|Home | About | Journals | Submit | Contact Us | Français|
Endovascular recanalization (guidewire traversal) of peripheral artery chronic total occlusion (CTO) can be challenging. X-Ray angiography resolves CTO poorly. Virtually “blind” device advancement during X-ray-guided interventions can lead to procedure failure, perforation and hemorrhage. Alternatively, magnetic resonance imaging (MRI) may delineate the artery within the occluded segment to enhance procedural safety and success. We hypothesized that real-time MRI (rtMRI) guided CTO recanalization can be accomplished in an animal model.
Carotid artery CTO was created by balloon injury in 19 lipid overfed swine. After 6–8 weeks, two underwent direct necropsy analysis for histology, three underwent primary X-ray-guided CTO recanalization attempts, and the remaining 14 underwent rtMRI-guided recanalization attempts in a 1.5T interventional MRI system. rtMRI intervention used custom CTO catheters and guidewires that incorporated MRI receiver antennae to enhance device visibility.
The mean length of the occluded segments was 13.3 ± 1.6cm. rtMRI-guided CTO recanalization was successful in 11/14 swine and only 1/3 swine using X-ray alone. After unsuccessful rtMRI (n = 3), X-ray-guided attempts also were all unsuccessful.
Recanalization of long CTO is feasible entirely using rtMRI guidance. Low profile clinical-grade devices will be required to translate this experience to humans.
Endovascular recanalization of chronic total arterial occlusion (CTO) is challenging under conventional X-ray guidance because devices are advanced almost blindly. MRI can image CTO borders and luminal contents, and could potentially guide these procedures. We test the feasibility of real-time MRI guided wire traversal in a swine model of peripheral artery CTO using custom active MRI catheters.
Chronic total occlusion (CTO) of peripheral arteries can cause disabling intermittent claudication or critical limb ischemia. Surgical revascularization offers symptom relief and limb salvage in eligible patients; however, perioperative morbidity, lengthy hospital stay and slow functional recovery limit wider application1, 2. Endovascular techniques are appealing because they offer prompt return of function and require neither general anesthesia nor extended hospital stay. However, endovascular recanalization of long, tortuous, occluded peripheral arteries remains challenging and risky, and thus, surgical or conservative management are typically recommended1.
X-Ray fluoroscopic angiography with radiocontrast identifies the occluded inflow artery and possibly collateral-dependent outflow beyond the occlusion, but cannot discriminate CTO arterial wall and lumen. In these cases, the entire occluded arterial segment remains invisible to the operator. Virtually “blind” device manipulation risks procedural failure and arterial perforation. Furthermore, ionizing radiation and nephrotoxic contrast exposure may become excessive during lengthy CTO recanalization procedures.
Magnetic resonance imaging (MRI) can image the entire peripheral artery occlusion including arterial wall, lumen content, and adjacent structures, without ionizing radiation or iodinated radiocontrast. Advances in rapid imaging, combined with development of catheter devices visible under MRI, have made real-time MRI (rtMRI) guided therapeutic interventions feasible3–19. We hypothesize that rtMRI can guide catheter navigation within an occluded artery, and that combined CTO recanalization and percutaneous transluminal angioplasty can be conducted entirely under rtMRI guidance using custom MRI-visible devices in a swine model.
Animal protocols were approved by the National Heart, Lung, and Blood Institute Animal Care and Use Committee. NIH mini-swine (mean weight, 50 ± 12 kg) were fed 2% high cholesterol diet for 1 week prior to CTO model creation, and thereafter until recanalization was attempted 6 to 8 weeks later. After transfemoral access, an angioplasty balloon (Agiltrac, Guidant) sized 1.5 to 2 times artery diameter was positioned in the mid left carotid artery (LCA), inflated to nominal pressure, and rapidly withdrawn up to six times. Contrast-enhanced magnetic resonance angiography (CEMRA) was performed at 2 weeks to confirm CTO. If the LCA remained patent (n = 4), balloon injury was repeated.
