Comprehensive clinical application analysis of artificial intelligence-enabled electrocardiograms for screening multiple valvular heart diseases.

BACKGROUND: Valvular heart disease (VHD) is becoming increasingly important to manage the risk of future complications. Electrocardiographic (ECG) changes may be related to multiple VHDs, and (AI)-enabled ECG has been able to detect some VHDs. We aimed to develop five deep learning models (DLMs) to identify aortic stenosis, aortic regurgitation, pulmonary regurgitation, tricuspid regurgitation, and mitral regurgitation. METHODS: Between 2010 and 2021, 77,047 patients with echocardiography and 12-lead ECG performed within 7 days were identified from an academic medical center to provide DLM development (122,728 ECGs), and internal validation (7,637 ECGs). Additional 11,800 patients from a community hospital were identified to external validation. The ECGs were classified as with or without moderate-to-severe VHDs according to transthoracic echocardiography (TTE) records, and we also collected the other echocardiographic data and follow-up TTE records to identify new-onset valvular heart diseases. RESULTS: AI-ECG adjusted for age and sex achieved areas under the curves (AUCs) of >0.84, >0.80, >0.77, >0.83, and >0.81 for detecting aortic stenosis, aortic regurgitation, pulmonary regurgitation, tricuspid regurgitation, and mitral regurgitation, respectively. Since predictions of each DLM shared similar components of ECG rhythms, the positive findings of each DLM were highly correlated with other valvular heart diseases. Of note, a total of 37.5-51.7% of false-positive predictions had at least one significant echocardiographic finding, which may lead to a significantly higher risk of future moderate-to-severe VHDs in patients with initially minimal-to-mild VHDs. CONCLUSION: AI-ECG may be used as a large-scale screening tool for detecting VHDs and a basis to undergo an echocardiography.

AI-enabled electrocardiography alert intervention and all-cause mortality: a pragmatic randomized clinical trial.

The early identification of vulnerable patients has the potential to improve outcomes but poses a substantial challenge in clinical practice. This study evaluated the ability of an artificial intelligence (AI)-enabled electrocardiogram (ECG) to identify hospitalized patients with a high risk of mortality in a multisite randomized controlled trial involving 39 physicians and 15,965 patients. The AI-ECG alert intervention included an AI report and warning messages delivered to the physicians, flagging patients predicted to be at high risk of mortality. The trial met its primary outcome, finding that implementation of the AI-ECG alert was associated with a significant reduction in all-cause mortality within 90 days: 3.6% patients in the intervention group died within 90 days, compared to 4.3% in the control group (4.3%) (hazard ratio (HR) = 0.83, 95% confidence interval (CI) = 0.70-0.99). A prespecified analysis showed that reduction in all-cause mortality associated with the AI-ECG alert was observed primarily in patients with high-risk ECGs (HR = 0.69, 95% CI = 0.53-0.90). In analyses of secondary outcomes, patients in the intervention group with high-risk ECGs received increased levels of intensive care compared to the control group; for the high-risk ECG group of patients, implementation of the AI-ECG alert was associated with a significant reduction in the risk of cardiac death (0.2% in the intervention arm versus 2.4% in the control arm, HR = 0.07, 95% CI = 0.01-0.56). While the precise means by which implementation of the AI-ECG alert led to decreased mortality are to be fully elucidated, these results indicate that such implementation assists in the detection of high-risk patients, prompting timely clinical care and reducing mortality. ClinicalTrials.gov registration: NCT05118035 .

A Deep-Learning Algorithm-Enhanced Electrocardiogram Interpretation for Detecting Pulmonary Embolism.

