 % and >40 % were r = 0.87, 0.90, 0.91, 0.89, and 0.94, respectively, with bias ± 2SD of 0.0 ± 3.0 kg, -2.6 ± 1.7 kg, -1.5 ± 2.8 kg, 0.5 ± 2.7 kg, and 2.0 ± 2.9 kg, respectively. The correlation coefficients between FFMBIA and FFMDXA in male subjects with a total-subjects BF%DXA of <10 %, 10 %-20 %, 20 %-30 %, and >30 % were r = 0.89, 0.89, 0.90, 0.93, and 0.91, respectively, with bias ± 2SD of 0.0 ± 3.2 kg, -2.3 ± 2.5 kg, -0.5 ± 3.2 kg, 0.4 ± 3.1 kg, and 2.1 ± 3.2 kg, respectively. CONCLUSIONS: The standing foot-to-foot BIA method developed in this study can accurately predict FFM in healthy Asian individuals with different levels of body fat.

Sparse principal component analysis in cancer research.

A critical challenging component in analyzing high-dimensional data in cancer research is how to reduce the dimension of data and how to extract relevant features. Sparse principal component analysis (PCA) is a powerful statistical tool that could help reduce data dimension and select important variables simultaneously. In this paper, we review several approaches for sparse PCA, including variance maximization (VM), reconstruction error minimization (REM), singular value decomposition (SVD), and probabilistic modeling (PM) approaches. A simulation study is conducted to compare PCA and the sparse PCAs. An example using a published gene signature in a lung cancer dataset is used to illustrate the potential application of sparse PCAs in cancer research.

Prognostic and predictive value of a malignancy-risk gene signature in early-stage non-small cell lung cancer.

BACKGROUND: The malignancy-risk gene signature is composed of numerous proliferative genes and has been applied to predict breast cancer risk. We hypothesized that the malignancy-risk gene signature has prognostic and predictive value for early-stage non-small cell lung cancer (NSCLC) patients. METHODS: The ability of the malignancy-risk gene signature to predict overall survival (OS) of early-stage NSCLC patients was tested using a large NSCLC microarray dataset from the Director's Challenge Consortium (n = 442) and two independent NSCLC microarray datasets (n = 117 and 133, for the GSE13213 and GSE14814 datasets, respectively). An overall malignancy-risk score was generated by principal component analysis to determine the prognostic and predictive value of the signature. An interaction model was used to investigate a statistically significant interaction between adjuvant chemotherapy (ACT) and the gene signature. All statistical tests were two-sided. RESULTS: The malignancy-risk gene signature was statistically significantly associated with OS (P < .001) of NSCLC patients. Validation with the two independent datasets demonstrated that the malignancy-risk score had prognostic and predictive values: Of patients who did not receive ACT, those with a low malignancy-risk score had increased OS compared with a high malignancy-risk score (P = .007 and .01 for the GSE13212 and GSE14814 datasets, respectively), indicating a prognostic value; and in the GSE14814 dataset, patients receiving ACT survived longer in the high malignancy-risk score group (P = .03), and a statistically significant interaction between ACT and the signature was observed (P = .02). CONCLUSIONS: The malignancy-risk gene signature was associated with OS and was a prognostic and predictive indicator. The malignancy-risk gene signature could be useful to improve prediction of OS and to identify those NSCLC patients who will benefit from ACT.