Stress Perfusion Cardiac Magnetic Resonance Research Articles

Serial quantitative stress perfusion cardiac magnetic resonance in patients undergoing coronary artery bypass grafting; prediction of symptomatic benefit

Abstract Background Serial multiparametric cardiovascular magnetic resonance (CMR) in patients undergoing coronary artery bypass grafting (CABG) may provide mechanistic insight into dynamic and persistent abnormalities of the myocardium in patients with stable coronary artery disease (CAD). Objectives To assess for dynamic CMR markers of myocardial hypoperfusion and cardiac remodelling in patients undergoing CABG. Methods Patients with CAD awaiting elective CABG were recruited into an observational cohort study. Baseline evaluation included bloods, Seattle Angina Questionnaire-7, and adenosine stress perfusion CMR. Automated fully quantitative stress and rest myocardial blood flow was calculated, alongside assessment of the visual ischaemic burden. Native and post-contrast septal T1 indices were also measured. Repeat evaluation at 6-12 months post CABG was performed. Results Of 43 patients who underwent serial evaluation with CMR (mean age 62.1±9.3, median LVEF 68% [IQR: 62-73%]), there was improvement in the median visual ischaemic burden (18% [IQR: 11-21] vs 42% [IQR: 27-51], P<0.001), mean global stress myocardial blood flow (1.5±0.5 ml/min/g vs 1.3±0.5 ml/min/g, P=0.002) and median global myocardial perfusion reserve (2.4 [IQR: 1.7-2.8] vs 1.7 [IQR: 1.4-2.2], P=0.002) following CABG. Greater improvement in the SAQ-7 summary score was associated with a larger visual ischaemic burden at baseline (ρ=0.44, P=0.004) and a greater decrease in the visual ischaemic burden following CABG (ρ=-0.38, P=0.02). Quantitative MBF metrics did not associate with baseline or change in SAQ-7 summary score. Mean LV extracellular matrix volume decreased following CABG (16.6±3.6ml/m2 vs 18.2±4.6ml/m2, P=0.009). Conclusion Multiparametric CMR identifies dynamic changes in markers of myocardial perfusion and interstitial fibrosis in patients following CABG. Visual assessment of inducible myocardial hypoperfusion may identify patients who will derive the most symptomatic benefit from CABG. The results, however, suggest an important degree of residual inducible ischaemia and an unclear clinical role for CMR-derived MBF calculations.Serial CMR perfusion dataAssociation of SAQ with perfusion

Fractional flow reserve in saphenous vein grafts for the detection of myocardial ischemia determined by perfusion imaging

Abstract Background Fractional flow reserve (FFR) is recommended to guide revascularization of native coronary arteries. Revascularization decision-making for saphenous vein grafts (SVGs), however, is based on angiographic estimation of lesion severity, as studies exploring the value of FFR to detect perfusion imaging defined myocardial ischemia in SVGs are scarce. Purpose We aimed to define an optimal FFR threshold and compare its diagnostic performance with the traditional FFR cutoff and with percentage diameter stenosis (%DS) in patients evaluated for SVG disease using myocardial perfusion imaging (MPI) as the reference standard. Methods Symptomatic patients with prior coronary artery bypass grafting (CABG) who underwent single-photon emission computed tomography, positron emission tomography or stress perfusion cardiac magnetic resonance imaging, and had FFR measurements of one or more SVGs were included. Angiographic lesion severity of the SVG was measured using 2D quantitative invasive coronary angiography. We allocated and matched the myocardial territory subtended by the SVG to compare the values of FFR and %DS with ischemia determined by MPI. Results Included were 77 patients (mean age 72 ± 7 years; 86% male) with 91 SVGs. A total of 38 SVGs (42%) supplied an ischemic coronary territory. FFR and %DS correlated moderately (r= -0.51, P<0.001). The optimal cutoff value of FFR was ≤0.94, resulting in a sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of 63%, 75%, 65%, 74% and 70%, respectively (Figure 1). Corresponding values of FFR≤0.80 and DS%≥50 are shown in Figure 1 as well. Sensitivity, NPV and accuracy of FFR ≤0.94 outperformed that of FFR≤0.80, whereas specificity was lower. FFR≤0.94 outperformed %DS≥50 only in terms of sensitivity. Conclusions The optimal cutoff value of FFR for the detection of MPI-defined disease in SVGs is higher than the traditional FFR cutoff used to guide native coronary artery revascularization. Importantly, the increased sensitivity led to a higher diagnostic performance of the adjusted cutoff. However, the high prevalence of ischemia in post-CABG patients questions whether the optimal cutoff reflects ischemia caused by SVG obstruction.Diagnostic performance of FFR and %DS

The relative importance of ischaemia and scar in the genesis of ventricular arrhythmia in ischaemic cardiomyopathy

