Introduction

Percutaneous pulmonary valve implantation (PPVI) is a nonsurgical option for treating right ventricular (RV) outflow tract and/or pulmonary trunk dysfunction, representing one of the most exciting approaches of interventional cardiology as the first in-human application of this technique [1]. In particular, Bonhoeffer et. al. [2] developed a new approach to treat RV conduit dysfunction using a bovine valve sutured on a platinum-iridium alloy stent, with welded joints of gold (Melody, Medtronic, Inc., Minneapolis, MN, US) [2].

Being less invasive than surgical pulmonary valve replacement, PPVI avoids the risks of ventriculotomy and cardiopulmonary bypass as well as the risks of bleeding and infections associated with reoperation, avoiding postoperative intensive care reduces costs. The procedure would have to provide feasibility and safety comparable to those of surgical procedures, which are burdened by conduit failures and require multiple re-interventions [3]. The most common indications for surgical pulmonary valve replacement are pulmonary regurgitation with or without stenosis, RV dilation and hemodynamically significant residual lesions [1].

The majority of patients referred for surgical pulmonary valve replacement or PPVI have a diagnosis of tetralogy of Fallot with pulmonary stenosis or pulmonary atresia or a previous surgical treatment for critical pulmonary valve stenosis [4–6]. In these patients, after repairing, the chronic RV volume overload leads to ventricular dilation and impairment of systolic and diastolic function. In the long term, this causes atrial and ventricular arrhythmia as well as heart failure and increases the risk of sudden death [7]. Restoration of pulmonary valve competence at an appropriate time has resulted in improvement in RV function and effort tolerance and reduces the incidence of arrhythmias [8, 9]. Thus, patient selection and correct timing for PPVI plays a key role in the success of this technique.

Follow-up after PPVI may be performed using echocardiography as a first-line technique [10]. This modality is relatively inexpensive, widely available, and is not associated with ionizing radiation exposure. However, its limitations include poor acoustic windows in assessing RV morphology, function and valve regurgitation [11]. In recent years, studies have been reporting the use of cardiac magnetic resonance (CMR) to evaluate patients before and after PPVI [12–14]. CMR has been demonstrated to outperform echocardiography in the evaluation of cardiac chambers, especially the RV, and it is considered as a first-choice technique in this setting [15–17].

Our aim was to report the 4-year CMR follow-up of patients treated with PPVI.

Materials and methods

Study design and population

After institutional review board approval waived specific patient consent for data analysis, we retrospectively retrieved prospective reports of CMR examinations performed before and after PPVI from January 2008 to June 2014 in 40 consecutive patients treated with a Melody valve stent (Medtronic, Inc., Minneapolis, MN, US). Demographics, age, and disease spectrum are shown in Table 1. CMR had been planned at least 1 month before PPVI, and 1, 3, 6, 12, 24, 36, and 48 months after PPVI. Paediatric cardiologists obtained informed consent for the intervention and use of this type of valve for PPVI. Informed consent was obtained by each patient for the injection of gadolinium (Gd)-based contrast material, when administered.

Table 1 Patient characteristics, disease spectrum, and conduit type
Full size table

CMR protocol

We used a 1.5-T unit with 40-mT/m gradient power (Magnetom Sonata Maestro Class, Siemens, Erlangen, Germany) and a four-channel cardio-thoracic phased-array coil. The same imaging protocol was used before and after PPVI. The mean CMR duration was 45 ± 10 min.

Electrocardiographically-triggered short-axis, four-chamber and long-axis images were obtained with a breath-hold steady-state free precession (SSFP) sequence (true fast imaging with SSFP, true-FISP) acquired with the following technical parameters: TR/TE, 4.0/1.5 ms; FA, 80°; slice thickness, 8 mm; temporal resolution, 45 ms; mean acquisition time, 14 ± 4 s. A breath-hold turbo spoiled gradient echo sequence (fast low-angle shot, FLASH) was performed for phase-velocity mapping of pulmonary flow with the following technical parameters: TR/TE, 4.0/3.2 ms; slice thickness, 5 mm; velocity encoding (VENC) from 150 to 350 ms; temporal resolution, 41 ms; mean acquisition time, 15 ± 4 s. Firstly we acquired the phase-contrast sequence with a VENC of 150 ms; in the presence of aliasing, we modified the VENC adding 50 ms for each new sequence, step-by-step up to the complete disappearance of the aliasing artefact.

