An assessment of attenuation correction of SPECT MPI images generated by deep learning model

This article evaluates the effectiveness of using a deep learning network model to generate reliable attenuation corrected the single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI). The authors collected myocardial perfusion imaging data of 88 patients from a SPECT/CT machine, with an average age of 62.47 years. Then, two datasets are created from the original data: set A includes the deep learning-based attenuation corrected images (Generated Attenuation Correction - GenAC), and the non-attenuation corrected images; set B contains only non-attenuation corrected images. These datasets were diagnosed by two doctors (in which, one has 7 years of experience and the other has 10 years of experience in reading SPECT MPI). The doctors diagnose based on the image data without knowing which dataset it belongs to. The sensitivity, specificity, diagnostic accuracy, and lesion rate were evaluated between the two data sets. Results: The average specificity, sensitivity, and accuracy of the set with the deep learning-based attenuation corrected images were 0.87, 0.86, 0.86, while the results with the non-attenuation corrected images are 0.69, 0.83, and 0.78.

Impact of age, sex, and cardiac size on the diagnostic performance of myocardial perfusion single-photon emission computed tomography: insights from the REFINE SPECT registry

Background Single-photon emission computed tomography myocardial perfusion imaging (SPECT MPI) is a well-validated non-invasive method for detecting coronary artery disease (CAD). Variations in diagnostic performance due to age and sex have been thoroughly investigated in the literature yet have demonstrated conflicting results. Several studies have associated female sex with reduced accuracy, although others have discovered no significant difference (1). Similarly, while SPECT MPI in the elderly has shown prognostic utility, cardiac event rates are elevated compared to younger patients despite a normal study (2). Additional analyses have suggested that cardiac chamber size may contribute to these observed differences due to its relationship with spatial resolution; however, the interaction of age, sex, and cardiac size remains unknown. Purpose We aimed to leverage a large, multicenter, international registry to assess the impact of age, sex, and left ventricular size on the diagnostic accuracy of contemporary SPECT MPI. Methods In 9 centers, 2067 patients (67% male, 64.7±11.2 years) in the REFINE SPECT database (REgistry of Fast Myocardial Perfusion Imaging with NExt Generation SPECT) underwent MPI with new generation solid-state scanners followed by invasive coronary angiography within 6 months (3). Stress total perfusion deficit was quantified automatically, and obstructive CAD was defined as >70% stenosis or >50% for left main. Receiver-operating characteristic curves and corresponding areas under the curve (AUC) were computed to compare diagnostic performance between cohorts created based on age (<75 vs. ≥75 years), sex, and end-diastolic volume (EDV; ≥20th vs. <20th sex-specific percentile). Results Female and elderly patients had a significantly lower EDV than male and younger patients respectively (p<0.001, Figure 1). Diagnostic accuracy of SPECT was similar by sex (p=0.63). Elderly patients (AUC 0.72 vs. 0.78, p=0.025) and patients with reduced volumes (AUC 0.72 vs. 0.79, p=0.009) exhibited significantly worse performance. When isolating male patients with reduced volumes, a significant difference in accuracy was observed (AUC 0.69 vs. 0.79, p=0.001; Figure 2A), while female patients trended towards significance (p=0.32). Likewise, SPECT performed poorly for elderly patients with reduced volumes (AUC 0.64 vs. 0.78, p=0.01; Figure 2B). If patients possessed any two characteristics of male sex, age ≥75, or low EDV, prediction of CAD with SPECT was significantly decreased (p=0.002; Figure 2C). Conclusions Our findings suggest that men with reduced cardiac volumes display worse diagnostic SPECT performance, although it is uncertain whether a pathophysiologic reason exists or further investigation is required for female patients. Patients age ≥75 tended to have lower cardiac volumes as well as lower diagnostic performance. Given these results, alternative diagnostic modalities may better diagnose CAD in patients with these characteristics. FUNDunding Acknowledgement Type of funding sources: None. Figure 1 Figure 2

Diagnostic accuracy of dynamic CZT-SPECT in coronary artery disease. A systematic review and meta-analysis

