Using Deep Transfer Learning to Detect Hyperkalemia From Ambulatory Electrocardiogram Monitors in Intensive Care Units: Personalized Medicine Approach.

BACKGROUND: Hyperkalemia is a critical condition, especially in intensive care units. So far, there have been no accurate and noninvasive methods for recognizing hyperkalemia events on ambulatory electrocardiogram monitors. OBJECTIVE: This study aimed to improve the accuracy of hyperkalemia predictions from ambulatory electrocardiogram (ECG) monitors using a personalized transfer learning method; this would be done by training a generic model and refining it with personal data. METHODS: This retrospective cohort study used open source data from the Waveform Database Matched Subset of the Medical Information Mart From Intensive Care III (MIMIC-III). We included patients with multiple serum potassium test results and matched ECG data from the MIMIC-III database. A 1D convolutional neural network-based deep learning model was first developed to predict hyperkalemia in a generic population. Once the model achieved a state-of-the-art performance, it was used in an active transfer learning process to perform patient-adaptive heartbeat classification tasks. RESULTS: The results show that by acquiring data from each new patient, the personalized model can improve the accuracy of hyperkalemia detection significantly, from an average of 0.604 (SD 0.211) to 0.980 (SD 0.078), when compared with the generic model. Moreover, the area under the receiver operating characteristic curve level improved from 0.729 (SD 0.240) to 0.945 (SD 0.094). CONCLUSIONS: By using the deep transfer learning method, we were able to build a clinical standard model for hyperkalemia detection using ambulatory ECG monitors. These findings could potentially be extended to applications that continuously monitor one's ECGs for early alerts of hyperkalemia and help avoid unnecessary blood tests.

Impact of coronary risk scores on disposition decision in emergency patients with chest pain.

BACKGROUND: Coronary risk scores (CRS) including History, Electrocardiogram, Age, Risk Factors, Troponin (HEART) score and Emergency Department Assessment of Chest pain Score (EDACS) can help identify patients at low risk of major adverse cardiac events. In the emergency department (ED), there are wide variations in hospital admission rates among patients with chest pain. OBJECTIVE: This study aimed to evaluate the impact of CRS on the disposition of patients with symptoms suggestive of acute coronary syndrome in the ED. METHODS: This retrospective cohort study included 3660 adult patients who presented to the ED with chest pain between January and July in 2019. Study inclusion criteria were age > 18 years and a primary position International Statistical Classification of Diseases and Related Health Problems-10th revision coded diagnosis of angina pectoris (I20.0-I20.9) or chronic ischemic heart disease (I25.0-I25.9) by the treating ED physician. If the treating ED physician completed the electronic structured variables for CRS calculation to assist disposition planning, then the patient would be classified as the CRS group; otherwise, the patient was included in the control group. RESULTS: Among the 2676 patients, 746 were classified into the CRS group, whereas the other 1930 were classified into the control group. There was no significant difference in sex, age, initial vital signs, and ED length of stay between the two groups. The coronary risk factors were similar between the two groups, except for a higher incidence of smokers in the CRS group (19.6% vs. 16.1%, p = 0.031). Compared with the control group, significantly more patients were discharged (70.1% vs. 64.6%) directly from the ED, while fewer patients who were hospitalized (25.9% vs. 29.7%) or against-advise discharge (AAD) (2.6% vs. 4.0%) in the CRS group. Major adverse cardiac events and mortality at 60 days between the two groups were not significantly different. CONCLUSIONS: A higher ED discharge rate of the group using CRS may indicate that ED physicians have more confidence in discharging low-risk patients based on CRS.

Fine particulate matter constituents associated with emergency room visits for pediatric asthma: a time-stratified case-crossover study in an urban area.

