Authors: Joshua Hawson, Subodh Joshi, Ahmed Al-kaisey, Souvik K. Das, Robert D. Anderson, Joseph Morton, Saurabh Kumar, Peter Kistler, Jonathan Kalman, Geoffrey Lee
Categories: Original Research Article, Cardiac imaging, Ventricular arrhythmias, Ventricular tachycardia
Source: Indian Pacing and Electrophysiology Journal
Ventricular tachycardia (VT) is a life-threatening arrhythmia that may be idiopathic or result from structural heart disease. Cardiac imaging is critical in the diagnostic workup and risk stratification of patients with VT. Data gained from cardiac imaging provides information on likely mechanisms and sites of origin, as well as risk of intervention. Pre-procedural imaging can be used to plan access route(s) and identify patients where post-procedural intensive care may be required. Integration of cardiac imaging into electroanatomical mapping systems during catheter ablation procedures can facilitate the optimal approach, reduce radiation dose, and may improve clinical outcomes. Intraprocedural imaging helps guide catheter position, target substrate, and identify complications early. This review summarises the contemporary imaging modalities used in patients with VT, and their uses both pre-procedurally and intra-procedurally.
Keywords: Ventricular arrhythmias, Ventricular tachycardia, Cardiac imaging
Ventricular tachycardia (VT) is a major cause of morbidity and mortality. VT often occurs in the setting of structural heart disease, whereby re-entry is facilitated by slow conduction within electrically insulated channels [1]. VT may also be focal in origin, due to autonomic or triggered activity, often arising from stereotyped sites [2]. The use of cardiac imaging in patients with VT will often provide the clinician with information to suggest the underlying substrate, VT mechanism, and prognosis. This will usually inform and optimal treatment strategy, including indication for intracardiac defibrillator (ICD) and catheter ablation procedure. There are, however, many different available imaging modalities to be considered in the evaluation of a patient with VT. In the following report we will review the different imaging modalities available, their uses, their strengths and weaknesses, and how they may be used as part of a VT ablation workflow.
Pre-procedural imaging to evaluate cardiac structure and function is standard of care in the vast majority of cases. Evaluation for the presence of structural heart disease provides clues as to the likely mechanism of VT – idiopathic VT is more likely to be focal in origin whereas VT in the setting of structural heart disease is more likely to be re-entrant [2]. Pre-procedural imaging may also reveal peri-procedural risk and/or anatomical limitations that help guide choice of therapy and plan intervention.
Historically, transthoracic echocardiography (TTE) and coronary angiography were recommended as first line investigations in patients presenting with VT [3]. Transthoracic imaging is typically an inexpensive and readily accessible tool, and so remains a first-line investigation [4]. TTE generally provides an accurate assessment of left ventricular ejection fraction and valve function. Areas of wall thinning and echo density may identify areas of scar, and asymmetric areas of hypertrophy may reveal hypertrophic cardiomyopathy as the underlying diagnosis. However, there are limitations to TTE that need to be considered. Other imaging modalities are superior to TTE in tissue characterization, identifying viability, and evaluating the size and function of the right ventricle [5]. Additionally, poor image quality limits the diagnostic utility of TTE in 10–15% of cases [6].
In recent years computed tomography (CT) has evolved to be a highly accurate method of assessing for coronary artery disease (CAD), with radiation exposure in a similar range to invasive coronary angiography. As such, CT coronary angiography (CTCA) is now recommended as a first line investigation to exclude coronary artery stenosis in those deemed low risk for CAD (4). Invasive angiography remains the investigation of choice in patients where coronary stenosis is suspected. The heart rhythm is an important consideration when deciding on the most appropriate modality of coronary imaging, as tachycardia or an irregular rhythm may reduce the quality and diagnostic yield of CT [7].
In addition to evaluating the coronary arteries, there are other diagnostic benefits that cardiac CT may add. The primary advantage of CT over other imaging modalities is higher spatial resolution and a true three-dimensional dataset that can be viewed in any imaging plane. Additionally, cardiac CT acquisition takes only seconds, making a CT scanner a far safer environment than an MRI scanner for a patient with VT. CMR resolution, even at 3 T, is limited to a voxel size of 1.4 mm^3^, whereas CT can achieve voxel resolution of <0.4 mm^3^ [[8], [9], [10]]. Due to the excellent resolution, CT is ideal for defining anatomical pathology, such as ventricular aneurysms, pseudoaneurysms, vascular anomalies and congenital abnormalities (Fig. 1) [11]. However, there are several limitations of CT to be considered. The lower contrast-to-noise ratio within myocardial tissue of CT compared with CMR results in poorer tissue characterization [5]. CT also offers less functional information compared with CMR, and has the added risks of radiation exposure and iodinated contrast administration [12].
