Search

Background Atrial fibrillation (AF) is the most common cardiac arrhythmia. It is often treated using catheter ablation, which aims to isolate arrhythmogenic regions by delivering localized energy. Whereas an insufficient energy delivery can lead to AF recurrence, excessive power can cause potentially lethal complications. Knowledge of atrial wall thickness can help select an optimal amount of energy, but clinical imaging does not currently provide atrial wall thickness data for patients. Even ex-vivo studies have only measured the thickness of the atrial wall in discrete locations using CT or post-mortem samples. In this study, we apply black-blood MR imaging to reconstruct both epiand endocardial surfaces of the entire atria and compute wall thickness maps in healthy volunteers. To our knowledge, this is the first study to provide a complete 3D map of the wall thickness of both right and left atria.

Background Catheter-myocardium contact force (CF) is a major determinant of ablation lesion generation in the atrium, with higher CF associated with greater energy transfer. Previous publications have suggested that acute atrial injury can be visualized with T2-weighted (T2W) cardiac magnetic resonance (CMR). This study therefore sought to compare T2W images of acute ablation lesions created at low and high CF. Methods Under general anesthesia, femoral venous access was obtained in eight male Gottingen minipigs. In two animals, pre-ablation T2W CMR imaging was performed. Using fluoroscopy, a decapolar reference catheter was placed in the coronary sinus and a force-sensing ablation catheter (SmartTouch, D curve, Biosense Webster) was advanced into the right atrium. Using an electroanatomical mapping system (Carto3-MEM) the geometry of the right atrium was obtained with a 20-pole circular mapping catheter (Lasso Nav, D curve, Biosense Webster). A linear ablation lesion was created from the SVC to the IVC (30W, 48°C, 8 ml/min irrigation). Different target contact forces (>20 g (high force) or

Background Late Gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging can be used to visualise regions of fibrosis and scarring in the left atrium (LA) myocardium. This can be important for treatment stratification of patients with atrial fibrillation (AF) and for assessment of treatment after radio frequency catheter ablation (RFCA). In this paper we present a standardised evaluation benchmarking framework for algorithms segmenting fibrosis and scar from LGE CMR images. The algorithms reported are the response to an open challenge that was put to the medical imaging community through an ISBI (IEEE International Symposium on Biomedical Imaging) workshop. Methods The image database consisted of 60 multicenter, multivendor LGE CMR image datasets from patients with AF, with 30 images taken before and 30 after RFCA for the treatment of AF. A reference standard for scar and fibrosis was established by merging manual segmentations from three observers. Furthermore, scar was also quantified using 2, 3 and 4 standard deviations (SD) and full-width-at-half-maximum (FWHM) methods. Seven institutions responded to the challenge: Imperial College (IC), Mevis Fraunhofer (MV), Sunnybrook Health Sciences (SY), Harvard/Boston University (HB), Yale School of Medicine (YL), King’s College London (KCL) and Utah CARMA (UTA, UTB). There were 8 different algorithms evaluated in this study. Results Some algorithms were able to perform significantly better than SD and FWHM methods in both pre- and post-ablation imaging. Segmentation in pre-ablation images was challenging and good correlation with the reference standard was found in post-ablation images. Overlap scores (out of 100) with the reference standard were as follows: Pre: IC = 37, MV = 22, SY = 17, YL = 48, KCL = 30, UTA = 42, UTB = 45; Post: IC = 76, MV = 85, SY = 73, HB = 76, YL = 84, KCL = 78, UTA = 78, UTB = 72. Conclusions The study concludes that currently no algorithm is deemed clearly better than others. There is scope for further algorithmic developments in LA fibrosis and scar quantification from LGE CMR images. Benchmarking of future scar segmentation algorithms is thus important. The proposed benchmarking framework is made available as open-source and new participants can evaluate their algorithms via a web-based interface.