Histology was performed on two uninstrumented CTO arteries, 8 weeks after balloon injury. Formalin fixed, paraffin embedded, transverse sections (4 μm) were taken at 5 mm intervals along the artery length, and stained with Hematoxylin and Eosin and Masson’s Trichrome.
Procedures were performed in a previously described combined X-ray-MRI interventional suite (Axiom Artis/1.5T Sonata, Siemens) 20. In-room hemodynamics, scan control and rendered images were displayed using shielded LCD projectors. Optical microphones (Phone-Or) and filtered headsets (Magnacoustics) permitted staff communication.
Digital subtraction X-ray and CEMRA of the arch and great vessels were performed immediately before and after intervention. Typical parameters [TR = repetitition time (ms) /TE = echo time (ms) /FA = flip angle (degrees) /BW = bandwidth (Hz/per pixel) /VOX = voxel size (mm)]: [TR/TE/FA/BW/VOX = 3.7/1.4/25/375/0.3x0.5x2, slab 300x244x128mm]. Real-time steady state free precession (rtSSFP) and 2D T1 weighted (T1w) spin echo axial images representing key anatomic planes were stored for uninterrupted retrieval during the intervention. Typical parameters:[rtSSFP: TR/TE/FA/BW/VOX = 3.5/1.7/45/975/1.0x1.4x4] and [T1w:TR/TE/FA/BW/VOX = 800/12/180/130/0.5x0.5x3, 2 averages].
During rapid device manipulation in the aortic arch, rtSSFP frame rate was accelerated 3:1 using echo-sharing wherein image data are interleaved over adjacent frames. Saturation pre-pulses suppressed background tissue to better visualize inflation of balloons filled with gadopentate dimeglumine (0.8% Gd-DTPA, Magnevist, Berlex) and of selective intra-arterial angiography (2.4% Gd-DTPA). Uninterrupted real-time subtraction angiography was achieved by displaying the difference between maximum intensity projections in time before and after the injection. Voxel size was reduced to increase spatial resolution, and temporal image filtering (averaging) was applied to improve signal-to-noise ratio (SNR) during slow device advancement within the CTO. Multiple oblique slices could be acquired simultaneously, repositioned interactively and individually disabled and enabled as desired. Custom scan features allowed adjustment of pulse sequence, reconstruction and display parameters without scan interruption. Finally, signals from separate device antennae could be independently displayed along their entire length, color-highlighted and combined in real-time with anatomic images from surface coils.
Active devices were prepared and tested in our laboratory. The intent was to create devices that were conspicuous along their entire length, including the tip.
Distal 1cm microcoils were mounted on 6Fx100cm angled, tapered, custom tungsten-braided catheters (Minnesota MedTec). Coil signals are transmitted along the catheter shaft under insulating polyimide. Tuning, matching and decoupling circuitry were attached at the proximal hub, in order to prevent heating during MRI. The device was connected to a separate MR scanner receiver channel, so that device-related images could be displayed in color.
Consecutive gold-silver-gold-plated nitinol core wires were inserted within nitinol hypotubes, separated by polymer insulation. Closely wound MP35N (cobalt-chromium alloy) microcoils were attached distally to add wire flexibility, and 2.5cm distal microcoils enhanced MRI receiver sensitivity at the tip. The final receiver-coil guidewire was 0.032” x 155cm, straight, and stiff. The proximal end of the guidewire had detachable connectors permitting catheter exchange.
Flexible tip nitinol guidewires (Nitrex, ev3, 0.032” x 155 cm) were altered to serve as active, “loopless” design21 receiver coils that were visible along their entire length. These were used as “workhorse” wires to safely deliver the CTO catheter to the origin of the occluded carotid artery.