BACKGROUND: Global asthma-related mortality tallies at around 2.5 million annually. Although asthma may be triggered or exacerbated by particulate matter (PM) exposure, studies investigating the relationship of PM and its components with emergency department (ED) visits for pediatric asthma are limited. This study aimed to estimate the impact of short-term exposure to PM constituents on ED visits for pediatric asthma. METHODS: We retrospectively evaluated non-trauma patients aged younger than 17 years who visited the ED with a primary diagnosis of asthma. Further, measurements of PM with aerodynamic diameter of < 10 µm (PM10), PM with aerodynamic diameter of < 10 µm (PM2.5), and four PM2.5 components (i.e., nitrate (NO3-), sulfate (SO42-), organic carbon (OC), and elemental carbon (EC)) were collected between 2007 and 2010 from southern particulate matter supersites. These included one core station and two satellite stations in Kaohsiung City, Taiwan. A time-stratified case-crossover study was conducted to analyze the hazard effect of PM. RESULTS: Overall, 1597 patients were enrolled in our study. In the single-pollutant model, the estimated risk increase for pediatric asthma incidence on lag 3 were 14.7% [95% confidence interval (CI), 3.2-27.4%], 13.5% (95% CI, 3.3-24.6%), 14.8% (95% CI, 2.5-28.6%), and 19.8% (95% CI, 7.6-33.3%) per interquartile range increments in PM2.5, PM10, nitrate, and OC, respectively. In the two-pollutant models, OC remained significant after adjusting for PM2.5, PM10, and nitrate. During subgroup analysis, children were more vulnerable to PM2.5 and OC during cold days (< 26 °C, interaction p = 0.008 and 0.012, respectively). CONCLUSIONS: Both PM2.5 concentrations and its chemical constituents OC and nitrate are associated with ED visits for pediatric asthma. Among PM2.5 constituents, OC was most closely related to ED visits for pediatric asthma, and children are more vulnerable to PM2.5 and OC during cold days.

Physicians' Risk Tolerance and Head Computed Tomography Use for Pediatric Patients With Minor Head Injury.

OBJECTIVES: Traumatic brain injury is the leading cause of death and disability in children worldwide. The objective of this study was to determine the association between physician risk tolerance and head computed tomography (CT) use in patients with minor head injury (MHI) in the emergency department (ED). METHODS: We retrospectively analyzed pediatric patients (<17 years old) with MHI in the ED and then administered 2 questionnaires (a risk-taking subscale [RTS] of the Jackson Personality Inventory and a malpractice fear scale [MFS]) to attending physicians who had evaluated these patients and made decisions regarding head CT use. The primary outcome was head CT use during ED evaluation; the secondary outcome was ED length of stay and final diagnosis of intracranial injury (ICI). RESULTS: Of 523 patients with MHI, 233 (44.6%) underwent brain CT, and 16 (3.1%) received a final diagnosis of ICI. Among the 16 emergency physicians (EPs), the median scores of the MFS and RTS were 22 (interquartile range, 17-26) and 23 (interquartile range, 19-25), respectively. Emergency physicians who were most risk averse tended to order more head CT scans compared with the more risk-tolerant EPs (56.96% vs 37.37%; odds ratio, 8.463; confidence interval, 2.783-25.736). The ED length of stay (P = 0.442 and P = 0.889) and final diagnosis (P = 0.155 and P = 0.835) of ICI were not significantly associated with the RTS and MFS scores. CONCLUSIONS: Individual EP risk tolerance, as measured by RTS, was predictive of CT use in pediatric patients with MHI.

Ambient air pollution is associated with pediatric pneumonia: a time-stratified case-crossover study in an urban area.

BACKGROUND: Pneumonia, the leading reason underlying childhood deaths, may be triggered or exacerbated by air pollution. To date, only a few studies have examined the association of air pollution with emergency department (ED) visits for pediatric pneumonia, with inconsistent results. Therefore, we aimed to elucidate the impact of short-term exposure to particulate matter (PM) and other air pollutants on the incidence of ED visits for pediatric pneumonia. METHODS: PM2.5, PM10, and other air pollutant levels were measured at 11 air quality-monitoring stations in Kaohsiung City, Taiwan, between 2008 and 2014. Further, we extracted the medical records of non-trauma patients aged ≤17 years and who had visited an ED with the principal diagnosis of pneumonia. A time-stratified case-crossover study design was employed to determine the hazard effect of air pollution in a total of 4024 patients. RESULTS: The single-pollutant model suggested that per interquartile range increment in PM2.5, PM10, nitrogen dioxide (NO2), and sulfur dioxide (SO2) on 3 days before the event increased the odds of pediatric pneumonia by 14.0% [95% confidence interval (CI), 5.1-23.8%], 10.9% (95% CI, 2.4-20.0%), 14.1% (95% CI, 5.0-24.1%), and 4.5% (95% CI, 0.8-8.4%), respectively. In two-pollutant models, PM2.5 and NO2 were significant after adjusting for PM10 and SO2. Subgroup analyses showed that older children (aged ≥4 years) were more susceptible to PM2.5 (interaction p = 0.024) and children were more susceptible to NO2 during warm days (≥26.5 °C, interaction p = 0.011). CONCLUSIONS: Short-term exposure to PM2.5 and NO2 possibly plays an important role in pediatric pneumonia in Kaohsiung, Taiwan. Older children are more susceptible to PM2.5, and all children are more susceptible to NO2 during warm days.