Background T1 imaging based on pixel-wise quantification of longitudinal relaxation has the potential to differentiate between normal and abnormal myocardium. The accuracy of T1 measurement has not been established nor systematically tested in the presence of health and disease. Methods Intra-observer, inter-observer and inter-study reproducibility of T1 imaging was assessed in subjects with left ventricular hypertrophy (LVH, n = 25) or dilated cardiomyopathy (DCM, n = 43). Thirty-eight subjects with low-pretest likelihood of cardiomyopathy served as a control group. T1 values were acquired in a single mid-ventricular short axis slice using modified Look-Locker imaging prior and after the application of gadolinium contrast at 1.5 and 3 T. Analysis was performed with regions of interest (ROI) placed conservatively within the septum or to include the whole short axis (SAX) myocardium. Results Intra-observer, inter-observer and inter-study repeated measurements within the septum showed smaller mean differences and narrower 95% confidence intervals than repeated short axis ROI measurements. Native T1 values were higher in septal ROIs compared with SAX values at both field strengths (1.5 T: 976 ± 37 vs. 952 ± 41, p Conclusion Conservative septal ROI T1 measurement is a robust technique with excellent intra-observer, inter-observer and inter-study reproducibility for native and post-contrast T1 value and partition coefficient measurements. Moreover, native septal T1 values reveal the greatest difference between normal and abnormal myocardium, which is independent of geometrical alterations of cardiac chamber and wall thickness. We propose the use of native T1 measurements using conservative septal technique as the standardized approach to distinguish health from disease assuming diffuse myocardial involvement.

Background A good correlation has been found between the extent of late gadolinium enhancement (LGE) and clinical outcome in patients who have undergone radiofrequency (RF) ablation for atrial fibrillation (AF). However, there can be strong residual blood signal in the traditional inversion-recovery (IR) sequence, which hampers scar visualization and causes poor repeatability of scar size measurements. The dual-IR sequence has previously been shown to improve blood suppression in LGE images of ventricular scar. We aimed to assess whether the superior blood suppression in the dual-IR pre-pulse improved the inter-observer variability of left atrial scar measurements compared with the IR sequence.

Background Radiofrequency ablation causes a combination of irreversible and reversible atrial injury. Previous publications have suggested that this can be visualized with a combination of T2-weighted (T2W) and late gadolinium enhancement (LGE) CMR. However, signal intensity (SI) thresholds have not been validated in the atrium. This study sought to compare T2W and LGE with histological findings in an animal model of atrial ablation.

Background Technical advances in perfusion cardiovascular magnetic resonance (CMR), particularly accelerated data acquisition methods, allow myocardial perfusion imaging with unprecedented spatial resolution. However, it is not clear how implementation of these recent advances affects perfusion image quality, signal and contrast to noise ratios (SNR & CNR) and the occurrence of important artefacts in routine clinical imaging. The objective of this study was therefore to compare a standard and an advanced, high-resolution perfusion sequence. Methods A standard ultrafast gradient echo perfusion sequence (st-GrE) was compared with an advanced kt-accelerated steady state free precession sequence (ktBLAST-SSFP) at 1.5 T in healthy volunteers (n = 16) and in patients (n = 32) with known or suspected coronary artery disease. Volunteers were imaged with both sequences at rest and patients underwent stress and rest imaging with either st-GrE or ktBLAST-SSFP prior to X-ray coronary angiography. A blinded expert scored image quality and respiratory artefact severity and also classified patients for the presence of CAD. The extent, transmurality and duration of dark rim artefacts (DRA) as well as signal to noise (SNR) and contrast to noise (CNR) were quantified. Results In normal hearts ktBLAST-SSFP imaging resulted in significantly improved image quality (p = 0.003), SNR (21.0 ± 6.7 vs. 18.8 ± 6.6; p = 0.009), CNR (15.4 ± 6.1 vs. 14.0 ± 6.0; p = 0.034) and a reduced extent (p = BLAST-SSFP imaging resulted in significantly improved image quality (p = 0.012), and a reduced extent (p = BLAST-SSFP in both patients and volunteers. Conclusions Advanced high resolution perfusion CMR using a k-t-accelerated SSFP technique results in significantly improved image quality, SNR and CNR and a reduction in the extent and transmurality of DRA compared to a standard sequence. These findings support the use of advanced perfusion sequences for clinical perfusion imaging however further studies exploring whether this results in improved diagnostic accuracy are required.