Abstract Background Ventricular arrhythmias are a leading cause of death in patients with ischaemic cardiomyopathy (ICM). Ischaemia and scar are thought to contribute to arrhythmogenesis; as scar is considered fixed, strategies to reduce ischaemia such as revascularisation are undertaken to attempt to reduce risk, despite limited evidence behind this practice. Purpose The study rationale was to understand the relative contributions of ischaemia and scar on the genesis of arrhythmia in an ICM population, by undertaking a detailed assessment of arrhythmic substrate in a highly characterised population and secondly by validating these findings in a large cohort. Methods We assessed the relationship between scar, ischaemic burden and impact of ischaemia modulation via revascularisation on the arrhythmic substrate (study 1) and major arrhythmic events (study 2). S1: Patients were eligible for enrolment if they had an LVEF≤40%, extensive coronary disease (BCIS Jeopardy Score (JS) ≥6) and were due to undergo CABG or PCI. Arrhythmic substrate was characterised by the LV activation recovery interval (ARI) dispersion recorded by electrocardiographic imaging (ECGi). Scar and ischaemic burden were assessed via stress perfusion cardiac magnetic resonance (pCMR) imaging, calculated as a percentage of total LV myocardial volume. ECGi and pCMR were repeated 3 months post revasc. S2: Arrhythmia data were collected from device interrogation reports and blinded core lab analyses performed of CMR scans and angiograms of patients enrolled in the REVIVED-BCIS2 trial. Extent of revasc (quantified by change in JS/baseline JS) was used to estimate change in ischaemic burden. The primary outcome was death or aborted sudden death. Results S1: 30 patients were recruited; age 67±10 years, 87% male, BMI 28±5. Baseline LV ARI dispersion at rest was correlated with scar burden (r=0.37, p=0.04) but not ischaemic burden (r=0.25,p=0.21). Changes in metrics are shown(Table 1). Ischaemic burden reduced after revascularisation. Delta ischaemic burden was correlated with delta LV ARI dispersion (r=0.51, p=0.007 unadjusted). S2: Data was collected on 479 patients from REVIVED-BCI2 who underwent CMR; age 69±9 yrs, 88% male, BMI 28±6, LVEF 26±8%. 237 were randomised to PCI and 242 to OMT; no difference in rates of all-cause death, CV death, aborted sudden death, appropriate device therapy or arrhythmia was identified between the groups. Scar burden was associated with death or aborted sudden death, extent of revascularisation was not(Figure 1). Conclusion In our studies, the volume of scar was strongly related to both the arrhythmic substrate and occurrence of arrhythmic events. Whilst ischaemia reduction affected the arrhythmic substrate, this did not translate to a reduction in clinical events. These findings suggest that scar burden, rather than ischaemia reduction by revascularisation, should be the focus of risk stratification and therapeutic decision making in patients with ICM.Table 1Figure 1

Quantitative perfusion by cardiac magnetic resonance analysis reveals compromised myocardial perfusion in patients with angina with non-obstructive coronary artery disease

Abstract Introduction Stress perfusion cardiac magnetic resonance (CMR) visually detects myocardial ischemia through first-pass contrast-enhanced imaging by identifying perfusion defects. However, in generalized myocardial diseases like angina with non-obstructed coronary arteries (ANOCA), visual assessment becomes challenging due to diffuse sub-endocardial perfusion defects, or the findings may even falsely appear normal. Fully automated quantitative perfusion (QP), allowing absolute quantification of myocardial blood flow (MBF) and deriving myocardial perfusion reserve (MPR), might increase diagnostic accuracy in ANOCA patients. Therefore, the purpose of this study is to test the use of QP in detecting ANOCA diagnosed by the invasive golden standard. Methods This study is a cross-sectional cohort study assessing the use of fully automated QP CMR in patients with ANOCA compared with age- and sex-matched healthy controls. All participants underwent adenosine stress perfusion CMR, including visual assessment and quantification of MBF and MPR. Furthermore, all ANOCA patients underwent intracoronary function testing (ICFT) to idenity vasospasm and/or coronary microvascular dysfunction. Results This study included 24 ANOCA patients (83% women, 56.8±8.9 years) and 25 healthy controls (80% women, 56.2±7.2 years). Visual perfusion assessment showed no group differences in terms of perfusion defects (p=0.536). However, in QP, ANOCA patients had higher relative MBF than healthy controls (0.83±0.15 vs 0.74±0.17, p=0.046), while global perfusion was significantly lower during stress (2.43±0.72 vs 2.99±0.65 ml/g/min, p=0.006). Moreover, ANOCA patients had significantly lower MPR than healthy controls (2.24 ±0.79 vs 2.68 ±0.64, respectively p=0.036), with a higher prevalence of an impared MPR (50% vs. 20%, p=0.027). MPR was not significantly different between the various ANOCA entities as stratified during ICFT (p=0.766). Conclusions ANOCA patients displayed significantly reduced MPR, indicating impaired perfusion. The prevalence of diminished MPR was higher in the ANOCA group compared to the healthy controls, constrasting with no differences in visual assessment. These results are promising and underscores the possibilities of the use of QP in detecting vasomotor dysfunction non-invasively in ANOCA patients.Quantitative Perfusion by CMR results

Stress CMR Perfusion Imaging in the Medicare-Eligible Population: Insights From the SPINS Study.

