ArticlePDF Available

Deep learning for myocardial ischemia auxiliary diagnosis using CZT SPECT myocardial perfusion imaging

October 2022
Journal of the Chinese Medical Association Publish Ahead of Print(1)

October 2022
Publish Ahead of Print(1)

DOI:10.1097/JCMA.0000000000000833

License
CC BY-NC-ND 4.0

Authors:

Wei-Shiang Chen

National Yang Ming Chiao Tung University

Show all 5 authorsHide

Background: The World Health Organization reported that cardiovascular disease is the most common cause of death worldwide. On average, one person dies of heart disease every 26-min worldwide. Deep learning approaches are characterised by the appropriate combination of abnormal features based on numerous annotated images. The constructed CNN model can identify normal states of reversible and irreversible myocardial defects and alert physicians for further diagnosis. Methods: Cadmium zinc telluride single-photon emission computed tomography myocardial perfusion resting-state images were collected at Chang Gung Memorial Hospital, Kaohsiung Medical Center, Kaohsiung, Taiwan, and were analysed with a deep learning convolutional neural network to classify myocardial perfusion images for coronary heart diseases. Results: In these grey-scale images, the heart blood flow distribution was the most crucial feature. The deep learning technique of You Only Look Once was used to determine the myocardial defect area and crop the images. After surrounding noise had been eliminated, a three-dimensional convolutional neural network model was used to identify patients with coronary heart diseases. The prediction area under the curve, accuracy, sensitivity, and specificity were 90.97%, 87.08%, 86.49%, and 87.41%, respectively. Conclusion: Our prototype system can considerably reduce the time required for image interpretation and improve the quality of medical care. It can assist clinical experts by offering accurate coronary heart disease diagnosis in practice.

Available via license: CC BY-NC-ND 4.0

Content may be subject to copyright.

Supplemental Table 1.

Diagnostic performance of software package Quantitative Perfusion SPECT (QPS,

Cedars-Sinai Medical Center, Los Angeles, CA, USA) in Stress-State Testing Set

(N=294, Criteria: SSS>4)

Subject Sensitivity Specificity PPV NPV Accuracy

294

0.8673

0.8508

0.784

0.9112

0.8571

Supplemental Table 2.

Diagnostic performance of software package Quantitative Perfusion SPECT (QPS,

Cedars-Sinai Medical Center, Los Angeles, CA, USA) in Stress/Rest-State Testing Set

(Criteria: SSS>4 or SDS≥2)

Subject

Sensitivity

Specificity

PPV

NPV

Accuracy

209 0.8784 0.7556 0.6632 0.9189 0.7790

Cardiac Allograft Vasculopathy: Challenges and Advances in Invasive and Non-Invasive Diagnostic Modalities

Article

Full-text available

Mar 2024

Cardiac allograft vasculopathy (CAV) is a distinct form of coronary artery disease that represents a major cause of death beyond the first year after heart transplantation. The pathophysiology of CAV is still not completely elucidated; it involves progressive circumferential wall thickening of both the epicardial and intramyocardial coronary arteries. Coronary angiography is still considered the gold-standard test for the diagnosis of CAV, and intravascular ultrasound (IVUS) can detect early intimal thickening with improved sensitivity. However, these tests are invasive and are unable to visualize and evaluate coronary microcirculation. Increasing evidence for non-invasive surveillance techniques assessing both epicardial and microvascular components of CAV may help improve early detection. These include computed tomography coronary angiography (CTCA), single-photon emission computed tomography (SPECT), positron emission tomography (PET), and vasodilator stress myocardial contrast echocardiography perfusion imaging. This review summarizes the current state of diagnostic modalities and their utility and prognostic value for CAV and also evaluates emerging tools that may improve the early detection of this complex disease.