Summary Cardiac magnetic resonance (CMR) is increasingly applied as the first line investigation into the causes of myocardial dysfunction and remodelling. Although regional fibrosis is easily imaged on late gadolinium enhancement, characterisation of myocardial conditions with diffuse fibrosis remains challenging. Several studies proposed T1 mapping of gadolinium-enhanced myocardium supported by an inverse relationship between T1 relaxation times and amount of fibrotic tissue on histology. We performed high-resolution modified LookLocker inversion recovery (hrMOLLI) imaging of native and postcontast myocardium in thirty-six subjects who underwent a routine CMR protocol for investigation of cardiomyopathy.

Summary Several methodologies were proposed to evaluate myocardial T1 values by cardiovascular magnetic resonance based on Gadolinium contrast enhancement. Whereas potentially valuable for evaluation of myocardial fibrosis, none of these methodologies have been sufficiently robust and consistently translated into clinical routine. We propose that imaging with a high-resolution modified Look-Locker inversion recovery sequence (hrMOLLI) of native and post-contrast myocardium can provide a robust tool in the clinical setting. Background Myocardial fibrosis is associated with myocardial dysfunction and remodelling leading to adverse outcomes. We propose that imaging with a high-resolution modified Look-Locker inversion recovery sequence (hrMOLLI) of native and post-contrast myocardium can provide a robust tool in the clinical setting. Methods

Background The variability of pulse wave velocity (PWV) with changing arterial wall stress may reflect the degree of aortic stiffness independent of arterial pressure. We implemented a real-time MR flow technique so as to be able to assess acute changes in PWV due to affected arterial blood and/or intra thoracic pressure. The aim of this study was to assess the accuracy of rapid PWV assessment using the real time MR technique.

We aim to explore tissue heterogeneity assessment using T1 maps generated with the modified Look Locker (MOLLI) sequence in patients with previous myocardial infarct. Conclusion: Differences between healthy myocardium and scarred tissues can be reliably distinguished from the R1 values derived from pre-contrast T1 maps.Potentially, patients without scarred myocardium do not need post-contrast imaging. Inpatients with scarred tissues, ΔR1 value derived from both thepre- and the post-contrast T1 maps provides better distinction between grey zone and scar core than either pre-contrast or post-contrast R1 value alone.

Navigator techniques allow the patient to breath freely during a whole heart coronary magnetic resonance angiography (WHCMRA) scan. However, long measuring times, caused by the necessity to synchronize the cardiac and the breathing cycle, and complex motion patterns lead to the suboptimal image quality. To overcome this problem, an abdominal belt (BELT), which can suppress the abdominal breathing motion and, thus, improve WHCMRA image quality, has been suggested by a Japanese group1. However, its feasibility has never been shown for a Western population.

Background Novel cardiovascular magnetic resonance (CMR) techniques and imaging biomarkers are often validated in small animal models or empirically in patients. Direct translation of small animal CMR protocols to humans is rarely possible, while validation in humans is often difficult, slow and occasionally not possible due to ethical considerations. The aim of this study is to overcome these limitations by introducing an MR-compatible, free beating, blood-perfused, isolated pig heart model for the development of novel CMR methodology. Methods 6 hearts were perfused outside of the MR environment to establish preparation stability. Coronary perfusion pressure (CPP), coronary blood flow (CBF), left ventricular pressure (LVP), arterial blood gas and electrolyte composition were monitored over 4 hours. Further hearts were perfused within 3T (n = 3) and 1.5T (n = 3) clinical MR scanners, and characterised using functional (CINE), perfusion and late gadolinium enhancement (LGE) imaging. Perfusion imaging was performed globally and selectively for the right (RCA) and left coronary artery (LCA). In one heart the RCA perfusion territory was determined and compared to infarct size after coronary occlusion. Results All physiological parameters measured remained stable and within normal ranges. The model proved amenable to CMR at both field strengths using typical clinical acquisitions. There was good agreement between the RCA perfusion territory measured by selective first pass perfusion and LGE after coronary occlusion (37% versus 36% of the LV respectively). Conclusions This flexible model allows imaging of cardiac function in a controllable, beating, human-sized heart using clinical MR systems. It should aid further development, validation and clinical translation of novel CMR methodologies, and imaging sequences.