Patients aged ≥65 years account for a disproportionately large portion of cardiovascular (CV) events and pose a challenge for noninvasive detection of coronary artery disease. This study sought to determine the prognostic value of stress cardiac magnetic resonance (CMR) in a Medicare-eligible group of patients in a multicenter setting in the United States. From a multicenter U.S. registry, the study identified patients aged≥65 years who were referred for stress CMR for evaluation of myocardial inducible ischemia. The primary outcome was defined as CV death or nonfatal myocardial infarction, whereas the secondary outcome was defined as any primary outcome, hospitalization for unstable angina, hospitalization for congestive heart failure, and unplanned late coronary artery bypass grafting. The associations of CMR findings with CV outcomes adjusted to clinical risk markers and health care cost spending were determined. Among 1,780 patients (aged 73 ± 5.7 years; 46% female), study investigators observed 144 primary eventsand 323 secondary events, over a median follow-up of 4.8 years. The presence of inducible ischemia and late gadolinium enhancement (LGE) was associated with incrementally higher event rates. Patients with neither inducible ischemia nor LGE experienced a<1% annualized rate of primary outcome. In a multivariable model adjusted for CV risk factors, inducible ischemia and LGE maintained an independent association with primary (HR: 2.80 [95%CI: 1.93-4.05]; P< 0.001; and HR: 1.85 [95%CI: 1.21-2.82]; P = 0.004, respectively) and secondary (HR: 2.46 [95%CI: 1.90-3.19]; P< 0.001; and HR: 1.72 [95%CI: 1.30-2.27]; P< 0.001, respectively) outcomes. Rates of revascularization, as well as downstream costs for patients without CMR-detected inducible ischemia, remained low throughout the follow-up period. In a multicenter cohort of Medicare-eligible older patients, stress CMR was effective in providing risk stratification. (Stress CMR Perfusion Imaging in the United States [SPINS] study; NCT03192891).

Plasma ceramides as biomarkers for microvascular disease and clinical outcomes in diabetes and myocardial infarction

BackgroundCeramides have recently been identified as novel biomarkers associated with diabetes mellitus (DM) and major adverse cardiac and cerebrovascular events (MACCE). This study aims to explore their utility in diagnosing microvascular disease.MethodsThis study prospectively enrolled 309 patients from 2018 to 2020 into three groups: healthy controls (Group 1, N = 51), DM patients without acute myocardial infarction (AMI) (Group 2, N = 150), and DM patients with AMI (Group 3, N = 108). We assessed outcomes using stress perfusion cardiac magnetic resonance (CMR) imaging for coronary microvascular disease (CMD) (Outcome 1), retinography for retinal microvascular disease (RMD) (Outcome 2), both CMD and RMD (Outcome 3), and absence of microvascular disease (w/o MD) (outcome 4). We evaluated the classification performance of ceramides using receiver operating characteristic (ROC) analysis and multiple logistic regression. 11-ceramide panel previously identified by our research group as related to macrovascular disease were used.ResultsAverage glycated hemoglobin (HbA1c) values were 5.1% in Group 1, 8.3% in Group 2, and 7.6% in Group 3. Within the cohort, CMD was present in 59.5% of patients, RMD in 25.8%, both CMD and RMD in 18.8%, and w/o MD in 38.5%. The AUC values for the reference ceramide ratios were as follows: CMD at 0.66 (p = 0.012), RMD at 0.61 (p = 0.248), CMD & RMD at 0.64 (p = 0.282), and w/o MD at 0.67 (p = 0.010). In contrast, the AUC values using 11-ceramide panel showed significant improvement in the outcomes prediction: CMD at 0.81 (p = 0.001), RMD at 0.73 (p = 0.010), CMD & RMD at 0.73 (p = 0.04), and w/o MD at 0.83 (p = 0.010). Additionally, the plasma concentration of C14.0 was notably higher in the w/o MD group (p < 0.001).ConclusionsPlasma ceramides serve as potential predictors for health status and microvascular disease phenotypes in diabetic patients.

Interdependence between myocardial deformation and perfusion in patients with T2DM and HFpEF: a feature-tracking and stress perfusion CMR study