Background: The early diagnosis of pulmonary embolism (PE) remains a challenge. Electrocardiograms (ECGs) and D-dimer levels are used to screen potential cases. Objective: To develop a deep learning model (DLM) to detect PE using ECGs and investigate the clinical value of false detections in patients without PE. Methods: Among patients who visited the emergency department between 2011 and 2019, PE cases were identified through a review of medical records. Non-PE ECGs were collected from patients without a diagnostic code for PE. There were 113 PE and 51,456 non-PE ECGs in the training and validation sets for developing the DLM, respectively, and 27 PE and 13,105 non-PE cases in an independent testing set for performance validation. A human-machine competition was conducted from the testing set to compare the performance of the DLM with that of physicians. Receiver operating characteristic (ROC) curves, sensitivity, and specificity were used to determine the diagnostic value. Survival analysis was used to assess the prognosis of the patients without PE, stratified by DLM prediction. Results: The DLM was as effective as physicians in diagnosing PE, with 70.8% sensitivity and 69.7% specificity. The area under the ROC curve of DLM was 0.778 in the testing set and up to 0.9 with D-dimer and demographic data. The non-PE patients whose ECG was misclassified as PE by DLM had higher all-cause mortality [hazard ratio (HR) 2.13 (1.51-3.02)] and risk of non-cardiovascular hospitalization [HR 1.55 (1.42-1.68)] than those correctly classified. Conclusions: A DLM-enhanced ECG system may prompt PE recognition and provide prognostic outcomes in patients with false-positive predictions.

Mortality risk prediction of the electrocardiogram as an informative indicator of cardiovascular diseases.

Background: The electrocardiogram (ECG) may be the most popular test in the management of cardiovascular disease (CVD). Although wide applications of artificial intelligence (AI)-enabled ECG have been developed, an integrating indicator for CVD risk stratification was not investigated. Since mortality may be the most important global outcome, this study aimed to develop a survival deep learning model (DLM) to establish a critical ECG value and explore the associations with various CVD events. Methods: We trained a DLM with 451,950 12-lead resting ECGs obtained from 210,552 patients, for whom 23,592 events occurred. The internal validation set included 27,808 patients with one ECG for each patient. The external validations were performed in a community hospital with 33,047 patients and two transnational data sets with 233,647 and 1631 ECGs. We distinguished the cause of mortality and additionally investigated CVD-related outcomes, including new-onset acute myocardial infarction (AMI), stroke (STK), and heart failure (HF). Results: The DLM achieved C-indices of 0.858/0.836 in internal/external validation sets by using ECG over a 10-year period. The high-mortality-risk group identified by the proposed DLM presented a hazard ratio (HR) of 14.16 (95% confidence interval (CI): 11.33-17.70) compared to the low-risk group in the internal validation and presented a higher risk of cardiovascular (CV) mortality (HR: 18.50, 95% CI: 9.82-34.84), non-CV mortality (HR: 13.68, 95% CI: 10.76-17.38), AMI (HR: 4.01, 95% CI: 2.24-7.17), STK (HR: 2.15, 95% CI: 1.70-2.72), and HF (HR: 6.66, 95% CI: 4.54-9.77), which was consistent in an independent community hospital. The transnational validation also revealed HRs of 4.91 (95% CI: 2.63-9.16) and 2.29 (95% CI: 2.15-2.44) for all-cause mortality in the SaMi-Trop and Clinical Outcomes in Digital Electrocardiography 15% (CODE15) cohorts. Conclusions: The mortality risk by AI-enabled ECG may be applied in passive electronic-health-record-based CVD risk screening, which may identify more asymptomatic and unaware high-risk patients.

Development and validation of a dynamic deep learning algorithm using electrocardiogram to predict dyskalaemias in patients with multiple visits.

Aims: Deep learning models (DLMs) have shown superiority in electrocardiogram (ECG) analysis and have been applied to diagnose dyskalaemias. However, no study has explored the performance of DLM-enabled ECG in continuous follow-up scenarios. Therefore, we proposed a dynamic revision of DLM-enabled ECG to use personal pre-annotated ECGs to enhance the accuracy in patients with multiple visits. Methods and results: We retrospectively collected 168 450 ECGs with corresponding serum potassium (K+) levels from 103 091 patients as development samples. In the internal/external validation sets, the numbers of ECGs with corresponding K+ were 37 246/47 604 from 13 555/20 058 patients. Our dynamic revision method showed better performance than the traditional direct prediction for diagnosing hypokalaemia [area under the receiver operating characteristic curve (AUC) = 0.730/0.720-0.788/0.778] and hyperkalaemia (AUC = 0.884/0.888-0.915/0.908) in patients with multiple visits. Conclusion: Our method has shown a distinguishable improvement in DLMs for diagnosing dyskalaemias in patients with multiple visits, and we also proved its application in ejection fraction prediction, which could further improve daily clinical practice.

