Advances in Artificial Intelligence-Assisted Coronary Computed Tomographic Angiography for Atherosclerotic Plaque Characterization.

Coronary artery disease is a leading cause of death worldwide. Major adverse cardiac events are associated not only with coronary luminal stenosis but also with atherosclerotic plaque components. Coronary computed tomography angiography (CCTA) enables non-invasive evaluation of atherosclerotic plaque along the entire coronary tree. However, precise and efficient assessment of plaque features on CCTA is still a challenge for physicians in daily practice. Artificial intelligence (AI) refers to algorithms that can simulate intelligent human behavior to improve clinical work efficiency. Recently, cardiovascular imaging has seen remarkable advancements with the use of AI. AI-assisted CCTA has the potential to facilitate the clinical workflow, offer objective and repeatable quantitative results, accelerate the interpretation of reports, and guide subsequent treatment. Several AI algorithms have been developed to provide a comprehensive assessment of atherosclerotic plaques. This review serves to highlight the cutting-edge applications of AI-assisted CCTA in atherosclerosis plaque characterization, including detecting obstructive plaques, assessing plaque volumes and vulnerability, monitoring plaque progression, and providing risk assessment. Finally, this paper discusses the current problems and future directions for implementing AI in real-world clinical settings.

Impact of coronary CT image quality on the accuracy of the FFRCT Planner.

OBJECTIVE: To assess the accuracy of a virtual stenting tool based on coronary CT angiography (CCTA) and fractional flow reserve (FFR) derived from CCTA (FFRCT Planner) across different levels of image quality. MATERIALS AND METHODS: Prospective, multicenter, single-arm study of patients with chronic coronary syndromes and lesions with FFR ≤ 0.80. All patients underwent CCTA performed with recent-generation scanners. CCTA image quality was adjudicated using the four-point Likert scale at a per-vessel level by an independent committee blinded to the FFRCT Planner. Patient- and technical-related factors that could affect the FFRCT Planner accuracy were evaluated. The FFRCT Planner was applied mirroring percutaneous coronary intervention (PCI) to determine the agreement with invasively measured post-PCI FFR. RESULTS: Overall, 120 patients (123 vessels) were included. Invasive post-PCI FFR was 0.88 ± 0.06 and Planner FFRCT was 0.86 ± 0.06 (mean difference 0.02 FFR units, the lower limit of agreement (LLA) - 0.12, upper limit of agreement (ULA) 0.15). CCTA image quality was assessed as excellent (Likert score 4) in 48.3%, good (Likert score 3) in 45%, and sufficient (Likert score 2) in 6.7% of patients. The FFRCT Planner was accurate across different levels of image quality with a mean difference between FFRCT Planner and invasive post-PCI FFR of 0.02 ± 0.07 in Likert score 4, 0.02 ± 0.07 in Likert score 3 and 0.03 ± 0.08 in Likert score 2, p = 0.695. Nitrate dose ≥ 0.8mg was the only independent factor associated with the accuracy of the FFRCT Planner (95%CI - 0.06 to - 0.001, p = 0.040). CONCLUSION: The FFRCT Planner was accurate in predicting post-PCI FFR independent of CCTA image quality. CLINICAL RELEVANCE STATEMENT: Being accurate in predicting post-PCI FFR across a wide spectrum of CT image quality, the FFRCT Planner could potentially enhance and guide the invasive treatment. Adequate vasodilation during CT acquisition is relevant to improve the accuracy of the FFRCT Planner. KEY POINTS: • The fractional flow reserve derived from coronary CT angiography (FFRCT) Planner is a novel tool able to accurately predict fractional flow reserve after percutaneous coronary intervention. • The accuracy of the FFRCT Planner was confirmed across a wide spectrum of CT image quality. Nitrates dose at CT acquisition was the only independent predictor of its accuracy. • The FFRCT Planner could potentially enhance and guide the invasive treatment.

A calcification subtraction method for postmortem coronary computed tomography angiography.

