Search

Prognostic markers for long-term outcomes are lacking in patients with deferred (nonculprit) coronary artery lesions. This study aimed to identify the morphological criteria for predicting adverse outcomes and validate their clinical impact. Using deep learning models, we extracted geometrical parameters and maximal attenuation (or calcium) burden index (ABI-max or CBI-max) from the intravascular ultrasound (IVUS) images of nonculprit vessels in 1115 patients. The endpoints included cardiac death, myocardial infarction, and target vessel revascularization of nonculprit vessel. Cardiac death occurred in 27 (2.4 %) patients at 3 years and 39 (3.5 %) patients at 5 years. At 5 years, the cardiac death-free survival rate was significantly lower with ABP-max ≥11.37 % vs. < 11.37 % (90.0 % vs. 98.7 %), CBI-max ≥13.40 % vs. < 13.40 % (92.8 % vs. 98.4 %), and percent atheroma volume ≥ 51.35 % vs. < 51.35 % (94.0 % vs. 97.7) (all log-rank p < 0.001). The independent predictors of 5-year cardiovascular mortality were age (hazard ratio [HR] 1.21), female sex (HR 0.33), history of heart failure (HR 6.06), chronic kidney disease (HR 18.28), ABI-max (HR 1.04), and CBI-max (HR 1.05). The independent determinants of 5-year target vessel revascularization of nonculprit vessel were fractional flow reserve (HR 0.95 per 0.01 increase), minimal lumen area (HR 0.63), and plaque burden (HR 1.15). In patients with nonculprit coronary artery lesions, a large burden of attenuated or calcified plaques predicted cardiac mortality, while IVUS geometry was associated with repeat revascularization. Thus, deep learning–based IVUS analysis of the whole target vessel may help clinicians identify high-risk lesions. Overall flow of the study. The Kaplan–Meier curves compare 3-year and 5-year cardiac death-free survivals according to the model-derived IVUS criteria of nonculprit vessels. ABI-max = maximal attenuation burden index, CBI-max = maximal calcium burden index, PAV = percent atheroma volume. [Display omitted] • In whole frame analysis of IVUS, the extent of attenuated and calcified plaques determined long-term cardiac mortality in patients with deferred coronary lesion. • Geometrical features representing the stenosis severity also predicted target vessel revascularization. • The deep learning-based IVUS analysis may help identify high-risk lesion subsets. [ABSTRACT FROM AUTHOR]

Although plaque characterization by intravascular ultrasound (IVUS) is important for risk stratification, frame-by-frame analysis of a whole vascular segment is time-consuming. The aim was to develop IVUS-based algorithms for classifying attenuation and calcified plaques. IVUS image sets of 598 coronary arteries from 598 patients were randomized into training and test sets with 5:1 ratio. Each IVUS frame at a 0.4-mm interval was circumferentially labeled as one of three classes: attenuated plaque, calcified plaque, or plaque without attenuation or calcification. The model was trained on multi-class classification with 5-fold cross validation. By converting from Cartesian to polar coordinate images, the class corresponding to each array from 0 to 360° was plotted. At the angle-level, Dice similarity coefficients for identifying calcification vs. attenuation vs. none by using ensemble model were 0.79, 0.74 and 0.99, respectively. Also, the maximal accuracy was 98% to classify those groups in the test set. At the frame-level, the model identified the presence of attenuation with 80% sensitivity, 96% specificity, and 93% overall accuracy, and the presence of calcium with 86% sensitivity, 97% specificity, and 96% overall accuracy. In the per-vessel analysis, the attenuation and calcification burden index closely correlated with human measurements (r = 0.89 and r = 0.95, respectively), as did the maximal attenuation and calcification burden index over 4 mm (r = 0.82 and r = 0.91, respectively). The inference times were 0.05 s per frame and 7.8 s per vessel. Our deep learning algorithms for plaque characterization may assist clinicians in recognizing high-risk coronary lesions. [Display omitted] • Deep learning algorithms allow automatic learning without explicit programming, potentially improving diagnostic accuracy. • The IVUS-based model demonstrated good performance in classifying and quantifying tissue attenuation and calcification. • It supports clinicians in identifying high-risk lesions that potentially lead to stent under expansion and adverse events. [ABSTRACT FROM AUTHOR]