Convolutional Neural Network in the Evaluation of Myocardial Ischemia from CZT SPECT Myocardial Perfusion Imaging: Comparison to Automated Quantification

Article

Full-text available

Jan 2021

This study analyzes CZT SPECT myocardial perfusion images that are collected at Chang Gung Memorial Hospital, Kaohsiung Medical Center in Kaohsiung. This study focuses on the classification of myocardial perfusion images for coronary heart diseases by convolutional neural network techniques. In these gray scale images, heart blood flow distribution contains the most important features. Therefore, data-driven preprocessing is developed to extract the area of interest. After removing the surrounding noise, the three-dimensional convolutional neural network model is utilized to classify whether the patient has coronary heart diseases or not. The prediction accuracy, sensitivity, and specificity are 87.64%, 81.58%, and 92.16%. The prototype system will greatly reduce the time required for physician image interpretation and write reports. It can assist clinical experts in diagnosing coronary heart diseases accurately in practice.

Deep convolutional neural network based medical image classification for disease diagnosis

Article

Full-text available

Dec 2019

Abstract Medical image classification plays an essential role in clinical treatment and teaching tasks. However, the traditional method has reached its ceiling on performance. Moreover, by using them, much time and effort need to be spent on extracting and selecting classification features. The deep neural network is an emerging machine learning method that has proven its potential for different classification tasks. Notably, the convolutional neural network dominates with the best results on varying image classification tasks. However, medical image datasets are hard to collect because it needs a lot of professional expertise to label them. Therefore, this paper researches how to apply the convolutional neural network (CNN) based algorithm on a chest X-ray dataset to classify pneumonia. Three techniques are evaluated through experiments. These are linear support vector machine classifier with local rotation and orientation free features, transfer learning on two convolutional neural network models: Visual Geometry Group i.e., VGG16 and InceptionV3, and a capsule network training from scratch. Data augmentation is a data preprocessing method applied to all three methods. The results of the experiments show that data augmentation generally is an effective way for all three algorithms to improve performance. Also, Transfer learning is a more useful classification method on a small dataset compared to a support vector machine with oriented fast and rotated binary (ORB) robust independent elementary features and capsule network. In transfer learning, retraining specific features on a new target dataset is essential to improve performance. And, the second important factor is a proper network complexity that matches the scale of the dataset.

Enlarging smaller images before inputting into convolutional neural network: zero-padding vs. interpolation

Article

Full-text available

Nov 2019

Mahdi Hashemi

Abstract The input to a machine learning model is a one-dimensional feature vector. However, in recent learning models, such as convolutional and recurrent neural networks, two- and three-dimensional feature tensors can also be inputted to the model. During training, the machine adjusts its internal parameters to project each feature tensor close to its target. After training, the machine can be used to predict the target for previously unseen feature tensors. What this study focuses on is the requirement that feature tensors must be of the same size. In other words, the same number of features must be present for each sample. This creates a barrier in processing images and texts, as they usually have different sizes, and thus different numbers of features. In classifying an image using a convolutional neural network (CNN), the input is a three-dimensional tensor, where the value of each pixel in each channel is one feature. The three-dimensional feature tensor must be the same size for all images. However, images are not usually of the same size and so are not their corresponding feature tensors. Resizing images to the same size without deforming patterns contained therein is a major challenge. This study proposes zero-padding for resizing images to the same size and compares it with the conventional approach of scaling images up (zooming in) using interpolation. Our study showed that zero-padding had no effect on the classification accuracy but considerably reduced the training time. The reason is that neighboring zero input units (pixels) will not activate their corresponding convolutional unit in the next layer. Therefore, the synaptic weights on outgoing links from input units do not need to be updated if they contain a zero value. Theoretical justification along with experimental endorsements are provided in this paper.

Ability of artificial intelligence to diagnose coronary artery stenosis using hybrid images of coronary computed tomography angiography and myocardial perfusion SPECT

