Two leads are placed inside the heart during a typical CRT procedure: one in the right ventricle (RV) and one in the left ventricle (LV). Implanting the LV lead in an optimal position, however, is known to be particularly problematic using current X-ray fluoroscopic guidance, as the lead must be guided through the heart's coronary sinus and into one of its sub-branches.
Unfortunately 30% of patients do not respond to CRT, and sub-optimal LV lead placement is thought to be a contributing factor. Now, researchers in the UK have teamed up with Philips Healthcare in a bid to improve this situation. The team has developed a complete imaging platform that firstly overlays pre-procedural MR data on top of live X-ray fluoroscopic images and secondly displays the complete anatomy of the LV in two dimensions on a screen during CRT to aid the placement of the crucial LV lead (Phys. Med. Biol. 57 2953).
"Our method gives the cardiologist both the structural and functional information they require to best place the LV lead," Ying Liang Ma from King's College London told medicalphysicsweb. "Functional information such as cardiac wall motion and cardiac scarring can be combined with anatomical information to guide the implant to target late contraction and avoid scar tissue. We also allow the whole of the LV to be visualized at once with all the corresponding functional information required to achieve successful placement."
Testing the platform
The group tested its platform on 21 patients undergoing CRT in a standard catheterization laboratory. Each patient underwent cardiac MRI prior to their CRT procedure in order to extract the exact morphology of their heart (and in particular, the LV and the coronary venous system), the regional motion of their heart and the position of any regions of scarring.
All of this information was then used to generate a single 3D model of each patient's heart and identify the optimal position for the LV lead. This 3D "roadmap" of the patient's heart was then overlaid on the live X-ray image during the CRT procedure, with alignment being maintained automatically by tracking the diaphragm in the X-ray images. The team reported a mean registration error of 1.2±0.7 mm.
A further step, and a unique aspect of this work, was to produce a 2D representation of this 3D roadmap through a process that the team calls cardiac unfold. "The information contained in the 3D model and the unfold is identical and both are used to augment the conventional X-ray pictures during the CRT procedure," explained Ma. "Unfolding the LV is a common practice within cardiology. However, for the first time, we combine the unfold with live device tracking to enable this representation to be used for procedure guidance."
Following this successful trial, the team is now carrying out a larger scale multi-centre randomized clinical trial to see if the technique can make a clinical difference. According to Kawal Rhode, a senior lecturer in image processing at King's who is leading this work, the new trial will involve nearly 300 patients with half having conventional CRT and the other half undergoing image-guided CRT.
The long-term objectives, however, are to commercialize the technology and develop a standard CRT guidance product. "The underlying software was implemented in conjunction with engineers at Philips Healthcare in The Netherlands," commented Rhode. "The anticipation is that this technology will be commercialized through this partnership so that it can be available in any cardiac catheterization laboratory."