Machine learning-based prediction of coronary care unit readmission: A multihospital validation study.

Objective: Readmission to the coronary care unit (CCU) has significant implications for patient outcomes and healthcare expenditure, emphasizing the urgency to accurately identify patients at high readmission risk. This study aims to construct and externally validate a predictive model for CCU readmission using machine learning (ML) algorithms across multiple hospitals. Methods: Patient information, including demographics, medical history, and laboratory test results were collected from electronic health record system and contributed to a total of 40 features. Five ML models: logistic regression, random forest, support vector machine, gradient boosting, and multilayer perceptron were employed to estimate the readmission risk. Results: The gradient boosting model was selected demonstrated superior performance with an area under the receiver operating characteristic curve (AUC) of 0.887 in the internal validation set. Further external validation in hold-out test set and three other medical centers upheld the model's robustness with consistent high AUCs, ranging from 0.852 to 0.879. Conclusion: The results endorse the integration of ML algorithms in healthcare to enhance patient risk stratification, potentially optimizing clinical interventions, and diminishing the burden of CCU readmissions.

Explainable Deep Learning Model for Predicting Serious Adverse Events in Hospitalized Geriatric Patients Within 72 Hours.

Background: The global aging population presents a significant challenge, with older adults experiencing declining physical and cognitive abilities and increased vulnerability to chronic diseases and adverse health outcomes. This study aims to develop an interpretable deep learning (DL) model to predict adverse events in geriatric patients within 72 hours of hospitalization. Methods: The study used retrospective data (2017-2020) from a major medical center in Taiwan. It included non-trauma geriatric patients who visited the emergency department and were admitted to the general ward. Data preprocessing involved collecting prognostic factors like vital signs, lab results, medical history, and clinical management. A deep feedforward neural network was developed, and performance was evaluated using accuracy, sensitivity, specificity, positive predictive value (PPV), and area under the receiver operating characteristic curve (AUC). Model interpretation utilized the Shapley Additive Explanation (SHAP) technique. Results: The analysis included 127,268 patients, with 2.6% experiencing imminent intensive care unit transfer, respiratory failure, or death during hospitalization. The DL model achieved AUCs of 0.86 and 0.84 in the validation and test sets, respectively, outperforming the Sequential Organ Failure Assessment (SOFA) score. Sensitivity and specificity values ranged from 0.79 to 0.81. The SHAP technique provided insights into feature importance and interactions. Conclusion: The developed DL model demonstrated high accuracy in predicting serious adverse events in geriatric patients within 72 hours of hospitalization. It outperformed the SOFA score and provided valuable insights into the model's decision-making process.

Prospective clinical evaluation of deep learning for ultrasonographic screening of abdominal aortic aneurysms.

Abdominal aortic aneurysm (AAA) often remains undetected until rupture due to limited access to diagnostic ultrasound. This trial evaluated a deep learning (DL) algorithm to guide AAA screening by novice nurses with no prior ultrasonography experience. Ten nurses performed 15 scans each on patients over 65, assisted by a DL object detection algorithm, and compared against physician-performed scans. Ultrasound scan quality, assessed by three blinded expert physicians, was the primary outcome. Among 184 patients, DL-guided novices achieved adequate scan quality in 87.5% of cases, comparable to the 91.3% by physicians (p = 0.310). The DL model predicted AAA with an AUC of 0.975, 100% sensitivity, and 97.8% specificity, with a mean absolute error of 2.8 mm in predicting aortic width compared to physicians. This study demonstrates that DL-guided POCUS has the potential to democratize AAA screening, offering performance comparable to experienced physicians and improving early detection.

RAPID-ED: A predictive model for risk assessment of patient's early in-hospital deterioration from emergency department.

