Predicting Postoperative Mortality With Deep Neural Networks and Natural Language Processing: Model Development and Validation.

BACKGROUND: Machine learning (ML) achieves better predictions of postoperative mortality than previous prediction tools. Free-text descriptions of the preoperative diagnosis and the planned procedure are available preoperatively. Because reading these descriptions helps anesthesiologists evaluate the risk of the surgery, we hypothesized that deep learning (DL) models with unstructured text could improve postoperative mortality prediction. However, it is challenging to extract meaningful concept embeddings from this unstructured clinical text. OBJECTIVE: This study aims to develop a fusion DL model containing structured and unstructured features to predict the in-hospital 30-day postoperative mortality before surgery. ML models for predicting postoperative mortality using preoperative data with or without free clinical text were assessed. METHODS: We retrospectively collected preoperative anesthesia assessments, surgical information, and discharge summaries of patients undergoing general and neuraxial anesthesia from electronic health records (EHRs) from 2016 to 2020. We first compared the deep neural network (DNN) with other models using the same input features to demonstrate effectiveness. Then, we combined the DNN model with bidirectional encoder representations from transformers (BERT) to extract information from clinical texts. The effects of adding text information on the model performance were compared using the area under the receiver operating characteristic curve (AUROC) and the area under the precision-recall curve (AUPRC). Statistical significance was evaluated using P<.05. RESULTS: The final cohort contained 121,313 patients who underwent surgeries. A total of 1562 (1.29%) patients died within 30 days of surgery. Our BERT-DNN model achieved the highest AUROC (0.964, 95% CI 0.961-0.967) and AUPRC (0.336, 95% CI 0.276-0.402). The AUROC of the BERT-DNN was significantly higher compared to logistic regression (AUROC=0.952, 95% CI 0.949-0.955) and the American Society of Anesthesiologist Physical Status (ASAPS AUROC=0.892, 95% CI 0.887-0.896) but not significantly higher compared to the DNN (AUROC=0.959, 95% CI 0.956-0.962) and the random forest (AUROC=0.961, 95% CI 0.958-0.964). The AUPRC of the BERT-DNN was significantly higher compared to the DNN (AUPRC=0.319, 95% CI 0.260-0.384), the random forest (AUPRC=0.296, 95% CI 0.239-0.360), logistic regression (AUPRC=0.276, 95% CI 0.220-0.339), and the ASAPS (AUPRC=0.149, 95% CI 0.107-0.203). CONCLUSIONS: Our BERT-DNN model has an AUPRC significantly higher compared to previously proposed models using no text and an AUROC significantly higher compared to logistic regression and the ASAPS. This technique helps identify patients with higher risk from the surgical description text in EHRs.

Using Text Content From Coronary Catheterization Reports to Predict 5-Year Mortality Among Patients Undergoing Coronary Angiography: A Deep Learning Approach.

Background: Current predictive models for patients undergoing coronary angiography have complex parameters which limit their clinical application. Coronary catheterization reports that describe coronary lesions and the corresponding interventions provide information of the severity of the coronary artery disease and the completeness of the revascularization. This information is relevant for predicting patient prognosis. However, no predictive model has been constructed using the text content from coronary catheterization reports before. Objective: To develop a deep learning model using text content from coronary catheterization reports to predict 5-year all-cause mortality and 5-year cardiovascular mortality for patients undergoing coronary angiography and to compare the performance of the model to the established clinical scores. Method: This retrospective cohort study was conducted between January 1, 2006, and December 31, 2015. Patients admitted for coronary angiography were enrolled and followed up until August 2019. The main outcomes were 5-year all-cause mortality and 5-year cardiovascular mortality. In total, 11,576 coronary catheterization reports were collected. BioBERT (bidirectional encoder representations from transformers for biomedical text mining), which is a BERT-based model in the biomedical domain, was utilized to construct the model. The area under the receiver operating characteristic curve (AUC) was used to assess model performance. We also compared our results to the residual SYNTAX (SYNergy between PCI with TAXUS and Cardiac Surgery) score. Results: The dataset was divided into the training (60%), validation (20%), and test (20%) sets. The mean age of the patients in each dataset was 65.5 ± 12.1, 65.4 ± 11.2, and 65.6 ± 11.2 years, respectively. A total of 1,411 (12.2%) patients died, and 664 (5.8%) patients died of cardiovascular causes within 5 years after coronary angiography. The best of our models had an AUC of 0.822 (95% CI, 0.790-0.855) for 5-year all-cause mortality, and an AUC of 0.858 (95% CI, 0.816-0.900) for 5-year cardiovascular mortality. We randomly selected 300 patients who underwent percutaneous coronary intervention (PCI), and our model outperformed the residual SYNTAX score in predicting 5-year all-cause mortality (AUC, 0.867 [95% CI, 0.813-0.921] vs. 0.590 [95% CI, 0.503-0.684]) and 5-year cardiovascular mortality (AUC, 0.880 [95% CI, 0.873-0.925] vs. 0.649 [95% CI, 0.535-0.764]), respectively, after PCI among these patients. Conclusions: We developed a predictive model using text content from coronary catheterization reports to predict the 5-year mortality in patients undergoing coronary angiography. Since interventional cardiologists routinely write reports after procedures, our model can be easily implemented into the clinical setting.