Extensive deep learning model to enhance electrocardiogram application via latent cardiovascular feature extraction from identity identification.

BACKGROUND AND OBJECTIVE: Deep learning models (DLMs) have been successfully applied in biomedicine primarily using supervised learning with large, annotated databases. However, scarce training resources limit the potential of DLMs for electrocardiogram (ECG) analysis. METHODS: We have developed a novel pre-training strategy for unsupervised identity identification with an area under the receiver operating characteristic curve (AUC) >0.98. Accordingly, a DLM pre-trained with identity identification can be applied to 70 patient characteristic predictions using transfer learning (TL). These ECG-based patient characteristics were then used for cardiovascular disease (CVD) risk prediction. The DLMs were trained using 507,729 ECGs from 222,473 patients and validated using two independent validation sets (n = 27,824/31,925). RESULTS: The DLMs using our method exhibited better performance than directly trained DLMs. Additionally, our DLM performed better than those of previous studies in terms of gender (AUC [internal/external] = 0.982/0.968), age (correlation = 0.886/0.892), low ejection fraction (AUC = 0.942/0.951), and critical markers not addressed previously, including high B-type natriuretic peptide (AUC = 0.921/0.899). Additionally, approximately 50% of the ECG-based characteristics provided significantly more prediction information for cardiovascular risk than real characteristics. CONCLUSIONS: This is the first study to use identity identification as a pre-training task for TL in ECG analysis. An extensive exploration of the relationship between ECG and 70 patient characteristics was conducted. Our DLM-enhanced ECG interpretation system extensively advanced ECG-related patient characteristic prediction and mortality risk management for cardiovascular diseases.

Artificial intelligence-enabled electrocardiogram screens low left ventricular ejection fraction with a degree of confidence.

Background: Artificial intelligence-enabled electrocardiogram has become a substitute tool for echocardiography in left ventricular ejection fraction estimation. However, the direct use of artificial intelligence-enabled electrocardiogram may be not trustable due to the uncertainty of the prediction. Objective: The study aimed to establish an artificial intelligence-enabled electrocardiogram with a degree of confidence to identify left ventricular dysfunction. Methods: The study collected 76,081 and 11,771 electrocardiograms from an academic medical center and a community hospital to establish and validate the deep learning model, respectively. The proposed deep learning model provided the point estimation of the actual ejection fraction and its standard deviation derived from the maximum probability density function of a normal distribution. The primary analysis focused on the accuracy of identifying patients with left ventricular dysfunction (ejection fraction ≤ 40%). Since the standard deviation was an uncertainty indicator in a normal distribution, we used it as a degree of confidence in the artificial intelligence-enabled electrocardiogram. We further explored the clinical application of estimated standard deviation and followed up on the new-onset left ventricular dysfunction in patients with initially normal ejection fraction. Results: The area under receiver operating characteristic curves (AUC) of detecting left ventricular dysfunction were 0.9549 and 0.9365 in internal and external validation sets. After excluding the cases with a lower degree of confidence, the artificial intelligence-enabled electrocardiogram performed better in the remaining cases in internal (AUC = 0.9759) and external (AUC = 0.9653) validation sets. For the application of future left ventricular dysfunction risk stratification in patients with initially normal ejection fraction, a 4.57-fold risk of future left ventricular dysfunction when the artificial intelligence-enabled electrocardiogram is positive in the internal validation set. The hazard ratio was increased to 8.67 after excluding the cases with a lower degree of confidence. This trend was also validated in the external validation set. Conclusion: The deep learning model with a degree of confidence can provide advanced improvements in identifying left ventricular dysfunction and serve as a decision support and management-guided screening tool for prognosis.

Artificial Intelligence-Enabled Electrocardiography Predicts Left Ventricular Dysfunction and Future Cardiovascular Outcomes: A Retrospective Analysis.