Fig. 1 CT For Defining Anatomy. 32 y F with history of double outlet right ventricle, transposition of the great arteries and VSD. Initial repair included arterial switch, closure of VSD and pulmonary artery reconstruction. Has since undergone patch closure of residual VSD and Bentall's procedure in adulthood. Panel A: 12 lead ECG of clinical VT. Panel B: Coronal plane of CT demonstrating VSD extending inferiorly from pulmonary artery. ICD is seen in the RV with significant associated artifact. Panel C: Sagittal plane CT demonstrating VSD patch immediately inferior to the pulmonary artery and prosthetic AoV. Panel D: Voltage map of the RV demonstrating an area of low voltage in the area corresponding to the VSD patch. A potential channel can be seen between the tricuspid annulus and the VSD patch. Panel E: VT activation demonstrating conduction through the channel between the VSD patch and tricuspid annulus. This location was dangerously close the His; successful ablation was performed by performing a line from the pulmonary artery to the top of the tricuspid annulus.
CMR provides excellent characterization of the ventricular myocardium and accurate assessment of function. RV dilatation, scarring, and poor function may point towards arrhythmogenic right ventricular cardiomyopathy (ARVC) as a diagnosis [13]. LV wall thickness and distribution of hypertrophy may reveal hypertrophic cardiomyopathy (HCM) as the underlying pathology [14]. In addition to quantifying RV and LV ejection fraction, CMR is generally considered the gold standard when assessing ventricular scar distribution and burden [15]. Myocardial scar is typically evaluated using the signal intensity of late gadolinium enhancement (LGE) (Fig. 2) [16]. However, gadolinium administration may not be possible in patients with poor renal function, due to the risk of nephrogenic systemic fibrosis, although the absolute risk is low [17]. Other modalities to define myocardial scar without the use of gadolinium contrast have been developed.
Fig. 2 Multimodality Imaging in Sarcoidosis. 48 y M with cardiac sarcoidosis and VT. **Panel A: -**LGE-CMR demonstrates scar in the anterior and inferior walls (white arrows). The distribution is characteristically subepicardial and in a non-coronary distribution. Panel B: PET/CT demonstrated activate inflammation in the areas of LGE seen on CMR. Additionally, there was extensive areas of inflammation in the RV free wall. Panel C: Tc^99^ sestamibi scan merged with the PET/CT demonstrates matched perfusion defect. White areas correspond with inflammation seen on PET/CT. Panel D: VT study demonstrated an extensive area of low unipolar voltage around the RVOT and RV free wall. A channel at the base of the RVOT was seen with bipolar voltage that corresponded to both the primary deceleration zone and site of entrainment IN (green dot), defined as concealed fusion with PPI-TCL of <30 ms. RV = right ventricle, RVOT = right ventricular outflow tract, VT = ventricular tachycardia.
The T1 relaxation time is a measure of how fast the nuclear spin magnetization returns to equilibrium after a radiofrequency pulse from the MRI scanner. “T1 mapping” refers to visualizing these properties, generally in a colour-coded scheme, whereby pixel values represent the T1 in each voxel [18]. Using T1 mapping it is possible to measure a parameter called the extracellular volume (ECV), which quantifies the relative expansion of the extracellular matrix [5]. Even without contrast (“native” T1 mapping) this process is sensitive to myocardial oedema, iron overload, and the presence of scar [18]. T2 is a measure of transverse relaxation time. Elevated T2 is specific for myocardial water content and is used as a marker of myocardial oedema. T2 mapping is a useful tool for diagnosing myocarditis as a potential cause of VT (15).
The distribution of myocardial scar offers critical insight into the underlying pathology and helps guide therapy. The classic subendocardial distribution of scar seen in ischaemic cardiomyopathy (ICM) often suggests than an endocardial approach to ablation is optimal [19]. However, transmural or epicardial scar, more typically seen in nonischaemic cardiomyopathy (NICM), may lead the operator to consider a combined endocardial/epicardial approach upfront [20]. In some patients thought to have purely NICM, CMR may revealed concomitant subendocardial scar indicative or infarction [19]. Apical HCM with an apical aneurysm may suggest than an endocardial approach is futile, as the aneurysm neck and HCM with a very thick septum may lead the clinician to consider alternate therapies, such as septal alcohol injection or bipolar ablation [21,22].
CMR plays a critical role in the evaluation of patients presenting with VT. One previous study has demonstrated that the added information provided by CMR alters the underlying diagnosis in 50% of patients [23]. However, there are several limitations to CMR imaging that need to be considered. One of the most common limitations is incompatibility of the implantable cardiac defibrillator (ICDs). Over the past several years, MRI-compatible devices have been developed. However, these still result in device-related artifacts that can limit the diagnostic utility of the study. Scar qualification is particularly susceptible to ICD artifact, with many of the myocardial segments often affected [24]. For this reason, when a patient is presenting with a new diagnosis of VT, it is desirable to perform the MRI prior to device implantation where possible. When an ICD is present, use of wideband LGE-CMR sequence can be used to decrease artifact [25].