Article

Full-text available

Mar 2019

Background Detecting culprit coronary arteries in patients with ischemia using only myocardial perfusion single-photon emission computed tomography (SPECT) can be challenging. This study aimed to improve the detection of culprit regions using an artificial neural network (ANN) to analyze hybrid images of coronary computed tomography angiography (CCTA) and myocardial perfusion SPECT. Methods This study enrolled 59 patients with stable coronary artery disease (CAD) who had been assessed by coronary angiography within 60 days of myocardial perfusion SPECT. Two nuclear medicine physicians interpreted the myocardial perfusion SPECT and hybrid images with four grades of confidence, then drew regions on polar maps to identify culprit coronary arteries. The gold standard was determined by the consensus of two other nuclear cardiology specialist based on coronary angiography findings and clinical information. The ability to detect culprit coronary arteries was compared among experienced nuclear cardiologists and the ANN. Receiver operating characteristics (ROC) curves were analyzed and areas under the ROC curves (AUC) were determined. Results Using hybrid images, observer A detected CAD in the right (RCA), left anterior descending (LAD), and left circumflex (LCX) coronary arteries with 83.6%, 89.3%, and 94.4% accuracy, respectively and observer B did so with 72.9%, 84.2%, and 89.3%, respectively. The ANN was 79.1%, 89.8%, and 89.3% accurate for each coronary artery. Diagnostic accuracy was comparable between the ANN and experienced nuclear medicine physicians. The AUC was significantly improved using hybrid images in the RCA region (observer A: from 0.715 to 0.835, p = 0.0031; observer B: from 0.771 to 0.843, p = 0.042). To detect culprit coronary arteries in perfusion defects of the inferior wall without using hybrid images was problematic because the perfused areas of the LCX and RCA varied among individuals. Conclusions Hybrid images of CCTA and myocardial perfusion SPECT are useful for detecting culprit coronary arteries. Diagnoses using artificial intelligence are comparable to that by nuclear medicine physicians.

Clinical Assistant Diagnosis for Electronic Medical Record Based on Convolutional Neural Network

Article

Full-text available

Apr 2018

Automatically extracting useful information from electronic medical records along with conducting disease diagnoses is a promising task for both clinical decision support(CDS) and neural language processing(NLP). Most of the existing systems are based on artificially constructed knowledge bases, and then auxiliary diagnosis is done by rule matching. In this study, we present a clinical intelligent decision approach based on Convolutional Neural Networks(CNN), which can automatically extract high-level semantic information of electronic medical records and then perform automatic diagnosis without artificial construction of rules or knowledge bases. We use collected 18,590 copies of the real-world clinical electronic medical records to train and test the proposed model. Experimental results show that the proposed model can achieve 98.67% accuracy and 96.02% recall, which strongly supports that using convolutional neural network to automatically learn high-level semantic features of electronic medical records and then conduct assist diagnosis is feasible and effective.

Convolutional Neural Networks as a Model of the Visual System: Past, Present, and Future

Article

Feb 2020

Grace W. Lindsay

Convolutional neural networks (CNNs) were inspired by early findings in the study of biological vision. They have since become successful tools in computer vision and state-of-the-art models of both neural activity and behavior on visual tasks. This review highlights what, in the context of CNNs, it means to be a good model in computational neuroscience and the various ways models can provide insight. Specifically, it covers the origins of CNNs and the methods by which we validate them as models of biological vision. It then goes on to elaborate on what we can learn about biological vision by understanding and experimenting on CNNs and discusses emerging opportunities for the use of CNNs in vision research beyond basic object recognition.

The Study of The Convolutional Neural Networks Applications

Article

Dec 2019

At present, deep learning is widely used in a broad range of arenas. A convolutional neural networks (CNN) is becoming the star of deep learning as it gives the best and most precise results when cracking real-world problems. In this work, a brief description of the applications of CNNs in two areas will be presented: First, in computer vision, generally, that is, scene labeling, face recognition, action recognition, and image classification; Second, in natural language processing, that is, the fields of speech recognition and text classification.

Data Augmentation Using Random Image Cropping and Patching for Deep CNNs

Article

Aug 2019

Deep convolutional neural networks (CNNs) have achieved remarkable results in image processing tasks. However, their high expression ability risks overfitting. Consequently, data augmentation techniques have been proposed to prevent overfitting while enriching datasets. Recent CNN architectures with more parameters are rendering traditional data augmentation techniques insufficient. In this study, we propose a new data augmentation technique called random image cropping and patching (RICAP) which randomly crops four images and patches them to create a new training image. Moreover, RICAP mixes the class labels of the four images, resulting in an advantage of the soft labels. We evaluated RICAP with current state-of-the-art CNNs (e.g., the shake-shake regularization model) by comparison with competitive data augmentation techniques such as cutout and mixup. RICAP achieves a new state-of-the-art test error of 2.19% on CIFAR-10. We also confirmed that deep CNNs with RICAP achieve better results on classification tasks using CIFAR-100 and ImageNet, an image-caption retrieval task using Microsoft COCO, and other computer vision tasks.

