Coronary CTA-based Vascular Radiomics Predicts Atherosclerosis Development Proximal to LAD Myocardial Bridging.

BACKGROUND: Cardiac cycle morphological changes can accelerate plaque growth proximal to myocardial bridging (MB) in the left anterior descending artery (LAD). OBJECTIVE: To assess coronary CT angiography (CCTA)-based vascular radiomics for predicting proximal plaque development in LAD MB. METHODS: Patients with repeated CCTA scans showing LAD MB without proximal plaque in index CCTA were included from Jinling Hospital as development set. They were divided into training and internal testing in an 8:2 ratio. Patients from 4 other tertiary hospitals were set as external validation set. The endpoint was proximal plaque development of LAD MB in follow-up CCTA. Four vascular radiomics models were built: MB centerline (MB CL), proximal MB CL (pMB CL), MB cross section (MB CS), and proximal MB CS (pMB CS), whose performances were evaluated using area under the curve (AUC), integrated discrimination improvement (IDI) and net reclassification improvement (NRI). RESULTS: 295 patients were included in the development (n=192; median age, 54±11 years; 137 men) and external validation sets (n=103; median age, 57±9 years; 57 men). The pMB CS vascular radiomics model exhibited higher AUCs in training, internal test, and external sets (AUC=0.78, 0.75, 0.75) than the clinical and anatomical model (all p<0.05). Integration of the pMB CS vascular radiomics model significantly raised the AUC of the clinical and anatomical model from 0.56 to 0.75 (p=0.002), along with enhanced NRI (0.76 [0.37-1.14], p<0.001) and IDI (0.17 [0.07-0.26], p<0.001) in the external validation set. CONCLUSION: The CCTA-based pMB CS vascular radiomics model can predict plaque development in LAD MB.

Aberrant cortical morphology patterns are associated with cognitive impairment in patients with chronic heart failure.

A mounting body of evidences suggests that patients with chronic heart failure (HF) frequently experience cognitive impairments, but the neuroanatomical mechanism underlying these impairments remains elusive. In this retrospective study, 49 chronic HF patients and 49 healthy controls (HCs) underwent brain structural MRI scans and cognitive assessments. Cortical morphology index (cortical thickness, complexity, sulcal depth and gyrification) were evaluated. Correlations between cortical morphology and cognitive scores and clinical variables were explored. Logistic regression analysis was employed to identify risk factors for predicting 3-year major adverse cardiovascular events. Compared with HCs, patients with chronic HF exhibited decreased cognitive scores (p < .001) and decreased cortical thickness, sulcal depth and gyrification in brain regions involved cognition, sensorimotor, autonomic nervous system (family-wise error correction, all p values <.05). Notably, HF duration and New York Heart Association (NYHA) demonstrated negative correlations with abnormal cortex morphology, particularly HF duration and thickness in left precentral gyrus (r = -.387, p = .006). Cortical morphology characteristics exhibited positive associations with global cognition, particularly cortical thickness in left pars opercularis (r = .476, p < .001). NYHA class is an independent risk factor for adverse outcome (p = .001). The observed correlation between abnormal cortical morphology and global cognition suggested that cortical morphology may serve as a promising imaging biomarker and provide insights into neuroanatomical underpinnings of cognitive impairment in patients with chronic HF.

FFR CT and Static Computed Tomography Myocardial Perfusion Imaging for Therapeutic Decision-making and Prognosis in Patients With Coronary Artery Disease.

PURPOSE: The purpose of this study was to investigate the effect of integrated evaluation of resting static computed tomography perfusion (CTP) and coronary computed tomography angiography (CCTA)-derived fractional flow reserve (FFR CT ) on therapeutic decision-making and predicting major adverse cardiovascular events (MACEs) in patients with suspected coronary artery disease. MATERIALS AND METHODS: In this post hoc analysis of a prospective trial of CCTA in patients assigned to either CCTA or CCTA plus FFR CT arms, 500 patients in the CCTA plus FFR CT arm were analyzed. Both resting static CTP and FFR CT were evaluated by using the conventional CCTA. Perfusion defects in the myocardial segments with ≥50% degree of stenosis in the supplying vessels were defined as resting static CTP positive, and any vessel with an FFR CT value of ≤0.80 was considered positive. Patients were divided into 3 groups: (1) negative CTP-FFR CT match group (resting static CTP-negative and FFR CT -negative group); (2) mismatch CTP-FFR CT group (resting static CTP-positive and FFR CT -negative or resting static CTP-negative and FFR CT -positive group); and (3) positive CTP-FFR CT match group (resting static CTP-positive and FFR CT -positive group). We compared the revascularization-to-invasive coronary angiography ratio and the MACE rate among 3 subgroups at 1- and 3-year follow-ups. The adjusted Cox hazard proportional model was used to assess the prognostic value of FFR CT and resting static CTP to determine patients at risk of MACE. RESULTS: Patients in the positive CTP-FFR CT match group were more likely to undergo revascularization at the time of invasive coronary angiography compared with those in the mismatch CTP-FFR CT group (81.4% vs 57.7%, P =0.033) and the negative CTP-FFR CT match group (81.4% vs 33.3%, P= 0.001). At 1- and 3-year follow-ups, patients in the positive CTP-FFR CT match group were more likely to have MACE than those in the mismatch CTP-FFR CT group (10.5% vs 4.2%, P= 0.046; 35.6% vs 9.4%, P <0.001) and the negative CTP-FFR CT match group (10.5% vs 0.9%, P <0.001; 35.6% vs 5.4%, P <0.001). A positive CTP-FFR CT match was strongly related to MACE at 1-year (hazard ratio=8.06, P= 0.003) and 3-year (hazard ratio=6.23, P <0.001) follow-ups. CONCLUSION: In patients with suspected coronary artery disease, the combination of FFR CT with resting static CTP could guide therapeutic decisions and have a better prognosis with fewer MACE in a real-world scenario.