Although coronary computed tomography (CT) angiography is a useful tool for evaluating coronary artery lesions both ante- and postmortem, accurate evaluation of the lumen is difficult when highly calcified lesions are present, owing to overestimation of stenosis caused by blooming and partial volume artifacts. In clinical practice, to overcome this diagnostic problem, a subtraction method has been devised to remove calcification by subtracting the precontrast image from the contrast image. In this report, we describe a calcification subtraction method using image analysis software for postmortem coronary CT angiography. This method was devised based on preliminary experimental results showing that the most accurate subtraction was achieved using images reconstructed with a narrower field of view and bone kernel, resulting in higher spatial resolution. This subtraction method allowed evaluation of lumen patency and the degree of stenosis on contrast-enhanced images in a verification using actual specimens where evaluation of the lumen had been difficult because of high calcification. The results were morphologically similar to the macroscopic findings. This method allows more rapid and reliable lesion retrieval and is expected to be useful for postmortem coronary angiography in forensic practice.

The role of coronary CT angiography in athletes.

The sudden death of a young or high-level athlete or adolescent during recreational sports is one of the events with the greatest impact on public opinion in modern society. Sudden cardiac death (SCD) is the principal medical cause of death in athletes and can be the first and last clinical presentation of underlying disease. To prevent such episodes, pre-participation screening has been introduced in many countries to guarantee cardiovascular safety during sports and has become a common target among medical sports/governing organizations. Different cardiac conditions may cause SCD, with incidence depending on definition, evaluation methods, and studied populations, and a prevalence and etiology changing according to the age of athletes, with CAD most frequent in master athletes, while coronary anomalies and non-ischemic causes prevalent in young. To detect silent underlying causes early would be of considerable clinical value. This review summarizes the pre-participation screening in athletes, the specialist agonistic suitability visit performed in Italy, the anatomical characteristics of malignant coronary anomalies, and finally, the role of coronary CT angiography in such arena. In particular, the anatomical conditions suggesting potential disqualification from sport, the post-treatment follow-up to reintegrate young athletes, the diagnostic workflow to rule-out CAD in master athletes, and their clinical management are analyzed.

Motion artifact removal in coronary CT angiography based on generative adversarial networks.

OBJECTIVES: Coronary motion artifacts affect the diagnostic accuracy of coronary CT angiography (CCTA), especially in the mid right coronary artery (mRCA). The purpose is to correct CCTA motion artifacts of the mRCA using a GAN (generative adversarial network). METHODS: We included 313 patients with CCTA scans, who had paired motion-affected and motion-free reference images at different R-R interval phases in the same cardiac cycle and included another 53 CCTA cases with invasive coronary angiography (ICA) comparison. Pix2pix, an image-to-image conversion GAN, was trained by the motion-affected and motion-free reference pairs to generate motion-free images from the motion-affected images. Peak signal-to-noise ratio (PSNR), structural similarity (SSIM), Dice similarity coefficient (DSC), and Hausdorff distance (HD) were calculated to evaluate the image quality of GAN-generated images. RESULTS: At the image level, the median of PSNR, SSIM, DSC, and HD of GAN-generated images were 26.1 (interquartile: 24.4-27.5), 0.860 (0.830-0.882), 0.783 (0.714-0.825), and 4.47 (3.00-4.47), respectively, significantly better than the motion-affected images (p < 0.001). At the patient level, the image quality results were similar. GAN-generated images improved the motion artifact alleviation score (4 vs. 1, p < 0.001) and overall image quality score (4 vs. 1, p < 0.001) than those of the motion-affected images. In patients with ICA comparison, GAN-generated images achieved accuracy of 81%, 85%, and 70% in identifying no, < 50%, and ≥ 50% stenosis, respectively, higher than 66%, 72%, and 68% for the motion-affected images. CONCLUSION: Generative adversarial network-generated CCTA images greatly improved the image quality and diagnostic accuracy compared to motion-affected images. KEY POINTS: • A generative adversarial network greatly reduced motion artifacts in coronary CT angiography and improved image quality. • GAN-generated images improved diagnosis accuracy of identifying no, < 50%, and ≥ 50% stenosis.

Current and Future Applications of Artificial Intelligence in Cardiac CT.