Frequent ventricular ectopy, common in patients being considered for ablation, also degrades image quality of CMR. Most standard sequences in CMR, including cine and LGE, are a composite of many cardiac cycles and rely on a consistent regular heart rate. Similarly, CMR images are typically acquired during a breath-hold of approximately 10 s, which can be difficult for some patients with cardiomyopathy. Real-time cine imaging, where each image is constructed from only one heartbeat, can overcome some of these obstacles. However, this is at the expense of temporal and spatial resolution. Lower resolution single-shot LGE sequences remain robust in the setting of arrhythmia. Newer motion-corrected free breathing LGE can be used in patients who are unable to breath-hold. Perhaps the most important consideration with regard to CMR is whether a patient is stable enough to attend MRI, as the scan may take >45 min. The magnetic field causes artifact on the ECG, so some experience is required to recognize ventricular arrhythmias whilst the patient is in the scanner.
Nuclear imaging may provide some incremental diagnostic information during the evaluation of a patient with VT. Nuclear imaging techniques to quantify areas of myocardial ischaemia and subsequent reversibility may reveal the utility of revascularization [26]. Positron emission topography (PET) scanning with ^18^fluorodeoxyglucose (FDG) is useful for identifying active inflammation [5]. In one retrospective study, 103 patients with unexplained cardiomyopathy and ventricular arrhythmias were evaluated with FDG-PET/CT scans. Nearly 50% demonstrated focal myocardial inflammation [27]. FDG-PET is also a useful tool in the diagnosis of cardiac sarcoidosis. In patients with a CMR result suggestive of sarcoidosis, FDG-PET/CT is used to determine disease activity and role for immunosuppression [28]. In patients where MRI is not possible, FDG-PET/CT can be used in the diagnostic workup for sarcoidosis. Here, in the absence of coronary artery disease, FDG-PET typically demonstrates a resting perfusion defect that has increased FDG uptake (Fig. 2) [29]. The advantages and disadvantages of imaging modalities are summarized in Table 1.
The presence and severity of structural heart disease are major contributors to the long term prognosis of patients with VT. Current guidelines recommend implantation of an ICD in a patient with sustained VT, which is haemodynamically compromising and/or occurs in the context of structural heart disease [30]. In the setting of significant abnormalities, TTE often provides sufficient imaging to guide device therapy. However, multimodality imaging often provides incremental information that allows for more patient-tailored decision making. As defined on CMR, the presence and burden of scar has been consistently shown to be reflective of risk in multiple conditions [[31], [32], [33], [34]].
Scar pattern has also been implicated as a marker of risk, with higher-risk features including scar heterogeneity, presence of conduction channels within scar, and number of core scar islands [[35], [36], [37]]. A specific pattern of LGE, termed “ring-like enhancement”, has been shown to be a high-risk feature, with a 50% risk of death, cardiac arrest, or appropriate ICD therapy at 5 years. This risk is present even in patients without sustained VT as an index event [38]. Conversely, the absence of LGE confers a relatively good prognosis with patients with frequent ventricular ectopy and/or non-sustained VT (38). In sarcoidosis, the presence of arrhythmogenic scar can be found despite a normal ejection fraction and portends a worse prognosis [34].
Methods have been developed to determine the risk of death following VT ablation, in order to appropriately select and inform patients for this procedure. In the PAAINESD risk score reported by Santangeli et al., the presence of ICM and/or a LV ejection fraction (LVEF) of <25% was found to be predictive of higher mortality [39]. The I-VT score, subsequently reported by Vergara et al., found the LVEF to be the strongest predictor, with an LVEF of <30% conveying a higher risk of death [40]. Adequate cardiac imaging is crucial in the assessment of patients with VT who are being considered for catheter ablation. The assessment of per-procedural risk may guide type of therapy, but also identifies patients who are likely to require haemodynamic support and/or specialized post-operative care. In patients at particularly high risk, noninvasive ablation (stereotactic body radiotherapy; SABR) may be favoured [41].
Diagnostic imaging usually provides information about which surface(s) of the ventricle need to be targeted during catheter ablation. ICM, often demonstrating subendocardial scar, is frequently suitable for an endocardial-only approach. NICM patients with intramural scar frequently require a combination of endocardial and epicardial approaches [20]. Some pathologies may lend themselves to epicardial-only approaches, such apical HCM with aneurysm formation or LV summit VT with a clear epicardial morphology on the 12-lead VT [42].
Imaging may help plan procedural approach in the setting of previous procedures (Fig. 3). In the setting of mechanical valves, multipolar catheters are generally contraindicated due to risk of entrapment and should be avoided [43]. Mechanical valves cannot be crossed with catheters and so limit the approaches available. The presence of large occlusion devices on the septum suggests a transseptal approach is likely to be challenging, causing the operator to consider a retro-aortic approach.
Fig. 3 CT For Pre-procedural Planning and Integration. 72 y M with NICM**. Panel A:** Pre-procedural CT showed hepatomegaly obstructing a standard subxiphoid epicardial approach. Epicardial access was subsequently performed with the assistance of a cardiothoracic surgeon. Panel B: EAM demonstrated a large area of basolateral scar**. Panel C:** CARTOUNIVU was used to define coronary anatomy in the area of interest. Panel D: CT, merged with the aid of CARTOUNIVU, helped to define anatomy and coronary arteries. Sustained VT was not inducible, and ablation was targeted to the primary deceleration zone and abnormal EGMs (LPs and LAVAs). LP = late potential, LAVA = local abnormal ventricular activity, LAD = left anterior descending artery, RCA = right coronary artery.