10F delivery sheaths (Fast Cath, St. Jude; Brite Tip, Cordis) were trimmed to 75cm. Stainless steel MRI markers (1mm x 0.014 inch) were bonded to the sheath tip creating passive markers (discrete signal void) during MRI.
The primary study endpoint was successful traversal of occluded arteries from their patent origins to their patent outflow regions. The secondary study endpoint was restoration of antegrade flow by adjunctive balloon angioplasty.
Animals were pretreated with aspirin 325 mg and clopidogrel 300mg. Intravenous heparin 100U/kg bolus was given initially and supplemented with 50U/kg every 2 hours. Following sheath insertion, the active CTO support catheter was navigated to the brachiocephalic artery over the active floppy guidewire and positioned proximal to the occluded LCA. The floppy wire was exchanged for the active CTO wire. The CTO guidewire and catheter were steered intraluminally through the CTO, using multi-slice rtMRI. Two parallel slices orthogonal to the CTO artery centerline were used to ensure an intraluminal device course. One slice tracked the wire tip, and the second advanced ahead of the wire tip providing a target for the operator. Once the CTO was traversed, a 260cm PTFE guidewire (Glidewire, Terumo) was used to exchange the CTO guide catheter for a 100mm long peripheral angioplasty balloon sized 1:1 to the artery wall. Following several balloon inflations with 5mM Gd-DTPA, selective angiography was performed using 5mL of 10mM Gd-DTPA. Intra-arterial nitroglycerin 200–400mcg was administered as needed to treat spasm.
X-ray guided recanalization was attempted in three swine, by experienced interventionists to assess the suitability and procedural difficulty of the CTO model. Contemporary techniques and devices were used (Shinobi 0.014”, Cordis; Cross-It 0.010–0.014”, Guidant; Glidewire Straight Stiff and Super-Stiff 0.035”; 4F Glidecath, Terumo). Each attempt used high-performance, X-ray imaging (Artis FC, Siemens) with 13cm field of view, 15fps, appropriate collimation, and at least 60 minutes of fluoroscopy. In three additional swine, after unsuccessful MRI-guided recanalization, X-ray-guided recanalization also was attempted.
Continuous variables were reported as mean ± standard deviation and were compared using a two-tailed Student’s t-test (Excel 2003, Microsoft). Discrete variables (procedure success) were compared using Fisher’s exact test24. The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
No significant heating was observed in phantom experiments during MRI of active CTO support catheters and guidewires (Appendix).
Native LCA diameter by X-ray was 4.6 ± 0.7mm before balloon injury, measured 70mm cranial to the carotid bifurcation. Immediate pre-intervention CTO diameter (T1w) at this level, and CTO length (CEMRA) were 4.7 ± 1.3mm and 133 ± 16mm. By comparison, the contralateral carotid artery diameter (T1w) at this level, the CTO minimal luminal diameter(T1w) and CTO maximal external wall diameter(T1w) were 5.0 ± 0.7mm, 3.1 ± 1.3mm, and 6.7 ± 1.2mm respectively. These measurements suggest zones of positive (outward) and negative (inward) remodeling in this swine CTO model and are consistent with histology. MRI identified the entire CTO artery wall endoluminal and external boundaries, and occasional proximal dissection flaps resulting from initial balloon injury (Figure 2). Histology of the explanted, naïve CTO arteries revealed dense luminal collagen deposition, disrupted intima, microchannels and distal organized thrombus (Figure 3).
The primary endpoint, rtMRI-guided CTO recanalization, was achieved in 11/14 animals. The shaft profile and tips of the active devices were conspicuous while navigating the aorta and brachiocephalic arteries (Figure 4). The occluded artery walls and active interventional devices were easily visualized under MRI (Figure 5). Multi-planar imaging facilitated continuous assessment of device-anatomic relationships enabling intraluminal device traversal. Mean time to wire traversal was 55 ± 22 min (spread 25–99 min), with an observed learning curve between the first and last third of interventions performed (69 ± 23 min vs. 34 ± 7 min, p = 0.048). After balloon angioplasty, real-time selective MRA demonstrated brisk antegrade flow in 6/11 of those successfully recanalized with a guidewire (Figure 6). Gross necropsy of all successfully recanalized animals revealed mild artery thickening with no extramural hematoma (Figure 7) and appeared similar to the uninstrumented CTO lesions.
