In patients with severe aortic stenosis (aortic valve area <1cm2 or <0.6cm2/m2) and high surgical risk (haemodynamic instability or significant co-morbidities), surgical aortic valve replacement is often rejected and, unfortunately, balloon valvuloplasty cannot provide a sustained improvement. Percutaneous heart valve (PHV) implantation is an excellent alternative treatment in such cases. The first successful human PHV implantation was performed in April 2002 by Cribier’s group at Rouen University Hospital in France, via the antegrade (venous) transseptal approach.1 Because of the complexity of the antegrade approach, several changes were introduced to implantation protocols.
Subsequently, the arterial transfemoral and transapical delivery routes were developed.2,3 The transapical approach is the most recently developed technique for transcatheter aortic valve replacement. This procedure involves a limited left lateral thoracotomy and requires a direct cardiac puncture and sheath insertion into the left ventricle. The transfemoral retrograde procedure is performed under local anaesthesia and mild sedation. The common femoral artery is first exposed by the surgeon, then catheterised. After retrograde catheterisation of the aortic valve balloon, pre-dilation of the aortic valve is performed. The femoral artery is then pre-dilated with a series of dilators of increasing size to facilitate the entry of the sheath. The PHV is advanced over the extra-stiff guidewire, placed within the native valve, and deployed using rapid ventricular pacing.4 The device used at our institution is an Edwards SAPIEN transcatheter aortic prosthesis (Edwards Lifescience, Irvine, CA, US) mounted on a balloon-expandable stainless steel stent. The trileaflet bovine pericardial prosthesis is attached to the stent and coated with an anticalcification treatment. The stent has a polyethylene terephthalate fabric skirt that reduces perivalvular leaks (see Figure 1). Owing to its high procedural success rate and the simplicity of the delivery technique, since 2005 the transfemoral antegrade approach has become the most commonly used procedure.4 This article will therefore examine this technique only.
Accurate assessment of the iliofemoral arteries prior to valve implantation is mandatory with this technique. Multidetector-row computed tomography (MDCT) is helpful since it offers detailed imaging of calcifications of the arterial wall and depicts vessel tortuosity in any plane with high spatial resolution. This article will describe the role of MDCT in planning and performing PHV implantation. Its contribution after implantation is still under evaluation.
Multidetector Computed Tomography Before Percutaneous Heart Valve Implantation
CT works as a complement to echocardiography and conventional angiography. At our institution, the examination comprises two acquisitions: a cardiac assessment of the aortic valve immediately followed by abdominal and pelvic CT angiography, which is essential for assessing the retrograde route of the device. Other authors suggest performing only one acquisition.5
Review of the Technique
At our institution, MDCT is performed on a Discovery 750HD 64-row MDCT scanner (GE Healthcare, Milwaukee, WI, US) and the axial data sets of images are sent to the Advantage Workstation version 4.5 (GE Healthcare, Milwaukee, WI, US), which is equipped with commercial software for cardiac and iliofemoral analysis. CT angiography of the aorta and iliofemoral arteries is almost always feasible even in tired, elderly patients. The cardiac examination can be more difficult to perform for some patient-related reasons (difficulties with holding the breath, deafness, tachycardia or tachyarrhythmia).
A good MDCT examination requires high-quality vein access. Renal failure is frequent in this population, but the risk posed to the kidneys by the use of iodinated contrast agent needs to be balanced against the natural history of the aortic stenosis if left untreated. Estimating and following up the creatinine clearance using the modification of the diet in renal disease (MDRD) equation is essential. At our institution, we try to space the injections and limit the volume and concentration of contrast agents. Patient hydration and administration of bicarbonates are well-known renal-protection techniques that can be implemented. Radiation dose control is not the first priority in this elderly population with high-risk disease. However, the rules of good practice are always applied (collimation, minimal time and zone exposure), and adaptive statistical iterative reconstruction (ASiR) is routinely used to lower the dose while maintaining equivalent image quality.
A comfortable position on the table, clear and controlled explanations, and apnoea tests prior to acquisition are other factors that help to guarantee the quality of the examination. The patients never receive beta-blockers because of concerns regarding the critical aortic stenosis. Pre-operative explorations are confronted with and validated by clinical data from other imaging modalities during a multidisciplinary meeting with interventionalists, radiologists and surgeons.
Cardiac Exploration
A collimation of 64 x 0.625mm and a rotation of 35ms are used. The tube voltage is usually set at 120kVp because of calcium accumulations, except in patients with a low body mass index (BMI), in whom 100kVp is used. Patients are scanned with retrospective gating with a similar field of view (FOV) to that used for a coronary artery examination; however, the FOV needs to be extended when mammary bypass grafts are present. As the entire systolic–diastolic motion of the aortic valve needs to be analysed, electrocardiogram (ECG) dose modulation is avoided. The tube current is manually set at between 450 and 550mA. About 70–80ml of iodine contrast material is administered, then flushed by saline. The flow rate is typically set at 5ml/s.
Peak enhancement detection of the ascending aorta is essential because of varying haemodynamic parameters and different degrees of aortic stenosis among these patients. Data acquisition is performed during an inspiratory breath-hold. Reconstruction covers at least 10 phases of the cardiac cycle (RR interval), allowing analysis of the entire systolic–diastolic set and assessment of the left ventricular ejection fraction as well as mitral motion. Filters adapted for highly attenuating structures are recommended, as is a high-resolution reconstruction.