Background With the appearance of cadmium-zinc-telluride (CZT) cameras, dynamic myocardial perfusion imaging (MPI) has been introduced, but comparable data to other MPI modalities, such as quantitative coronary angiography (CAG) with fractional flow reserve (FFR) and positron emission tomography (PET), are lacking. This study aimed to evaluate the diagnostic accuracy of dynamic CZT single-photon emission tomography (SPECT) in coronary artery disease compared to quantitative CAG, FFR, and PET as reference. Materials and Methods Different databases were screened for eligible citations performing dynamic CZT-SPECT against CAG, FFR, or PET. PubMed, OvidSP (Medline), Web of Science, the Cochrane Library, and EMBASE were searched on the 5th of July 2020. Studies had to meet the following pre-established inclusion criteria: randomized controlled trials, retrospective trails or observational studies relevant for the diagnosis of coronary artery disease, and performing CZT-SPECT and within half a year the methodological references. Studies which considered coronary stenosis between 50% and 70% as significant based only on CAG were excluded. Data extracted were sensitivity, specificity, likelihood ratios, and diagnostic odds ratios. Quality was assessed with QUADAS-2 and statistical analysis was performed using a bivariate model. Results Based on our criteria, a total of 9 studies containing 421 patients were included. For the assessment of CZT-SPECT, the diagnostic value pooled analysis with a bivariate model was calculated and yielded a sensitivity of 0.79 (% CI 0.73 to 0.85) and a specificity of 0.85 (95% CI 0.74 to 0.92). Diagnostic odds ratio (DOR) was 17.82 (95% CI 8.80 to 36.08, P < 0.001). Positive likelihood ratio (PLR) and negative likelihood ratio (NLR) were 3.86 (95% CI 2.76 to 5.38, P < 0.001) and 0.21 (95% CI 0.13 to 0.33, P < 0.001), respectively. Conclusion Based on the results of the current systematic review and meta-analysis, dynamic CZT-SPECT MPI demonstrated a good sensitivity and specificity to diagnose CAD as compared to the gold standards. However, due to the heterogeneity of the methodologies between the CZT-SPECT MPI studies and the relatively small number of included studies, it warrants further well-defined study protocols.

Exploration of the Efficiency of Radiomics Applied to Left Ventricular Tomograms Obtained From D-SPECT MPI for the Auxiliary Diagnosis of Myocardial Ischemia in CAD

Purpose: To explore the feasibility and efficacy of radiomics with left ventricular tomograms obtained from D-SPECT myocardial perfusion imaging (MPI) for auxiliary diagnosis of myocardial ischemia in coronary artery disease (CAD).Methods: The images of 103 patients with CAD myocardial ischemia between September 2020 and April 2021 were retrospectively selected. After information desensitization processing, format conversion, annotation using the Labelme tool on an open-source platform, lesion classification, and establishment of a database, the images were cropped for analysis. The ResNet18 model was used to automate two steps (classification and segmentation) with five randomization, training and validation steps. Sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, positive predictive value, negative predictive value, Youden’s index, agreement rate, and kappa value were calculated as evaluation indexes of the classification results for each training-validation step; then, receiver operating characteristics (ROC) curves were drawn, and the areas under the curve (AUCs) were calculated. The Dice coefficient, intersection over union, and Hausdorff distance were calculated as evaluation indexes of the segmentation results for each training-validation step; then, the predicted images were exported.Results: Under the existing conditions, the radiomics model can distinguish myocardial ischemia quite accurately, with AUCs all exceeding 0.95, and predict the areas of myocardial ischemia quite accurately; all evaluated indexes were close to those of the gold standard.Conclusion: Radiomics can be feasibly applied to left ventricular tomograms obtained from D-SPECT MPI for auxiliary diagnosis, with quite good results. Patients may benefit from this approach as technology evolves and associated software is developed.