Electrochemical SERS on 2D Mapping for Metabolites Detection.

Surface-enhanced Raman scattering (SERS) has been widely used for bioanalysis because it provides a high sensitivity for detecting analytes of ultralow concentrations. However, the clinical application of a 2D SERS-active substrate remains challenging because of the difficulty of obtaining accurate quantification, especially at low concentration. In this study, we proposed an analytical method that integrates an optimized sample mapping strategy with an electrochemical SERS (EC-SERS) technique to resolve this problem. We adopted this method to detect two metabolites of azathioprine, namely 6-thioguanine nucleotides (6-TGNs) and 6-methylmercaptopurine (6-MMP), as our proof-of-concept experiment. We first prepared a conductive SERS-active substrate by electrochemically depositing Au nanoparticles (AuNPs) on indium tin oxide glass. The two metabolites were then randomly absorbed on the surface of the AuNPs of the SERS-active substrates. When we applied a negative potential on the substrate, we observed a large enhancement of Raman intensity for both metabolites, which was attributed to both the charge transfer effect and reorientation of metabolites on the substrate surface, leading to the formation of Au-S bonds. In addition, by optimizing the mapping range, we were able to efficiently reduce the standard deviation of SERS intensity and achieve a consistent standard deviation lower than 10%. With these two features, we were able to achieve quantitative analysis of 6-TGNs and 6-MMP with a detection limit of 10 and 100 nM, respectively. The integration of EC-SERS and the mapping method provided a reliable and quantitative analytical platform for analytes, which can be electrochemically modulated, like 6-TGNs and 6-MMP.

Electrochemical SERS for in Situ Monitoring the Redox States of PEDOT and Its Potential Application in Oxidant Detection.

In response to recent developments for applying conducting polymers on various biomedical applications, the development of characterization techniques for evaluating the states of conducting polymers in liquids is beneficial to the applications of these materials. In this study, we propose a platform using electrochemical surface-enhanced Raman scattering (EC-SERS) technology, which allows a direct measurement of the redox states of conducing polymers in liquids. A thiophene-based conducting polymer, hydroxymethyl poly(3,4-ethylenedioxythiophene) or poly(EDOT-OH), was used to demonstrate this concept. Poly(EDOT-OH) films were coated on Au nanoparticle-coated ITO glass as SERS-active substrates. Taking the advantage of Raman enhancement, we can in situ and clearly monitor the redox behavior of poly(EDOT-OH) in aqueous solutions. The Raman peak intensity decreases as the poly(EDOT-OH) film is oxidized. Furthermore, we demonstrated our idea to utilize this phenomenon as the sensing mechanism for oxidant detection. The Raman intensity of conducting polymers reduces faster when oxidants exist, and we obtain a quantitative analysis for the detection of oxidants. Moreover, the oxidized poly(EDOT-OH) films can be reused for detection of oxidants simply by applying a reduction potential to activate the poly(EDOT-OH) films. The film stability was also confirmed, and the detection of two other oxidants, namely ammonium persulfate and iron chloride, were also demonstrated. The results show different SERS spectra of poly(EDOT-OH) films oxidized by using different oxidants. Besides, the oxidized films can be easily recovered simply by applying a cathodic potential, which allows repeating usage and makes it possible for continuous monitoring applications. To the best of our knowledge, this is the first time to apply PEDOT's Raman feature for detection purposes.