Victoria Stoll

Dilated Cardiomyopathy: Phosphorus 31 MR Spectroscopy at 7 T

Cardiac phosphorus spectroscopy is demonstrated to be feasible in patients at 7 T, giving higher signal-to-noise ratios and more precise quantification of the phosphocreatine to adenosine triphosphate concentration ratio than at 3 T in a group of 25 patients with dilated cardiomyopathy.

The kinetic energies of left ventricular 4D flow components correlate with established markers of prognosis and represent novel imaging biomarkers in both ischaemic and dilated cardiomyopathy

Background Despite different aetiology of myocardial damage in dilated cardiomyopathy (DCM) and ischaemic cardiomyopathy (IHD), cardiac remodelling occurs in both, culminating in the end result of a dilated left ventricle with impaired function. Cardiac remodelling is a complex process in which numerous cellular, mechanical and flow processes become deranged. Insights into changes of multidimensional flow patterns and kinetic energy (KE) within the left ventricle are now afforded by the use of 4D flow CMR. We hypothesised that greater derangements in 4D flow measures would relate to: 1) decreased mechanical cardiac function and dilatation, as assessed by LV ejection fraction (LVEF), LV myocardial strain and LV volumes, 2) increased levels of biochemical remodelling markers and 3) worsening patient symptoms and functional capacity. We hypothesised these changes to be independent of the initial aetiology of the myocardial damage, instead reflecting the self-propagating nature of cardiac remodelling.

7T versus 3T phosphorous magnetic resonance spectroscopy in patients with dilated cardiomyopathy

Background Phosphorous magnetic resonance spectroscopy (PMRS) allows insight into cardiac energetics in vivo, but it is a technique with an intrinsically low signal to noise ratio (SNR). The 2.4x increased SNR, that is predicted by theory at 7T compared to 3T, should allow detection of smaller changes in metabolite concentrations or measurement of changes in smaller patient groups, which will further our understanding of cardiac energetics. This study represents the first cardiac P-MRS patient data acquired at a 7 Tesla field strength.

Reproducibility and variability of left ventricular 4D flow in healthy volunteers

Background Blood flow through the heart is a fundamental aspect of the function of the cardiovascular system. Left ventricular intra-cardiac flow, as assessed by retrospectively gated 4D flow, can be divided into 4 functional components; direct flow (DF), delayed ejection flow (DEF), retained inflow (RI) and residual volume (RV). Additionally the kinetic energy of these flow components can be calculated throughout the cardiac cycle. Previous studies have demonstrated differences in the proportions and kinetic energy of flow components between healthy volunteers and patients with dilated cardiomyopathy. This study aims to assess the inter-scan reproducibility and variability of the LV 4D flow components in healthy volunteers.

112 Evaluation of patients with left ventricular thrombus using intra-cardiac blood visualisation with 4d flow

Introduction Ventricular thrombus is a serious complication in a subgroup of left ventricular dysfunction (LVD) patients because of the risk of embolisation. Thrombus occurs as a consequence of stasis, hypercoagulability and endothelial injury. There are no reliable predictors for which patients will develop thrombus. 4D flow CMR may allow insights into thrombus formation by intra-cardiac blood flow visualisation. We hypothesise that in patients with LV dysfunction and thrombus, compared to those without thrombus, the residual volume would constitute a similar proportion of the LV end diastolic volume (EDV) but possess less kinetic energy, thereby predisposing the blood to stasis and therefore thrombus formation. Methods 100 participants (47 LV dysfunction but no thrombus (LVD) patients, 17 LV dysfunction and thrombus patients and 36 controls underwent CMR (Table 1)). LV flow was analysed as 4 components; direct flow, retained inflow, delayed ejection flow and residual volume. Each components volume was calculated in proportion to the EDV. The kinetic energy of the blood per millilitre was summed throughout the cardiac cycle and divided by the cycle length to calculate the average kinetic power. 25 controls, 14 LVD and 14 thrombus patients returned for an interval scan to assess the stability of flow parameters. Results Both patient groups had significantly increased residual volume (LVD 50±10%, thrombus 51±12% vs 30±4% controls, p 0.001) and decreased direct flow (LVD 11±7%, thrombus 16±10% vs 38±4% controls, p 0.001). There was no difference between the 2 patient groups (Fig 1A). The average kinetic power of the residual volume was significantly higher in the LVD group (0.55±0.30 microJ/ml) compared to the thrombus group (0.38±0.16 microJ/ml, p 0.02) (Fig 1B). No difference between patient groups was seen for the direct flow average kinetic power (Fig 1C). 4D flow parameters were similar between visits with no significant change on paired t-tests (Table 2). The average kinetic power of the residual volume was higher in the LV dysfunction than thrombus group at visit 1 and 2, but failed to reach statistical significance with the smaller cohorts. Discussion The residual volume blood of thrombus patients possessed less kinetic power than that of LV dysfunction patients with a well matched LV size, impairment and proportion of residual volume. Residual volume blood resides within the ventricle for at least two cardiac cycles; reduced movement of this blood component may be a contributing factor to stasis and hence thrombus formation. Similar results at interval studies propose that the residual volume average kinetic power is a temporally stable parameter. This study suggests that the average kinetic power of the residual volume is a novel imaging biomarker which may allow identification, monitoring and potentially aid anticoagulation decisions in patients with LV dysfunction at higher risk of thrombus formation.Abstract 112 Table 1 Clinical characteristics of controls, LV dysfunction and thrombus patients.Abstract 112 Table 2 Results for conventional and 4D flow cardic remodelling parameters for participants who attended two study visits. There were no significant differences between visit 1 and visit 2 when paired t-tests were performed for each participant for the parameters shown above.Abstract 112 Figure 1 A) The volume of the four flow components by mean percentage ± SD in relation to the end diastolic volume. There is increased residual volume and decreased direct flow in the LV dysfunction and thrombus groups compared to controls. B) The residual volume summed kinetic energy per millilitre, normalised to duration of heart cycle for each group. The LV thrombus group have significantly lower residual volume summed kinetic energy, compared to the LV dysfunction group. C) Direct flow summed kinetic energy; no difference is seen between the LV dysfunction and thrombus groups.Abstract 112 Figure 2 ROC analysis of MPR assessed using CMR for detecting microvascular ischaemia as defined by high IMR (> 40) in the absence of significant epicardial stenosis (FFR >0.8). True positives: high IMR>40; true negatives: normal IMR<20. Area under the curve 0.87,±0.06.

