Multimodality Lead Positioning for Cardiac Resynchronization Therapy: How Much Imaging do we Need?

This article refers to ‘Multimodality imaging-guided left ventricular lead placement in cardiac resynchronization therapy: a randomized controlled trial’ by A. Sommer et al. and to ‘Cardiac resynchronization therapy guided by multimodality cardiac imaging’ by M. Bertini et al. published in this issue on pages 1365–1382. Cardiac resynchronization therapy (CRT) has been of great benefit to many heart failure (HF) patients with reduced ejection fraction (EF) and widened QRS complexes; however, response is variable. Patient response to CRT appears to be affected by the complex interaction of left ventricular (LV) electrical delay, regional discoordination and stretch, underlying myocardial substrate/scar, and pacing lead location (Figure 1).1-4 This issue contains two interesting studies that apply advanced multimodality imaging strategies to guide placement of the LV lead as a means to improve CRT response. The usefulness of cardiac imaging to optimize LV lead positioning and affect patient outcomes was first demonstrated by Khan et al.5 in the Targeted Left Ventricular Lead Placement to Guide Cardiac Resynchronization Therapy (TARGET) trial and Saba et al.3 in the Speckle Tracking Assisted Resynchronization Therapy for Electrode Region (STARTER) trial. These two randomized clinical trials were performed independently with nearly identical methods and nearly identical results. TARGET and STARTER both used speckle tracking radial strain with the latest segment of peak strain as the desired LV lead location and showed a clinical benefit compared with routine care. Specifically, STARTER demonstrated reductions in HF hospitalizations and deaths as the primary end-point and TARGET showed similar findings as a secondary end-point. Both studies also demonstrated an advantage of the imaging-guided strategy to promote LV reverse remodelling. There remains continued interest in further improving CRT response. In the first of these studies published in the present issue, Sommer et al.6 performed the ImagingCRT trial as a double-blind randomized controlled clinical trial of multimodality imaging-guided LV lead placement for CRT vs. routine care. They randomized 182 patients to routine care vs. multimodality imaging guidance, consisting of three imaging methods: cardiac computerized axial tomography (CT) for venography, 99mTechnetium myocardial perfusion imaging, and speckle-tracking echocardiography. Their specific approach was radial strain from mid-LV short axis view, similar to TARGET and STARTER. Myocardial perfusion imaging at rest was performed with 99mTechnetium-sestamibi single photon emission computed tomography (SPECT). Finally, helical high-pitch electrocardiogram (ECG)-gated cardiac CT was performed to visualize cardiac venous anatomy. From these multimodality data, the optimal coronary sinus branch closest to the non-scarred myocardial segment with latest mechanical activation was targeted. The control group positioned the LV lead using both fluoroscopic guidance directed toward non-apical segments of the posterolateral wall. In addition, control group lead placement was with a preferred electrical delay of ≥50% of the QRS duration, as measured intraprocedurally from QRS onset to the first sharp deflection of the sensed signal in the LV electrogram. The authors used a complicated primary endpoint of clinical non-response 6 months after CRT, defined as at least one of the following: (i) death, (ii) HF hospitalization, or (iii) lack of improvement in New York Heart Association class and <10% increase in 6-min walk distance. Secondary outcomes included LV remodelling and the combination of all-cause mortality and hospitalization owing to HF during 1.8 ± 0.9 years. Although the imaging-assisted group showed fewer clinical non-responders (26% vs. 42%, P = 0.02), there were no differences in LV reverse remodelling or clinical end-points of death or HF hospitalizations. Why did the multimodality ImagingCRT trial have different results from those of the TARGET or STARTER trials? Both TARGET and STARTER used echocardiography as a single modality and one would have anticipated that multimodality imaging should have shown a greater, rather than a lesser clinical impact. One of the likely reasons is that ImagingCRT only enrolled patients with left bundle branch block (LBBB), of which approximately 75% had wide QRS duration ≥150 ms.4, 7 Patients with these ECG criteria (LBBB >150 ms) alone have the highest probability of a favourable CRT response. A substudy of the STARTER trial demonstrated imaging-guided LV lead placement to have the greatest clinical benefit in patients with QRS 120–149 ms or non-LBBB morphologies.8 Furthermore, the control group in ImagingCRT had LV lead positioning by intraprocedural measurement of electrical delay, or QLV interval, which was not employed in the control groups of the TARGET or STARTER trials. Targeting the maximum QLV has been demonstrated to be associated with favourable LV reverse remodelling.9 Routine fluoroscopic LV lead placement without measurement of electrical delays remains the most commonly employed in clinical practice by most centres. In summary, these two major factors of patient selection by LBBB and QLV measurement in the control group made it more difficult to detect differences in CRT response in imaging-guided group versus control group. The potential favourable impact of multimodality imaging-guided LV lead placement in patients with QRS 120–149 ms or non-LBBB compared with routine fluoroscopic LV lead placement alone remains unknown. In the second study in this issue, by Bertini et al.,10 a series of 50 consecutive CRT patients were studied by multimodality imaging using speckle tracking longitudinal strain and cardiac magnetic resonance (CMR) imaging. The imaging-guided strategy was to target sites of latest time to peak longitudinal strain and avoid regions of scar identified by CMR late gadolinium enhancement (LGE). Echocardiographic longitudinal strain time-to-peak delay was determined from three apical views using a 16-segment model. A 1.5-T CMR system was used with LGE >75% of myocardial wall