Identifying the region of culprit scar on CMR can help the operator choose the best mode of access. Inferobasal and lateral scar is likely to be easily accessed via a transseptal approach. Scar around the LV ostium favours retro-aortic access. CMR may also provide anatomical information about surrounding structures that may impede epicardial access, although CT is generally a superior test for this due to the higher resolution (Fig. 3). The complementary use of a QRS axis-based algorithm in conjunction with CMR will often point the operator towards the culprit scar likely to be harbouring the putative VT isthmus channel [44,45]. This may be of use where there are multiple areas of patchy scar, and the culprit area is not clear.
Due to the higher resolution, CT is generally the optimal imaging modality for defining anatomical structures when considering epicardial approach [8]. The traditional method of accessing the epicardial space is via a percutaneous subxiphoid approach [46]. However, patient specific anatomy may make this difficult or impossible (Fig. 3). The presence of a large xiphoid process may obstruct the access angle. The present of an elevated hemidiaphragm may displace the heart cranially, creating a large distance from the xiphisternum to the pericardium. In these cases, surgical assistance in accessing the epicardial space is advisable, either via a subcostal approach or mini-thoracotomy [47]. Thickened or calcified pericardium seen on CT may suggest the presence of adhesions that will limit or prevent epicardial access [48].
Electrocardiographic imaging (ECGi) has emerged as a novel technology to localise VT origin. This technology uses mathematical algorithms to analyse unipolar electrograms from 252 sites over the chest wall. Activation is then displayed on a 3D reconstruction of the heart, generated from a cardiac CT [49]. Whilst the utility of ECGi is relatively niche, there are some occasions where it may be of benefit. We have previously described a case where mechanical aortic (bileaflet) and mitral (caged ball) valves prevented endocardial access to the LV cavity via traditional routes [47]. ECGi was used to demonstrated VT breakout at the basal inferolateral wall. Transapical access was then performed rather than a transseptal puncture, as it was thought the angle of approach would be more favourable (Fig. 4). In cases where access is difficult, knowing the area of VT origin can be critical in procedural planning and ECGi may provide a valuable tool. Graham et al. previously compared simultaneous ECGi with invasive EAM [50]. The minimum distance between pacing sites and the region of earliest activation on ECGi was 12.3 (0.0–28.2) mm. Operators must be mindful that ECGi maps are a guide to the VT exit and do not localise the diastolic isthmus. It is also important to note that the utility of ECGi is limited when the VT is arising from the ventricular septum or an intracardial structure [51].
Fig. 4 ECGi to Plan Procedural Approach. 52 y M with mechanical aortic and mitral valves in the setting of non-ischaemic cardiomyopathy. Panel A: ECGi of clinical VT demonstrating breakout at the basolateral area, adjacent to the mitral annulus. Panel B: Diagrammatic representation of access planning. In the setting of mechanical aortic and mitral valves, endocardial access to the LV was only possible via puncture from the RA into the LV or via apical puncture. Given the basolateral origin of the VT and the presence of a caged ball mechanical mitral valve, the angle of approach was deemed more favourable with transapical access.
Preprocedural exclusion of intracardiac thrombus is generally indicated due to the potential risk of embolic complications. Transthoracic imaging is commonly used to exclude LV thrombus, although its sensitivity is poor, particular in patients with severe LV dysfunction [52]. Echocardiographic contrast administration in the form of microbubbles improves sensitivity substantially, however only protuberant thrombi can be reliably identified [53]. CMR is generally considered the gold standard imaging modality for detecting intracardiac thrombus, and small laminated thrombi can usually only be seen with CMR [54]. Current evidence suggests that the presence of mobile intracardiac thrombus is a contraindication to ablation, whereas ablation may still be performed in the presence of laminated thrombus [55,56]. In the presence of mobile thrombi, epicardial-only ablation has been described as a viable strategy [57].
Multiple imaging modalities are now able to be integrated into electroanatomical mapping (EAM) systems. The primary advantage of image integration is to delineate anatomical structures and areas of abnormal tissue to facilitate and guide ablation. It has been demonstrated that integration of pre-procedural imaging reduces procedure time and improves safety in catheter ablation procedures [5]. Studies evaluating integration of imaging are summarized in Supplemental Tables 2–4.
Integration of CT is a valuable tool in VT ablation. Features on CT that can be used to delineate scar include wall thickness, hypoattenuation, and decreased perfusion [58,59]. Several studies have compared CT-defined scar and EAM-defined scar following integration [59,60]. The agreement is typically good in patients with ICM, although less so in patients with NICM [61]. This is likely due to the fact that A) intramural and epicardial scar correlates poorly with LV wall thickness, and B) intramural scar may demonstrate preserved endocardial voltages [59]. Image integration has been shown to impact procedural management, particularly in ARCV and other causes of NICM(61).
