Novel dark-blood versus conventional bright-blood late gadolinium enhancement CMR: A pilot study comparing impact on myocardial ischaemic burden

Funding Acknowledgements Type of funding sources: Foundation. Main funding source(s): British Heart Foundation Background In patients with coronary artery disease (CAD), increasing myocardial ischaemic burden (MIB) is a strong predictor of adverse events. When measured by cardiovascular magnetic resonance (CMR), a MIB ≥12.5% is considered significant and often used as a threshold to guide revascularisation. Ischaemic scar can cause stress perfusion defects which do not represent ischaemia and should be excluded from the MIB calculation. Conventional bright-blood late gadolinium enhancement (LGE) is able to identify ischaemic scar but can suffer from poor scar-to-blood contrast, making accurate assessment of scar volume difficult. Dark-blood LGE methods increase scar-to-blood contrast and improve scar conspicuity which may impact the calculated scar burden and consequently the estimation of MIB when read in conjunction with perfusion images. Purpose To evaluate the impact of dark-blood LGE versus conventional bright-blood LGE on the estimation of MIB in patients with CAD. Methods 37 patients with suspected or known CAD who had evidence of CMR stress perfusion defects and ischaemic scar on LGE imaging were recruited. Patients underwent adenosine stress perfusion imaging followed by dark-blood LGE then conventional bright-blood LGE imaging at 3T. For dark-blood LGE, phase sensitive inversion recovery imaging with a shorter inversion time to null the LV blood-pool was used without any additional magnetization preparation. For each patient, three short-axis LGE slices were selected to match the three perfusion slice locations. Images were anonymised and analysed in random order. Ischaemic scar burden (ISB) was quantified for both LGE methods using a threshold >5 standard deviations above remote myocardium. Perfusion defect burden (PDB) was quantified by manual contouring of perfusion defects. MIB was calculated by subtracting the ISB from the PDB. Results MIB calculated using dark-blood LGE was 19% less compared to bright-blood LGE (15.7 ± 15.2% vs 19.4 ± 15.2%, p < 0.001). There was a strong positive correlation between the two LGE methods (rs = 0.960, p < 0.001, Figure 1A). Bland-Altman analysis revealed a significant fixed bias (mean bias = -3.6%, bias 95% CI: -2.6 to -4.7%, 95% limits of agreement: -9.8 to 2.5%) with no proportional bias (Figure 1B). MIB was calculated ≥12.5% and <12.5% by both LGE methods in 19 (51%) and 12 (32%) patients respectively. In 6 patients (16%), MIB was ≥12.5% using bright-blood LGE and <12.5% using dark-blood LGE (Figure 1A – orange data points). Overall, when used to classify MIB as <12.5% or ≥12.5%, there was only substantial agreement between the two LGE methods (κ=0.67, 95% CI: 0.45 to 0.90). Conclusions The use of dark-blood LGE in conjunction with perfusion imaging results in a lower estimate of MIB compared to conventional bright-blood LGE. This can cause disagreement around the threshold of clinically significant ischaemia which could impact clinical management in patients being considered for coronary revascularisation. Abstract Figure. Linear regression with corresponding B&A

Defining the prognostic value of [15O]H2O positron emission tomography-derived myocardial ischaemic burden

Aims Myocardial ischaemic burden (IB) is used for the risk stratification of patients with coronary artery disease (CAD). This study sought to define a prognostic threshold for quantitative [15O]H2O positron emission tomography (PET)-derived IB. Methods and results A total of 623 patients with suspected or known CAD who underwent [15O]H2O PET perfusion imaging were included. The endpoint was a composite of death and non-fatal myocardial infarction (MI). A hyperaemic myocardial blood flow (hMBF) and myocardial flow reserve (MFR)-derived IB were determined. During a median follow-up time of 6.7 years, 62 patients experienced an endpoint. A hMBF IB of 24% and MFR IB of 28% were identified as prognostic thresholds. Patients with a high hMBF or MFR IB (above threshold) had worse outcome compared to patients with a low hMBF IB [annualized event rates (AER): 2.8% vs. 0.6%, P < 0.001] or low MFR IB [AER: 2.4% vs. 0.6%, P < 0.001]. Patients with a concordant high IB had the worst outcome (AER: 3.1%), whereas patients with a concordant low or discordant IB result had similar and low AERs of 0.5% and 0.9% (P = 0.953), respectively. Both thresholds were of prognostic value beyond clinical characteristics, however, only the hMBF IB threshold remained predictive when adjusted for clinical characteristics and combined use of the hMBF and MFR thresholds. Conclusion A hMBF IB ≥24% was a stronger predictor of adverse outcome than an MFR IB ≥28%. Nevertheless, classifying patients according to concordance of IB result allowed for the identification of low- and high-risk patients.

Artificial intelligence and echocardiography

Echocardiography plays a crucial role in the diagnosis and management of cardiovascular disease. However, interpretation remains largely reliant on the subjective expertise of the operator. As a result inter-operator variability and experience can lead to incorrect diagnoses. Artificial intelligence (AI) technologies provide new possibilities for echocardiography to generate accurate, consistent and automated interpretation of echocardiograms, thus potentially reducing the risk of human error. In this review, we discuss a subfield of AI relevant to image interpretation, called machine learning, and its potential to enhance the diagnostic performance of echocardiography. We discuss recent applications of these methods and future directions for AI-assisted interpretation of echocardiograms. The research suggests it is feasible to apply machine learning models to provide rapid, highly accurate and consistent assessment of echocardiograms, comparable to clinicians. These algorithms are capable of accurately quantifying a wide range of features, such as the severity of valvular heart disease or the ischaemic burden in patients with coronary artery disease. However, the applications and their use are still in their infancy within the field of echocardiography. Research to refine methods and validate their use for automation, quantification and diagnosis are in progress. Widespread adoption of robust AI tools in clinical echocardiography practice should follow and have the potential to deliver significant benefits for patient outcome.