Late gadolinium enhanced cardiac MR derived radiomics approach for predicting all-cause mortality in cardiac amyloidosis: a multicenter study.

OBJECTIVES: To evaluate the prognostic value of radiomics features based on late gadolinium enhancement (LGE) cardiac magnetic resonance (CMR) images in patients with cardiac amyloidosis (CA). METHODS: This retrospective study included 120 CA patients undergoing CMR at three institutions. Radiomics features were extracted from global and three different segments (base, mid-ventricular, and apex) of left ventricular (LV) on short-axis LGE images. Primary endpoint was all-cause mortality. The predictive performance of the radiomics features and semi-quantitative and quantitative LGE parameters were compared by ROC. The AUC was used to observe whether Rad-score had an incremental value for clinical stage. The Kaplan-Meier curve was used to further stratify the risk of CA patients. RESULTS: During a median follow-up of 12.9 months, 30% (40/120) patients died. There was no significant difference in the predictive performance of the radiomics model in different LV sections in the validation set (AUCs of the global, basal, middle, and apical radiomics model were 0.75, 0.77, 0.76, and 0.77, respectively; all p > 0.05). The predictive performance of the Rad-score of the base-LV was better than that of the LGE total enhancement mass (AUC:0.77 vs. 0.54, p < 0.001) and LGE extent (AUC: 0.77 vs. 0.53, p = 0.004). Rad-score combined with Mayo stage had better predictive performance than Mayo stage alone (AUC: 0.86 vs. 0.81, p = 0.03). Rad-score (≥ 0.66) contributed to the risk stratification of all-cause mortality in CA. CONCLUSIONS: Compared to quantitative LGE parameters, radiomics can better predict all-cause mortality in CA, while the combination of radiomics and Mayo stage could provide higher predictive accuracy. CLINICAL RELEVANCE STATEMENT: Radiomics analysis provides incremental value and improved risk stratification for all-cause mortality in patients with cardiac amyloidosis. KEY POINTS: • Radiomics in LV-base was superior to LGE semi-quantitative and quantitative parameters for predicting all-cause mortality in CA. • Rad-score combined with Mayo stage had better predictive performance than Mayo stage alone or radiomics alone. • Rad-score ≥ 0.66 was associated with a significantly increased risk of all-cause mortality in CA patients.

Machine learning based ischemia-specific stenosis prediction: A Chinese multicenter coronary CT angiography study.

OBJECTIVES: To evaluate the performance of coronary computed tomography angiography (CCTA) derived characteristics including CT derived fractional flow reserve (CT-FFR) with FFR as a reference standard in identifying the lesion-specific ischemia by machine learning (ML) algorithms. METHODS: The retrospective analysis enrolled 596 vessels in 462 patients (mean age, 61 years ± 11 [SD]; 71.4 % men) with suspected coronary artery disease who underwent CCTA and invasive FFR. The data were divided into training cohort, internal validation cohort, external validation cohorts 1 and 2 according to participating centers. All CCTA-derived parameters, which contained 10 qualitative and 33 quantitative plaque parameters, were collected to establish ML model. The Boruta and unsupervised clustering algorithm were implemented to select important and non-redundant parameters. Finally, the eight features with the highest mean importance were included for further ML model establishment and decision tree building. Five models were built to predict lesion-specific ischemia: stenosis degree from CCTA, CT-FFR, ΔCT-FFR, ML model and nested model. RESULTS: Low-attenuation plaque, bend and lesion length were the main predictors of ischemia-specific lesions. Of 5 models, the ML model showed favorable discrimination for ischemia-specific lesions in the training and three validation sets (area under the curve [95 % confidence interval], 0.93 [0.90-0.96], 0.86 [0.79-0.94], 0.88 [0.83-0.94], and 0.90 [0.84-0.96], respectively). The nested model which combined the ML model and CT-FFR showed better diagnostic efficacy (AUC [95 %CI], 0.96 [0.94-0.99], 0.92 [0.86-0.99], 0.92 [0.86-0.99] and 0.94 [0.91-0.98], respectively; all P < 0.05), and net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were significantly higher than CT-FFR alone. CONCLUSIONS: Comprehensive CCTA-derived multiparameter model could better predict the ischemia-specific lesions by ML algorithms compared to stenosis degree from CTA, CT-FFR and ΔCT-FFR. Decision tree can be used to predict myocardial ischemia effectively.