In 3/14 animals, the active guidewire would not advance beyond the proximal one-third of the occluded segment despite an apparently intraluminal trajectory. Following careful MRI interrogation to rule out any adverse sequelae such as hematoma, X-ray-guided attempts in all 3 were similarly unsuccessful.
In 2/11 successfully recanalized under rtMRI, antegrade blood flow was not achieved despite intra-arterial nitroglycerin and several prolonged balloon inflations, due to refractory recoil, spasm, and unavailability of suitable stents. Intraluminal position was confirmed by selective real-time MR and X-ray angiography through the wire port of balloons positioned distal to the lesion.
Home-made active CTO support catheters would not advance over active CTO guidewires that successfully traversed the occlusions in 3/11, preventing balloon exchange and subsequent angioplasty. In one of these three (the very first experimental animal), over-aggressive attempts at advancing the active catheter over the wire resulted in clear real-time visualization of extraluminal CTO wire exit followed by extravascular hematoma accumulation. This complication was immediately recognized in rtMRI, and the procedure was aborted.
Primary X-ray guided CTO recanalization was successful in only one of 3 different animals attempted (requiring 45min of fluoroscopy time). Cross-over X-ray recanalization was also attempted in 3/14 animals in after unsuccessful recanalization under rtMRI; none were successful. Combining these, 1/6 X-ray guided procedures were successful.
Successful CTO recanalization was more likely (79% vs. 17%, p = 0.02) during rtMRI attempts (11/14) compared with all X-ray-guided attempts (1/6). During extended primary X-ray recanalization attempts, wire exit caused contrast extravasation in one animal, and dissection was associated with lethal mediastinal hematoma in another.
Subtraction real-time MR angiography improved visibility of intra-arterial Gd-DTPA injections compared with saturation preparation alone (Figure 6E and F). We increased spatial resolution by reducing the voxel size but this degraded the signal-to-noise ratio (SNR). We used temporal image filtering (averaging) to improve this reduced SNR during periods of slow device motion.
Percutaneous peripheral artery CTO intervention can be highly challenging, time-consuming and risky. Failure rates for iliac and femoropopliteal CTO are 3–36% 25–33 and 12–25%34, 35 respectively. Novel alternative devices and techniques have had varied success for coronary and peripheral artery CTO recanalization, though to date none have prevailed clinically7, 36–43.
This experience demonstrates successful wholly rtMRI-guided endovascular recanalization of long and challenging CTO using custom devices, in a suitable animal model. In this model, MRI was successful in 79% of attempts using homemade devices, and in 17% of all X-ray attempts using higher-performance commercial devices.
The model CTO artery diameter and length resemble long (albeit rigid) human iliofemoral artery occlusions. These lesions would be considered TASC D or best suited for surgical rather than endovascular recanalization1. Unlike X-ray, MRI delineates the occluded artery walls permitting truly image guided wire navigation. Custom “active” MRI CTO devices prove conspicuous and distinct from adjacent tissues. Procedure duration declined throughout this experience due to a “learning” phenomenon.
Device conspicuity, simultaneous display of multiple real-time imaging slices, and uninterrupted interactive features contributed to procedural success. Real-time selective MRA techniques using small volumes of dilute Gd-DTPA confirmed artery patency. This work represents an advance because a broad assortment of imaging and device features are fully integrated to support complex and completely rtMRI-guided vascular interventions.
Long conductive devices may heat significantly during MRI radiofrequency excitation, and pose a safety concern. The custom “active” devices used here are appropriately tuned, matched and decoupled and do not heat during scanning in a suitable test phantom (see Appendix).