This article reports the first application of a convolutional neural network (CNN) to in vivo X-ray fluorescence (XRF) images of gold nanoparticles (GNPs) obtained by a benchtop X-ray system to eliminate Compton-scattered photons. The XRF imaging system comprises a 2-D cadmium zinc telluride (CZT) gamma camera, a pinhole collimator, and fan-beam polychromatic X-rays. An architecture of the 2-D CNN model for Compton background elimination was optimally designed and trained with data sets obtained by the measurements of water only and GNP-embedded imaging phantoms. The XRF images generated by the trained 2-D CNN were compared with those generated by the direct subtraction method. The developed 2-D CNN was also applied to generate in vivo XRF images of gold in living mice exposed to GNPs. The in vivo XRF images generated by the two methods were then compared in terms of the difference in GNP concentrations. The results demonstrate that the 2-D CNN model was successfully applied to in vivo XRF images for detecting the concentration and location of GNPs in living mice. This also implies that in vivo XRF images can be acquired using the 2-D CNN without preinjection scanning; thus, the image acquisition time and imaging dose can be reduced by half. Furthermore, in vivo XRF images of GNPs can be acquired at any time after the injection, as the 2-D CNN eliminates the need to restrict a living object during the interscans. [ABSTRACT FROM AUTHOR]

Abstract We present a rapid and label-free method for hematologic screening for diseases and syndromes, utilizing quantitative phase imaging (QPI) and machine learning. We aim to establish an efficient blood examination framework that does not suffer from the drawbacks of conventional blood assays, which are incapable of profiling single cells or require labeling procedures. Our method involves the synergistic employment of QPI and machine learning. The high-dimensional refractive index information arising from the QPI-based profiling of single red blood cells is processed to screen for diseases and syndromes using machine learning, which can utilize high-dimensional data beyond the human level. Accurate screening for iron-deficiency anemia, reticulocytosis, hereditary spherocytosis, and diabetes mellitus is demonstrated (>98% accuracy) using the proposed method. Furthermore, we highlight the synergy between QPI and machine learning in the proposed method by analyzing the performance of the method. Highlights • Multiple properties of individual red blood cells are obtained using quantitative phase imaging. • The pathophysiological condition can be predicted based on the properties of a red blood cell, using neural networks. • Neural networks learn the disease-specific alterations of red blood cells during training. • Accurate (> 98% accuracy) screening of hematological disorders are achieved. [ABSTRACT FROM AUTHOR]

This study aims to compare the effectiveness of using discrete heartbeats versus an entire 12-lead electrocardiogram (ECG) as the input for predicting future occurrences of arrhythmia and atrial fibrillation using deep learning models. Experiments were conducted using two types of inputs: a combination of discrete heartbeats extracted from 12-lead ECG and an entire 12-lead ECG signal of 10 s. This study utilized 326,904 ECG signals from 134,447 patients and categorized them into three groups: true–normal sinus rhythm (T-NSR), atrial fibrillation–normal sinus rhythm (AF-NSR), and clinically important arrhythmia–normal sinus rhythm (CIA-NSR). The T-NSR group comprised patients with at least three normal rhythms in a year and no atrial fibrillation or arrhythmias history. Clinically important arrhythmia included atrial fibrillation, atrial flutter, atrial premature contraction, atrial tachycardia, ventricular premature contraction, ventricular tachycardia, right and left bundle branch block, and atrioventricular block over the second degree. The AF-NSR group included normal sinus rhythm paired with atrial fibrillation or atrial flutter within 14 days, and the CIA-NSR group comprised normal sinus rhythm paired with CIA occurring within 14 days. Three deep learning models, ResNet-18, LSTM, and Transformer-based models, were utilized to distinguish T-NSR from AF-NSR and T-NSR from CIA-NSR. The experiments demonstrated the potential of using discrete heartbeats in predicting future arrhythmia and atrial fibrillation incidences extracted from 12-lead electrocardiogram (ECG) signals alone, without any additional patient information. The analysis reveals that these discrete heartbeats contain subtle patterns that deep learning models can identify. Focusing on discrete heartbeats may lead to more timely and accurate diagnoses of these conditions, improving patient outcomes and enabling automated diagnosis using ECG signals as a biomarker. [ABSTRACT FROM AUTHOR]