Introduction: The objective of this multi-center retrospective cohort study was to devise a predictive tool known as RAPID-ED. This model identifies non-traumatic adult patients at significant risk for cardiac arrest within 48 hours post-admission from the emergency department. Methods: Data from 224,413 patients admitted through the emergency department (2016-2020) was analyzed, incorporating vital signs, lab tests, and administered therapies. A multivariable regression model was devised to anticipate early cardiac arrest. The efficacy of the RAPID-ED model was evaluated against traditional scoring systems like National Early Warning Score (NEWS) and Modified Early Warning Score (MEWS) and its predictive ability was gauged via the area under the receiver operating characteristic curve (AUC) in both hold-out validation set and external validation set. Results: RAPID-ED outperformed traditional models in predicting cardiac arrest with an AUC of 0.819 in the hold-out validation set and 0.807 in the external validation set. In this critical care update, RAPID-ED offers an innovative approach to assessing patient risk, aiding emergency physicians in post-discharge care decisions from the emergency department. High-risk score patients (≥13) may benefit from early ICU admission for intensive monitoring. Conclusion: As we progress with advancements in critical care, tools like RAPID-ED will prove instrumental in refining care strategies for critically ill patients, fostering an improved prognosis and potentially mitigating mortality rates.

Explainable deep learning model to predict invasive bacterial infection in febrile young infants: A retrospective study.

BACKGROUND: Machine learning models have demonstrated superior performance in predicting invasive bacterial infection (IBI) in febrile infants compared to commonly used risk stratification criteria in recent studies. However, the black-box nature of these models can make them difficult to apply in clinical practice. In this study, we developed and validated an explainable deep learning model that can predict IBI in febrile infants ≤ 60 days of age visiting the emergency department. METHODS: We conducted a retrospective study of febrile infants aged ≤ 60 days who presented to the pediatric emergency department of a medical center in Taiwan between January 1, 2011 and December 31, 2019. Patients with uncertain test results and complex chronic health conditions were excluded. IBI was defined as the growth of a pathogen in the blood or cerebrospinal fluid. We used a deep neural network to develop a predictive model for IBI and compared its performance to the IBI score and step-by-step approach. The SHapley Additive Explanations (SHAP) technique was used to explain the model's predictions at different levels. RESULTS: Our study included 1847 patients, 53 (2.7%) of whom had IBI. The deep learning model performed similarly to the IBI score and step-by-step approach in terms of sensitivity and negative predictive value, but provided better specificity (54%), positive predictive value (5%), and area under the receiver-operating characteristic curve (0.87). SHapley Additive exPlanations identified five influential predictive variables (absolute neutrophil count, body temperature, heart rate, age, and C-reactive protein). CONCLUSION: We have developed an explainable deep learning model that can predict IBI in febrile infants aged 0-60 days. The model not only performs better than previous scoring systems, but also provides insight into how it arrives at its predictions through individual features and cases.

Development and validation of a deep learning pipeline to measure pericardial effusion in echocardiography.

Objectives: The aim of this study was to develop a deep-learning pipeline for the measurement of pericardial effusion (PE) based on raw echocardiography clips, as current methods for PE measurement can be operator-dependent and present challenges in certain situations. Methods: The proposed pipeline consisted of three distinct steps: moving window view selection (MWVS), automated segmentation, and width calculation from a segmented mask. The MWVS model utilized the ResNet architecture to classify each frame of the extracted raw echocardiography files into selected view types. The automated segmentation step then generated a mask for the PE area from the extracted echocardiography clip, and a computer vision technique was used to calculate the largest width of the PE from the segmented mask. The pipeline was applied to a total of 995 echocardiographic examinations. Results: The proposed deep-learning pipeline exhibited high performance, as evidenced by intraclass correlation coefficient (ICC) values of 0.867 for internal validation and 0.801 for external validation. The pipeline demonstrated a high level of accuracy in detecting PE, with an area under the receiving operating characteristic curve (AUC) of 0.926 (95% CI: 0.902-0.951) for internal validation and 0.842 (95% CI: 0.794-0.889) for external validation. Conclusion: The machine-learning pipeline developed in this study can automatically calculate the width of PE from raw ultrasound clips. The novel concepts of moving window view selection for image quality control and computer vision techniques for maximal PE width calculation seem useful in the field of ultrasound. This pipeline could potentially provide a standardized and objective approach to the measurement of PE, reducing operator-dependency and improving accuracy.

Association between Air Pollution and Short-Term Outcome of ST-Segment Elevation Myocardial Infarction in a Tropical City, Kaohsiung, Taiwan.