CT has the advantage of being the most effective technique for defining detailed anatomy due its high resolution (<0.4 mm^3^). Coronary arterial and venous systems can be segmented and visualized, which can be useful when determining appropriate ablation locations on the epicardium (Fig. 3). Integration of the phrenic nerve location may prevent injury [62]. Identifying areas of epicardial fat may reveal the cause of low bipolar voltage [63]. Identifying epicardial fat is also crucial in determining where ablation may be futile, as ablation at sites with >10 mm of fat thickness is ineffective [63].
Whereby the VT is suspected to be coming from an area of complex anatomy, CT is useful for defining anatomical variation in that area. For example, there is significant variation in papillary muscle anatomy that should be considered when planning an ablation of papillary muscle VT [64,65]. CT is also the modality of choice when planning a procedural approach in the setting of congenital heart disease [11]. The higher spatial resolution makes it superior in defining anatomical structures, baffle pathways and abnormal connections when considering access routes.
It has been repeatedly shown that putative VT channels often occur in CMR-defined scar [66,67]. Interestingly, the correlation of EAM-defined and CMR-defined scar is variable, with a number of studies demonstrating a mismatch between these two modalities [68,69]. Attempts to use CMR integration to target ablation have focused on scar regions with ‘high risk’ features, such as increased transmurality, scar border-zones, and regions at the junction of scar core and scar border-zone [35,36,70]. Identifying arrhythmogenic substrate on CMR and integration of CMR into EAM systems helps to plan an appropriate ablation strategy [71]. However, the clinical benefit of CMR integration in the absence of dedicated imaging processing and segmentation (see below) remains unclear, as evidence in this space is lacking.
Combined PET/CT imagines may also be integrated into EAM systems to help target ablation. Previous reports have demonstrated that critical isthmuses are frequently located in PET-defined scar [72]. PET can also be used to identify channels of viable myocardium in areas of low voltage on EAM that may otherwise be interpreted as scar [73]. A major limitation of PET is the lack of anatomical detail, and PET typically requires fusion with either CT or CMR prior to integration. Current PET/CT systems offer a resolution of 2–5 mm [74]. Historically, cardiac PET scans were ungated, producing a source of inaccuracy. More recently, dual cardiac and respiratory gating of cardiac PET has been incorporated and shown to reduce the diameter of hot spots [75]. Similar to CMR, the impact that PET/CT integration has on clinical outcomes has not been adequately assessed.
Automatic Detection of Arrhythmogenic Substrate (ADAS) software has been developed with the aim of delineating the ventricular architecture and integrating this into EAM systems [25]. ADAS processes either CMR or CT data to generate ventricular scar maps. ADAS scar maps depict potential conduction channels of preserved myocardium within dense scar – so-called “border-zone corridors” (Fig. 5). These corridors can be identified either manually, or automatically via algorithms built into the software (central illustration). Processed imaging can be integrated into EAM systems to facilitate procedures (Fig. 5).
Fig. 5 Integration of Processed Imaging to Guide Ablation. 72 y M with ICM. Panel A: ADAS-processed CT demonstrating areas of wall thinning anterobasally and laterally. A channel of borderzone tissue between two areas of denser scar is seen. Panel B: Integration of ADAS-processed CT into EAM. Panel C: EAM demonstrated anterobasal and lateral scar with a channel between these two areas, consistent with the findings of the ADAS-processed CT. Panel D: VT activation mapping confirmed this channel was the putative VT isthmus.
For CT, ADAS uses wall thickness as a surrogate for scar. Standard cutoffs in the left ventricle are <2 mm for dense scar and >5 mm for normal tissue, although validation of these values in large studies has not been performed [76]. The threshold values can be manually adjusted by the operator to reveal conduction channels. In CMR analysis, ADAS uses pixel signal intensity (PSI) to identify scar. The myocardium is segmental using a semi-automated approach. The LV is then divided into a number of layers (typically 9) from the endocardium to the epicardium. PSI maps are then generated at each layer, with pixel intensity of >60 ± 5% used as the threshold for dense scar and <40 ± 5% as the threshold for normal tissue [77]. Reconstruction of each layer map is then performed to create a 3D model of scar [78].
There have been multiple studies assessing the benefit of integrating ADAS-processed images into EAM systems. In a non-randomised prospective study, CMR with ADAS processing was used to define continuous corridors of surviving myocardium, bordered by scare core, that connected two areas of healthy tissue [78]. In a subset of patients, ablation was targeted to these corridors without complementary EAM (CMR-only group). Compared to the groups utilizing EAM-only or EAM + CMR, the CMR-only group had lower procedural and fluoroscopy times. The VT-free survival was also improved compared with the EAM-only group, although it was not significantly different compared with the EAM + CMR group [78].