Our CTO model differs from tortuous, lipid and calcium-rich atherosclerotic human iliofemoral disease. Severe arterial recoil and spasm prevented antegrade flow (n = 2) or guiding catheter traversal (n = 3) despite successful wire recanalization. The lesions were sufficiently long and challenging for a feasibility experiment. Of note, X-ray-guided recanalization attempts were less successful, even though they employed higher-performance commercial clinical devices. Indeed, MRI-guidance offers the potential to visualize the entire occluded segment and luminal boundaries, particularly in severely tortuous lesions often found in human peripheral CTO.
This experience adds to that of other published reports of rtMRI-guided procedures that traverse anatomic boundaries such as transjugular intrahepatic portosystemic shunt (TIPS)14, 44 and atrial transeptal puncture16, 19. CTO recanalization is a more complex procedure because the vascular lumen is smaller, the diseased vessels tend to be tortuous, and the risk of lethal performation is higher when manipulating catheters within the high pressure arterial system. These support the principle that MRI has potential utility to guide conventional and novel catheter-based interventions because of superior tissue visualization.
Because the SNR of proton MRI is very low, there is inadequate temporal and spatial resolution to image coronary arteries in real-time using MRI. Barring unforeseen technical breakthroughs, it is unlikely that such recanalization procedures can be conducted on occluded coronary arteries.
The comparison of X-ray and MRI-guided recanalization attempts should be interpreted with caution because of the imbalance in the number of animals in each group, and because primary and cross-over attempts were combined for analysis.
Repeated use of recycled devices, abrupt transition points, high friction and poor mechanical responsiveness account for failure to recanalize and deliver catheter devices into the target arterial occlusion in a few animals. Our home-made devices had inferior pushability, torquability, and tip-bending performance compared with clinical X-ray devices in bench testing (data not shown). Low-profile, tapered, trackable, steerable, and hydrophilic-coated active MRI devices can be manufactured using available technology, but would have been prohibitively expensive to manufacture in the small quantities tested here. In addition, suitable stents would have treated flow-limiting artery recoil and spasm, but were not employed here. Metal alloys such as nitinol, platinum-iridium, or nickel-cobalt-chromium have less MRI susceptibility artifact compared with stainless steel. Numerous reports already have been published of experimental rtMRI-guided deployment of such devices7, 17, 18, 45. Nevertheless, our primary endpoint, rtMRI-guided CTO recanalization was successfully accomplished using our rudimentary device prototypes. Subsequent therapy to maintain antegrade flow and patency such as stenting could be conducted under X-ray, immediately after successful rtMRI CTO recanalization.
This work demonstrates the feasibility of rtMRI guidance to recanalize arterial CTO using custom “active” MRI visible devices in a suitably challenging swine model. Future image-guided human peripheral artery CTO interventions may be possible using this technology. However, clinical-grade catheter devices will be need to be manufactured to translate these findings into patients.
Traditionally, surgical bypass is preferred over X-ray guided endovascular repair for long, tortuous CTO. Endovascular guidewire traversal under X-ray is conducted almost “blindly” through occluded vessels that are not opacified by contrast. The guidewires and catheters required to recanalize these lesions risk failure, perforation and life-threatening hemorrhage. We demonstrate successful real-time MRI-guided recanalization of a long peripheral artery CTO in a suitable pig model. Real-time MRI may offer an advantage in artery visualization during CTO traversal, while obviating ionizing radiation and nephrotoxic radiocontrast.
The authors thank Kathryn Hope, Kathy Lucas and Joni Taylor for their animal care and support. Supported by NIH Z01-HL005062-03 (RJL).
Journal Subject Head for Circulation
 Endovascular coronary and valvular interventions: other
 Computerized tomography and Magnetic Resonance Imaging
 Animal models of human disease
Conflicts of Interest Disclosures