ST-segment elevation myocardial infarction (STEMI), one of the primary factors leading to global mortality, has been shown through epidemiological studies to have a relationship with short-term exposure to air pollutants; however, the association between air pollutants and the outcome of STEMI has not been well studied. The aim of this study was to estimate the impact of air pollutants on the outcomes of STEMI. Data on particulate matter <2.5 µm (PM2.5), <10 µm (PM10), nitrogen dioxide (NO2), and ozone (O3) at each of the 11 air monitoring stations in Kaohsiung City were collected between 1 January 2012 and 31 December 2017. Medical records of non-trauma patients aged > 20 years who had presented to the Emergency Department (ED) with a principal diagnosis of STEMI were extracted. The primary outcome measure was in-hospital mortality. After adjusting for potential confounders and meteorological variables, we found that an increase in the interquartile range (IQR) in NO2 was associated with an elevated risk of in-hospital mortality in patients with STEMI. Moreover, there was an observed higher risk of in-hospital mortality associated with an increase in the IQR of NO2 during the warm season, specifically in lag 3 (3 days prior to the onset, OR = 3.266; 95%CI: 1.203-8.864, p = 0.02). Conversely, an IQR increase in PM10 was associated with an increased risk of in-hospital mortality in patients with STEMI in lag 3 (OR = 2.792; 95%CI: 1.115-6.993, p = 0.028) during the cold season. Our study suggests that exposure to NO2 (during the warm season) and PM10 (during the cold season) may contribute to a higher risk of poor prognosis in patients with STEMI.

Correction: Yau et al. Risk Factors for Early Return Visits to the Emergency Department in Patients Presenting with Nonspecific Abdominal Pain and the Use of Computed Tomography Scan. Healthcare 2021, 9, 1470.

The authors would like to make corrections to their published paper [...].

Utilization of Personalized Machine-Learning to Screen for Dysglycemia from Ambulatory ECG, toward Noninvasive Blood Glucose Monitoring.

Blood glucose (BG) monitoring is important for critically ill patients, as poor sugar control has been associated with increased mortality in hospitalized patients. However, constant BG monitoring can be resource-intensive and pose a healthcare burden in clinical practice. In this study, we aimed to develop a personalized machine-learning model to predict dysglycemia from electrocardiogram (ECG) data. We used the Medical Information Mart for Intensive Care III database as our source of data and obtained more than 20 ECG records from each included patient during a single hospital admission. We focused on lead II recordings, along with corresponding blood sugar data. We processed the data and used ECG features from each heartbeat as inputs to develop a one-class support vector machine algorithm to predict dysglycemia. The model was able to predict dysglycemia using a single heartbeat with an AUC of 0.92 ± 0.09, a sensitivity of 0.92 ± 0.10, and specificity of 0.84 ± 0.04. After applying 10 s majority voting, the AUC of the model's dysglycemia prediction increased to 0.97 ± 0.06. This study showed that a personalized machine-learning algorithm can accurately detect dysglycemia from a single-lead ECG.

Association between Ambient Air Pollution and Emergency Room Visits for Pediatric Respiratory Diseases: The Impact of COVID-19 Pandemic.

The level and composition of air pollution have changed during the coronavirus disease 2019 (COVID-19) pandemic. However, the association between air pollution and pediatric respiratory disease emergency department (ED) visits during the COVID-19 pandemic remains unclear. The study was retrospectively conducted between 2017 and 2020 in Kaohsiung, Taiwan, from 1 January 2020 to 1 May 2020, defined as the period of the COVID-19 pandemic, and 1 January 2017 to 31 May 2019, defined as the pre-COVID-19 pandemic period. We enrolled patients under 17 years old who visited the ED in a medical center and were diagnosed with respiratory diseases such as pneumonia, asthma, bronchitis, and acute pharyngitis. Measurements of particulate matter (PM) with aerodynamic diameters of <10 µm (PM10) and < 2.5 µm (PM2.5), nitrogen dioxide (NO2), and Ozone (O3) were collected. During the COVID-19 pandemic, an increase in the interquartile range of PM2.5, PM10, and NO2 levels was associated with increases of 72.5% (95% confidence interval [CI], 50.5−97.7%), 98.0% (95% CI, 70.7−129.6%), and 54.7% (95% CI, 38.7−72.6%), respectively, in the risk of pediatric respiratory disease ED visits on lag 1, which were greater than those in the pre-COVID-19 pandemic period. After adjusting for temperature and humidity, the risk of pediatric respiratory diseases after exposure to PM2.5 (inter p = 0.001) and PM10 (inter p < 0.001) was higher during the COVID-19 pandemic. PM2.5, PM10, and NO2 may play important roles in pediatric respiratory events in Kaohsiung, Taiwan. Compared with the pre-COVID-19 pandemic period, the levels of PM2.5 and PM10 were lower; however, the levels were related to a greater increase in ED during the COVID-19 pandemic.

Machine learning models for predicting in-hospital mortality in patient with sepsis: Analysis of vital sign dynamics.