Development and Validation of CCTA-based Radiomics Signature for Predicting Coronary Plaques With Rapid Progression.

BACKGROUND: Rapid plaque progression (RPP) is associated with a higher risk of acute coronary syndromes compared with gradual plaque progression. We aimed to develop and validate a coronary computed tomography angiography (CCTA)-based radiomics signature (RS) of plaques for predicting RPP. METHODS: A total of 214 patients who underwent serial CCTA examinations from 2 tertiary hospitals (development group, 137 patients with 164 lesions; validation group, 77 patients with 101 lesions) were retrospectively enrolled. Conventional CCTA-defined morphological parameters (eg, high-risk plaque characteristics and plaque burden) and radiomics features of plaques were analyzed. RPP was defined as an annual progression of plaque burden ≥1.0% on lesion-level at follow-up CCTA. RS was built to predict RPP using XGBoost method. RESULTS: RS significantly outperformed morphological parameters for predicting RPP in both the development group (area under the receiver operating characteristic curve, 0.82 versus 0.74; P=0.04) and validation group (area under the receiver operating characteristic curve, 0.81 versus 0.69; P=0.04). Multivariable analysis identified RS (odds ratio, 2.35 [95% CI, 1.32-4.46]; P=0.005) as an independent predictor of subsequent RPP in the validation group after adjustment of morphological confounders. Unlike unchanged RS in the non-RPP group, RS increased significantly in the RPP group at follow-up in the whole dataset (P<0.001). CONCLUSIONS: The proposed CCTA-based RS had a better discriminative value to identify plaques at risk of rapid progression compared with conventional morphological plaque parameters. These data suggest the promising utility of radiomics for predicting RPP in a low-risk group on CCTA.

Burden and distribution of monosodium urate deposition in patients with hyperuricemia and gout: a cross-sectional Chinese population-based dual-energy CT study.

Background: To investigate the distribution and burden of monosodium urate (MSU) deposition in hyperuricemia and gout patients with dual-energy computed tomography (DECT). Methods: A total of 1,936 consecutive patients from January 1, 2009, to September 15, 2017, underwent DECT examinations in Jinling Hospital. Of these, 1,294 patients were excluded due to other clinical diagnoses (n=1,041), inappropriate locations (n=82), poor-quality images (n=105), training cases (n=30) and duplicated data (n=36). Finally, 642 patients were included in this study. We retrospectively analyzed 1,127 DECT examinations in 642 consecutive patients (hyperuricemia group, n=121; gout group, n=521) and recorded the volume and number of MSU deposits. For each anatomical location, we recorded MSU deposition in the soft tissue and joint cavity. MSU deposition was analyzed and compared between groups. For normally distributed data, independent sample t-tests were used for comparison between the two groups. The independent samples nonparametric test was used to analyze nonnormally distributed data. Results: (I) The burden of MSU deposition in the gout group {volume [0.14 (0.04-1.36)] and numbers [10.00 (5.00-19.00)]} was significantly higher than that {volume [0.08 (0.02-0.47), P=0.003] and numbers [9.50 (2.00-16.00), P=0.01]} in the hyperuricemia group. (II) The burden of MSU deposition in the knees {volume [0.24 (0.01-1.79), P=0.002] and quantity [6.00 (2.00-12.00), P=0.04]} and feet {volume [0.10 (0.04-0.66)] and number [9.00 (5.00-15.00)]} was significantly higher in the gout group than those {knees: the volume [0.03 (0.00-0.27), P=0.002] and the quantity [4.00 (0.00-9.00), P=0.04]; feet: the volume [0.07 (0.02-0.19), P=0.003)] and number [8.00 (2.25-12.00), P=0.04]} in the hyperuricemia group, respectively. (III) In the hyperuricemia group, the volume of MSU deposition was significantly higher in the soft tissues of the knee (0.022±0.042) and ankle (0.062±0.305) than in those (knee: 0.001±0.005, P=0.02; ankle: 0.027±0.234, P=0.02) in the joint cavity. Conclusions: Although subclinical urate deposition can occur in patients with asymptomatic hyperuricemia, the burden of urate deposition is greater in patients with symptomatic gout, and the distribution is more pronounced in the foot/knee. Thus, more effective patient management and monitoring can be achieved by measuring the burden of MSU deposits in the patient's feet/knees. These data suggest that a threshold for urate crystal volume at typical sites may be required before symptomatic disease develops.