identified as scar. Unlike TARGET, STARTER, and ImagingCRT, this was not a randomized clinical trial. Patients with imaging-guided LV lead placement were compared with a group of 50 CRT patients from the previous year matched for age, gender and EF who had routine LV fluoroscopic lead placement. The authors reported that in 8% speckle-tracking was not feasible, and in 14% CMR was not possible owing to claustrophobia, severe renal dysfunction, or contrast allergy. The results showed that imaging-guided CRT patients had a greater proportion with LV reverse remodelling (LV end-systolic volume reduction of 15% or more) at 6 months (78% of imaging-guided patients vs. 56% in the comparison group; P = 0.019). Interestingly, they reported a significantly higher rate of super responders defined as having a LV end-systolic volume reduction of 30% or more (54% of imaging-guided patients vs. 32% of comparison group; P = 0.026). Advantages of this study are that CMR LGE is considered the gold standard for scarring, and that longitudinal strain may produce more stable signals than radial strain. The relatively small sample size, the use of a comparison group rather than a randomized design, and the lack of clinical outcome data combine to suggest that this is a pilot rather than a definitive study. Because the results are intriguing, a larger prospective randomized trial would be of great interest. There is mechanistic support and clinical evidence that LV pacing within a region of scar is mechanically disadvantageous with respect to pump function and is a potential source of genesis of ventricular arrhythmias.11-14 Accordingly, avoidance of LV lead placement within or adjacent to a region of LV scar is beneficial. Use of CMR with LGE is considered the gold standard to localize and quantify myocardial fibrosis because of its superior spatial resolution. The clinical practical limitation is that CMR services are usually only available at larger medical centres with specialized expertise. Furthermore, many CRT patients cannot be scanned because of existing implanted hardware, renal dysfunction where gadolinium is contraindicated, claustrophobia, or other reasons. A large proportion of patients who were not eligible for CMR have been reported in a recent large series.15 Imaging by SPECT is widely available to identify myocardial scar by its lack of radionuclide uptake. A disadvantage is the poorer spatial resolution of SPECT imaging, which may interfere with precise localization of scar. Another limitation is that potential ambiguities may exist in differentiating significant ischaemia from scar. Other limitations include the need for an additional imaging test and the costs and radiation exposure associated with SPECT. Echocardiographic strain imaging studies have demonstrated reduced amplitude as a marker for scar.13 However, speckle tracking strain imaging currently cannot confidently determine myocardial scar in several important scenarios. Although scar typically will have reduced amplitude of deformation, echocardiography often cannot reliably differentiate profound hypocontractility without scar from myocardial fibrosis. Accordingly, the most clinically applicable means currently to assess for myocardial scar remain CMR with LGE or SPECT imaging. The two new studies discussed demonstrate the potential of multimodality imaging-guided LV lead placement to improve CRT response. The TARGET and STARTER trials previously provided evidence that an echocardiographic strain-assisted strategy of mapping latest mechanical activation may provide clinical benefit over routine fluoroscopic lead placement. The less impressive differences in ImagingCRT, which studied only LBBB patients, may suggest that multimodality imaging guidance may have had more significant clinical results in patients with non-LBBB or with a QRS width of 120–149 ms, as shown before.8 It is impossible to individually assess the additive benefit of CT venography in ImagingCRT, although the current evidence for routine use of CT venography before CRT is not convincing, considering the disadvantages of cost, radiation, and exposure to contrast. Furthermore, the potential additive benefit of intraprocedural measurement of electrical delay to imaging guidance remains to be tested.9 The most evidence and support for a second imaging modality to be added to echo strain imaging mapping is CMR with LGE or SPECT to assess for scar, in particular for patients with ischaemic disease. The current state-of-the-art approach to a patient with ischaemic cardiomyopathy who is a candidate for CRT often includes either CMR with LGE or SPECT, in addition to echocardiography. In our experience, CMR or SPECT studies have usually been performed in CRT candidates as part of a revascularization or viability work-up.1 These two new studies combine with TARGET and STARTER to use speckle tracking strain echocardiography to demonstrate a benefit of imaging-guided LV lead placement for CRT. However, the reality is that few clinical centres are employing an imaging-assisted strategy for LV lead placement in CRT. A possibility for lack of adoption is the perception that speckle tracking strain imaging mapping is too complicated for routine patient care. Another possibility is that CRT patient subgroups that may benefit the most from imaging-guided LV lead placement, such as those with ischaemic disease or QRS width 120–149 ms or non-LBBB, have not been clearly established.8 Furthermore, proof is lacking of the incremental benefit of LV lead placement over other variables that have an impact on CRT response, namely differentiating the electromechanical substrate responsive to CRT from scar burden and irreversible myocardial disease.2 The 12-lead ECG is considered by many as the only diagnostic information needed to determine CRT candidacy and predict response.4, 7 Evidence continues to accumulate that patient selection by ECG alone is overly simplified. Improvement in CRT response with critically important information provided by multimodality cardiac imaging continues to evolve and future enhancements are anticipated. Conflict of interest: J.G. receives research grant support from: Medtronic, Biotronik, Toshiba, and Hitachi-Aloka.