[Development of a Deep Learning Model for Judging Late Gadolinium-enhancement in Cardiac MRI].

PURPOSE: To verify the usefulness of a deep learning model for determining the presence or absence of contrast-enhanced myocardium in late gadolinium-enhancement images in cardiac MRI. METHODS: We used 174 late gadolinium-enhancement myocardial short-axis images obtained from contrast-enhanced cardiac MRI performed using a 3.0T MRI system at the University of Tokyo Hospital. Of these, 144 images were used for training, extracting a region of interest targeting the heart, scaling signal intensity, and data augmentation were performed to obtain 3312 training images. The interpretation report of two cardiology specialists of our hospital was used as the correct label. A learning model was constructed using a convolutional neural network and applied to 30 test data. In all cases, the acquired mean age was 56.4±12.1 years, and the male-to-female ratio was 1 : 0.82. RESULTS: Before and after data augmentation, sensitivity remained consistent at 93.3%, specificity improved from 0.0% to 100.0%, and accuracy improved from 46.7% to 96.7%. CONCLUSION: The prediction accuracy of the deep learning model developed in this research is high, suggesting its high usefulness.

Prediction of pacemaker-induced cardiomyopathy using a convolutional neural network based on clinical findings prior to pacemaker implantation.

Risk factors for pacemaker-induced cardiomyopathy (PICM) have been previously reported, including a high burden of right ventricular pacing, lower left ventricular ejection fraction, a wide QRS duration, and left bundle branch block before pacemaker implantation (PMI). However, predicting the development of PICM remains challenging. This study aimed to use a convolutional neural network (CNN) model, based on clinical findings before PMI, to predict the development of PICM. Out of a total of 561 patients with dual-chamber PMI, 165 (mean age 71.6 years, 89 men [53.9%]) who underwent echocardiography both before and after dual-chamber PMI were enrolled. During a mean follow-up period of 1.7 years, 47 patients developed PICM. A CNN algorithm for prediction of the development of PICM was constructed based on a dataset prior to PMI that included 31 variables such as age, sex, body mass index, left ventricular ejection fraction, left ventricular end-diastolic diameter, left ventricular end-systolic diameter, left atrial diameter, severity of mitral regurgitation, severity of tricuspid regurgitation, ischemic heart disease, diabetes mellitus, hypertension, heart failure, New York Heart Association class, atrial fibrillation, the etiology of bradycardia (sick sinus syndrome or atrioventricular block) , right ventricular (RV) lead tip position (apex, septum, left bundle, His bundle, RV outflow tract), left bundle branch block, QRS duration, white blood cell count, haemoglobin, platelet count, serum total protein, albumin, aspartate transaminase, alanine transaminase, estimated glomerular filtration rate, sodium, potassium, C-reactive protein, and brain natriuretic peptide. The accuracy, sensitivity, specificity, and area under the curve of the CNN model were 75.8%, 55.6%, 83.3% and 0.78 respectively. The CNN model could accurately predict the development of PICM using clinical findings before PMI. This model could be useful for screening patients at risk of developing PICM, ensuring timely upgrades to physiological pacing to avoid missing the optimal intervention window.

Optimization of treatment strategy by using a machine learning model to predict survival time of patients with malignant glioma after radiotherapy.

The purpose of this study was to predict the survival time of patients with malignant glioma after radiotherapy with high accuracy by considering additional clinical factors and optimize the prescription dose and treatment duration for individual patient by using a machine learning model. A total of 35 patients with malignant glioma were included in this study. The candidate features included 12 clinical features and 192 dose-volume histogram (DVH) features. The appropriate input features and parameters of the support vector machine (SVM) were selected using the genetic algorithm based on Akaike's information criterion, i.e. clinical, DVH, and both clinical and DVH features. The prediction accuracy of the SVM models was evaluated through a leave-one-out cross-validation test with residual error, which was defined as the absolute difference between the actual and predicted survival times after radiotherapy. Moreover, the influences of various values of prescription dose and treatment duration on the predicted survival time were evaluated. The prediction accuracy was significantly improved with the combined use of clinical and DVH features compared with the separate use of both features (P < 0.01, Wilcoxon signed rank test). Mean ± standard deviation of the leave-one-out cross-validation using the combined clinical and DVH features, only clinical features and only DVH features were 104.7 ± 96.5, 144.2 ± 126.1 and 204.5 ± 186.0 days, respectively. The prediction accuracy could be improved with the combination of clinical and DVH features, and our results show the potential to optimize the treatment strategy for individual patients based on a machine learning model.

Enantioselective synthesis of (S)-2-cyano-2-methylpentanoic acid by nitrilase.

The nitrilase gene of Rhodococcus rhodochrous J1 was expressed in Escherichia coli using the expression vector, pKK223-3. The recombinant E. coli JM109 cells hydrolyzed enantioselectively 2-methyl-2-propylmalononitrile to form (S)-2-cyano-2-methylpentanoic acid (CMPA) with 96 % e.e. Under optimized conditions, 80 g (S)-CMPA l(-1) was produced with a molar yield of 97 % at 30 °C after a 24 h without any by-products.