PURPOSE OF REVIEW: In this review, we aim to summarize state-of-the-art artificial intelligence (AI) approaches applied to cardiovascular CT and their future implications. RECENT FINDINGS: Recent studies have shown that deep learning networks can be applied for rapid automated segmentation of coronary plaque from coronary CT angiography, with AI-enabled measurement of total plaque volume predicting future heart attack. AI has also been applied to automate assessment of coronary artery calcium on cardiac and ungated chest CT and to automate the measurement of epicardial fat. Additionally, AI-based prediction models integrating clinical and imaging parameters have been shown to improve prediction of cardiac events compared to traditional risk scores. Artificial intelligence applications have been applied in all aspects of cardiovascular CT - in image acquisition, reconstruction and denoising, segmentation and quantitative analysis, diagnosis and decision assistance and to integrate prognostic risk from clinical data and images. Further incorporation of artificial intelligence in cardiovascular imaging holds important promise to enhance cardiovascular CT as a precision medicine tool.

Screening for Coronary Artery Disease in Patients with Diabetes.

PURPOSE OF REVIEW: The study aims to describe methods for detecting subclinical coronary artery disease (CAD) and their potential implications in asymptomatic patients with diabetes. RECENT FINDINGS: Imaging tools can assess non-invasively the presence and severity of CAD, based on myocardial ischemia, coronary artery calcium score, and coronary computed tomography coronary angiography. Subclinical CAD is common in the general population ageing 50 to 64 years with any coronary atherosclerosis present in 42.1% and obstructive CAD in 5.2%. In patients with diabetes, an even higher prevalence has been noted. The presence of myocardial ischemia, obstructive CAD, and the extent of coronary atherosclerosis provide powerful risk stratification regarding the risk of cardiovascular events. However, randomized trials evaluating systematic screening in the general population or patients with diabetes have demonstrated only moderate impact on management and no significant impact on patient outcomes. Despite providing improved risk stratification, systematic screening of CAD is not recommended in patients with diabetes.

Performance of a Novel CT-Derived Fractional Flow Reserve Measurement to Detect Hemodynamically Significant Coronary Stenosis.

BACKGROUND: Fractional flow reserve (FFR) based on computed tomography (CT) has been shown to better identify ischemia-causing coronary stenosis. However, this current technology requires high computational power, which inhibits its widespread implementation in clinical practice. This prospective, multicenter study aimed at validating the diagnostic performance of a novel simple CT based fractional flow reserve (CT-FFR) calculation method in patients with coronary artery disease. METHODS: Patients who underwent coronary CT angiography (CCTA) within 90 days and invasive coronary angiography (ICA) were prospectively enrolled. A hemodynamically significant lesion was defined as an FFR ≤ 0.80, and the area under the receiver operating characteristic curve (AUC) was the primary measure. After the planned analysis for the initial algorithm A, we performed another set of exploratory analyses for an improved algorithm B. RESULTS: Of 184 patients who agreed to participate in the study, 151 were finally analyzed. Hemodynamically significant lesions were observed in 79 patients (52.3%). The AUC was 0.71 (95% confidence interval [CI], 0.63-0.80) for CCTA, 0.65 (95% CI, 0.56-0.74) for CT-FFR algorithm A (P = 0.866), and 0.78 (95% CI, 0.70-0.86) for algorithm B (P = 0.112). Diagnostic accuracy was 0.63 (0.55-0.71) for CCTA alone, 0.66 (0.58-0.74) for algorithm A, and 0.76 (0.68-0.82) for algorithm B. CONCLUSION: This study suggests the feasibility of automated CT-FFR, which can be performed on-site within several hours. However, the diagnostic performance of the current algorithm does not meet the a priori criteria for superiority. Future research is required to improve the accuracy.

A patient- and acquisition-tailored injection approach for improving consistency of CT enhancement towards a target CT value in coronary CT angiography.