Optimal Measurement Sites of Coronary-Computed Tomography Angiography-derived Fractional Flow Reserve: The Insight From China CT-FFR Study.

OBJECTIVES: To investigate the optimal measurement site of coronary-computed tomography angiography-derived fractional flow reserve (FFR CT ) for the assessment of coronary artery disease (CAD) in the whole clinical routine practice. MATERIALS AND METHODS: This retrospective multicenter study included 396 CAD patients who underwent coronary-computed tomography angiography, FFR CT , and invasive FFR. FFR CT was measured at 1 cm (FFR CT -1 cm), 2 cm (FFR CT -2 cm), 3 cm (FFR CT -3 cm), and 4 cm (FFR CT -4 cm) distal to coronary stenosis, respectively. FFR CT and invasive FFR ≤0.80 were defined as lesion-specific ischemia. The diagnostic performance of FFR CT to detect ischemia was obtained using invasive FFR as the reference standard. Reduced invasive coronary angiography rate and revascularization efficiency were calculated. After a median follow-up of 35 months in 267 patients for major adverse cardiovascular events (MACE), Cox hazard proportional models were performed with FFR CT values at each measurement site. RESULTS: For discriminating lesion-specific ischemia, the areas under the curve of FFR CT -1 cm (0.91) as well as FFR CT -2 cm (0.91) were higher than those of FFR CT -3 cm (0.89) and FFR CT -4 cm (0.88), respectively (all P <0.05). The higher reduced invasive coronary angiography rate (81.6%) was found at FFR CT -1 cm than FFR CT -2 cm (81.6% vs. 62.6%, P <0.05). Revascularization efficiency did not differ between FFR CT -1 cm and FFR CT -2 cm (80.8% vs. 65.5%, P =0.019). In 12.4% (33/267) MACE occurred and only values of FFR CT -2 cm were independently predictive of MACE (hazard ratio: 0.957 [95% CI: 0.925-0.989]; P =0.010). CONCLUSIONS: This study indicates FFR CT -2 cm is the optimal measurement site with superior diagnostic performance and independent prognostic role.

The prognostic value of CT-derived fractional flow reserve in coronary artery bypass graft: a retrospective multicenter study.

OBJECTIVES: To investigate the predictive value of CT-derived fractional flow reserve (FFRCT) in anastomosis occlusion after coronary artery bypass graft (CABG) surgery. METHODS: Patients undergoing CABG with both pre- and post-operative coronary computed tomographic angiography (CCTA) were retrospectively included. Preoperative CCTA studies were used to evaluate anatomical and FFRCT information of target vessels. A diameter stenosis (DS) ≥ 70% or left main > 50% was considered to be anatomically severe, while FFRCT value ≤ 0.80 be functionally significant. The primary endpoint was anastomosis occlusion evaluated on post-operative CCTA during follow-up. Predictors of anastomosis occlusion were assessed by the multivariate binary logistic regression with generalized estimating equations. RESULTS: A total of 270 anastomoses were identified in 88 enrolled patients. Forty-one anastomoses from 30 patients exhibited occlusion during a follow-up of 15.3 months after CABG. The occluded group had significantly increased prevalence of non-severe DS (58.5% vs. 40.2%; p = 0.023) and non-significant FFRCT (48.8% vs. 10.0%; p < 0.001). Multivariable analysis indicated FFRCT ≤ 0.80 (odds ratio [OR]: 0.10, 95% CI: 0.03-0.33; p < 0.001) and older age (OR: 0.92, 95% CI: 0.87-0.97; p = 0.001) were predictors for bypass patency during follow-up, while myocardial infarction history and anastomosis to a local lesion or bifurcation (all p value < 0.05) were predictors of occlusion. Adding FFRCT into the model based on the clinical and anatomical predictors had an improved AUC of 0.848 (p = 0.005). CONCLUSIONS: FFRCT ≤ 0.80 was associated with a significant risk reduction of anastomosis occlusion after CABG. Preoperative judgment of the hemodynamic significance may improve the CABG surgery strategy and reduce graft failure. KEY POINTS: • FFRCT ≤ 0.80 was associated with a significant risk reduction of anastomosis occlusion after CABG. • The addition of FFRCT into the integrated model including clinical (age and history of myocardial infarction) and anatomical CCTA indicators (local lesion and bifurcation) significantly improved the model performance with an AUC of 0.848 (p = 0.005). • Preoperative judgment of the hemodynamic significance may help improve the decision-making and surgery planning in patients indicated for CABG and significantly reduce graft failure, without an extra radiation exposure and risk of invasive procedure.

Cardiovascular Imaging in China: Yesterday, Today, and Tomorrow.