Before PPVI, the slice of the phase-contrast sequence was located on the pulmonary conduit, perpendicular to the vessel long axis, immediately distal to the valve; after PPVI, the slice was located at least 4 mm distal to the stent artefact to avoid contamination of images from the metallic artefact. Mr angiography with a three-dimensional gradient-echo fast low-angle shot sequence (FLASH, TR = 3.32 ms, TE = 1.21 ms, flip angle 25°, slice thickness = 1.5 mm, 72 partitions, acquisition time 21 s) after injection of 0.1 mmol/kg of gadobenate dimeglumine (MultiHance, Bracco Imaging SpA, Milan, Italy) was performed only at the baseline CMR before PPVI [18]. MR angiography was not performed after PPVI due to the presence of the metal artefact in the pulmonary conduit.

Image analysis

Cine true-FISP images in the short-axis plane, from heart base to apex, were segmented semiautomatically (Syngo-Argus, version VE32B, Siemens Medical Solutions, Erlangen, Germany) by a reader with 7 years of experience in CMR imaging, to obtain end-diastolic volume (EDV), normalized to body surface area (EDV index, EDVI), end-systolic volume index (ESVI), stroke volume (SVI), and ejection fraction (EF) of the RV and left ventricle (LV) [19]. In Fig. 1 is shown an example of RV segmentation. Papillary muscles were excluded from the volume. The basis was chosen with the references of a four-chamber view [20].

Fig. 1
figure 1

Example of RV volume analysis using CMR cine sequences in a short-axis view of the basis (a, b), the mid-ventricle (c, d) and the apex (e, f). In the left column are shown images of the short axis at the end-diastolic phase (a, c, e) and in the right column at the end-systolic phase (b, d, f)

Full size image

Reverse flow, forward flow and peak velocity data of the pulmonary valve were obtained by segmenting phase contrast images with the same software. We estimated pulmonary regurgitation fraction (RF) by flow data [21]. The pressure gradient (PG) was estimated from peak flow velocities using Bernoulli’s equation as described [22].

Statistical analysis

A Friedman test was used to compare EDVI, ESVI, SVI, EF, RF and PG before and after 1, 3, 6, 12, 24, 36, and 48 months from PPVI. All testing was performed using SPSS statistical software for Windows, release 14.0 (SPSS Inc., Chicago, IL, US). P values lower than 0.05 were considered statistically significant.

Results

Overall results of CMR before and during the follow-up after PPVI are shown in Table 2. Overall, PG (31 ± 06 to 16 ± 4 mmHg) and RF (16 ± 17 to 0.3 ± 1 %) declined significantly (p < 0.001). A significant reduction of RV volumes (EDVI from 82 ± 38 to 58 ± 12 ml/m2, p < 0.001; ESVI from 44 ± 12 to 30 ± 13 ml/m2, p < 0.001), as well as a significant increase of RVEF (from 49 ± 13 to 58 ± 12 %, p < 0.001) and SVI (from 38 ± 14 to 40 ± 8 ml/m2, p < 0.001) were observed; RF and PG were significantly reduced at the last follow-up (p < 0.001). LV analysis showed a significant increase in EDVI (from 67 ± 17 to 73 ± 18 ml/m2, p = 0.034) and SVI (from 37 ± 11 to 43 ± 10 ml/m2, p < 0.001) while no significant changes were observed for ESVI and EF. In Fig. 2, RV and LV volumes, EF, RF and PG are represented during the follow-up as plotted versus time. In Figs. 3 and 4, CMR images before PPVI and during the follow-up are shown.