BACKGROUND: Unoptimized coronary CT angiography (CTA) exams typically result in a highly variable arterial enhancement (HUa ) across patients. This study aimed at harmonizing arterial enhancement by implementing a patient-, contrast- and kV-tailored injection protocol. METHODS: First, the optimal body size metric to predict HUa was identified by retrospectively analysing images of 76 patients, acquired with 70 ml contrast media (G1). Second, using phantom experiments, correction factors for the effect of kV and contrast concentration on HUa were determined. Third, a model was developed, prescribing the optimal contrast dose to be injected to obtain a diagnostically appropriate arterial target enhancement HUtarget . The model was then validated on 278 prospectively collected patients, in two groups with two different HUtarget : 525 HU (207 patients, G2A) and 425 HU (71 patients, G2B). The HUa histograms were compared among groups and to the target enhancement through their mean and standard deviation (SD) at 100 kVp reference level. Also, signal-to-noise ratio was obtained and compared among the groups. RESULTS: Fat free mass (FFM) showed the highest correlation with HUa (r = 0.69). KVp correction factors ranged from 0.65 at 70 kVp to 1.22 at 140 kVp. The obtained model reduced the group heterogeneity (SD) from 101HU for reference G1 to 75HU (p < 0.001) for G2A and 68HU (p < 0.001) for G2B. The mean HUa of 506HU in G2A was slightly below HUtarget  = 525HU (p = 0.01) whereas in G2B, the mean HUa of 414HU was not significantly different from HUtarget  = 425HU (p = 0.54). The total iodine dose was lowered from 19.5 g-I to 17.6 g-I and 14.2 g-I from G1 to G2A and G2B, on average. CONCLUSION: A contrast injection model, based on patient's fat free mass and accounting for the contrast agent concentration and the planned CT-scan tube voltage, harmonized arterial enhancement among patients towards a predefined target enhancement in coronary CTA scanning, without affecting the bolus timing.

Diagnostic performance of deep learning-based vessel extraction and stenosis detection on coronary computed tomography angiography for coronary artery disease: a multi-reader multi-case study.

BACKGROUND: Post-processing and interpretation of coronary CT angiography (CCTA) imaging are time-consuming and dependent on the reader's experience. An automated deep learning (DL)-based imaging reconstruction and diagnosis system was developed to improve diagnostic accuracy and efficiency. METHODS: Our study including 374 cases from five sites, inviting 12 radiologists, assessed the DL-based system in diagnosing obstructive coronary disease with regard to diagnostic performance, imaging post-processing and reporting time of radiologists, with invasive coronary angiography as a standard reference. The diagnostic performance of DL system and DL-assisted human readers was compared with the traditional method of human readers without DL system. RESULTS: Comparing the diagnostic performance of human readers without DL system versus with DL system, the AUC was improved from 0.81 to 0.82 (p < 0.05) at patient level and from 0.79 to 0.81 (p < 0.05) at vessel level. An increase in AUC was observed in inexperienced radiologists (p < 0.05), but was absent in experienced radiologists. Regarding diagnostic efficiency, comparing the DL system versus human reader, the average post-processing and reporting time was decreased from 798.60 s to 189.12 s (p < 0.05). The sensitivity and specificity of using DL system alone were 93.55% and 59.57% at patient level and 83.23% and 79.97% at vessel level, respectively. CONCLUSIONS: With the DL system serving as a concurrent reader, the overall post-processing and reading time was substantially reduced. The diagnostic accuracy of human readers, especially for inexperienced readers, was improved. DL-assisted human reader had the potential of being the reading mode of choice in clinical routine.

[Computed tomography in cardiology: history and perspectives].

The review article highlights the main stages of the formation of computed tomography (CT) as a key method used in modern cardiology. The progress of CT scanners is directly related to the increase in the number of detectors, and thus, with an increase in the number of simultaneously collected projections. Modern developments and future technologies in the field of further development of the technique, including CT angiography and other new methods for assessing coronary blood flow, are discussed. The use of artificial intelligence technologies may make it possible to improve and accelerate the interpretation of the resulting images in the future, especially if it is economically justified.

3D computed tomography intravascular endoscopy.