Automatic Diagnosis of Coronary Artery Disease in SPECT Myocardial Perfusion Imaging Employing Deep Learning

Focusing on coronary artery disease (CAD) patients, this research paper addresses the problem of automatic diagnosis of ischemia or infarction using single-photon emission computed tomography (SPECT) (Siemens Symbia S Series) myocardial perfusion imaging (MPI) scans and investigates the capabilities of deep learning and convolutional neural networks. Considering the wide applicability of deep learning in medical image classification, a robust CNN model whose architecture was previously determined in nuclear image analysis is introduced to recognize myocardial perfusion images by extracting the insightful features of an image and use them to classify it correctly. In addition, a deep learning classification approach using transfer learning is implemented to classify cardiovascular images as normal or abnormal (ischemia or infarction) from SPECT MPI scans. The present work is differentiated from other studies in nuclear cardiology as it utilizes SPECT MPI images. To address the two-class classification problem of CAD diagnosis, achieving adequate accuracy, simple, fast and efficient CNN architectures were built based on a CNN exploration process. They were then employed to identify the category of CAD diagnosis, presenting its generalization capabilities. The results revealed that the applied methods are sufficiently accurate and able to differentiate the infarction or ischemia from healthy patients (overall classification accuracy = 93.47% ± 2.81%, AUC score = 0.936). To strengthen the findings of this study, the proposed deep learning approaches were compared with other popular state-of-the-art CNN architectures for the specific dataset. The prediction results show the efficacy of new deep learning architecture applied for CAD diagnosis using SPECT MPI scans over the existing ones in nuclear medicine.

High Frequency QRS analysis to supplement ST-Analysis for Myocardial Ischemia: multi-center prospective cohort

Funding Acknowledgements Type of funding sources: None. Background Exercise testing is a well-known non-invasive assessment method for myocardial ischemia in patients with suspected coronary artery disease (CAD). Stress electrocardiography (ECG) alone is underutilized in this population despite guideline recommendations in part due to poor diagnostic accuracy. High frequency QRS analysis (HF-QRS) is a novel tool to supplement standard ST-analysis during stress ECG and has been shown in single-center retrospective analyses to identify any and substantial ischemia with high diagnostic accuracy. We sought to compare the diagnostic accuracy of HF-QRS + standard ST-analysis compared to standard ST-analysis alone for the identification of moderate to severe myocardial ischemia by exercise SPECT MPI. Methods The study population included 388 consecutive patients who underwent exercise SPECT MPI. An ischemic HF-QRS pattern was defined as an absolute reduction of ≥1 μV and a relative reduction of ≥50% between maximal and minimal values of the mean root square of the 150-250 Hz band signal in ≥3 leads. The diagnostic accuracy of HF-QRS + ST-analysis was compared with ST-analysis alone for moderate to severe myocardial ischemia using chi-square analysis and semi-quantitative gated SPECT MPI as the gold standard. The incremental diagnostic value of HF-QRS was assessed by logistic regression analysis. The likelihood of any ischemia by number of leads positive for HF-QRS was also determined. Results The study cohort was 71% male and 84% Caucasian with a mean age of 58.3 ± 11.8 years. ST- and HF-QRS analyses were positive in 96 (24.7%) and 121 (31.2%) of patients, respectively. HF-QRS had a substantially higher sensitivity than ST-analysis for moderate-severe ischemia (66.7% vs. 40.0%, p &lt;0.003). There was no statistically significant difference in specificities for HF-QRS vs ST-analysis for moderate-severe ischemia. (70.5% vs 75.7%, p = 0.08). There was a stepwise increase in ischemia as number of positive HF-QRS leads increased (p = 0.0004). HF-QRS demonstrated incremental diagnostic value to clinical risk factors without ST-analysis (p = 0.006) compared to Clinical + ST depressions (p &lt; 0.001) versus clinical factors. Conclusions This multicenter, prospective study expands the literature showing the benefit of HF-QRS analysis. HF-QRS analysis substantially improves detection of moderate-severe ischemia over ST-analysis and clinical risk factors in patients undergoing exercise stress ECG. This noninvasive adjunct may improve CAD risk stratification and encourage use of stress ECG without imaging, reducing costs and radiation exposure.