InHeart software similarly analyses CMR or CT imaging data to create 3D cardiac models. Segmentation of important anatomic features is performed, such as wall thinning, epicardial fat, myocardial fatty metaplasia, phrenic nerves, and coronary arteries. For CT, wall thickness ≤5 mm or hypoattenuation of ≤10 Hounsfield Units (HU) is segmented as scar, with wall thickness ≤2 mm considered dense scar. For CMR, signal intensity of >35% of maximum is used to denote scar. Using InHeart segmentation, channels are typically defined as either 1) regions of ≤5 mm thickness, with a difference of ≥1 mm with surrounding tissue, with channels traversing part of the scar or the entire scar, or 2) channels of healthy tissue between 2 regions of CMR-defined scar. In a small series, 100% of putative isthmus channels were identified using InHeart segmentation [79].
Ghannam et al. examined 15 consecutive patients with previous myocardial infarction who were referred for VT ablation [80]. Cardiac CT was performed on all patients and, following InHeart processing, were integrated into EAM systems. Ablation target sites, identified with either pace-mapping or entrainment mapping, were then correlated with wall thickness. Ablation target sites were found to be located on CT-defined ridges (wall thickness 4.2 ± 1.2 mm) bordered by areas of thinning (wall thickness 2.6 ± 1.1 mm) in 14 (93%) of the patients, and in 49 of 58 (84%) of the mappable VTs identified. This led to the conclusion that wall thickness mapping using InHeart software could identify ridges of myocardium that serves as putative conduction channels [80].
Berte et al. examined 49 consecutive patients undergoing image integration-guided VT ablation, utilizing either CMR or CT (81). Patients were divided into two group 1 had detailed EAM during the procedure, whilst group 2 had only limited EAM and ablation was predominantly imaging guided. Group 2, undergoing limited EAM, was shown to have shorter procedural times (151 ± 22 min vs 180 ± 53 min, P < 0.01) despite there being no difference between groups with regards to acute VT non-inducibility (71% vs 74%, P = 0.8) or VT-free survival at 19 ± 8 months follow-up [81]. Randomized studies to evaluate clinical outcomes using InHeart software are ongoing.
Subendocardial fat deposition is a common part of the healing process following infarction, and is seen in up to 84% of post-infarction scars on histology [82]. However, in recent years, infiltrative fat within myocardial scar has been shown to correlate with critical sites of re-entry. Cheniti et al. identified fatty deposition, defined as intramyocardial attenuation of ≤ -30 HU, in 44 (64%) of 69 consecutive patients undergoing VT ablation [83]. InHeart software was used for segmentation. On EAM, areas of fat deposition were characterized by lower bipolar amplitude and prolonged electrogram duration. At a mean follow-up of 26 months, patients with fatty deposition had worse all-cause mortality and VT-free survival (P < 0.01).
Xu et al. studied 30 patients undergoing VT ablation who had undergone pre-procedural CMR and CT. Images were analysed using ADAS software. Lipomatous metaplasia (fatty deposition) was defined from CT images as myocardial areas with attenuation ranging from −180 to 0 HU and volume >1 mm^2^. ADAS was used for automatic detection of potential conduction channels. A total of 381 corridors were found, with 84 proven to participate in VT re-entrant circuits. Among these putative channels, 83 (99%) traversed or were adjacent to lipomatous metaplasia, significantly more than the number of non-putative channels (4%).
One proposed mechanism to explain these findings is the lower impedance of adipose tissue compared with scar. Whereas the lower impedance of scar allows for current leak, adipose tissue causes more effective electrical insulation and facilitates propagation of re-entry [84]. However, the role of adipose deposition may be more diverse, and further studies validating this mechanism are required [85]. Adipose tissue is a potential source of inflammatory mediators that may affect conduction properties. Pouliopoulos et al. found that fatty deposition was associated with an increased in connexin 43 lateralisation and slower conduction in neighbouring myocytes [86]. A large prospective cohort study of mechanistic associations between intra-myocardial far deposition and ventricular tachycardia in cardiomyopathy is ongoing [84].
The major limitation of image integration is the propensity for registration error. Imperfect image integration is an ongoing and unsolved issue. Currently there are no widely accepted standardized approaches to merging, although various methods have been described. Even in highly regimented studies, registration error is typically 3–5 mm [25]. Changes in anatomy between the time of scanning and the time of ablation may also provide a source of registration error [87]. Chamber volume and cardiac orientation are subject to change over time based on fluid status, heart rate and heart rhythm [87,88]. Inaccurate co-registration integration leads to incorrect identification of substrate ablation targets.
Radiation exposure during ablation procedures is an ongoing concern. Consequently, there have been extensive efforts to minimize radiation during VT ablation, including procedures performed with zero fluoroscopy [89]. The advent of 3D EAM systems has facilitated the reduction in intraprocedural radiation, and use of these EAM systems during a VT ablation is now considered standard practice. However, these systems derive 3D reconstructions from catheter electrode contact at the myocardial surface. The extrapolation of data using this method can be prone to error from a number of A) wall excursion, B) geometric shift in different rhythms, C) confounding of far-field signals, and D) patient movement [5,88]. The ability of fluoroscopic imaging to provide rapid information on catheter position is excellent, and fluoroscopy remains an important tool when performing ablation procedures.