Using coronary computed tomography angiography (CCTA), coronary plaques can be characterized based on both their morphology and composition. Coronary plaques are generally assessed on 2D axial and multiplanar reformatted images. Nevertheless, these visualization tools are limited to observing extraluminal changes in the coronary artery. The presence of plaques prevents them from providing a visual representation of the intraluminal coronary wall. Since its invention in 2000, coronary fly-through or virtual angioscopy (VA) has been extensively studied. However, its application was limited because it required an optimal CT scan and time-consuming post-processing. In recent years, advances in post-processing software have made construction of VA easier, but until recently the quality of the images was insufficient for most patients. Using 3D intravascular endoscopy (3DIE) visualization, we present various intraluminal appearances of the coronary wall and plaque in relation to various types of plaque.

Predicting Chronic Myocardial Ischemia Using CCTA-Based Radiomics Machine Learning Nomogram.

BACKGROUND: Coronary computed tomography angiography (CCTA) is a well-established non-invasive diagnostic test for the assessment of coronary artery diseases (CAD). CCTA not only provides information on luminal stenosis but also permits non-invasive assessment and quantitative measurement of stenosis based on radiomics. PURPOSE: This study is aimed to develop and validate a CT-based radiomics machine learning for predicting chronic myocardial ischemia (MIS). METHODS: CCTA and SPECT-myocardial perfusion imaging (MPI) of 154 patients with CAD were retrospectively analyzed and 94 patients were diagnosed with MIS. The patients were randomly divided into two sets: training (n = 107) and test (n = 47). Features were extracted for each CCTA cross-sectional image to identify myocardial segments. Multivariate logistic regression was used to establish a radiomics signature after feature dimension reduction. Finally, the radiomics nomogram was built based on a predictive model of MIS which in turn was constructed by machine learning combined with the clinically related factors. We then validated the model using data from 49 CAD patients and included 18 MIS patients from another medical center. The receiver operating characteristic curve evaluated the diagnostic accuracy of the nomogram based on the training set and was validated by the test and validation set. Decision curve analysis (DCA) was used to validate the clinical practicability of the nomogram. RESULTS: The accuracy of the nomogram for the prediction of MIS in the training, test and validation sets was 0.839, 0.832, and 0.816, respectively. The diagnosis accuracy of the nomogram, signature, and vascular stenosis were 0.824, 0.736 and 0.708, respectively. A significant difference in the number of patients with MIS between the high and low-risk groups was identified based on the nomogram (P < .05). The DCA curve demonstrated that the nomogram was clinically feasible. CONCLUSION: The radiomics nomogram constructed based on the image of CCTA act as a non-invasive tool for predicting MIS that helps to identify high-risk patients with coronary artery disease.

Transient ischemic dilation or transient RV visualization in patients with normal SPECT stress myocardial perfusion imaging: Correlation with CT coronary artery calcium scoring and coronary angiography.

BACKGROUND: Ancillary findings on MPI, such as transient ischemic dilation (TID) and transient right ventricular visualization (TRV), are recognized as markers of extensive CAD and predictive of adverse outcomes. They usually occur in association with stress-induced regional MPI abnormalities. However, the clinical significance of these ancillary markers in the presence of normal stress MPI is incompletely understood. METHODS: From a cohort of 564 consecutive patients referred for clinical SPECT stress MPI, 44 patients had normal stress SPECT MPI and either TID (n = 28) or TRV (n = 16). These imaging findings were correlated with CT coronary calcium (CAC), CT coronary angiography (CTA), and invasive coronary angiography (ICA) in patients with severe CAC ≥ 1000 HU. TID and TRV were quantified as stress/rest ratios. Severe CAD was defined as > 70% luminal stenosis on CTA or ICA. RESULTS: The median TID ratio was 1.23, with a range of 1.13-1.48; the median TRV ratio was 1.30, with a range of 1.20-1.48. Of 44 patients with TID or TRV, only 9 patients (20.5%) had severe obstructive > 70% CAD by angiography (6 of 28 patients (21.5%) with TID and 3 of 16 patients (19%) with TRV). Severe multi-vessel CAD occurred in only 2 of 44 patients (4.5%). In contrast, of 9 patients with CAC > 1000 HU, 6 (67%) had severe obstructive CAD. CONCLUSION: In patients with normal stress SPECT MPI and TID or TRV, the incidence of severe obstructive CAD was relatively low and predominantly single-vessel CAD. These findings do not support the concept that TID or TRV with normal stress MPI is predictive of high-risk CAD.