Methods Ten patients with Type B aortic dissection underwent MRI at 3.0 T (Philips Healthcare) with a 6-element cardiac coil. After an initial survey and reference scan, direct thrombus imaging was performed by an inversion recovery (IR) ECG-triggered respiratory-navigated 3D TFE sequence (FOV: 300 × 255 × 60 mm3 with 2 × 2 × 2.5 mm3 resolution, TI = 490 ms, TFE-factor = 36, TR/TE = 3.2/1 ms). A blood-pool agent (Gadofosveset) (dose 0.12 ml/kg at 4 ml/s) was injected and first-pass imaging performed by 3D CE MRA (FOV = 420 × 280 × 135 mm3, resolution = 1.8 mm3, FA = 35°, TR/TE = 6.0/1.8 ms), breathhold, without cardiac gating. Two-dimensional flow analyses were performed at 4 aortic levels (FOV = 2 × 2 × 10 mm3, FA = 10°, TR/TE = 5.0/2.7 ms, 25 cardiac-phases, VENC = 250 cm/s). For blood-pool imaging, a respiratory-navigated ECG-triggered IR-3D SSFP sequence was used (FOV = 400 × 253 × 156 mm3, resolution = 1.5 mm3, FA = 20°, TI = 350 ms, TR/TE = 4.0/1.3 ms, TFE-factor = 22). The volume of thrombus was extracted from the different datasets (first pass MRA, blood pool MRI and first pass CT) by an expert using manual segmentation (ViewForum, Philips Healthcare). Areas of low signal on blood pool images were correlated with direct thrombus images Figure 1. from 13th Annual SCMR Scientific Sessions Phoenix, AZ, USA. 21-24 January 2010

Introduction Delayed-enhancement (DE) in the left atrium following radiofrequency (RF) ablation has been previously shown [1]. However, current DE techniques do not reach sufficient resolution to determine if ablation lesions are transmural, a factor that can determine the success of an ablation procedure. Myocardial thickness in the left atrium can reach approximately 1 mm [2]. Therefore a delayed enhancement imaging sequence must be sub-millimetre in resolution. For this requirement, we develop a high resolution inversion recovery (IR) sequence with a spiral k-space trajectory. A spiral trajectory is a very efficient sampling scheme that allows for a very short acquisition window (minimising cardiac blurring) at a much reduced cost to overall imaging time compared to Cartesian.

Introduction For cardiac resynchronization therapy (CRT), knowledge of the coronary venous anatomy in relation to areas of myocardial scar is important to plan the optimal LV lead position during the CRT-procedure. Cardiac magnetic resonance imaging (CMR) has previously been used to assess coronary veins without a contrast agent (CA) using MTC-prepulse or injecting an intravascular CA. However, neither of these methods provide information about myocardial scar.

Introduction Although great technical advances have been made in coronary MRA, long scan times have played a part in the slow translation to routine clinical practice. The recent advent of array coil technology with high numbers of elements allows for higher SENSE factors, which lowers scanning time to a more acceptable level [1,2]. The trade-off of higher acceleration is lower SNR, which can be mediated by going to a higher field strength. Previous studies have investigated the use of a 32-channel coil at 1.5 T [1,2], however, only recently has this technology become available at 3 T.

The use of electro-anatomical mapping systems (EAMS) for catheter guidance and ablation point recording is widespread for the treatment of atrial arrhythmias using radio frequency ablation (RFA). Also, the evolution of RFA lesions over time may be an important factor in the reoccurrence of arrhythmias. In this study, we use late enhancement (LE) magnetic resonance imaging (MRI) as a tool to measure RFA lesions and present a novel visualisation method to represent this information in an intuitive way. This allows the validation of the in-vivo accuracy of EAMS and the examination of the evolution of RFA lesions from the acute to the chronic timescales.