Accuracy of an artificial neural network for detecting a regional abnormality in myocardial perfusion SPECT

Article

Oct 2018

Objectives The patient-based diagnosis with an artificial neural network (ANN) has shown potential utility for the detection of coronary artery disease; however, the region-based accuracy of the detected regions has not been fully evaluated. The aim of this study was to demonstrate the accuracy of all detected regions compared with expert interpretation. Methods A total of 109 abnormal regions including 33 regions with stress defects and 76 regions with ischemia were examined, which were derived from 21 patients who underwent myocardial perfusion SPECT within 45 days of coronary angiography. The gray and color scale images, a polar map of stress, rest and difference, and left ventricular function were displayed on the monitor to score the extent and severity of stress defect and ischemia. Two experienced nuclear medicine physicians (Observers A and B) scored the abnormality with a 4-point scale and draw abnormal regions on a polar map. The gold standard was determined by the final judgment of normal or abnormal by the consensus of two other independent expert nuclear cardiologists, and was compared with the stress defect and ischemia derived from ANN. Results The concordance rate of ANN to the gold standard was higher than that of two observers. Furthermore, the κ coefficient indicated moderate to substantial agreement for stress defect and slight to the fair agreement for ischemia. The area under the curve (AUC) of ANN was the highest for both stress defect and ischemia; in particular, the ANN of ischemia showed significantly higher AUC than Observer A (p = 0.005). The ANN of stress defect showed higher specificity compared with two observers, while the ANN of ischemia showed higher sensitivity. Consequently, the accuracy of ANN showed the highest in this study. Conclusion The ANN-based regional diagnosis showed a high concordance rate with the gold standard and comparable or even higher than the interpretation by nuclear medicine physicians.

Deep Learning for Prediction of Obstructive Disease From Fast Myocardial Perfusion SPECT

Article

Mar 2018

Objectives: The study evaluated the automatic prediction of obstructive disease from myocardial perfusion imaging (MPI) by deep learning as compared with total perfusion deficit (TPD). Background: Deep convolutional neural networks trained with a large multicenter population may provide improved prediction of per-patient and per-vessel coronary artery disease from single-photon emission computed tomography MPI. Methods: A total of 1,638 patients (67% men) without known coronary artery disease, undergoing stress99mTc-sestamibi or tetrofosmin MPI with new generation solid-state scanners in 9 different sites, with invasive coronary angiography performed within 6 months of MPI, were studied. Obstructive disease was defined as ≥70% narrowing of coronary arteries (≥50% for left main artery). Left ventricular myocardium was segmented using clinical nuclear cardiology software and verified by an expert reader. Stress TPD was computed using sex- and camera-specific normal limits. Deep learning was trained using raw and quantitative polar maps and evaluated for prediction of obstructive stenosis in a stratified 10-fold cross-validation procedure. Results: A total of 1,018 (62%) patients and 1,797 of 4,914 (37%) arteries had obstructive disease. Area under the receiver-operating characteristic curve for disease prediction by deep learning was higher than for TPD (per patient: 0.80 vs. 0.78; per vessel: 0.76 vs. 0.73: p < 0.01). With deep learning threshold set to the same specificity as TPD, per-patient sensitivity improved from 79.8% (TPD) to 82.3% (deep learning) (p < 0.05), and per-vessel sensitivity improved from 64.4% (TPD) to 69.8% (deep learning) (p < 0.01). Conclusions: Deep learning has the potential to improve automatic interpretation of MPI as compared with current clinical methods.

Deep learning for myocardial ischemia auxiliary diagnosis using CZT SPECT myocardial perfusion imaging

Abstract

Recommended publications

Convolutional Neural Network in the Evaluation of Myocardial Ischemia from CZT SPECT Myocardial Perf...

Automatic classification of temporomandibular joint disorders by magnetic resonance imaging and conv...

Deep Learning Analysis of Polar Maps from SPECT Myocardial Perfusion Imaging for Prediction of Coron...

Diagnostic accuracy of stress-only myocardial perfusion SPECT improved by deep learning