The high prevalence and mortality of cardiovascular diseases in China's large population has increased the use of cardiovascular imaging for the assessment of conditions in recent years. In this study, we review the past 20 years of cardiovascular imaging in China, the increasingly important role played by cardiovascular computed tomography in coronary artery disease and pulmonary embolism assessment, magnetic resonance imaging's use for cardiomyopathy assessment, the development and application of artificial intelligence in cardiovascular imaging, and the future of Chinese cardiovascular imaging.

Diagnosis of Cardiac Amyloidosis Using a Radiomics Approach Applied to Late Gadolinium-Enhanced Cardiac Magnetic Resonance Images: A Retrospective, Multicohort, Diagnostic Study.

Objectives: To assess the potential of a radiomics approach of late gadolinium enhancement (LGE) cardiac magnetic resonance (CMR) in the diagnosis of cardiac amyloidosis (CA). Materials and Methods: This retrospective study included 200 patients with biopsy-proven light-chain (AL) amyloidosis. CA was diagnosed on the basis of systemic amyloidosis confirmed with evidence of cardiac involvement by imaging and clinical biomarkers. A total of 139 patients [54 ± 8 years, 75 (54%) men] in our institution were divided into training cohort [n = 97, mean age of 53 ± 8 years, 54 (56%) men] and internal validation cohort [n = 42, mean age: 56 ± 8 years, 21 (50%) men] with a ratio of 7:3, while 61 patients [mean age: 60 ± 9 years, 42 (69%) men] from the other two institutions were enrolled for external validation. Radiomics features were extracted from global (all short-axis images from base-to-apex) left ventricular (LV) myocardium and three different segments (basal, midventricular, and apex) on short-axis LGE images using the phase-sensitive reconstruction (PSIR) sequence. The Boruta algorithm was used to select the radiomics features. This model was built using the XGBoost algorithm. The two readers performed qualitative and semiquantitative assessment of the LGE images based on the visual LGE patterns, while the quantitative assessment was measured using a dedicated semi-automatic CMR software. The diagnostic performance of the radiomics and other qualitative and quantitative parameters were compared by a receiver operating characteristic (ROC) curve analysis. A correlation between radiomics and the degree of myocardial involvement by amyloidosis was tested. Results: A total of 1,906 radiomics features were extracted for each LV section. No statistical significance was indicated between any two slices for diagnosing CA, and the highest area under the curve (AUC) was found in basal section {0.92 [95% confidence interval (CI), 0.86-0.97] in the LGE images in the training set, 0.89 (95% CI, 0.79-1.00) in the internal validation set, and 0.92 (95% CI, 0.85-0.99) in the external validation set}, which was superior to the visual assessment and quantitative LGE parameters. Moderate correlations between global or basal radiomics scores (Rad-scores) and Mayo stage in all patients were reported (Spearman's Rho = 0.61, 0.62; all p < 0.01). Conclusion: A radiomics analysis of the LGE images provides incremental information compared with the visual assessment and quantitative parameters on CMR to diagnose CA. Radiomics was moderately correlated with the severity of CA. Further studies are needed to assess the prognostic significance of radiomics in patients with CA.

Functional CAD-RADS using FFRCT on therapeutic management and prognosis in patients with coronary artery disease.

OBJECTIVES: To propose a novel functional Coronary Artery Disease-Reporting and Data System (CAD-RADS) category system integrated with coronary CT angiography (CCTA)-derived fractional flow reserve (FFRCT) and to validate its effect on therapeutic decision and prognosis in patients with coronary artery disease (CAD). METHODS: Firstly, we proposed a novel functional CAD-RADS and evaluated the performance of functional CAD-RADS for guiding treatment strategies with actual clinical treatment as a reference standard in a retrospective multicenter cohort with CCTA and invasive FFR performed in all patients (n = 466). Net reclassification improvement (NRI) of functional CAD-RADS over anatomical CAD-RADS was calculated. Secondly, the prognostic value of functional CAD-RADS in a prospective two-arm cohort (566 [FFRCT arm] vs. 567 [CCTA arm]) was calculated, after a 1-year follow-up, functional CAD-RADS in FFRCT arm (n = 513) and anatomical CAD-RADS in CCTA arm (n = 511) to determine patients at risk of adverse outcomes were compared with a Cox hazard proportional model. RESULTS: Functional CAD-RADS demonstrated superior value over anatomical CAD-RADS (AUC: 0.828 vs. 0.681, p < 0.001) and comparable performance to FFR (AUC: 0.828 vs. 0.848, p = 0.253) in guiding therapeutic decisions. Functional CAD-RADS resulted in the revision of management plan as determined by anatomical CAD-RADS in 30.0% of patients (n = 140) (NRI = 0.369, p < 0.001). Functional CAD-RADS was an independent predictor for 1-year outcomes with indexes of concordance of 0.795 and the corresponding value was 0.751 in anatomical CAD-RADS. CONCLUSION: The novel functional CAD-RADS gained incremental value in guiding therapeutic decision-making compared with anatomical CAD-RADS and comparable power in 1-year prognosis with anatomical CAD-RADS in a real-world scenario. KEY POINTS: • The novel functional CAD-RADS category system with FFRCT integrated into the anatomical CAD-RADS categories was originally proposed. • The novel functional CAD-RADS category system was validated superior value over anatomical CAD-RADS (AUC: 0.828 vs. 0.681, p < 0.001) in guiding therapeutic decisions and revised management plan in 30.0% of patients as determined by anatomical CAD-RADS (net reclassification improvement index = 0.369, p < 0.001). • Functional CAD-RADS was an independent predictor with an index of concordance of 0.795 and 0.751 in anatomical CAD-RADS for 1-year prognosis of adverse outcomes.