Introduction Late presentation of a large left to right shunt or the presence of significant respiratory co-morbidity may be associated with the development of pulmonary vascular disease. In this situation, invasive assessment of the pulmonary vascular resistance (PVR) is required in order to determine operability. Phase contrast MRI (PC-MRI) allows for the non-invasive assessment of arterial flow and left to right shunt. When combined with simultaneous cardiac catheterization it is possible to accurately determine the PVR.

Introduction We have previously developed an augmented reality system that provides an anatomical roadmap derived from MR imaging that is continually aligned to X-ray fluoroscopy images in the XMR hybrid imaging environment [1]. We have used this system to guide cardiac catheterisation for more than 50 clinical cases. This system provides an accuracy of 2 mm, but respiratory motion introduces errors that are typically much greater than this. In this abstract, we describe a novel technique for correcting for respiratory motion using a patient-specific motion model derived from MR imaging. Validation was performed on four volunteer and three patient datasets.

Results Figure 1 shows the results of the three dimensional shape analysis. Those receiving a RVPA shunt had highly significant differences (p

Background Aortic Coarctation (AoCo) accounts for 5-8% of the children with CHD. Even after successful early repair, life expectancy is still markedly reduced (80% at 50 years after surgery) compared to normal population due to long term complications (hypertension). Usually, invasive diagnostic catheter investigations are required to evaluate the pressure gradient across the aorta at rest, or unmask such a gradient by use of isoprenaline stress to mimic physical exercise. The application of imagebased computational fluid dynamics (CFD) in patients with AoCo appears promising as an alternative noninvasive diagnostic tool, as it may allow the avoidance of cardiac catheterization to determine pressure gradients. The motivation of this research is to know if a MRI based CFD model can accurately predict the pressure

Background Hypertrophic cardiomyopathy (HCM) is associated with significant associated morbidity and mortality. Increased maximal left ventricular wall thickness (LVWT) has been postulated as major risk factor of sudden death; however, relatives with normal LVWT are also at risk. Genetically driven interstitial collagenesis has been proposed as a possible mechanism of diffuse myocardial fibrosis and increased extracellular volume fractions (ECV) has been demonstrated in genotype positive subjects. We investigated whether native T1 can also separate genotype positive subjects in the absence of increase in LVWT and how does it relate to ECV measurements.

Background Cardiac magnetic resonance (CMR) increasingly adds to clinical conformation of the diagnosis in patients with suspected myocarditis. The proposed Lake Louise Consensus Criteria ("any-two” approach) can separate from chronic forms of myocardial inflammation. However, both global enhancement ratio (GRE) and T2-weighted imaging are underutilized, due to poor reproducibility and high susceptibility to artefacts. T1 and T2 mapping by CMR projects tissue-dependent relaxation times. T2 mapping has been recently proposed as a robust and accurate technique to identify areas of focal myocardial oedema. Our aim was to investigate the value of quantitative T2 values in discrimination between health and disease, and separation between active/acute myocarditis and chronic convalescent stage of the disease. Methods Twenty-four patients with acute presentation of viral myocarditis and twenty-three subjects in clinical convalescence were recruited. Thirty-three healthy subjects were served as controls. All subjects underwent CMR study for routine assessment of myocardial oedema, function and scar by at 3-Tesla scanner. T2 values were acquired in midventricular short-axis slice (mSAX) using GraSE sequence. We examined regional T2 values in patients and controls. T2 values are presented as an average of the six segments per mSAX. Secondly we investigated the differences between visually involved and remote myocardium (involved myocardium = areas by LGE, remote = areas with no LGE) Results Patients with acute myocarditis and chronic myocarditis showed significantly raised T2 values [controls vs. acute vs. chronic myocarditis, T2 myocardium (msec): 48 ± 3 vs. 59 ± 9 vs. 53 ± 5, p < 0.0001). Compared to controls, T2 values in remote myocardium were significantly different for acute myocarditis only (T2, msec: 48 ± 3 vs. 51 ± 5 vs. 49 ± 2, p < 0.01), whereas T2 values in involved myocardium differed between all groups (T2, msec: 48 ± 3 vs. 69 ± 13 vs. 57 ± 8, p < 0.0001]. Using T2 values of complete mSAX or involved were identified as the independent discriminators between active and chronic myocarditis. T2 values were concordant with T2 oedema ratio (T2 involved, r = 0.42, p < 0.0001) and with native T1 (r = 0.55, p < 0.001). Conclusions We demonstrate that quantitative T2 values are increased in patients with myocarditis. We further demonstrate that average mSAX and involved T2 values can discriminate between acute and chronic stage of the disease. Funding