BackgroundPatients with diabetes have an increased risk of developing heart failure with preserved ejection fraction (HFpEF). This study aimed to compare indices of myocardial deformation and perfusion between patients with type 2 diabetes mellitus (T2DM) with and without HFpEF and to investigate the relationship between myocardial strain and perfusion reserve.MethodsThis study included 156 patients with T2DM without obstructive coronary artery disease (CAD) and 50 healthy volunteers who underwent cardiac magnetic resonance (CMR) examination at our center. Patients with T2DM were subdivided into the T2DM–HFpEF (n = 74) and the T2DM–non-HFpEF (n = 82) groups. The parameters of left ventricular (LV) and left atrial (LA) strain as well as stress myocardial perfusion were compared. The correlation between myocardial deformation and perfusion parameters was also assessed. Mediation analyses were used to evaluate the direct and indirect effects of T2DM on LA strain.ResultsPatients with T2DM and HFpEF had reduced LV radial peak systolic strain rate (PSSR), LV circumferential peak diastolic strain rate (PDSR), LA reservoir strain, global myocardial perfusion reserve index (MPRI), and increased LA booster strain compared to patients with T2DM without HFpEF (all P < 0.05). Furthermore, LV longitudinal PSSR, LA reservoir, and LA conduit strain were notably impaired in patients with T2DM without HFpEF compared to controls (all P < 0.05), but LV torsion, LV radial PSSR, and LA booster strain compensated for these alterations (all P < 0.05). Multivariate linear regression analysis demonstrated that LA reservoir and LA booster strain were independently associated with global MPRI (β = 0.259, P < 0.001; β = − 0.326, P < 0.001, respectively). Further, the difference in LA reservoir and LA booster strain between patients with T2DM with and without HFpEF was totally mediated by global MPRI. Global stress PI, LA booster, global rest PI, and global MPRI showed high accuracy in diagnosing HFpEF among patients with T2DM (areas under the curve [AUC]: 0.803, 0.790, 0.740, 0.740, respectively).ConclusionsPatients with T2DM and HFpEF exhibited significant LV systolic and diastolic deformation, decreased LA reservoir strain, severe impairment of myocardial perfusion, and elevated LA booster strain that is a compensatory response in HFpEF. Global MPRI was identified as an independent influencing factor on LA reservoir and LA booster strain. The difference in LA reservoir and LA booster strain between patients with T2DM with and without HFpEF was totally mediated by global MPRI, suggesting a possible mechanistic link between microcirculation impairment and cardiac dysfunction in diabetes. Myocardial perfusion and LA strain may prove valuable for diagnosing and managing HFpEF in the future.

Diagnostic values of inferior Q-waves for myocardial scar identification detected by 3.0 T cardiac MRI

While Q-waves in inferior leads, particularly lead III, can be regarded as a minor abnormality, it can also indicate the presence of myocardial scar. This study assessed the diagnostic value of pathologic inferior Q-waves (lead II, III, aVF) for detecting ischemic scars using a high-resolution 3.0 T cardiac magnetic resonance (CMR). We retrospectively analyzed 1692 patients with suspected or known coronary artery disease who underwent stress CMR perfusion or viability assessment. Pathologic Q-waves were defined as duration of ≥ 30 ms and depth of ≥ 1 mm or QS-complex. Eleven models were created to evaluate the presence of Q-waves in different combinations of inferior leads. Of the 1692 patients, 436 (25.8%) had pathologic Q-waves. Models with Q-waves in leads II + aVF (model 7) and II + III + aVF (model 9) showed high specificity (100% and 99.6%), positive predictive value (PPV) (80.0% and 86.7%), and negative predictive value (NPV) (82.6% and 84.3%) but low sensitivity (1.3% and 13.1%). Other models also maintained high specificity and NPV but poor sensitivity and PPV. Notably, 21% of patients with an isolated pathologic Q-wave in lead III (model 4) exhibited scars. These findings highlight the need for careful clinical assessment when pathologic Q-waves are present.

Assessment of myocardial microvascular dysfunction in patients with different stages of diabetes mellitus: An adenosine stress perfusion cardiac magnetic resonance study

PurposeTo examine myocardial perfusion and T1 mapping indicesin individuals with type 2 diabetes mellitus (T2DM) at various stages of glycemic control and whether uncontrolled glycemic levels would worsen myocardial microvascular function. MethodCardiac magnetic resonance examinations were performed on 114 T2DM patients without obstructive coronary artery disease and 55 matched controls. Participants were further divided into four subgroups: Q1 (control); Q2 (prediabetes); Q3 (controlled T2DM) and Q4 (uncontrolled T2DM). The correlation between glycosylated hemoglobin (HbA1c) levels and myocardial perfusion parameters was evaluated. ResultsGlobal myocardial perfusion reserve index (MPRI) was significantly reduced in the Q3 and Q4 subgroups compared to the Q1 or Q2 subgroup (all P<0.001). Compared with the Q1 subgroup, global stress T1 reactivity (stress ΔT1) was significantly reduced in the Q3 and Q4 subgroups (P=0.004 and < 0.001, respectively), but elevated in the Q2 subgroup (P=0.018). Global extracellular volume (ECV) was considerably higher in the Q2 subgroup and gradually rose in the Q3 and Q4 subgroups compared to the Q1 subgroup (P=0.011, 0.001, and 0.007, respectively). HbA1c levels correlated negatively with global MPRI and stress ΔT1, but positively with global ECV (β = -1.993, P<0.001; β = -0.180, P<0.001; and β = 0.127, P<0.001, respectively). ConclusionsGlobal stress ΔT1 reduced in T2DM patients but rose in prediabetes patients. Compared to MPRI, the ECV parameter can indicate diabetes-induced coronary microvascular dysfunction earlier and persists throughout the disorder. Myocardial perfusion and T1 mapping at stress can be used to detect early signs of microvascular dysfunction and subclinical risk factors in patients with T2DM.