Optical diffraction tomography measures the three-dimensional refractive index map of a specimen and visualizes biochemical phenomena at the nanoscale in a non-destructive manner. One major drawback of optical diffraction tomography is poor axial resolution due to limited access to the three-dimensional optical transfer function. This missing cone problem has been addressed through regularization algorithms that use a priori information, such as non-negativity and sample smoothness. However, the iterative nature of these algorithms and their parameter dependency make real-time visualization impossible. In this article, we propose and experimentally demonstrate a deep neural network, which we term DeepRegularizer, that rapidly improves the resolution of a three-dimensional refractive index map. Trained with pairs of datasets (a raw refractive index tomogram and a resolution-enhanced refractive index tomogram via the iterative total variation algorithm), the three-dimensional U-net-based convolutional neural network learns a transformation between the two tomogram domains. The feasibility and generalizability of our network are demonstrated using bacterial cells and a human leukaemic cell line, and by validating the model across different samples. DeepRegularizer offers more than an order of magnitude faster regularization performance compared to the conventional iterative method. We envision that the proposed data-driven approach can bypass the high time complexity of various image reconstructions in other imaging modalities. [ABSTRACT FROM AUTHOR]

Although grayscale intravascular ultrasound (IVUS) is commonly used for assessing coronary lesion morphology and optimizing stent implantation, detection of vulnerable plaques by IVUS remains challenging. We aimed to develop machine learning (ML) models for predicting optical coherence tomography-derived thin-cap fibroatheromas (OCT-TCFAs). In 517 patients with angina, 414 and 103 coronary lesions were randomized into training vs. test sets. Each of the IVUS-OCT co-registered frames (including 32,807 for training and 8101 for test) was labeled according to the presence vs. absence of OCT-TCFA. Among 1449 computed IVUS features based on two-dimensional geometry and texture, 17 features were finally selected and used in supervised ML with artificial neural network (ANN), support vector machine (SVM), and naïve Bayes. IVUS sections with (vs. without) OCT-TCFA showed a larger plaque burden, and a smaller and eccentric lumen. TCFA-containing sections were characterized by increased ratios of variance, entropy, and kurtosis; reduced ratio of homogeneity within the superficial to the deeper plaque; and decreased smoothness within the fibrous cap. In addition, OCT-TCFA was associated with low ratios of gamma-beta, Nakagami-μ and Nakagami-ω, and a high ratio of Rayleigh-b within the superficial to the deeper region. With a 5-fold cross-validation, the averaged accuracies were 81 ± 5% for ANN (area under the curve [AUC] = 0.80 ± 0.08), 77 ± 4% for SVM (AUC = 0.74 ± 0.05), and 78 ± 2% for naïve Bayes (AUC = 0.77 ± 0.04) for predicting OCT-TCFA. In the test set, ANN and naïve Bayes showed the overall accuracies of >80%. Supervised ML algorithms with computed IVUS features predicted the presence of OCT-TCFA. This data-driven approach may help clinicians in recognizing high-risk coronary lesions. Image 1 • Machine learning provides the ability to automatically learn without being explicitly programmed, which enables improving diagnostic accuracies. • The models using computed IVUS features predict the presence of OCT-derived thin-cap fibroatheromas with an overall accuracy >80%. • The data-driven approach may help clinicians identify high-risk coronary lesions. [ABSTRACT FROM AUTHOR]