Background Myocardial T1 mapping is emerging as a promising means to non-invasively discriminate between normal and diseased myocardium. Original modified Look-Locker sequence (3’3’5-MOLLI) can induce a long breath-hold and is prone to cardiac and respiratory motion. Several shorter sequences (3’5MOLLI, shMOLLI and SASHA) have been proposed for acquisition of T1 values. We examined native and postcontrast T1 values derived by above sequences within the same subject and investigated their ability to separate between health and disease. We further examined reproducibility of measurements within and between observers.

Background T1 mapping by cardiovascular magnetic resonance (CMR) provides tissue-dependent relaxation times in line with the underlying myocardial composition. In this study, we examined the value of native T1 in differentiation between acute viral myocarditis and chronic stages of healing. Methods Fifty-three patients with a clinical diagnosis of acute viral myocarditis were compared to fifty-six subjects in clinical convalescence. All subjects underwent assessment of myocardial oedema, function and scar by CMR at 1.5 and 3-Tesla. T1 values were measured in midventricular short-axis slice prior to and after administration of 0.2 mmol/kg of gadobutrol. Results Compared to controls (n = 40), native T1 values were increased in both acute and chronic myocarditis at both field strengths (controls vs. acute vs. chronic myocarditis; 1.5T: native T1 (msec): 940 ± 17 vs. 1070 ± 34 vs. 998 ± 28, p < 0.001; 3T: 1045 ± 18 vs. 1182 ± 66 vs. 1094 ± 15, p < 0.001). Native T1 values were significantly higher in acute vs. chronic myocardium (p < 0.001). Native T1 was identified as the independent discriminator between health and disease, as well as acute and chronic myocarditis (Figure 1). Native T1 showed gradual reduction of values between acute and chronic stage of myocardial inflammation.

Background Increased inflammation has been linked to myocardial dysfunction and heart failure. We investigated whether T1 mapping by cardiovascular magnetic resonance (CMR) in patients with systemic lupus erythematosus (SLE) can detect subclinical cardiac involvement.

Background Classical modified Look-Locker sequence (MOLLI) can induce a long breath-hold and is prone to cardiac and respiratory motion. Several shorter sequences based on inversion (3’5MOLLI, shMOLLI) and saturation recovery (SASHA) of longitudinal relaxation have been proposed for derivation of T1 values or calculation of extracellular volume fraction or lambdas. Despite the validation in T1 gel phantoms, it has not been determined whether these novel sequences provide equivalent information on T1 when performed within the same individual. Methods Twenty-three subjects underwent T1 mapping in a midventricular equatorial short axis slice using above sequences on 3T clinical scanner prior and after gadolinium contrast (0.2 mmol/kg) administration. The images were analyzed using PRIDE tool with in-built automated motion correction. ROIs were drawn conservatively within the septal and lateral myocardium and the blood pool. Comparison of the native T1 values and lambdas were performed and expressed as percentage of mean difference from the values obtained with goldstandard MOLLI sequence.

Background T1 mapping enables true quantitative assessment of myocardial tissue, separating scars and healthy myocardium. The modified Look-Locker inversion recovery (MOLLI) sequence acquires eleven images over the duration of a minimum of fifteen heart beats to obtain the final T1 map, which is susceptible to undesired respiratory motion due to a relatively long breath-hold. We aimed to implement a motion correction method using a hierarchical local affine registration of MOLLI images and tested if it improved image quality and thus produced more accurate post contrast myocardial tissue T1 estimation.