83 Use of feature tracking to assess systemic right ventricles in congenital heart disease patients with both single and dual ventricular circulations

Introduction The management of congenital patient’s frequently utilises cardiac magnetic resonance imaging to assess changes in patients cardiac function. Ventricular function assessment is challenging in this cohort due to complex ventricular geometry. Feature tracking uses routinely acquired MRI images to assess ventricular strain, a measure of cardiac contractility, which has been found in non-congenital cohorts to be a more sensitive marker of ventricular dysfunction than conventional imaging parameters. We hypothesised that 1) patients with a systemic ventricle of right ventricular (RV) morphology would have impaired strain parameters compared to controls and 2) patients with a systemic RV in a single ventricle circulation would have more impaired contractility than those with a dual ventricular circulation. Methods 3 groups were analysed: 1) 16 patients with hypoplastic left heart syndrome (HLHS) resulting in a single ventricle circulation (mean age 20±2 years; 81% male) 2) 16 patients with a systemic RV (sRV) in a dual ventricle circulation (age 32±5 years; 63% male) and 3) 16 healthy controls (mean age 30±4 years; 56% male). Participants underwent CMR at 1.5T for ventricular function assessment, analysis was undertaken using Circle cvi42 (v5.3) to calculate RV mid ventricular circumferential strain and peak longitudinal strain. Results The two patient groups had increased RV volumes with reduced ejection fraction (EF), elevated RV mass, but similar stroke volumes compared to controls (Table 1). The peak longitudinal RV strain was significantly reduced between the patient groups (HLHS mean −13±4; sRV −12±3) compared to controls (−18±5, p<0.001). However, there was no difference between the HLHS and sRV groups. Mid circumferential RV strain was not significantly different between both patient groups and also compared to controls (HLHS −13±4, sRV −12±4, controls −13±3, p 0.6). However comparison of mid circumferential RV strain to controls LV strain showed a significant reduction (controls LV mid circumferential strain mean −16±3, p 0.004) (Figure 1). The peak longitudinal strain in patients correlated with the RV end diastolic volume index (r=0.38, p 0.03), RV end systolic volume index (r=0.52, p 0.002), RV EF (r=0.535, p?0.002) and RV mass index (r=0.43, p 0.01). The mid RV circumferential strain only correlated with RV EF (r=0.40, p 0.02) and no other remodelling parameters. Discussion RV longitudinal strain was reduced in patients with a systemic RV irrespective of whether the ventricular configuration was single or dual. No difference was seen between the single circulation systemic RV compared to the dual circulation systemic RV, suggesting that RV remodelling is mostly in response to the systemic position rather than the ventricular configuration. Longitudinal studies will be required to assess the utility of the longitudinal strain in the prediction of outcomes in the follow up of these patients.Abstract 83 Table 1 Conventional right ventricular imaging parameters for the three study groupsAbstract 83 Figure 1 A) Mean longitudinal strain for the RV of patients with a single ventricle (HLHS), a systemic RV in a dual ventricular system (systemic RV) and the RV of healthy controls. B) Mean mid ventricular circumferential strain for the RV of patients with a single ventricle (HLHS), a systemic RV in a dual ventricular system (systemic RV) and RV and LV strain for healthy controls.