Epicardial adipose tissue volume is greater in men with severe psoriasis, implying an increased cardiovascular disease risk: A cross-sectional study.

BACKGROUND: Patients with psoriasis have elevated risk of coronary artery disease. OBJECTIVE: Do patients with severe psoriasis have larger epicardial adipose tissue volumes (EAT-V) that are associated with cardiovascular risk? METHODS: For this cross-sectional study, we recruited dermatology patients with severe psoriasis and control patients without psoriasis or rheumatologic disease themselves or in a first-degree relative. Participants aged 34 to 55 years without known coronary artery disease or diabetes mellitus underwent computed tomography (CT); EAT-V was obtained from noncontrast CT heart images. RESULTS: Twenty-five patients with psoriasis (14 men, 11 women) and 16 controls (5 men, 11 women) participated. Groups had no statistical difference in age, body mass index, various cardiovascular risk factors (except high-sensitivity C-reactive protein in men), CT-determined coronary artery calcium scores or plaque, or family history of premature cardiovascular disease. Mean EAT-V was greater in the psoriasis group compared to controls (P = .04). There was no statistically significant difference among women; however, male patients with psoriasis had significantly higher EAT-V than controls (P = .03), even when corrected for elevated high-sensitivity C-reactive protein (P = .05). LIMITATIONS: A single-center convenience sample may not be representative. CONCLUSION: Males with psoriasis without known coronary disease or diabetes had greater EAT-V than controls. EAT-V may be an early identifier of those at increased risk for cardiovascular events.

Stable patients with suspected myocardial ischemia: comparison of machine-learning computed tomography-based fractional flow reserve and stress perfusion cardiovascular magnetic resonance imaging to detect myocardial ischemia.

BACKGROUND: Machine-Learning Computed Tomography-Based Fractional Flow Reserve (CT-FFRML) is a novel tool for the assessment of hemodynamic relevance of coronary artery stenoses. We examined the diagnostic performance of CT-FFRML compared to stress perfusion cardiovascular magnetic resonance (CMR) and tested if there is an additional value of CT-FFRML over coronary computed tomography angiography (cCTA). METHODS: Our retrospective analysis included 269 vessels in 141 patients (mean age 67 ± 9 years, 78% males) who underwent clinically indicated cCTA and subsequent stress perfusion CMR within a period of 2 months. CT-FFRML values were calculated from standard cCTA. RESULTS: CT-FFRML revealed no hemodynamic significance in 79% of the patients having ≥ 50% stenosis in cCTA. Chi2 values for the statistical relationship between CT-FFRML and stress perfusion CMR was significant (p < 0.0001). CT-FFRML and cCTA (≥ 70% stenosis) provided a per patient sensitivity of 88% (95%CI 64-99%) and 59% (95%CI 33-82%); specificity of 90% (95%CI 84-95%) and 85% (95%CI 78-91%); positive predictive value of 56% (95%CI 42-69%) and 36% (95%CI 24-50%); negative predictive value of 98% (95%CI 94-100%) and 94% (95%CI 90-96%); accuracy of 90% (95%CI 84-94%) and 82% (95%CI 75-88%) when compared to stress perfusion CMR. The accuracy of cCTA (≥ 50% stenosis) was 19% (95%CI 13-27%). The AUCs were 0.89 for CT-FFRML and 0.74 for cCTA (≥ 70% stenosis) and therefore significantly different (p < 0.05). CONCLUSION: CT-FFRML compared to stress perfusion CMR as the reference standard shows high diagnostic power in the identification of patients with hemodynamically significant coronary artery stenosis. This could support the role of cCTA as gatekeeper for further downstream testing and may reduce the number of patients undergoing unnecessary invasive workup.

The diagnostic performance of on-site workstation-based computed tomography-derived fractional flow reserve. Comparison with myocardium perfusion imaging.