Methods A standard ultrafast gradient echo perfusion sequence (st-GrE) was compared with an advanced kt-accelerated steady state free precession sequence (kt-SSFP) at 1.5T in a hardware perfusion phantom, healthy volunteers (n=16) and patients (n=31) with known or suspected coronary artery disease. Volunteers had both sequences at rest in alternating order. Patients underwent stress imaging with either st-GrE (15) or kt-SSFP (16) prior to X-ray coronary angiography. The phantom was used to generate signal intensity curves and noise maps for SNR and CNR analysis. Human images were analysed by a blinded observer using a nominal scale for quality (0 non-diagnostic, 1 poor, 2 moderate, 3 good, 4 excellent) and respiratory artefacts (0 non-diagnostic, 1 severe, 2 moderate, 3 minor, 4 nil) and also for the presence of CAD in patients. Other analyses included the extent (% affected segments), transmurality (1: 1-25%, 2: 26-50%, 3: 5175%, 4 76-100%) and duration (frames) of dark rim artefacts (DRA). Segmental SNR and CNR were also quantified using the mid ventricular slice in volunteers.

Background Delayed-enhancement (DE) MRI is an effective technique for imaging left atrial (LA) ablation lesions following radio-frequency ablation for atrial fibrillation (AF). Existing techniques for segmentation/visualization of lesions require manual interaction of an expert user making them prone to high observer variability. Oakes et al.[1] applied thresholding on the intensity histogram of the manually-outlined atrial wall. Peters et al.[2] used maximum intensity projections (MIPs) to generate volume lesion visualizations for expert-user interpretation. Knowles et al.[3] employed MIP followed by userinteractive thresholding for lesion surface visualization.

Background ICD implantation is known to improve prognosis in patients with ischemic cardiomyopathy (ICM). Determining which patients are at high risk of ventricular arrhythmias and sudden death remains difficult. The presence and extent of grey zone detected by CMR has been linked to ventricular arrhythmias and may be useful in determining which patients should have an ICD. Several processing methods to define the area of grey zone based on traditional late enhancement sequences (LGE) have been developed. Recently, a novel delayed enhancement sequence (MCDE) combined with an automated segmentation method was introduced to more accurately assess the peri-infarct area. 1,2 One limitation of this technique is the presence of cardiac motion in the image making the analysis cumbersome as the images from the same point in the cardiac cycle need to be filtered out of the dataset during postprocessing. An alternative option for the quantitative T1 measurement of the grey zone is the modified look-locker (MOLLI), which is not affected by cardiac motion. In the current study, the MCDE and MOLLI sequences were compared for the evaluation of the grey zone.

A single-phase 3d steady-state-free precession whole heart approach with respiratory navigator gating and ECG triggering (3d SSFP) is now frequently used for diagnostic imaging in children with congenital heart disease (CHD). However, certain cardiac structures may be better assessed during the systolic and others in the diastolic rest period. Therefore, a previously described dual phase 3d-SSFP whole heart sequence may offer clinical benefit in these patients. (Uribe et al 2008)

Methods Our MR-compatible phantom resembles the anatomy of the heart and of the thoracic vessels of a 60 kg subject. Water flow (2-4 l/min) is driven into the system by a pump located outside the MR room. Gadolinium injections (Gadovist, gadobutrol, Bayer Schering, Germany) are performed in the vena cava by a power-injector (Medrad Solaris, Germany). Progressive dilution of the Gadolinium bolus across the chambers generates the AIF. A fraction of the cardiac output perfuses two myocardial compartments (made by arrays of parallel tubes) where perfusion can be controlled precisely and independently. Outside the MR room, a control unit allows for precise measurement and control of cardiac output and myocardial perfusion. The phantom was tested in a 3T-scanner (Philips Achieva, Netherlands). Acquisition of the perfusion sequence was repeated multiple times with a k-t SENSE sequence (flip-angle=20°; TR2.2 ms; TE1.1 ms; k-t factor 5, 11 training profiles) using similar flow conditions to demonstrate the reproducibility of the measurements. Different flow conditions were used to assess the response of the system to different myocardial perfusion rates. Quantification was performed by Fermi deconvolution.