Background: Invasive fractional flow reserve (FFR) is a standard tool for identifying ischemia-producing coronary stenosis. However, in clinical practice, over 70% of treatment decisions still rely on visual estimation of angiographic stenosis, which has limited accuracy (about 60%-65%) for the prediction of FFR < 0.80. One of the reasons for the visual-functional mismatch is that myocardial ischemia can be affected by the supplied myocardial size, which is not always evident by coronary angiography. The aims of this study were to develop an angiography-based machine learning (ML) algorithm for predicting the supplied myocardial volume for a stenosis, as measured using coronary computed tomography angiography (CCTA), and then to build an angiography-based classifier for the lesions with an FFR < 0.80 versus ≥ 0.80.Methods and Findings: A retrospective study was conducted using data from 1,132 stable and unstable angina patients with 1,132 intermediate lesions who underwent invasive coronary angiography, FFR, and CCTA at the Asan Medical Center, Seoul, Korea, between 1 May 2012 and 30 November 2015. The mean age was 63 ± 10 years, 76% were men, and 72% of the patients presented with stable angina. Of these, 932 patients (assessed before 31 January 2015) constituted the training set for the algorithm, and 200 patients (assessed after 1 February 2015) served as a test cohort to validate its diagnostic performance. Additionally, external validation with 79 patients from two centers (CHA University, Seongnam, Korea, and Ajou University, Suwon, Korea) was conducted. After automatic contour calibration using the caliber of guiding catheter, quantitative coronary angiography was performed using the edge-detection algorithms (CAAS-5, Pie-Medical). Clinical information was provided by the Asan BiomedicaL Research Environment (ABLE) system. The CCTA-based myocardial segmentation (CAMS)-derived myocardial volume supplied by each vessel (right coronary artery [RCA], left anterior descending [LAD], left circumflex [LCX]) and the myocardial volume subtended to a stenotic segment (CAMS-%Vsub) were measured for labeling. The ML for (1) predicting vessel territories (CAMS-%LAD, CAMS-%LCX, and CAMS-%RCA) and CAMS-%Vsub and (2) identifying the lesions with an FFR < 0.80 was constructed. Angiography-based ML, employing a light gradient boosting machine (GBM), showed mean absolute errors (MAEs) of 5.42%, 8.57%, and 4.54% for predicting CAMS-%LAD, CAMS-%LCX, and CAMS-%RCA, respectively. The percent myocardial volumes predicted by ML were used to predict the CAMS-%Vsub. With 5-fold cross validation, the MAEs between ML-predicted percent myocardial volume subtended to a stenotic segment (ML-%Vsub) and CAMS-%Vsub were minimized by the elastic net (6.26% ± 0.55% for LAD, 5.79% ± 0.68% for LCX, and 2.95% ± 0.14% for RCA lesions). Using all attributes (age, sex, involved vessel segment, and angiographic features affecting the myocardial territory and stenosis degree), the ML classifiers (L2 penalized logistic regression, support vector machine, and random forest) predicted an FFR < 0.80 with an accuracy of approximately 80% (area under the curve [AUC] = 0.84-0.87, 95% confidence intervals 0.71-0.94) in the test set, which was greater than that of diameter stenosis (DS) > 53% (66%, AUC = 0.71, 95% confidence intervals 0.65-0.78). The external validation showed 84% accuracy (AUC = 0.89, 95% confidence intervals 0.83-0.95). The retrospective design, single ethnicity, and the lack of clinical outcomes may limit this prediction model's generalized application.Conclusion: We found that angiography-based ML is useful to predict subtended myocardial territories and ischemia-producing lesions by mitigating the visual-functional mismatch between angiographic and FFR. Assessment of clinical utility requires further validation in a large, prospective cohort study. [ABSTRACT FROM AUTHOR]