Non-invasive electrocardiographic assessment of arrhythmic risk in ischaemic cardiomyopathy patients undergoing revasularisation

Abstract Background Ventricular arrhythmias are a leading cause of death in patients with ischaemic cardiomyopathy (ICM). Ischaemia is thought to increase arrhythmic risk and coronary revascularisation is frequently undertaken to reduce this risk, although the recent REVIVED-BCIS2 trial did not show a difference in the incidence of death, device therapies or sustained arrhythmias in patients assigned to have percutaneous coronary intervention (PCI) compared to optimal medical therapy alone(1). Purpose We assessed the mechanistic effects of revascularisation on arrhythmic risk in patients with ICM using non-invasive electrocardiographic imaging (ECGi). Methods Patients were eligible if they had a left ventricular ejection fraction (LVEF) &amp;lt;40%, extensive coronary disease (BCIS Jeopardy Score≥6) and were scheduled to undergo revascularisation with coronary artery bypass grafting or PCI. ECGi assessment was performed using a 252-electrode vest with anatomical co-registration using a CT thorax. Patients underwent stress perfusion cardiac magnetic resonance (CMR) imaging; both ECGi and CMR were conducted at baseline and 3 months post revascularisation(Figure 1). The primary outcome measure was dispersion of repolarisation measured using global activation recovery interval (ARI) dispersion. Secondary outcome measures included repolarisation time (RT) dispersion, mean ARI and LV ARI dispersion. We also assessed the relationship between ARI dispersion, revascularisation and left ventricular scar burden. Data are presented as mean ± SD or median [IQR]. Results Thirty patients were recruited between 2020-2023; age 67±11 years, 90% male, BMI 28±5 and baseline LVEF was 29±8%. Twelve patients underwent CABG, 18 (60%) had PCI. The BCIS jeopardy score was 10 [9-11] at enrolment and decreased to 2 [0-4] after revascularisation. The LV scar burden at baseline was 15.1±6.5% and the change at 3 months was -0.5±5% (p=0.70). ARI dispersion and LV ARI dispersion were correlated with LV scar burden (r=0.62,p=0.01 and 0.71,p=0.002 respectively). No differences were seen in ARI dispersion (44.5 vs 44.1, p=0.86) or secondary outcome measures before and after revascularisation (RT dispersion: 45.5 vs 44.7, p=0.49; Mean ARI: 281.3 vs 282.6, p=0.74; ARI dispersion: 44.5 vs 44.1, p=0.86; LV ARI dispersion: 38.4 vs 36.7, p=0.39)(Figure 2). The LVEF change on CMR at 3 months was +6.9±10%; no relationship was seen between change in LVEF and ARI dispersion. Conclusion Arrhythmic risk assessed by non-invasive ECGi was not altered by coronary revascularisation. LV ARI dispersion was highly correlated with scar burden, which is not altered by revascularisation. Future work includes analysis of perfusion data. Our results support the findings of the REVIVED arrhythmia study that revascularisation does not reduce arrhythmic risk in patients with ICM(1). ECGi derived LV ARI dispersion correlates with scar and may represent a non-invasive arrhythmic risk assessment tool.Figure 1Figure 2

Hybrid artificial intelligence outcome prediction using features extraction from stress perfusion cardiac magnetic resonance images and electronic health records.

Prediction of clinical outcomes in coronary artery disease (CAD) has been conventionally achieved using clinical risk factors. The relationship between imaging features and outcome is still not well understood. This study aims to use artificial intelligence to link image features with mortality outcome. A retrospective study was performed on patients who had stress perfusion cardiac magnetic resonance (SP-CMR) between 2011 and 2021. The endpoint was all-cause mortality. Convolutional neural network (CNN) was used to extract features from stress perfusion images, and multilayer perceptron (MLP) to extract features from electronic health records (EHRs), both networks were concatenated in a hybrid neural network (HNN) to predict study endpoint. Image CNN was trained to predict study endpoint directly from images. HNN and image CNN were compared with a linear clinical model using area under the curve (AUC), F1 scores, and McNemar's test. Total of 1,286 cases were identified, with 201 death events (16%). The clinical model had good performance (AUC =80%, F1 score =37%). Best Image CNN model showed AUC =72% and F1 score =38%. HNN outperformed the other two models (AUC =82%, F1 score =43%). McNemar's test showed statistical difference between image CNN and both clinical model (P<0.01) and HNN (P<0.01). There was no significant difference between HNN and clinical model (P=0.15). Death in patients with suspected or known CAD can be predicted directly from stress perfusion images without clinical knowledge. Prediction can be improved by HNN that combines clinical and SP-CMR images.