To compare the diagnostic performance of on-site workstation-based computed tomography-derived fractional flow reserve (CT-FFR)Few data of CT-FFR were reported regarding the diagnostic performance for detecting hemodynamically significant coronary artery disease (CAD). This retrospective single-center analysis included 132 vessels in 77 patients who underwent CT angiography, myocardial perfusion imaging (MPI), and invasive FFR. The correlation coefficient between CT-FFR and invasive FFR and optimal cut-off value for CT-FFR to identify invasive FFR ≤ 0.8 were evaluated. The diagnostic accuracies of CT- FFR, and MPI were evaluated using an area under the receiver-operating characteristic curve (AUC) with invasive FFR as a reference standard. Diagnostic performance of CT-FFR was also evaluated concerning lesion characteristics, including intermediate lesions, left main lesions, tandem lesions, and/or diffuse lesions, and coronary calcium (Agatston score over 400). The Receiver Operating Characteristic curve analysis showed that the optimal cut-off value of CT-FFR for detecting invasive FFR ≤ 0.80 was 0.80 [AUC = 0.83, 95%CI: 0.76-0.90). Diagnostic sensitivity, specificity, positive and negative predictive value, and accuracy of CT-FFR when compared with those of MPI regarding per-patient analysis were 93% vs. 63%, 48% vs. 61%, 81% vs. 79%, 73% vs. 41%, and 79% vs. 62%, respectively, and for per-vessel analysis were 89% vs. 24%, 66% vs. 82%, 75% vs. 61%, 83% vs. 48%, and 78% vs. 51%, respectively. The AUC of the CT-FFR was significantly higher than MPI (0.83 vs. 0.57, p < 0.0001) regarding the per-vessel analysis. No differences in the diagnostic performance of CT-FFR were noted in the presence of intermediate lesions, left main lesions, tandem lesions, and/or diffuse lesions, and severe coronary calcium. On-site CT-FFR delivered a higher diagnostic performance than MPI for detecting CAD with invasive FFR ≤ 0.8, indicating the potential of CT-FFR as the gatekeeper of invasive coronary angiogram as well as percutaneous coronary intervention.

Motion artefact reduction in coronary CT angiography images with a deep learning method.

BACKGROUND: The aim of this study was to investigate the ability of a pixel-to-pixel generative adversarial network (GAN) to remove motion artefacts in coronary CT angiography (CCTA) images. METHODS: Ninety-seven patients who underwent single-cardiac-cycle multiphase CCTA were retrospectively included in the study, and raw CCTA images and SnapShot Freeze (SSF) CCTA images were acquired. The right coronary artery (RCA) was investigated because its motion artefacts are the most prominent among the artefacts of all coronary arteries. The acquired data were divided into a training dataset of 40 patients, a verification dataset of 30 patients and a test dataset of 27 patients. A pixel-to-pixel GAN was trained to generate improved CCTA images from the raw CCTA imaging data using SSF CCTA images as targets. The GAN's ability to remove motion artefacts was evaluated by the structural similarity (SSIM), Dice similarity coefficient (DSC) and circularity index. Furthermore, the image quality was visually assessed by two radiologists. RESULTS: The circularity was significantly higher for the GAN-generated images than for the raw images of the RCA (0.82 ± 0.07 vs. 0.74 ± 0.11, p < 0.001), and there was no significant difference between the GAN-generated images and SSF images (0.82 ± 0.07 vs. 0.82 ± 0.06, p = 0.96). Furthermore, the GAN-generated images achieved the SSIM of 0.87 ± 0.06, significantly better than those of the raw images 0.83 ± 0.08 (p < 0.001). The results for the DSC showed that the overlap between the GAN-generated and SSF images was significantly higher than the overlap between the GAN-generated and raw images (0.84 ± 0.08 vs. 0.78 ± 0.11, p < 0.001). The motion artefact scores of the GAN-generated CCTA images of the pRCA and mRCA were significantly higher than those of the raw CCTA images (3 [4-3] vs 4 [5-4], p = 0.022; 3 [3-2] vs 5[5-4], p < 0.001). CONCLUSIONS: A GAN can significantly reduce the motion artefacts in CCTA images of the middle segment of the RCA and has the potential to act as a new method to remove motion artefacts in coronary CCTA images.

