Development and Validation of Artificial Intelligence-based Method for Diagnosis of Mitral Regurgitation from Chest Radiographs.

Purpose: To develop an artificial intelligence-based model to detect mitral regurgitation on chest radiographs. Materials and Methods: This retrospective study included echocardiographs and associated chest radiographs consecutively collected at a single institution between July 2016 and May 2019. Associated radiographs were those obtained within 30 days of echocardiography. These radiographs were labeled as positive or negative for mitral regurgitation on the basis of the echocardiographic reports and were divided into training, validation, and test datasets. An artificial intelligence model was developed by using the training dataset and was tuned by using the validation dataset. To evaluate the model, the area under the curve, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were assessed by using the test dataset. Results: This study included a total of 10 367 images from 5270 patients. The training dataset included 8240 images (4216 patients), the validation dataset included 1073 images (527 patients), and the test dataset included 1054 images (527 patients). The area under the curve, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value in the test dataset were 0.80 (95% CI: 0.77, 0.82), 71% (95% CI: 67, 75), 74% (95% CI: 70, 77), 73% (95% CI: 70, 75), 68% (95% CI: 64, 72), and 77% (95% CI: 73, 80), respectively. Conclusion: The developed deep learning-based artificial intelligence model may possibly differentiate patients with and without mitral regurgitation by using chest radiographs.Keywords: Computer-aided Diagnosis (CAD), Cardiac, Heart, Valves, Supervised Learning, Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning Algorithms Supplemental material is available for this article. © RSNA, 2022.

Artificial intelligence-based detection of aortic stenosis from chest radiographs.

Aims: We aimed to develop models to detect aortic stenosis (AS) from chest radiographs-one of the most basic imaging tests-with artificial intelligence. Methods and results: We used 10 433 retrospectively collected digital chest radiographs from 5638 patients to train, validate, and test three deep learning models. Chest radiographs were collected from patients who had also undergone echocardiography at a single institution between July 2016 and May 2019. These were labelled from the corresponding echocardiography assessments as AS-positive or AS-negative. The radiographs were separated on a patient basis into training [8327 images from 4512 patients, mean age 65 ± (standard deviation) 15 years], validation (1041 images from 563 patients, mean age 65 ± 14 years), and test (1065 images from 563 patients, mean age 65 ± 14 years) datasets. The soft voting-based ensemble of the three developed models had the best overall performance for predicting AS with an area under the receiver operating characteristic curve, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of 0.83 (95% confidence interval 0.77-0.88), 0.78 (0.67-0.86), 0.71 (0.68-0.73), 0.71 (0.68-0.74), 0.18 (0.14-0.23), and 0.97 (0.96-0.98), respectively, in the validation dataset and 0.83 (0.78-0.88), 0.83 (0.74-0.90), 0.69 (0.66-0.72), 0.71 (0.68-0.73), 0.23 (0.19-0.28), and 0.97 (0.96-0.98), respectively, in the test dataset. Conclusion: Deep learning models using chest radiographs have the potential to differentiate between radiographs of patients with and without AS. Lay Summary: We created artificial intelligence (AI) models using deep learning to identify aortic stenosis (AS) from chest radiographs. Three AI models were developed and evaluated with 10 433 retrospectively collected radiographs and labelled from echocardiography reports. The ensemble AI model could detect AS in a test dataset with an area under the receiver operating characteristic curve of 0.83 (95% confidence interval 0.78-0.88). Since chest radiography is a cost-effective and widely available imaging test, our model can provide an additive resource for the detection of AS.

[Estimation of myocardial ischemia by diastolic strain analysis in exercise stress echocardiography: comparison with exercise thallium-201 single photon emission computed tomography].

OBJECTIVES: The feasibility of detecting persistent regional left ventricular abnormal relaxation due to myocardial ischemia using strain echocardiography several minutes after exercise was investigated. METHODS: Consecutive 27 patients (mean age 65 +/- 9 years, 21 males, 6 females) with suspected coronary artery disease were enrolled. Strain echocardiographic images were acquired at the mid segments of the left ventricular wall before and 5 min after exercise in the apical long-axis, two-chamber and four-chamber views. Strain curves were obtained at each segment, and peak values of strain at the closure of aortic valve (A) and at one third of diastolic duration (B) were measured. Strain diastolic index (SDI) was calculated as (A - B)/A X 100%. The ratio of SDI before exercise to that after exercise was defined as the SDI ratio and compared with exercise thallium-201 single photon emission computed tomography (SPECT) as the reference standard to detect myocardial ischemia. RESULTS: A total of 162 segments were evaluated. Based on the results of exercise SPECT, 119 segments were classified as non-ischemic segments, and 43 as ischemic segments. Ischemic segments showed significant decreases in SDI before and after exercise, whereas non-ischemic segments showed no significant differences in SDI before and after exercise. SDI ratio was significantly decreased in ischemic segments, but not in non-ischemic segments. SDI ratio with a cut off value of 0.51 had a sensitivity of 91% and a specificity of 89% to detect myocardial ischemia in the receiver-operating characteristics. Conclusions. Strain echocardiography can provide quantitative assessment of myocardial ischemia by detecting post-ischemic regional left ventricular delayed relaxation even 5 min after exercise.