One-year outcomes of CCTA alone versus machine learning-based FFRCT for coronary artery disease: a single-center, prospective study.

OBJECTIVES: To explore downstream management and outcomes of machine learning (ML)-based CT derived fractional flow reserve (FFRCT) strategy compared with an anatomical coronary computed tomography angiography (CCTA) alone assessment in participants with intermediate coronary artery stenosis. METHODS: In this prospective study conducted from April 2018 to March 2019, participants were assigned to either the CCTA or FFRCT group. The primary endpoint was the rate of invasive coronary angiography (ICA) that demonstrated non-obstructive disease at 90 days. Secondary endpoints included coronary revascularization and major adverse cardiovascular events (MACE) at 1-year follow-up. RESULTS: In total, 567 participants were allocated to the CCTA group and 566 to the FFRCT group. At 90 days, the rate of ICA without obstructive disease was higher in the CCTA group (33.3%, 39/117) than that (19.8%, 19/96) in the FFRCT group (risk difference [RD] = 13.5%, 95% confidence interval [CI]: 8.4%, 18.6%; p = 0.03). The ICA referral rate was higher in the CCTA group (27.5%, 156/567) than in the FFRCT group (20.3%, 115/566) (RD = 7.2%, 95% CI: 2.3%, 12.1%; p = 0.003). The revascularization-to-ICA ratio was lower in the CCTA group than that in the FFRCT group (RD = 19.8%, 95% CI: 14.1%, 25.5%, p = 0.002). MACE was more common in the CCTA group than that in the FFRCT group at 1 year (HR: 1.73; 95% CI: 1.01, 2.95; p = 0.04). CONCLUSION: In patients with intermediate stenosis, the FFRCT strategy appears to be associated with a lower rate of referral for ICA, ICA without obstructive disease, and 1-year MACE when compared to the anatomical CCTA alone strategy. KEY POINTS: • In stable patients with intermediate stenosis, ML-based FFRCT strategy was associated with a lower referral ICA rate, a lower normalcy rate of ICA, and higher revascularization-to-ICA ratio than the CCTA strategy. • Compared with the CCTA strategy, ML-based FFRCTshows superior outcome prediction value which appears to be associated with a lower rate of 1-year MACE. • ML-based FFRCT strategy as a non-invasive "one-stop-shop" modality may be the potential to change diagnostic workflows in patients with suspected coronary artery disease.

Influence of diabetes mellitus on the diagnostic performance of machine learning-based coronary CT angiography-derived fractional flow reserve: a multicenter study.

OBJECTIVES: To examine the diagnostic accuracy of machine learning-based coronary CT angiography-derived fractional flow reserve (FFRCT) in diabetes mellitus (DM) patients. METHODS: In total, 484 patients with suspected or known coronary artery disease from 11 Chinese medical centers were retrospectively analyzed. All patients underwent CCTA, FFRCT, and invasive FFR. The patients were further grouped into mild (25~49 %), moderate (50~69 %), and severe (≥ 70 %) according to CCTA stenosis degree and Agatston score < 400 and Agatston score ≥ 400 groups according to coronary artery calcium severity. Propensity score matching (PSM) was used to match DM (n  = 112) and non-DM (n  = 214) groups. Sensitivity, specificity, accuracy, and area under the curve (AUC) with 95 % confidence interval (CI) were calculated and compared. RESULTS: Sensitivity, specificity, accuracy, and AUC of FFRCT were 0.79, 0.96, 0.87, and 0.91 in DM patients and 0.82, 0.93, 0.89, and 0.89 in non-DM patients without significant difference (all p > 0.05) on a per-patient level. The accuracies of FFRCT had no significant difference among different coronary stenosis subgroups and between two coronary calcium subgroups (all p > 0.05) in the DM and non-DM groups. After PSM grouping, the accuracies of FFRCT were 0.88 in the DM group and 0.87 in the non-DM group without a statistical difference (p > 0.05). CONCLUSIONS: DM has no negative impact on the diagnostic accuracy of machine learning-based FFRCT. KEY POINTS: • ML-based FFRCT has a high discriminative accuracy of hemodynamic ischemia, which is not affected by DM. • FFRCT was superior to the CCTA alone for the detection of ischemia relevance of coronary artery stenosis in both DM and non-DM patients. • Coronary calcification had no significant effect on the diagnostic accuracy of FFRCT to detect ischemia in DM patients.