The auto segmentation for cardiac structures using a dual-input deep learning network based on vision saliency and transformer.

PURPOSE: Accurate segmentation of cardiac structures on coronary CT angiography (CCTA) images is crucial for the morphological analysis, measurement, and functional evaluation. In this study, we achieve accurate automatic segmentation of cardiac structures on CCTA image by adopting an innovative deep learning method based on visual attention mechanism and transformer network, and its practical application value is discussed. METHODS: We developed a dual-input deep learning network based on visual saliency and transformer (VST), which consists of self-attention mechanism for cardiac structures segmentation. Sixty patients' CCTA subjects were randomly selected as a development set, which were manual marked by an experienced technician. The proposed vision attention and transformer mode was trained on the patients CCTA images, with a manual contour-derived binary mask used as the learning-based target. We also used the deep supervision strategy by adding auxiliary losses. The loss function of our model was the sum of the Dice loss and cross-entropy loss. To quantitatively evaluate the segmentation results, we calculated the Dice similarity coefficient (DSC) and Hausdorff distance (HD). Meanwhile, we compare the volume of automatic segmentation and manual segmentation to analyze whether there is statistical difference. RESULTS: Fivefold cross-validation was used to benchmark the segmentation method. The results showed the left ventricular myocardium (LVM, DSC = 0.87), the left ventricular (LV, DSC = 0.94), the left atrial (LA, DSC = 0.90), the right ventricular (RV, DSC = 0.92), the right atrial (RA, DSC = 0.91), and the aortic (AO, DSC = 0.96). The average DSC was 0.92, and HD was 7.2 ± 2.1 mm. In volume comparison, except LVM and LA (p < 0.05), there was no significant statistical difference in other structures. Proposed method for structural segmentation fit well with the true profile of the cardiac substructure, and the model prediction results closed to the manual annotation. CONCLUSIONS: The adoption of the dual-input and transformer architecture based on visual saliency has high sensitivity and specificity to cardiac structures segmentation, which can obviously improve the accuracy of automatic substructure segmentation. This is of gr.

Diagnostic Performance of CT-Derived Fractional Flow Reserve in Australian Patients Referred for Invasive Coronary Angiography.

BACKGROUND: Non-invasive computed tomography (CT)-derived fractional flow reserve (FFRCT) is computed from standard coronary CT angiography (CTA) datasets and provides accurate vessel-specific ischaemia assessment of coronary artery disease (CAD). To date, the technique and its diagnostic performance has not been verified in the Australian clinical context. The aim of this study was to describe and compare the diagnostic performance of FFRCT and CTA for the detection of vessel-specific ischaemia as determined by invasive fractional flow reserve (FFR) in the Australian patient population. METHODS: One-hundred-and-nine patients (219 vessels) referred for clinically mandated invasive angiography were retrospectively assessed. Each patient underwent research mandated CTA and FFRCT within 3 months of invasive angiography and invasive FFR assessment. Independent core laboratory assessments were made to determine visual CTA stenosis, FFRCT and invasive FFR values. FFRCT values were matched with the corresponding invasive FFR measurement taken at the given wire position. Visual CTA stenosis ≥50%, FFRCT values ≤0.8 and invasive FFR values ≤0.8 were considered significant for ischaemia. RESULTS: Per vessel accuracy, sensitivity, specificity, positive predictive value and negative predictive value of FFRCT were 80.4%, 80.0%, 80.6%, 64.9% and 90.0% respectively. Corresponding values for CTA were 75.1%, 87.1%, 69.2%, 58.1% and 91.7% respectively. In receiver operating characteristic curve analysis, FFRCT demonstrated superior area under the curve (AUC) compared with CTA in both per vessel (0.87 vs 0.77, p=0.004) and per patient analysis (0.86 vs 0.74, p=0.011). Per vessel AUC of combined CTA and FFRCT was superior to CTA alone (0.89 vs 0.77, p<0.0001). CONCLUSION: In this cohort of Australian patients, the diagnostic performance of FFRCT was found to be comparable to existing international literature, with demonstrated improvement in performance compared with CTA alone for the detection of vessel-specific ischaemia.