Percutaneous cardiac interventions are currently performed under X-ray guidance. MR-guided cardiac interventions have been previously described but mainly in animals. Their translation into humans has been limited by the lack of MR-compatible and safe equipment, such as MR guidewires with mechanical characteristics similar to standard guidewires.

Background Perfusion imaging visualizes early changes within the ischaemic cascade and has been shown to accurately detect significant coronary artery stenoses. First pass myocardial perfusion CMR yields high spatial resolution and permits quantification of myocardial blood flow. However, perfusion-CMR has only been validated in a small number of previous studies and using relatively early perfusion methods. An isolated blood-perfused MR compatible porcine heart model which can be imaged by identical equipment used for humans has been recently proposed. The model permits full control of regional myocardial blood flow, thus providing a valuable tool for the validation of quantitative perfusion assessment.

It is known that 30-40% of heart failure patients that undergo cardiac resynchronization therapy (CRT) do not derive significant clinical benefit. Using cine imaging a measure of volume change over the cardiac cycle can be derived called the Systolic Dyssynchrony Index (SDI). This has the potential to improve patient selection pre CRT.

Background Following catheter ablation for atrial fibrillation (AF), late gadolinium enhancement magnetic resonance imaging (LGE MRI) may be able to visualise areas of fibrosis and therefore reduced endocardial voltage. Unlike previous studies describing qualitative visual comparisons, we have developed a new quantitative technique for comparison of left atrial (LA) endocardial voltage with 3D LGE MRI signal intensity and have applied it in patients undergoing repeat left atrial ablation. Methods Ten patients who had undergone previous catheter ablation for AF, and who represented with either paroxysmal AF (n=4) or atrial tachycardia (AT) (n=6) underwent pre-ablation LGE MRI. A 3D LA reconstruction was created by projecting the LGE data on to a segmented LA shell (signal intensities were displayed as the number of standard deviations (SD) from the mean signal intensity of the atrial blood pool). During the ablation procedure, high density (mean number of points: 368 ± 127) LA endocardial voltage maps were acquired in either sinus rhythm or AT using CARTO 3 (Biosense Webster). The 3D LGE MRI reconstructions and endocardial voltage maps were manually segmented into LA regions (left and right WACA, roof, mitral line, anterior, posterior, inferior, septum and lateral) using custom-written software to derive the mean signal intensity and endocardial voltage for each segment. Results

Background Selection of patients with congenital heart disease for surgical septation in biventricular repair or surgical palliation in functionally single ventricles requires low pulmonary vascular resistance (PVR). Where there is uncertainty, PVR can be assessed using hybrid cardiovascular magnetic resonance (CMR) and fluoroscopic (X-Ray) guided cardiac catheterizations (XMR). CMR/XMR catheterization is a validated technique for accurate assessment of pulmonary vascular resistance. However, data concerning its application in clinical practice is lacking. Methods PVR assessments were performed in 167 studies in 149 congenital heart disease patients by CMR/XMR catheterization. Data was collated on patient demographics, procedural data, complications and outcomes. Institutional ethics approval was obtained. Results Median age was 3.6 years (6 days - 67 years) and weight 13.8 kg (2.3 -122 kg). One hundred and eight studies were in biventricular circulations and 59 in functionally single ventricles. Median radiation dose was 0.72 mSv. A baseline Qp:Qs ≤2.75 in biventricular circulations with left-to-right shunts predicted a PVR ≥6 WU.m2 with 100% sensitivity and 48% specificity. Median follow up until death or last review was 4.2 years (4 days - 11 years). Eighty-four patients had a surgical or catheter intervention based on CMR/XMR catheterization findings at a median of 94 days after the study. This included successful biventricular repair at resting PVR values ≤6 WU.m2 and Fontan completion at ≤4 WU.m2. Conclusion PVR measured by CMR/XMR catheterization allows accurate stratification for intervention in patients with congenital heart disease in both, biventricular and univentricular circulations.