In the presence of a patent foramen ovale paroxysmal embolism risk increases with non-vortical right atrial blood flow

Background Despite the fact that stroke is a leading cause of disability, in up to 40% of cases no cause is found on routine clinical investigation. Although a patent foramen ovale (PFO) is an attractive mechanism to explain these cryptogenic strokes, using current imaging techniques, distinguishing between a causative rather than an incidental PFO remains elusive. We hypothesised that, in the presence of a PFO, non-vortical right atrial (RA) flow patterns would be linked to embolism risk by making it more likely to shunt blood, and as such thrombus through the PFO. In order to investigate this we assessed RA flow patterns and interatrial shunt size in patients with a PFO and investigated whether these metrics predicted the incidence of paradoxical embolism.

14 Peak turbulent kinetic energy assessed by cardiac magnetic resonance correlates better than aortic valve area with left ventricular parameters in aortic stenosis

Introduction Previous studies have shown only modest correlation between aortic valve area (AVA) and left ventricular mass in aortic stenosis (AS), likely due to the variable additional after-load from the aortic and systemic circulation. Cardiac magnetic resonance (CMR) has developed sequences to measure turbulent kinetic energy (TKE) in the proximal aorta, derived from time resolved three-dimensional (4D) flow imaging. It represents energy dissipation and irreversible pressure loss. Work in the same patient cohort demonstrated peak TKE is higher in bicuspid than tricuspid AS, probably due to larger ascending aortas. We assessed peak systolic TKE in patients with wide ranging AS (mild to severe) and correlated it with left ventricular (LV) parameters. Methods 22 patients with tricuspid AS (mean age 72.1 ± 8.6 years, mean indexed valve area (AVA) 0.55 ± 0.18cm2/m2) and 20 with bicuspid AS (age 65.2 ± 8.5, indexed AVA 0.77 ± 0.42 cm2/m2) were scanned. Peak systolic TKE was measured using specialist analysis software and systolic LV longitudinal strain with proprietary feature-tracking software. Results LV mass index was the only measure of LV function that correlated with AVA: inverse correlation (r) = −0.404; p 0.006. Peak TKE had a stronger correlation with LV mass (r 0.53; p 0.001), especially in bicuspid AS (r 0.83; p 0.0001). Peak TKE also correlated with indexed LV end diastolic volume (r 0.486; p 0.003), stroke volume (r 0.564; p 0.0004) and stroke work (r 0.58; p 0.0003). This correlation was especially strong in BAV AS (Table 1). No correlation with LV ejection fraction or systolic longitudinal strain was found. Abstract 14 Table 1 LV parameter correlates with peak systolic TKE Correlations with TKE AS TAV AS BAV AS r p r p r p LV mass index 0.530 0.001 0.203 0.391 0.830 0.0001 LVEDVi 0.486 0.003 −0.126 0.5976 0.688 0.005 SVi 0.564 0.0004 −.044 0.853 0.784 0.001 LVEF −0.087 0.621 0.072 0.762 0.026 0.928 Stroke work index 0.58 0.0003 0.21 0.373 0.833 0.0002 Conclusions Peak TKE is a promising tool to better understand the variable effects of AS on the LV, beyond simply severity of AS. The effect is particularly strong in bicuspid AS, who have larger aortas - the likely mechanism for increased TKE.