Fluoroscopic imaging can be used in conjunction with EAM systems in order to reduce the required fluoroscopic time and subsequent radiation dose. The CARTOUNIVU module (Biosense Webster) allows for the integration of fluoroscopy images into the CARTO 3 EAM system [90]. Once the registration process has been performed, the operator is able to acquire fluoroscopic images in different angles. The EAM is then created on top of the acquired fluoroscopy images (Fig. 3) [91]. This allows demonstration of the proximity of the catheter to important anatomical structures, such as the coronary arteries [92]. Use of the CARTOUNIVU module has been associated with low levels of radiation exposure during VT ablation procedures [92].
Intracardiac echocardiography (ICE) can typically be performed with two different 1) rotational (or radial) ICE, and 2) Phased-array ICE. Rotational ICE provides a 360° image perpendicular to the long axis of the catheter and is typically used in coronary intervention. Phased-array ICE usually consists of a 64-element transducer mounted on the distal end of a steerable catheter [93]. Phased-array ICE produces a wedge-shaped image that is displayed on a conventional ultrasound screen (Fig. 6). Phased-array ICE provides deeper tissue penetration (∼15 cm) than rotational ICE and is the modality of choice for interventional procedures in electrophysiology.
Fig. 6 Intracardiac Echocardiography to Guide Papillary Muscle VT Ablation. 40 y F with structurally normal heart and recurrent NSVT/PVC from anterolateral papillary muscle. Panel A: ICE was used to create LV and papillary muscle geometry. Outline of the LV cavity and papillary muscle through multiple slices (green lines) allow for construction of 3D anatomy. Panel B: ICE was used to confirm catheter position on the anterolaterally papillary muscle (AL PAP). Ablation at the earliest activation site rendered the NSVT quiescent.
The two most common locations to position the ICE catheter are the right atrium (RA) and the right ventricle (RV). An ICE catheter placed in RA pointed towards the tricuspid valve is typically referred to as the ‘home view’. Clocking the ICE catheter from the home view will bring the left atrial appendage into the imaging window for exclusion of thrombus. In the RV, the ICE catheter is typically placed at the base of the RVOT by flexing the catheter in the home view, advancing it across the tricuspid valve, and releasing the flexion [94]. Imaging the heart from other locations is possible – such as from the coronary sinus, pulmonary artery, and left atrium – although this typically requires an experienced operator due to difficult catheter manipulation and higher risk of complication [93].
Due to a lower requirement for tissue penetration, ICE typically operates with a higher frequency (5–12 MHz) than transoesophageal (4–8 MHz) and transthoracic (2–5 MHz) echocardiography, allowing for higher resolution imaging [[95], [96], [97]]. ICE is therefore superior in assessing the myocardial architecture than the other echocardiographic modalities. Myocardium that is thinned, akinetic, or has abnormal echodensity suggests an area of scar [98]. The degree of echo density can be used to differentiate scar core from border zones [99]. Identification of scar using ICE has been shown to correlate well with scar defined by EAM systems, CT and CMR [99,100]. However, ICE remains inferior to CT and CMR in the detailed assessment of scar structure [5].
The anatomic information provided by ICE can be incorporated into EAM systems. Use of an ICE catheter that contains a location sensor can be visualized in the EAM system. By using dedicated ICE catheters, the endocardial surface can be traced on the ICE images and used to create the 3D anatomical reconstruction. Scar can be delineated and incorporated [101]. Intracardial structures can also be traced and incorporated, such as the moderator band and papillary muscles. Electroanatomical structures may also be tagged on the integrated ICE images. For example, identification of the His may be tagged and displayed on ICE images. This can facilitate imaging of the surrounding structures and planning of the optimal approach [94].
ICE provides real-time visualization of the ablation catheter and adjacent structures. This allows for continuous assessment of both catheter position and contact (Fig. 6). This is critical for ablation at anatomical complex sites. In aortic cusp VT ablation, the ablation location can be confirmed with the ICE catheter (typically with the ICE catheter in the RV). Adequate ICE imaging may demonstrate that the catheter tip is > 10 mm away from coronary ostia, obviating the need for coronary angiography when in this area [93]. Use of doppler flow may confirm that the catheter is a safe distance from coronary arteries.
Use of ICE is essential in defining papillary muscle anatomy when ablating papillary muscle VT (Fig. 6). Anatomical variation can be examined, as well as adequate catheter contact with the structure. In the absence of ICE, it can be difficult to determine if the ablation catheter is truly on the papillary muscle or in contact with the ventricular wall [102]. ICE is also useful for defining the RV moderator band and confirming appropriate catheter position when ablation arrhythmias from this structure.
ICE can be used to assess real-time lesion formation during procedures. Lesions are typically visualized an increased echogenicity at the site of ablation [103].