Coronary microvascular dysfunction: prevalence and aetiology in patients with suspected myocardial ischaemia

To evaluate the prevalence, aetiology, and corresponding morbidity of coronary microvascular dysfunction (CMD) in patients with suspected myocardial ischaemia. The present study included 115 patients with suspected myocardial ischaemia who underwent stress perfusion cardiac magnetic resonance imaging. CMD was assessed visually based on the myocardial perfusion results. The CMR-derived myocardial perfusion reserve index (MPRI) and left ventricular (LV) strain parameters obtained using the post-processing software CVI42 were employed to evaluate LV myocardial perfusion and deformation. LV strain parameters included global longitudinal, circumferential, and radial strain (GLS, GCS, and GRS), global systolic/diastolic longitudinal, circumferential, and radial strain rates (SLSR, SCSR, SRSR, DLSR, DCSR, and DRSR). Of the 115 patients, 12 patients were excluded and 103 patients were finally included in the study. CMD was observed in 79 % (81 patients, aged 53±12 years) of patients. Regarding aetiology, 91 (88 %) patients had non-obstructive coronary artery disease (CAD), eight (8 %) had obstructive CAD, and four (4 %) had hypertrophic cardiomyopathy (HCM). The incidence of CMD was highest (100 %) in patients with HCM, followed by those with non-obstructive CAD (up to 79 %). There were no statistical differences between CMD and non-CMD groups in GCS, GRS, GLS, SRSR, SCSR, SLSR, DCSR, DRSR and DLSR. The incidence of CMD was higher in patients with signs and symptoms of ischaemia. CMD occurred with non-obstructive CAD, obstructive CAD, and HCM, with the highest prevalence of CMD in HCM.

High-resolution quantification of stress perfusion defects by cardiac magnetic resonance.

Quantitative stress perfusion cardiac magnetic resonance (CMR) is becoming more widely available, but it is still unclear how to integrate this information into clinical decision-making. Typically, pixel-wise perfusion maps are generated, but diagnostic and prognostic studies have summarized perfusion as just one value per patient or in 16 myocardial segments. In this study, the reporting of quantitative perfusion maps is extended from the standard 16 segments to a high-resolution bullseye. Cut-off thresholds are established for the high-resolution bullseye, and the identified perfusion defects are compared with visual assessment. Thirty-four patients with known or suspected coronary artery disease were retrospectively analysed. Visual perfusion defects were contoured on the CMR images and pixel-wise quantitative perfusion maps were generated. Cut-off values were established on the high-resolution bullseye consisting of 1800 points and compared with the per-segment, per-coronary, and per-patient resolution thresholds. Quantitative stress perfusion was significantly lower in visually abnormal pixels, 1.11 (0.75-1.57) vs. 2.35 (1.82-2.9) mL/min/g (Mann-Whitney U test P < 0.001), with an optimal cut-off of 1.72 mL/min/g. This was lower than the segment-wise optimal threshold of 1.92 mL/min/g. The Bland-Altman analysis showed that visual assessment underestimated large perfusion defects compared with the quantification with good agreement for smaller defect burdens. A Dice overlap of 0.68 (0.57-0.78) was found. This study introduces a high-resolution bullseye consisting of 1800 points, rather than 16, per patient for reporting quantitative stress perfusion, which may improve sensitivity. Using this representation, the threshold required to identify areas of reduced perfusion is lower than for segmental analysis.

Prognostic value of stress perfusion cardiac magnetic resonance in patients with prediabetes and suspected coronary artery disease.

Stress perfusion cardiac magnetic resonance (CMR) is an accurate and comprehensive modality for evaluating patients with suspected coronary artery disease (CAD), but its prognostic value in prediabetic patients is uncertain. This retrospective study included 452 consecutive prediabetic patients without prior diagnoses of CAD who underwent adenosine stress perfusion CMR. The primary endpoint was major adverse cardiovascular events (MACE), defined as cardiovascular death, nonfatal myocardial infarction (MI), hospitalization for heart failure, ischemic stroke, and late coronary revascularization (>90 days post-CMR). The secondary endpoint was a composite of cardiovascular death, nonfatal MI, and hospitalization for heart failure. The mean age was 68±11 years (49% male). Over a median follow-up time of 8.1 (IQR 5.7, 10.4) years, 55 patients experienced MACE, and 24 met the secondary endpoint. Patients with inducible ischemia had significantly greater annualized event rates for MACE (5.7% vs. 0.7%, p<0.001) and for the secondary endpoint (2.0% vs. 0.3%, p<0.001) than those without ischemia. Multivariable analysis revealed inducible ischemia as a consistent predictor for MACE (HR 3.36, 95%CI 1.90-5.94, p<0.001) and for the secondary endpoint (HR 2.89, 95%CI 1.22-6.80, p = 0.01). Late gadolinium enhancement (LGE) was an independent predictor of the secondary endpoint (HR 3.56, 95%CI 1.25-10.13; p = 0.02). Incorporating inducible ischemia and LGE data significantly improved the model's ability to discriminate MACE risk (C-statistic increase from 0.77 to 0.83; net reclassification improvement 0.42; integrated discrimination improvement 0.05). Stress perfusion CMR offers substantial independent prognostic value and effectively aids in reclassifying cardiovascular risk among prediabetic patients with suspected CAD.

Robust Myocardial Perfusion MRI Quantification with DeepFermi.