Deep learning powered coronary CT angiography for detecting obstructive coronary artery disease: The effect of reader experience, calcification and image quality.

OBJECTIVES: To investigate the effect of reader experience, calcification and image quality on the performance of deep learning (DL) powered coronary CT angiography (CCTA) in automatically detecting obstructive coronary artery disease (CAD) with invasive coronary angiography (ICA) as reference standard. METHODS: A total of 165 patients (680 vessels and 1505 segments) were included in this study. Three sessions were performed in order: (1) The artificial intelligence (AI) software automatically processed CCTA images, stenosis degree and processing time were recorded for each case; (2) Six cardiovascular radiologists with different experiences (low/ intermediate/ high experience) independently performed image post-processing and interpretation of CCTA, (3) AI + human reading was performed. Luminal stenosis ≥50% was defined as obstructive CAD in ICA and CCTA. Diagnostic performances of AI, human reading and AI + human reading were evaluated and compared on a per-patient, per-vessel and per-segment basis with ICA as reference standard. The effects of calcification and image quality on the diagnostic performance were also studied. RESULTS: The average post-processing and interpretation times of AI was 2.3 ± 0.6 min per case, reduced by 76%, 72%, 69% compared with low/ intermediate/ high experience readers (all P < 0.001), respectively. On a per-patient, per-vessel and per-segment basis, with ICA as reference method, the AI overall diagnostic sensitivity for detecting obstructive CAD were 90.5%, 81.4%, 72.9%, the specificity was 82.3%, 93.9%, 95.0%, with the corresponding areas under the curve (AUCs) of 0.90, 0.90, 0.87, respectively. Compared to human readers, the diagnostic performance of AI was higher than that of low experience readers (all P < 0.001). The diagnostic performance of AI + human reading was higher than human reading alone, and AI + human readers' ability to correctly reclassify obstructive CAD was also improved, especially for low experience readers (Per-patient, the net reclassification improvement (NRI) = 0.085; per-vessel, NRI = 0.070; and per-segment, NRI = 0.068, all P < 0.001). The diagnostic performance of AI was not significantly affected by calcification and image quality (all P > 0.05). CONCLUSIONS: AI can substantially shorten the post-processing time, while AI + human reading model can significantly improve the diagnostic performance compared with human readers, especially for inexperienced readers, regardless of calcification severity and image quality.

Feasibility and prognostic role of machine learning-based FFRCT in patients with stent implantation.

OBJECTIVES: To investigate the feasibility and prognostic implications of coronary CT angiography (CCTA) derived fractional flow reserve (FFRCT) in patients who have undergone stents implantation. METHODS: Firstly, the feasibility of FFRCT in stented vessels was validated. The diagnostic performance of FFRCT in identifying hemodynamically in-stent restenosis (ISR) in 33 patients with invasive FFR ≤ 0.88 as reference standard, intra-group correlation coefficient (ICC) between FFRCT and FFR was calculated. Secondly, prognostic value was assessed with 115 patients with serial CCTA scans after PCI. Stent characteristics (location, diameter, length, etc.), CCTA measurements (minimum lumen diameter [MLD], minimum lumen area [MLA], ISR), and FFRCT measurements (FFRCT, ΔFFRCT, ΔFFRCT/stent length) both at baseline and follow-up were recorded. Longitudinal analysis included changes of MLD, MLA, ISR, and FFRCT. The primary endpoint was major adverse cardiovascular events (MACE). RESULTS: Per-patient accuracy of FFRCT was 0.85 in identifying hemodynamically ISR. FFRCT had a good correlation with FFR (ICC = 0.84). 15.7% (18/115) developed MACE during 25 months since follow-up CCTA. Lasso regression identified age and follow-up ΔFFRCT/length as candidate variables. In the Cox proportional hazards model, age (hazard ratio [HR], 1.102 [95% CI, 1.032-1.177]; p = 0.004) and follow-up ΔFFRCT/length (HR, 1.014 [95% CI, 1.006-1.023]; p = 0.001) were independently associated with MACE (c-index = 0.856). Time-dependent ROC analysis showed AUC was 0.787 (95% CI, 0.594-0.980) at 25 months to predict adverse outcome. After bootstrap validation with 1000 resamplings, the bias-corrected c-index was 0.846. CONCLUSIONS: Noninvasive ML-based FFRCT is feasible in patients following stents implantation and shows prognostic value in predicting adverse events after stents implantation in low-moderate risk patients. KEY POINTS: • Machine-learning-based FFRCT is feasible to evaluate the functional significance of in-stent restenosis in patients with stent implantation. • Follow-up △FFRCT along with the stent length might have prognostic implication in patients with stent implantation and low-to-moderate risk after 2 years follow-up. The prognostic role of FFRCT in patients with moderate-to-high or high risk needs to be further studied. • FFRCT might refine the clinical pathway of patients with stent implantation to invasive catheterization.