Purpose: To build machine learning models for predicting the risk of in-hospital death in patients with sepsis within 48 h, using only dynamic changes in the patient's vital signs. Methods: This retrospective observational cohort study enrolled septic patients from five emergency departments (ED) in Taiwan. We adopted seven variables, i.e., age, sex, systolic blood pressure, diastolic blood pressure, heart rate, respiratory rate, and body temperature. Results: Among all 353,253 visits, after excluding 159,607 visits (45%), the study group consisted of 193,646 ED visits. With a leading time of 6 h, the convolutional neural networks (CNNs), long short-term memory (LSTM), and random forest (RF) had accuracy rates of 0.905, 0.817, and 0.835, respectively, and the area under the receiver operating characteristic curve (AUC) was 0.840, 0.761, and 0.770, respectively. With a leading time of 48 h, the CNN, LSTM, and RF achieved accuracy rates of 0.828, 0759, and 0.805, respectively, and an AUC of 0.811, 0.734, and 0.776, respectively. Conclusion: By analyzing dynamic vital sign data, machine learning models can predict mortality in septic patients within 6 to 48 h of admission. The performance of the testing models is more accurate if the lead time is closer to the event.

Using Machine Learning to Predict Invasive Bacterial Infections in Young Febrile Infants Visiting the Emergency Department.

BACKGROUND: The aim of this study was to develop and evaluate a machine learning (ML) model to predict invasive bacterial infections (IBIs) in young febrile infants visiting the emergency department (ED). METHODS: This retrospective study was conducted in the EDs of three medical centers across Taiwan from 2011 to 2018. We included patients age in 0-60 days who were visiting the ED with clinical symptoms of fever. We developed three different ML algorithms, including logistic regression (LR), supportive vector machine (SVM), and extreme gradient boosting (XGboost), comparing their performance at predicting IBIs to a previous validated score system (IBI score). RESULTS: During the study period, 4211 patients were included, where 126 (3.1%) had IBI. A total of eight, five, and seven features were used in the LR, SVM, and XGboost through the feature selection process, respectively. The ML models can achieve a better AUROC value when predicting IBIs in young infants compared with the IBI score (LR: 0.85 vs. SVM: 0.84 vs. XGBoost: 0.85 vs. IBI score: 0.70, p-value < 0.001). Using a cost sensitive learning algorithm, all ML models showed better specificity in predicting IBIs at a 90% sensitivity level compared to an IBI score > 2 (LR: 0.59 vs. SVM: 0.60 vs. XGBoost: 0.57 vs. IBI score >2: 0.43, p-value < 0.001). CONCLUSIONS: All ML models developed in this study outperformed the traditional scoring system in stratifying low-risk febrile infants after the standardized sensitivity level.

Using a Multiclass Machine Learning Model to Predict the Outcome of Acute Ischemic Stroke Requiring Reperfusion Therapy.

Prediction of functional outcome in ischemic stroke patients is useful for clinical decisions. Previous studies mostly elaborate on the prediction of favorable outcomes. Miserable outcomes, which are usually defined as modified Rankin Scale (mRS) 5-6, should be considered as well before further invasive intervention. By using a machine learning algorithm, we aimed to develop a multiclass classification model for outcome prediction in acute ischemic stroke patients requiring reperfusion therapy. This was a retrospective study performed at a stroke medical center in Taiwan. Patients with acute ischemic stroke who visited between January 2016 and December 2019 and who were candidates for reperfusion therapy were included. Clinical outcomes were classified as favorable outcome, intermediate outcome, and miserable outcome. We developed four different multiclass machine learning models (Logistic Regression, Supportive Vector Machine, Random Forest, and Extreme Gradient Boosting) to predict clinical outcomes and compared their performance to the DRAGON score. A sample of 590 patients was included in this study. Of them, 180 (30.5%) had favorable outcomes and 152 (25.8%) had miserable outcomes. All selected machine learning models outperformed the DRAGON score on accuracy of outcome prediction (Logistic Regression: 0.70, Supportive Vector Machine: 0.67, Random Forest: 0.69, and Extreme Gradient Boosting: 0.67, vs. DRAGON: 0.51, p < 0.001). Among all selected models, Logistic Regression also had a better performance than the DRAGON score on positive predictive value, sensitivity, and specificity. Compared with the DRAGON score, the multiclass machine learning approach showed better performance on the prediction of the 3-month functional outcome of acute ischemic stroke patients requiring reperfusion therapy.