Usefulness of noninvasive transient elastography for assessment of liver fibrosis stage in chronic hepatitis C.

AIM: To evaluate the method of noninvasive transient elastography for assessment of histological stage of liver fibrosis in patients with chronic hepatitis C (CHC). METHODS: Two hundred and thirty-seven patients with CHC were included in this study. Liver biopsy was performed under ultrasonography on 217 of the patients, excluding twenty with clear clinical evidence of liver cirrhosis. Fifty subjects without liver disease were enrolled as a control group (stage 0). Twenty-five patients with sustained virological response (SVR) to interferon (IFN) therapy were also enrolled. These patients underwent liver biopsy before IFN therapy. Examination of liver stiffness (LS) was performed by elastography. RESULTS: Medians (50% levels) of LS were 4.1 (3.5-4.9), 6.3 (4.8-8.5), 8.8 (6.8-12.0), 14.6 (10.5-18.6), and 22.2 (15.4-28.0), respectively, in the fibrosis stages 0-4 (P < 0.001). LS was significantly correlated with four serum fibrosis markers. LS values in patients with SVR were 3.8 (3.5-5.6), 5.2 (4.4-6.8), 6.8 (6.1-7.6), and 6.1 (3.6-7.9), respectively, in the fibrosis stages 1-4. In all stages, LS for patients with SVR was significantly lower than that for patients who did not undergo IFN therapy. LS was significantly correlated with serum concentrations of hyaluronic acid, type IV collagen, type IV collagen 7S, and type III procollagen N peptide. CONCLUSION: LS correlated well with the histological stage of fibrosis. Changes in liver fibrosis stage may thus be estimated noninvasively using transient elastography.

[Usefulness of transthoracic freehand three-dimensional echocardiography for the evaluation of mitral valve prolapse].

BACKGROUND: Two-dimensional (2D) echocardiography is routinely used in evaluating patients with mitral valve prolapse but requires a systematic examination for accurate assessment of the involved lesion of mitral valve prolapse, because the sonographer is required to mentally reconstruct two-dimensional images into three dimensions. Recently, freehand three-dimensional (3D) echocardiography has been introduced in the clinical setting for three-dimensional visualization of the mitral valve apparatus. OBJECTIVES: To evaluate the accuracy of the freehand 3D echocardiography system in assessing the involved lesion in patients with mitral valve prolapse. METHODS: This study consisted of 25 consecutive patients (15 men, 10 women, mean age 55 +/- 17 years) with mitral valve prolapse who were scheduled for 3D echocardiography. Mitral valve was reconstructed in the view from the left atrium (surgeon's view) by 3D echocardiography. The location of the involved lesion in mitral valve was classified as the medial, middle and lateral portions of the anterior leaflet, and the medial, middle and lateral scallops of the posterior leaflet, respectively. The results by 3D echocardiography were compared with those of 2D echocardiography as the clinical standard. RESULTS: An adequate three-dimensional display of the entire mitral valve for analysis of the involved lesion could be reconstructed in all 25 patients (feasibility 100%). The sensitivity of 3D echocardiography for detecting the lesions at the medial, middle and lateral portions of the anterior leaflet was 80%, 100% and 75%, and the medial, middle and lateral scallops of the posterior leaflet was 100%, 100% and 0%, respectively. The specificities were 100% at all locations in the mitral valve. CONCLUSIONS: These results indicate that freehand 3D echocardiography is useful for assessment of the involved lesion of the mitral valve in patients with mitral valve prolapse.

[Usefulness of left ventricular opacification with intravenous contrast echocardiography in patients with asymptomatic negative T waves on electrocardiography].

OBJECTIVES: Patients with electrocardiography (ECG) abnormalities sometimes present without obvious symptoms or abnormal physical findings. In some cases, echocardiography fails to reveal abnormalities compatible with the ECG findings because of poor echocardiographic image quality. New intravenous contrast agents now enable opacification of the left ventricular cavity in the clinical setting. The usefulness of left ventricular opacification (LVO) by ultrasonic contrast agent was examined in asymptomatic patients with negative T waves on ECG. METHODS: Thirty-four consecutive patients (23 males, 11 females, mean age 68 +/- 10 years) with negative T waves without symptoms or abnormal physical findings underwent routine transthoracic echocardiography and LVO by intravenous injection of Levovist using harmonic imaging. RESULTS: Without LVO, abnormal echocardiographic findings compatible with the negative T waves were identified in 22 (65%) of the 34 patients. With LVO, the endocardial border of the left ventricle was observed in all patients and findings compatible with negative T waves were detected in 32 (94%) of the 34 patients, which was significantly higher than that without LVO (chi 2 = 9.79, p = 0.0055). Furthermore, LVO revealed additional findings in nine patients such as mural thrombus or apical thinning. In total, LVO revealed abnormal apical findings in 22 patients (65%), which was significantly higher than that without LVO (chi 2 = 10.0, p = 0.0013). CONCLUSIONS: LVO is useful for identifying abnormal findings, especially in the apex in asymptomatic patients with negative T waves on ECG.