BACKGROUND: The ejection fraction (EF) provides critical information about heart failure (HF) and its management. Electrocardiography (ECG) is a noninvasive screening tool for cardiac electrophysiological activities that has been used to detect patients with low EF based on a deep learning model (DLM) trained via large amounts of data. However, no studies have widely investigated its clinical impacts. OBJECTIVE: This study developed a DLM to estimate EF via ECG (ECG-EF). We further investigated the relationship between ECG-EF and echo-based EF (ECHO-EF) and explored their contributions to future cardiovascular adverse events. METHODS: There were 57,206 ECGs with corresponding echocardiograms used to train our DLM. We compared a series of training strategies and selected the best DLM. The architecture of the DLM was based on ECG12Net, developed previously. Next, 10,762 ECGs were used for validation, and another 20,629 ECGs were employed to conduct the accuracy test. The changes between ECG-EF and ECHO-EF were evaluated. The primary follow-up adverse events included future ECHO-EF changes and major adverse cardiovascular events (MACEs). RESULTS: The sex-/age-matching strategy-trained DLM achieved the best area under the curve (AUC) of 0.9472 with a sensitivity of 86.9% and specificity of 89.6% in the follow-up cohort, with a correlation of 0.603 and a mean absolute error of 7.436. In patients with accurate prediction (initial difference < 10%), the change traces of ECG-EF and ECHO-EF were more consistent (R-square = 0.351) than in all patients (R-square = 0.115). Patients with lower ECG-EF (≤35%) exhibited a greater risk of cardiovascular (CV) complications, delayed ECHO-EF recovery, and earlier ECHO-EF deterioration than patients with normal ECG-EF (>50%). Importantly, ECG-EF demonstrated an independent impact on MACEs and all CV adverse outcomes, with better prediction of CV outcomes than ECHO-EF. CONCLUSIONS: The ECG-EF could be used to initially screen asymptomatic left ventricular dysfunction (LVD) and it could also independently contribute to the predictions of future CV adverse events. Although further large-scale studies are warranted, DLM-based ECG-EF could serve as a promising diagnostic supportive and management-guided tool for CV disease prediction and the care of patients with LVD.

Artificial Intelligence-Enabled Electrocardiogram Estimates Left Atrium Enlargement as a Predictor of Future Cardiovascular Disease.

BACKGROUND: Left atrium enlargement (LAE) can be used as a predictor of future cardiovascular diseases, including hypertension (HTN) and atrial fibrillation (Afib). Typical electrocardiogram (ECG) changes have been reported in patients with LAE. This study developed a deep learning model (DLM)-enabled ECG system to identify patients with LAE. METHOD: Patients who had ECG records with corresponding echocardiography (ECHO) were included. There were 101,077 ECGs, 20,510 ECGs, 7611 ECGs, and 11,753 ECGs in the development, tuning, internal validation, and external validation sets, respectively. We evaluated the performance of a DLM-enabled ECG for diagnosing LAE and explored the prognostic value of ECG-LAE for new-onset HTN, new-onset stroke (STK), new-onset mitral regurgitation (MR), and new-onset Afib. RESULTS: The DLM-enabled ECG achieved AUCs of 0.8127/0.8176 for diagnosing mild LAE, 0.8587/0.8688 for diagnosing moderate LAE, and 0.8899/0.8990 for diagnosing severe LAE in the internal/external validation sets. Notably, ECG-LAE had higher prognostic value compared to ECHO-LAE, which had C-indices of 0.711/0.714 compared to 0.695/0.692 for new-onset HTN, 0.676/0.688 compared to 0.663/0.677 for new-onset STK, 0.696/0.695 compared to 0.676/0.673 for new-onset MR, and 0.800/0.806 compared to 0.786/0.760 for new-onset Afib in internal/external validation sets, respectively. CONCLUSIONS: A DLM-enabled ECG could be considered as a LAE screening tool and provide better prognostic information for related cardiovascular diseases.

Detection of Left Ventricular Systolic Dysfunction Using an Artificial Intelligence-Enabled Chest X-Ray.