CT evaluation of the aortic root begins with an assessment of the aortic valve calcification: intensity, topography and extension to the interventricular septum and aortic valve. At our institution we routinely use the classification suggested by Willmann et al.,6 as follows (see Figure 2): grade 1, no calcification; grade 2, mild calcification (small isolated spots of calcification); grade 3, moderate calcification (multiple larger spots of calcification) and grade 4, heavy calcification (extensive calcification of all aortic valve leaflets). The distance between the annulus and the coronary ostiae is measured, as is the diameter of the annulus, the sinus of Vasalva, the sinotubular junction and the ascending tubular portion (see Figure 3). Measurement of the systolic aortic valve opening may be possible depending on the imaging quality. This measurement can be taken during systole at the summit of the aortic cone. Even if the severity of the stenosis is assessed from echocardiography and derived from flow measurements, MDCT will be able to measure the maximal valve opening with acceptable accuracy.7 The course and permeability of the coronary artery bypass grafts (frequent in this population) are also analysed.
Angiography of the Aorta and the Iliofemoral Arteries
Performed after a minimal delay (tube cooling) and taking advantage of the residual opacification of cardiac imaging, iliofemoral angiography requires a second injection of about 50ml of contrast medium at a rate of 4ml/s flushed by saline at a rate of 3.5ml/s. The FOV is large and covers the abdominal–pelvic region from the diaphragmatic hiatus to at least 3cm under the femoral tripod. Vascular reconstruction requires close attention since no blind segments can be accepted. The reconstruction is composed of luminal and short-axis views of the whole vessels. A 3D view (including anteroposterior, lateral and oblique views) of the iliofemoral arteries allows the sinuosity, angulations and global calcification of the vascular axis to be analysed, as shown in Figure 4.
At our institution, when uni- or bi-lateral hip prosthesis is present we prefer a dual-energy CT examination, which consists of a dynamic switching acquisition between two different energy levels of X-rays (80 and 140kVp). The dual-energy data are processed using Gemstone Spectral Imaging (GSI GE Healthcare, Milwaukee, WI, US) to generate ‘quality-check’ images from high-kVp data (140kVp) and monochromatic images with a metal artifact reduction algorithm. This technique allows better visualisation and analysis of the iliofemoral artery adjacent to the metal implant. Wall vessel depiction and automatic vessel analysis are also improved compared with a conventional acquisition (see Figure 5). On short-axis views, the arterial puncture segment is analysed in terms of calibre and calcification. Then, by simulating the passage of the prosthesis on the luminal view, we look for a critical calibre zone (based on the minimal and average diameter). This segment is optimally analysed in its short axis. A critical calibre with a non-calcified vessel (or with a punctiform calcification) is less problematic than a vessel with stenosis and a circumferential or mirror calcification (see Figure 4).
The ideal patient for PHV implantation will have large-calibre iliofemoral arteries free of calcification. The radiologist’s role is ultimately to infer the arterial plasticity from CT data, including calibres, calcification and tortuosity. Focal arterial dissection or aneurysms are not rare, and are considered contraindications to catheterisation. The 23mm PHV is compatible with a 22Fr sheath, and the minimum diameter of the iliofemoral arteries is 7mm. For the 26mm PHV (24Fr sheath), the minimal diameter of the iliofemoral arteries is 8mm.4
Non-cardiovascular Findings
Screening MDCT examinations can often reveal incidental non-cardiovascular findings such as renal or lung cancer, liver and gallbladder disease, inguinal hernia or other disease. These findings should be reported and taken into consideration in the selection of patients for PHV implantation.
Evaluation Following Prosthetic Heart Valve Implantation
MDCT is of potential interest for evaluating the correct positioning and sizing of the Cribier-Edwards transcutaneous PHV.8 The accurate positioning of the bioprosthesis and its relationship with the annulus, coronary ostiae, interventricular septum and mitral valve can be demonstrated by volume imaging, as can the position of the displaced native calcifications (see Figure 1). CT is used as a complement to echocardiography, which remains the reference technique to identify aortic regurgitation.9 Patients who have undergone valve replacement are imaged with a retrospectively gated cardiac CT protocol similar to that used for pre-operative assessment.
The Role of Cardiac Magnetic Resonance Examination
Magnetic resonance (MR) is not recommended when planning vascular access analysis because of its limited spatial resolution. Furthermore, MR examination is difficult to perform in elderly patients who have difficulties supporting a prolonged decubitus and repeat multiple breath-holds. At our institution we perform MR examinations to define the anatomy (bicuspid versus tricuspid aortic valve) and maximal systolic opening in discordant cases. MR is often more efficient than CT for assessing systolic aortic valve area and bicuspid valvular anomalies because it is not disturbed by calcifications and has a better temporal resolution. Phase-contrast sequences are mostly used in this area. Another area where MR may be of interest is the identification of fibrosis from delayed-enhancement sequences.10 MR is not adapted for post-implantation evaluation because the stent is a major source of artifacts. Nevertheless, the Edwards prosthesis is compatible with MR, and MR examination of other organs is not contraindicated.
Conclusion
PHV implantation has the potential to offer life-saving treatment to patients with inoperable severe aortic stenoses. MDCT plays a key role in patient selection and evaluation. Not only radiologists but also general physicians, interventionalists and surgeons should share and co-ordinate knowledge, skills and clinical wisdom to ensure the successful future development of PHV implantation.