Stress perfusion cardiac magnetic resonance is an important technique for examining and assessing the blood supply of the myocardium. Currently, the majority of clinical perfusion scans are evaluated based on visual assessment by experienced clinicians. This makes the process subjective, and to this end, quantitative methods have been proposed to offer a more user-independent assessment of perfusion. These methods, however, rely on time-consuming deconvolution analysis and are susceptible to data outliers caused by artifacts due to cardiac or respiratory motion. In our work, we introduce a novel deep-learning method that integrates the commonly used Fermi function with a neural network architecture for fast, accurate, and robust myocardial perfusion quantification. This approach employs the Fermi model to ensure that the perfusion maps are consistent with measured data, while also utilizing a prior based on a 3D convolutional neural network to generalize spatio-temporal information across different patient data. Our network is trained within a self-supervised learning framework, which circumvents the need for ground-truth perfusion labels that are challenging to obtain. Furthermore, we extended this training methodology by adopting a technique that ensures estimations are resistant to data outliers, thereby improving robustness against motion artifacts. Our simulation experiments demonstrated an overall improvement in the accuracy and robustness of perfusion parameter estimation, consistently outperforming traditional deconvolution analysis algorithms across varying Signal-to-Noise Ratio scenarios in the presence of data outliers. For the in vivo studies, our method generated robust perfusion estimates that aligned with clinical diagnoses, while being approximately five times faster than conventional algorithms.

Enhanced External Counterpulsation for Management of Postacute Sequelae of SARS-CoV-2 Associated Microvascular Angina and Fatigue: An Interventional Pilot Study.

Postacute sequelae of SARS-CoV-2 infection (PASC) are a novel clinical syndrome characterized in part by endothelial dysfunction. Enhanced external counterpulsation (EECP) produces pulsatile shear stress, which has been associated with improvements in systemic endothelial function. To explore the effects of EECP on symptom burden, physical capacity, mental health, and health-related quality of life (HRQoL) in patients with PASC-associated angina and microvascular dysfunction (MVD). An interventional pilot study was performed, including 10 patients (male = 5, mean age 50.3 years) recruited from a tertiary specialized PASC clinic. Patients with angina and MVD, defined as index of microcirculatory resistance (IMR) ≥25 and/or diagnosed through stress perfusion cardiac magnetic resonance imaging, were included. Patients underwent one modified EECP course (15 one-hour sessions over five weeks). Symptom burden, six-minute walk test, and validated generic self-reported instruments for measuring psychological distress and HRQoL were assessed before and one month after treatment. At baseline, most commonly reported PASC symptoms were angina (100%), fatigue (80%), and dyspnea (80%). Other symptoms included palpitations (50%), concentration impairment (50%), muscle pain (30%), and brain fog (30%). Mean IMR was 63.6. After EECP, 6MWD increased (mean 29.5 m, median 21 m) and angina and fatigue improved. Mean depression scores showed reduced symptoms (-0.8). Mean HRQoL scores improved in seven out of eight subscales (+0.2 to 10.5). Patients with PASC-associated angina and evidence of MVD experienced subjective and objective benefits from EECP. The treatment was well-tolerated. These findings warrant controlled studies in a larger cohort.

Machine learning outcome prediction using stress perfusion cardiac magnetic resonance reports and natural language processing of electronic health records

Machine learning outcome prediction using stress perfusion cardiac magnetic resonance reports and natural language processing of electronic health records

Hybrid Artificial Intelligence Outcome Prediction Using Feature Extraction from Stress Perfusion Cardiac Magnetic Resonance Images and Electronic Health Records

Abstract Background Prediction of mortality and cardiovascular (CVS) events in coronary artery disease (CAD) has been conventionally achieved using clinical risk factors. Conventional qualitative visual assessment and quantitative analysis of stress perfusion cardiac magnetic resonance (SP-CMR) has previously been shown to have independent prognostic value (1), but the relationship between hidden or recognised imaging features and mortality outcome is still not well understood. Purpose To use deep learning with hybrid neural networks to combine image pixel data and clinical information for outcome prediction. Methods A retrospective study was performed on patients who had SP-CMR between 2011 and 2021. Ethical approval was obtained under the Research Ethics Committee in compliance with the Declaration of Helsinki. Clinical characteristics and CVS risk factors were extracted by using CogStack (2), a natural language processing (NLP) application to extract and analyse clinical data. SP-CMR images were extracted from patients who had mortality events, and from equally random sample without the events for class balance. Data were split into 70% for training, 15% for validation, and 15% for testing. A hybrid neural network (HNN), which combined convolutional neural network (CNN) and multilayer perceptron (MLP), was developed to predict the mortality outcome. The CNN and MLP were trained to extract image features and clinical data, respectively, followed by prediction from concatenated features. This prediction was compared with two models, one CNN model trained on images only and another linear model trained on clinical data including CMR reports findings. DeLong test was used to compare AUC and F1 scores. All deep learning techniques was performed using Python 3.7 and Tensorflow library. Results A total of 4,188 patients were identified, a subset of 500 patients were included in training the networks, including 220 patients with positive mortality events (44%). There were more males than females (73% vs 27%), approximately 36% of patients had history of myocardial infarction (MI), and there was a significant number of patients who had positive ischaemia on SP-CMR (40%), and/or evidence of ischaemic scar (48%). Mean left ventricular ejection fraction (LVEF) was 51 ± 16 %. Area under the curve (AUC) was found to be higher in HNN (80%) followed by clinical data prediction model (78%), and superior to image CNN (63%). There was no significant AUC difference between HNN and clinical model (p=0.072). However, the F1 score showed significant improvement in precision in HNN model (72%) followed by image CNN model (52%), overtaking clinical data prediction model (21%). There was significant F1 score difference between HNN and clinical model (p&amp;lt;0.001). Conclusion Death in patients with suspected or known CAD can be better predicted by a hybrid neural network that combines clinical and SP-CMR image data than by using clinical data alone.Full pipeline for hybrid neural networkAUC and F1 Score comparisons