BACKGROUND: Assessment of left ventricular systolic dysfunction provides essential information related to the prognosis and management of cardiovascular diseases. The aim of this study was to develop a deep-learning model to identify left ventricular ejection fraction (LVEF) ≤ 35% via chest X-ray (CXR [CXR-EF≤35%]) features and investigate the performance and clinical implications. METHODS: This study collected 90,547 CXRs with the corresponding LVEF according to transthoracic echocardiography from the outpatient department in an academic medical center. Among these, 77,227 CXRs were used to develop the identification of CXR-EF≤35%. Another 13,320 CXRs were used to validate the performance, which was evaluated by area under the receiver operating characteristic curve (AUC). Furthermore, CXR-EF≤35% was tested to assess the long-term risks of developing LVEF ≤ 35% and cardiovascular outcomes, which were evaluated by Kaplan-Meier survival analysis and the Cox proportional hazards model. RESULTS: The AUCs of CXR-EF≤35% for the detection of LVEF ≤ 35% were 0.888 and 0.867 in the internal and external validation cohorts, respectively. Patients with baseline LVEF > 50% but detected as CXR-EF≤35% were at higher risk of long-term development of LVEF ≤ 35% (hazard ratio, internal validation cohort [HRi] 3.91, 95% CI 2.98-5.14; hazard ratio, external validation cohort [HRe] 2.49, 95% CI 1.89-3.27). Furthermore, patients detected as LVEF ≤ 35% by CXR-EF≤35% had significantly higher future risks of all-cause mortality (HRi 1.40, 95% CI 1.15-1.71; HRe 1.38, 95% CI 1.15-1.66), cardiovascular mortality (HRi 3.02, 95% CI 1.84-4.98; HRe 2.60, 95% CI 1.77-3.82), and new-onset atrial fibrillation (HRi 2.81, 95% CI 2.15-3.66; HRe 2.93, 95% CI 2.34-3.67) compared with those detected as no LVEF ≤ 35%. CONCLUSIONS: CXR-EF≤35% may serve as a screening tool for early detection of LVEF ≤ 35% and could independently contribute to predictions of long-term development of LVEF ≤ 35% and cardiovascular outcomes. Further prospective studies are needed to confirm the model performance.

Propofol-based total intravenous anesthesia is associated with better survival than desflurane anesthesia in glioblastoma surgery.

BACKGROUND: Previous research has shown that anesthetic techniques can influence patient outcomes following cancer surgery. However, the effects of anesthesia in patients undergoing glioblastoma surgery are still not known. We studied the relationship between the type of anesthesia and patient outcomes following elective glioblastoma surgery. METHODS: This was a retrospective cohort study of patients who underwent elective glioblastoma surgery between January 2008 and December 2018. Patients were grouped according to the anesthesia they received, desflurane or propofol. A Kaplan-Meier analysis was conducted, and survival curves were presented from the date of surgery to death. Univariable and multivariable Cox regression models were used to compare hazard ratios for death after propensity matching. RESULTS: A total of 50 patients (45 deaths, 90.0%) under desflurane anesthesia and 53 patients (38 deaths, 72.0%) under propofol anesthesia were included. Thirty-eight patients remained in each group after propensity matching. Propofol anesthesia was associated with improved survival (hazard ratio, 0.51; 95% confidence interval, 0.30-0.85; P = 0.011) in a matched analysis. Furthermore, patients under propofol anesthesia exhibited less postoperative recurrence than those under desflurane anesthesia (hazard ratio, 0.60; 95% confidence interval, 0.37-0.98; P = 0.040) in a matched analysis. CONCLUSIONS: In this limited sample size, we observed that propofol anesthesia was associated with improved survival and less postoperative recurrence in glioblastoma surgery than desflurane anesthesia. Further investigations are needed to examine the influence of propofol anesthesia on patient outcomes following glioblastoma surgery.

Deep Learning Algorithm for Management of Diabetes Mellitus via Electrocardiogram-Based Glycated Hemoglobin (ECG-HbA1c): A Retrospective Cohort Study.