Serial coronary CT angiography-derived fractional flow reserve and plaque progression can predict long-term outcomes of coronary artery disease.

OBJECTIVE: To investigate the utility of coronary CT angiography-derived fractional flow reserve (FFRCT) and plaque progression in patients undergoing serial coronary CT angiography for predicting major adverse cardiovascular events (MACE). METHODS: This retrospective study evaluated patients suspected or known coronary artery disease who underwent serial coronary CT angiography examinations between January 2006 and December 2017 and followed up until June 2019. The primary endpoint was MACE, defined as acute coronary syndrome, rehospitalization due to progressive angina, percutaneous coronary intervention, or cardiac death. FFRCT and plaque parameters were analyzed on a per-vessel and per-patient basis. Univariable and multivariable COX regression analysis determined predictors of MACE. The prognostic value of FFRCT and plaque progression were assessed in nested models. RESULTS: Two hundred eighty-four patients (median age, 61 years (interquartile range, 54-70); 202 males) were evaluated. MACE was observed in 45 patients (15.8%, 45/284). By Cox multivariable regression modeling, vessel-specific FFRCT ≤ 0.80 was associated with a 2.4-fold increased risk of MACE (HR (95% CI): 2.4 (1.3-4.4); p = 0.005) and plaque progression was associated with a 9-fold increased risk of MACE (HR (95% CI): 9 (3.5-23); p < 0.001) after adjusting for clinical and imaging risk factors. FFRCT and plaque progression improved the prediction of events over coronary artery calcium (CAC) score and high-risk plaques (HRP) in the receiver operating characteristics analysis (area under the curve: 0.70 to 0.86; p = 0.002). CONCLUSIONS: Fractional flow reserve and plaque progression assessed by serial coronary CT angiography predicted the risk of future MACE. KEY POINTS: • Vessel-specific CT angiography-derived fractional flow reserve (FFRCT) ≤ 0.80 and plaque progression improved the prediction of events over current risk factors. • Major adverse cardiovascular events (MACE) significantly increased with the presence of plaque progression at follow-up stratified by the FFRCT change group.

The effect of coronary calcification on diagnostic performance of machine learning-based CT-FFR: a Chinese multicenter study.

OBJECTIVE: To investigate the effect of coronary calcification morphology and severity on the diagnostic performance of machine learning (ML)-based coronary CT angiography (CCTA)-derived fractional flow reserve (CT-FFR) with FFR as a reference standard. METHODS: A total of 442 patients (61.2 ± 9.1 years, 70% men) with 544 vessels who underwent CCTA, ML-based CT-FFR, and invasive FFR from China multicenter CT-FFR study were enrolled. The effect of calcification arc, calcification remodeling index (CRI), and Agatston score (AS) on the diagnostic performance of CT-FFR was investigated. CT-FFR ≤ 0.80 and lumen reduction ≥ 50% determined by CCTA were identified as vessel-specific ischemia with invasive FFR as a reference standard. RESULTS: Compared with invasive FFR, ML-based CT-FFR yielded an overall sensitivity of 0.84, specificity of 0.94, and accuracy of 0.90 in a total of 344 calcification lesions. There was no statistical difference in diagnostic accuracy, sensitivity, or specificity of CT-FFR across different calcification arc, CRI, or AS levels. CT-FFR exhibited improved discrimination of ischemia compared with CCTA alone in lesions with mild-to-moderate calcification (AUC, 0.89 vs. 0.69, p < 0.001) and lesions with CRI ≥ 1 (AUC, 0.89 vs. 0.71, p < 0.001). The diagnostic accuracy and specificity of CT-FFR were higher than CCTA alone in patients and vessels with mid (100 to 299) or high (≥ 300) AS. CONCLUSION: Coronary calcification morphology and severity did not influence diagnostic performance of CT-FFR in ischemia detection, and CT-FFR showed marked improved discrimination of ischemia compared with CCTA alone in the setting of calcification. KEY POINTS: • CT-FFR provides superior diagnostic performance than CCTA alone regardless of coronary calcification. • No significant differences in the diagnostic performance of CT-FFR were observed in coronary arteries with different coronary calcification arcs and calcified remodeling indexes. • No significant differences in the diagnostic accuracy of CT-FFR were observed in coronary arteries with different coronary calcification score levels.