Intraprocedural imaging is critical in the rapid identification of complications. Fluoroscopy may demonstrate a globular cardiac silhouette with decreased excursion, suggesting pericardial effusion and impending tamponade [104]. Assessment of the pericardial space can also be immediately performed with ICE. Haemodynamic collapse may also occur due to air embolism, which most frequently affects the right coronary artery due to its superior position in the supine patient. Electrocardiographic changes, such an inferior ST-elevation, usually suggest the diagnosis [105]. ICE can be used to evaluate for segmental dysfunction of the myocardium in this setting. ICE may be used to detect thrombi on catheters before embolism has occurred, allowing for pre-emptive treatment with anticoagulation. In the setting of otherwise unexplained shock, a collapsed LV may suggest hypovolaemia from bleeding at a remote site. Finally, ICE can be used to detect valvular incompetence that may arise from iatrogenic causes intraprocedurally [93].
Due to its superiority in assessing myocardial tissue, CMR can be used to assess ablation lesion post-procedurally. It has been demonstrated that CMR with LGE can accurately evaluate radiofrequency ablation lesions, with good correlation to histologic findings [106]. Tissue remodelling is a dynamic process post ablation, and so the timing of lesion assessment needs to be considered. It is thought to be more accurate at approximately 3 months post ablation than at later time points [107]. The adoption of post-procedural CMR to assess ablation lesions has been lesser in VT ablation than in atrial ablations. This may be due to the presence of CIEDs causing artifact and reducing CMR quality and/or due to the lack of standardized endpoints in VT ablation [107]. Case reports have described CMR assessment of ablation lesions, with homogenous transmural lesions considered a surrogate for successful ablation [108]. However, to date, the clinical utility of post-procedural assessment of ablation lesions has not yet been investigated [107].
Catheter ablation with real-time CMR using MRI-compatible catheters have been described. The aim of this method is to eliminate radiation associated with fluoroscopy and facilitate real-time imaging of ablation lesions [109]. Whilst animal studies suggest this method is feasible, there are several limiting factors to consider. First, there is a requirement for all equipment to be MRI compatible [110]. Second, MRI machines are both expensive and cumbersome, potentially requiring construction of dedicated laboratories to house the equipment. As such, ablation utilizing real-time CMR is likely to remain a niche procedure.
Computational modelling based on cardiac imaging is an emerging technology of interest. Multimodality imaging provides detailed patient-specific data on both ventricular geometry and scar architecture. Incorporating this data into biophysical models allows for the performance of virtual electrophysiology studies. This can both identify channels and determine which channels are likely to be putative [111]. Several small proof-of-concept reports have demonstrated that this computer modelling-based approach may be used to predict VT circuits and arrhythmic events [112]. However, this technology remains in the experimental stages. It is not yet widely available, nor has it been prospectively validated.
There are many imaging modalities available during both the assessment and management of patients with VT. Each modality has its own strengths and weaknesses. As such, multiple different modalities are often performed for their complementary features. Utilization of the multimodality approach and combining information on anatomy, substrate, and electrical activation can provide a comprehensive assessment of cardiac arrhythmias and their mechanisms [113]. This allows the clinician to make the correct diagnosis of the underlying pathology and facilitates catheter ablation procedures as indicated [71]. A framework for a stepwise approach to imaging in VT is provided in Fig. 7.
Central figure Multimodality Imaging in 3D VT Circuit. 47 y M with previous transmural infarct of posterolateral branch of RCA who subsequently underwent an ENDO/EPI VT ablation**. Panel A:** CMR demonstrates LGE in the inferobasolateral area. Panel B: ADAS-processed CMR demonstrates heterogeneity of scar in this area with multiple border zone corridors (yellow) between patches of dense scar (red). Panel C: Endocardial EAM demonstrated abnormal bipolar and unipolar voltages at the inferobasolateral area. Panel D: Epicardial EAM also demonstrated abnormal bipolar and unipolar voltages in this area. Panel E: 3D reconstruction of the scar geometry using ADAS demonstrated transmural conducting channels. Panel F: VT activation demonstrated both endocardial and epicardial segments of the circuit. The segments of the circuit that are unmapped or demonstrate far-field signals are displayed in interrupted arrows. Panel G: Merged ADAS-processed CMR imaging with CMR and EAM. VT circuit demonstrated in which arrows, with the intramural portion of the circuit displayed by the interrupted arrow.
Fig. 7 Approach to imaging in a patient presenting with VT. A stepwise framework to help guide cardiac imaging in a patient with VT.
Cardiac imaging in patients with VT is critical in both forming a diagnosis and management plan. Cardiac imaging can be used pre-procedurally and intra-procedurally to facilitate catheter ablation procedures. Utilizing cardiac imaging improves management decisions and clinical outcomes in patients with VT.
Saurabh Kumar has received honoraria from Biosense Webster, Abbott Medical, Biotronik, and Sanofi Aventis. Jonathan Kalman is supported by a National Health and Medical Research Council of Australia practitioner fellowship, and has received research and fellowship support from Biosense Webster, Abbott and Medtronic. Geoffrey Lee has received consulting fees and speaker honoraria from Biosense Webster. Other authors have no discloses.
The following is the Supplementary data to this article.