BACKGROUND: glycated hemoglobin (HbA1c) provides information on diabetes mellitus (DM) management. Electrocardiography (ECG) is a noninvasive test of cardiac activity that has been determined to be related to DM and its complications. This study developed a deep learning model (DLM) to estimate HbA1c via ECG. METHODS: there were 104,823 ECGs with corresponding HbA1c or fasting glucose which were utilized to train a DLM for calculating ECG-HbA1c. Next, 1539 cases from outpatient departments and health examination centers provided 2190 ECGs for initial validation, and another 3293 cases with their first ECGs were employed to analyze its contributions to DM management. The primary analysis was used to distinguish patients with and without mild to severe DM, and the secondary analysis was to explore the predictive value of ECG-HbA1c for future complications, which included all-cause mortality, new-onset chronic kidney disease (CKD), and new-onset heart failure (HF). RESULTS: we used a gender/age-matching strategy to train a DLM to achieve the best AUCs of 0.8255 with a sensitivity of 71.9% and specificity of 77.7% in a follow-up cohort with correlation of 0.496 and mean absolute errors of 1.230. The stratified analysis shows that DM presented in patients with fewer comorbidities was significantly more likely to be detected by ECG-HbA1c. Patients with higher ECG-HbA1c under the same Lab-HbA1c exhibited worse physical conditions. Of interest, ECG-HbA1c may contribute to the mortality (gender/age adjusted hazard ratio (HR): 1.53, 95% conference interval (CI): 1.08-2.17), new-onset CKD (HR: 1.56, 95% CI: 1.30-1.87), and new-onset HF (HR: 1.51, 95% CI: 1.13-2.01) independently of Lab-HbA1c. An additional impact of ECG-HbA1c on the risk of all-cause mortality (C-index: 0.831 to 0.835, p < 0.05), new-onset CKD (C-index: 0.735 to 0.745, p < 0.01), and new-onset HF (C-index: 0.793 to 0.796, p < 0.05) were observed in full adjustment models. CONCLUSION: the ECG-HbA1c could be considered as a novel biomarker for screening DM and predicting the progression of DM and its complications.

A deep learning algorithm for detecting acute myocardial infarction.

BACKGROUND: Delayed diagnosis or misdiagnosis of acute myocardial infarction (AMI) is not unusual in daily practice. Since a 12-lead electrocardiogram (ECG) is crucial for the detection of AMI, a systematic algorithm to strengthen ECG interpretation may have important implications for improving diagnosis. AIMS: We aimed to develop a deep learning model (DLM) as a diagnostic support tool based on a 12-lead electrocardiogram. METHODS: This retrospective cohort study included 1,051/697 ECGs from 737/287 coronary angiogram (CAG)-validated STEMI/NSTEMI patients and 140,336 ECGs from 76,775 non-AMI patients at the emergency department. The DLM was trained and validated in 80% and 20% of these ECGs. A human-machine competition was conducted. The area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were used to evaluate the performance of the DLM. RESULTS: The AUC of the DLM for STEMI detection was 0.976 in the human-machine competition, which was significantly better than that of the best physicians. Furthermore, the DLM independently demonstrated sufficient diagnostic capacity for STEMI detection (AUC=0.997; sensitivity, 98.4%; specificity, 96.9%). Regarding NSTEMI detection, the AUC of the combined DLM and conventional cardiac troponin I (cTnI) increased to 0.978, which was better than that of either the DLM (0.877) or cTnI (0.950). CONCLUSIONS: The DLM may serve as a timely, objective and precise diagnostic decision support tool to assist emergency medical system-based networks and frontline physicians in detecting AMI and subsequently initiating reperfusion therapy.

Propofol-based total intravenous anesthesia did not improve survival compared to desflurane anesthesia in breast cancer surgery.