The left main coronary artery (LMCA) physiology study: comprehensive clinical haemodynamic evaluation of physiology and pathophysiology

Abstract Background &amp; Purpose At present, physiological evaluation of left main coronary artery (LMCA) disease is based on the assumption that pressure-based indices are agnostic to the vessel being assessed. We conducted mechanistic studies to determine whether pressure-based indices in the left anterior descending (LAD) and left circumflex (LCx) arteries are different and if so, the basis of this discrepancy, in patients with normal coronary arteries (LMCA Physiology Study) and in those with isolated LMCA disease (LMCA Pathophysiology Study). Methods Patients with exertional angina were prospectively enrolled after baseline investigation of computed tomography or invasive angiography and those with left ventricular impairment were excluded. Eligibility for the Physiology Study required the absence of epicardial disease on angiography and/or fractional flow reserve or significant coronary microvascular disease (LAD coronary flow reserve (CFR) &amp;lt;2.0). Simultaneous intracoronary pressure and Doppler measurements were performed in the LAD and LCx arteries, at rest and adenosine mediated hyperaemia. A subset also underwent 3-Tesla quantitative stress perfusion cardiac magnetic resonance imaging. Eligibility for the Pathophysiology Study required isolated LMCA disease, which was verified physiologically by pressure wire pullback according to predefined criteria of 1) ≥80% of total FFR arising from the left main coronary artery disease and 2) ≤0.05 FFR gradient distal to the left main coronary artery disease in both vessels. A paired 2-tailed test for significance was performed for all analyses. The protocols were approved by the UK National research ethics committee. Results 71 patients were prospectively enrolled, 28 in the Physiology Study and 43 in the Pathophysiology Study (Table 1). In the LMCA Physiology Study, CFR was higher and hyperaemic microvascular resistance (hMR) lower in the LAD compared to the LCx. In the LMCA Pathophysiology Study, the contribution to total FFR of downstream disease was negligible and comparable in both branches. Moreover, hyperaemic, and non-hyperaemic pressure-based indices were all significantly lower in the LAD, compared to the LCx. In addition, CFR was higher and hMR lower in the LAD compared to the LCx. Correlations are outlined in Figure 1. Differences in pressure-derived measurements remained after correction for the influence of hydrostatic pressures. However, when adjusted for subtended myocardial mass, there was no significant difference in invasive hemodynamic indices or quantitative perfusion analysis-derived myocardial perfusion reserve index. Conclusions Commonly used indices of stenosis severity are not agnostic to the vessel assessed due to differences in myocardial mass and microvascular resistances. The need for vessel-specific thresholds has significant implications on LMCA disease assessment and potentially, interpretation of the existing evidence base on physiological assessment of non-LMCA disease.Table 1Figure 1

Abstract 15210: Qualitative Stress Perfusion American Heart Association Plot and Outcome Prediction Using Artificial Intelligence

Introduction: Stress perfusion cardiac magnetic resonance (CMR) is a guidelines-backed non-invasive test for the assessment of coronary artery disease (CAD), and its prognostic value is well validated. However, image interpretation requires a high level of expertise. Furthermore, the direct relationship between image pixels and outcome is not well understood. Hypothesis: Qualitative stress perfusion map based on American Heart Association (AHA) segmentation and direct linkage with outcome can be achieved using artificial intelligence. Methods: Retrospective CMR data collection was performed in one center between 2011 and 2021. Cases with death events were identified. Mean time to event was considered as a follow-up period, and cases without events and with shorter duration from CMR to collection date were excluded. Stress perfusion images were selected from 3 left ventricular (LV) levels: basal, mid and apical. Clinical CMR reports with binary AHA perfusion values were available. Six convolutional neural networks (CNNs) were trained to predict AHA segments values for each level and merged into four CNNs for the apical level. Perfusion, scar image pixels and clinical data were transformed into features, and a hybrid CNN (HNN) was trained to predict mortality. Training, validation, and test split was 70%, 15%, and 15%, respectively. Evaluation metrics included area under curve (AUC). Results: Total number of cases included for AHA plot classification were 2139 (6417 perfusion images). The average AUC for AHA perfusion classifiers was 61%, an example with correct predictions is presented. Follow-up period was calculated at 999 days. Total number of cases included for HNN development was 1294. HNN achieved high prediction value with AUC of 81%. Conclusions: Qualitative AHA stress perfusion map was feasible using artificial intelligence. Outcome prediction with HNN was a powerful new approach for risk stratifying patients with known or suspected CAD.