PARADOXICAL EMBOLISM RISK INCREASES WITH ATYPICAL RIGHT ATRIAL BLOOD FLOW IN THE PRESENCE OF A PATENT FORAMEN OVALE

Background Stroke is a leading cause of disability, but in up to 40% of cases no cause is found. Although a patent foramen ovale (PFO) is an attractive mechanism to explain these cryptogenic strokes, using current imaging techniques, distinguishing between a causative rather than an incidental PFO remains elusive. We hypothesised that, in the presence of a PFO, atypical right atrial (RA) flow patterns would be linked to embolism risk by increasing the shunt of blood, and as such thrombus through the PFO. In order to investigate this we assessed RA flow patterns and interatrial shunt size in patients with a PFO and investigated whether these metrics predicted the incidence of paradoxical embolism. Methods 3 groups were recruited; 1) participants with a PFO but no embolism (n = 12), 2) patients with presumed paradoxical embolism (n = 20) and 3) participants without a PFO (n = 28). All underwent RA 4D flow assessment, and bubble transthoracic echocardiography to determine interatrial shunt size. Atypical RA flow was defined as any flow pattern that was not a classical vortex. Results Flow Patterns RA flow patterns were similar between the 2 groups with no embolic event irrespective of the presence of a PFO. In the PFO with embolism group, they were significantly different with a higher incidence of atypical RA flow (P = 0.0067, Figure 1B). Risk of embolism When considering all the subjects with a PFO (n = 32), the presence of an atypical flow pattern was 11.5 times more common in those who have had an embolic event (P = 0.002, Fisher’s exact test). To explore whether this effect was mediated by changing the degree of shunting, moderated multiple regression was performed. This showed that flow patterns were related to shunt grade (a pathway, β 34.0, p < 0.01), and that shunt grade was related to embolism incidence (b pathway, β 0.08, p < 0.01). As the a and b pathways were significant, mediation analysis was tested using 2,000 bootstrap resamples to generate a 95% confidence interval (bias corrected) of the indirect effect. This showed that the effect of an atypical flow pattern upon an embolic event is indeed mediated by increasing the shunt across the PFO (CI 0.45–18.42, Figure 2). As the direct effect of flow patterns on embolic risk becomes insignificant (c1 pathway, p = 0.06) this suggests full mediation. Conclusions Patients with a PFO and atypical RA flow pattern were 11.5 times more likely to have had an embolic event. This increased embolic risk seems to be mediated via increasing the shunt size across the PFO. As a result, not only will identification of the presence or absence of RA classical vortical flow in individuals presenting with a cryptogenic stroke help distinguish a causative PFO, but it may also identify patients with a PFO who are at elevated risk of future embolism.Abstract 137 Figure 1 (A)- i) Vortical flow: the IVC and SVC turn in a clockwise vortex. All other flow patterns were classified as atypical flow patterns: ii) Spiral-vortical flow: the IVC forms a vortex whilst the SVC passes laterally and is then enveloped in a spiral fashion by the IVC, iii) Spiral flow: where the IVC and SVC combine in a spiral, iv) Complex flow: involving multiple vortices arising from the IVC and SVC flow. (B)-Flow pattern frequencies in controls with and without a PFO versus patientsAbstract 137 Figure 2 Mechanistic model showing the contribution of an atypical RA blood flow pattern upon the PFO shunt severity and embolic event occurrence

YI-1 Changes in ascending aortic flow pattern after bicuspid aortic valve replacement differ with prosthesis type

Background Abnormal aortic flow patterns in bicuspid aortic valve disease (BAV) may be partly responsible for the associated aortic dilation. Aortic valve replacement (AVR) may therefore slow the growth of concomitant aortic dilation via normalisation of flow patterns. Methods and Results 90 participants underwent 4D flow cardiovascular magnetic resonance: 30 BAV patients with prior AVR (11 mechanical, 10 bioprosthetic, 9 Ross procedure), 30 BAV patients with a native aortic valve and 30 healthy volunteers. The majority of subjects with mechanical AVR or Ross showed normal flow pattern (73% and 67% respectively) with near normal rotational flow values (7.2 ± 3.9 and 10.6 ± 10.5 mm2/s respectively vs 3.8 ± 3.1 mm2/s for normal volunteers; both p > 0.05); and reduced in-plane wall shear stress (0.19 ± 0.13 N/m2 for mechanical AVR vs. 0.40 ± 0.28 N/m2 for native BAV, p < 0.05). By contrast, all subjects with bioprosthetic AVR had abnormal flow patterns (mainly marked right-handed helical flow), with similar rotational flow values to native BAV (20.7 ± 8.8mm2/s and 26.6 ± 16.6 mm2/s respectively, p > 0.05). Wall shear stress post-bioprosthetic AVR showed a similar pattern. Abstract YI-1 Figure 1 Ascending aortic flow patterns (A) healthy volunteer with a laminar flow pattern (B) native bicuspid aortic valve disease with a right-handed helical flow pattern (C) AVR mechanical with 2 laminar jets (D) AVR tissue with a right handed helical flow pattern (E) AVR-Ross with a laminary flow pattern. Data before and after valve replacement (n = 16) supported these findings: mechanical AVR showed a significant reduction in rotational flow (30.4 ± 16.3 → 7.3 ± 4.1 mm2/s; p < 0.05) and in-plane wall shear stress (0.47 ± 0.20 → 0.20 ± 0.13 N/m2; p < 0.05), whereas these remained unchanged in the bioprosthetic AVR group. Conclusion Abnormal flow patterns in bicuspid aortic valve disease are significantly reduced after mechanical AVR or Ross procedure, though remain similar after bioprosthetic AVR. The type of valve replacement may thus influence post-operative flow patterns, and could have important implications for future aortic growth.