BACKGROUND: Breast cancer is the most common cancer in women and several perioperative factors may account for tumor recurrence and metastasis. The anesthetic agents employed during cancer surgery might play a crucial role in cancer cell survival and patient outcomes. We conducted a retrospective cohort study to investigate the relationship between the type of anesthesia and overall survival in patients who underwent breast cancer surgery performed by one experienced surgeon. METHODS: All patients who underwent breast cancer surgery by an experienced surgeon between January 2006 and December 2010 were included in this study. Patients were separated into two groups according to the use of desflurane or propofol anesthesia during surgery. Locoregional recurrence and overall survival rates were assessed for the two groups (desflurane or propofol anesthesia). Univariable and multivariable Cox regression models and propensity score matching analyses were used to compare the hazard ratios for death and adjust for potential confounders (age, body mass index, American Society of Anesthesiologists physical status classification, TNM stage, neoadjuvant chemotherapy, Charlson Comorbidity Index, anesthesiologists, and functional status). RESULTS: Of the 976 breast cancer patients, 632 patients underwent breast cancer surgery with desflurane anesthesia, while 344 received propofol anesthesia. After propensity scoring, 592 patients remained in the desflurane group and 296 patients in the propofol group. The mortality rate was similar in the desflurane (38 deaths, 4%) and propofol (22 deaths, 4%; p = 0.812) groups in 5-year follow-up. The crude hazard ratio (HR) for all patients was 1.13 (95% confidence interval [CI] 0.67-1.92, p = 0.646). No significant difference in the locoregional recurrence or overall 5-year survival rates were found after breast surgery using desflurane or propofol anesthesia (p = 0.454). Propensity score-matched analyses demonstrated similar outcomes in both groups. Patients who received propofol anesthesia had a higher mortality rate than those who received desflurane anesthesia in the matched groups (7% vs 6%, respectively) without significant difference (p = 0.561). In the propensity score-matched analyses, univariable analysis showed an insignificant finding (HR = 1.23, 95% CI 0.72-2.11, p = 0.449). After adjustment for the time since the earliest included patient, the HR remained insignificant (HR = 1.23, 95% CI 0.70-2.16, p = 0.475). CONCLUSION: In our non-randomized retrospective analysis, neither propofol nor desflurane anesthesia for breast cancer surgery by an experienced surgeon can affect patient prognosis and survival. The influence of propofol anesthesia on breast cancer outcome requires further investigation.

Projection Word Embedding Model With Hybrid Sampling Training for Classifying ICD-10-CM Codes: Longitudinal Observational Study.

BACKGROUND: Most current state-of-the-art models for searching the International Classification of Diseases, Tenth Revision Clinical Modification (ICD-10-CM) codes use word embedding technology to capture useful semantic properties. However, they are limited by the quality of initial word embeddings. Word embedding trained by electronic health records (EHRs) is considered the best, but the vocabulary diversity is limited by previous medical records. Thus, we require a word embedding model that maintains the vocabulary diversity of open internet databases and the medical terminology understanding of EHRs. Moreover, we need to consider the particularity of the disease classification, wherein discharge notes present only positive disease descriptions. OBJECTIVE: We aimed to propose a projection word2vec model and a hybrid sampling method. In addition, we aimed to conduct a series of experiments to validate the effectiveness of these methods. METHODS: We compared the projection word2vec model and traditional word2vec model using two corpora sources: English Wikipedia and PubMed journal abstracts. We used seven published datasets to measure the medical semantic understanding of the word2vec models and used these embeddings to identify the three-character-level ICD-10-CM diagnostic codes in a set of discharge notes. On the basis of embedding technology improvement, we also tried to apply the hybrid sampling method to improve accuracy. The 94,483 labeled discharge notes from the Tri-Service General Hospital of Taipei, Taiwan, from June 1, 2015, to June 30, 2017, were used. To evaluate the model performance, 24,762 discharge notes from July 1, 2017, to December 31, 2017, from the same hospital were used. Moreover, 74,324 additional discharge notes collected from seven other hospitals were tested. The F-measure, which is the major global measure of effectiveness, was adopted. RESULTS: In medical semantic understanding, the original EHR embeddings and PubMed embeddings exhibited superior performance to the original Wikipedia embeddings. After projection training technology was applied, the projection Wikipedia embeddings exhibited an obvious improvement but did not reach the level of original EHR embeddings or PubMed embeddings. In the subsequent ICD-10-CM coding experiment, the model that used both projection PubMed and Wikipedia embeddings had the highest testing mean F-measure (0.7362 and 0.6693 in Tri-Service General Hospital and the seven other hospitals, respectively). Moreover, the hybrid sampling method was found to improve the model performance (F-measure=0.7371/0.6698). CONCLUSIONS: The word embeddings trained using EHR and PubMed could understand medical semantics better, and the proposed projection word2vec model improved the ability of medical semantics extraction in Wikipedia embeddings. Although the improvement from the projection word2vec model in the real ICD-10-CM coding task was not substantial, the models could effectively handle emerging diseases. The proposed hybrid sampling method enables the model to behave like a human expert.

Propofol-based Total Intravenous Anesthesia Is Associated with Better Survival Than Desflurane Anesthesia in Colon Cancer Surgery.

WHAT WE ALREADY KNOW ABOUT THIS TOPIC: WHAT THIS ARTICLE TELLS US THAT IS NEW: BACKGROUND:: Previous research has shown different effects of anesthetics on cancer cell growth. Here, the authors investigated the association between type of anesthetic and patient survival after elective colon cancer surgery. METHODS: A retrospective cohort study included patients who received elective colon cancer surgery between January 2005 and December 2014. Patients were grouped according to anesthesia received: propofol or desflurane. After exclusion of those who received combined propofol anesthesia with inhalation anesthesia or epidural anesthesia, survival curves were constructed from the date of surgery to death. After propensity matching, univariable and multivariable Cox regression models were used to compare hazard ratios for death. Subgroup analyses were performed for tumor-node-metastasis staging and postoperative metastasis. RESULTS: A total of 706 patients (307 deaths, 43.5%) with desflurane anesthesia and 657 (88 deaths, 13.4%) with propofol anesthesia were eligible for analysis. After propensity matching, 579 patients remained in each group (189 deaths, 32.6%, in the desflurane group vs. 87, 15.0%, in the propofol group). In the matched analyses, the propofol-treated group had a better survival, irrespective of lower tumor-node-metastasis stage (hazard ratio, 0.22; 95% CI, 0.11 to 0.42; P < 0.001) or higher tumor-node-metastasis stage (hazard ratio, 0.42; 95% CI, 0.32 to 0.55; P < 0.001) and presence of metastases (hazard ratio, 0.67; 95% CI, 0.51 to 0.86; P = 0.002) or absence of metastases (hazard ratio, 0.08; 95% CI, 0.01 to 0.62; P = 0.016). Simple propensity score adjustment produced similar findings. CONCLUSIONS: Propofol anesthesia for colon cancer surgery is associated with better survival irrespective of tumor-node-metastasis stage.

Artificial Intelligence Learning Semantics via External Resources for Classifying Diagnosis Codes in Discharge Notes.

BACKGROUND: Automated disease code classification using free-text medical information is important for public health surveillance. However, traditional natural language processing (NLP) pipelines are limited, so we propose a method combining word embedding with a convolutional neural network (CNN). OBJECTIVE: Our objective was to compare the performance of traditional pipelines (NLP plus supervised machine learning models) with that of word embedding combined with a CNN in conducting a classification task identifying International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis codes in discharge notes. METHODS: We used 2 classification methods: (1) extracting from discharge notes some features (terms, n-gram phrases, and SNOMED CT categories) that we used to train a set of supervised machine learning models (support vector machine, random forests, and gradient boosting machine), and (2) building a feature matrix, by a pretrained word embedding model, that we used to train a CNN. We used these methods to identify the chapter-level ICD-10-CM diagnosis codes in a set of discharge notes. We conducted the evaluation using 103,390 discharge notes covering patients hospitalized from June 1, 2015 to January 31, 2017 in the Tri-Service General Hospital in Taipei, Taiwan. We used the receiver operating characteristic curve as an evaluation measure, and calculated the area under the curve (AUC) and F-measure as the global measure of effectiveness. RESULTS: In 5-fold cross-validation tests, our method had a higher testing accuracy (mean AUC 0.9696; mean F-measure 0.9086) than traditional NLP-based approaches (mean AUC range 0.8183-0.9571; mean F-measure range 0.5050-0.8739). A real-world simulation that split the training sample and the testing sample by date verified this result (mean AUC 0.9645; mean F-measure 0.9003 using the proposed method). Further analysis showed that the convolutional layers of the CNN effectively identified a large number of keywords and automatically extracted enough concepts to predict the diagnosis codes. CONCLUSIONS: Word embedding combined with a CNN showed outstanding performance compared with traditional methods, needing very little data preprocessing. This shows that future studies will not be limited by incomplete dictionaries. A large amount of unstructured information from free-text medical writing will be extracted by automated approaches in the future, and we believe that the health care field is about to enter the age of big data.