Table 2 CMR results before and after percutaneous pulmonary valve implantation
Full size table
Fig. 2
figure 2

Graphs showing ventricle volumes, ejection fraction, regurgitation fraction and pressure gradient as plotted versus time. PPVI: percutaneous pulmonary valve replacement. RV: right ventricle. LV: left ventricle. EDVI: end-diastolic volume index. ESVI: end-systolic volume index. SVI: stroke volume index. EF: ejection fraction. RF: regurgitation fraction. PG: pressure gradient

Full size image
Fig. 3
figure 3

CMR images of a 14-year-old patient with congenital pulmonary atresia treated with a homograft. Due to homograft stenosis, a percutaneous pulmonary valve (PPVI) was implanted. In the figure are shown images of cine true-FISP sequences at the end-diastolic phase. Right ventricle outflow tract (a, b) and three short axis planes (c, d, e, f, g and h) from the basal portion of the heart to the apex before (upper line) and after 48 months from PPVI (lower line) are shown. Note the artefact of pulmonary valve stent after PPVI in b and the reduction of the right ventricle end-diastolic volume (from 80 ml/m2 to 55 ml/m2) and septal flattening. Septal flattening is visible on c and e. Reduction of the volume abd of septal flattenig is evident when comparing images e and f

Full size image
Fig. 4
figure 4

CMR of a 24-year-old patient with tetralogy of Fallot treated with a homograft. Due to homograft stenosis, a percutaneous pulmonary valve was implanted (PPVI). Cine true-FISP images at the end-diastolic phase of the right ventricle outflow tract before (a) and after PPVI (b) are shown. Reference lines (white lines) of phase-contrast sequences before (a) and after PPVI (b) are also shown

Full size image

One patient performed CMR with an intracardiac defibrillator with the housing implanted on the right anterior side of the chest (Fig. 5). Before CMR examinations, a dedicated cardiologist deactivated the device to avoid a wrong response during the examination.

Fig. 5
figure 5

CMR of a 31-year-old patient with tetralogy of Fallot with an implantable cardiac defibrillator (ICD), previously treated with a homograft. Cine FLASH sequences at the end-diastolic phase of the right ventricle outflow tract before (a) and after PPVI (b) are shown. Note the artefact created by the catheter in the right ventricle and that created by the pulmonary valve stent after PPVI. The ICD in the right position of the chest allowed for a diagnostic CMR without artefacts impairing heart evaluation

Full size image

Two patients needed a surgical conduit replacement 24 and 36 months after the PPVI due to a fracture of the pulmonary stents. One patient needed a second implantation of a stent in the pulmonary conduit after 24 months of PPVI due to a restenosis of the valve. Patient characteristics as well as baseline and follow-up CMR results of these three patients up to the adverse event are shown in Table 3. We did not find any signs or parameters of CMR that could predict the adverse events.

Table 3 CMR results before and after percutaneous pulmonary valve implantation in three cases of valve failure
Full size table

Discussion

In the last decade, PPVI has emerged as the preferred treatment in pulmonary conduit dysfunction in CHD patients, being less invasive than surgical repair. We investigate with CMR a consecutive series of patients before and after PPVI to evaluate the immediate and mid-term changes in cardiac function values.

We observed a significant and progressive reduction of RV volumes as well as a significant increase of RVEF and RVSVI during the investigated follow-up. The most important remodelling occurred 3 months after the PPVI. However, it was important to show what happened during the following four years, when a further progressive remodelling was observed with a continuous increase of RVEF. Regarding flow analysis, both RF and PG were strongly and immediately reduced (at 1 month follow-up) and were thereafter stable at lowered values. These results were in agreement with the literature [12–14, 16], showing a good effect of PPVI on the RV function. However, to our knowledge, there are no studies showing the follow-up of these patients up to 4 years after PPVI.

Interestingly, follow-up of LV volumes showed a significant increase in EDVI and SVI. The former result can be explained considering that the reduction of RVEDVI allows a better diastolic compliance of LV, in turn determining an increased LVSVI. The improvement of LVEDVI likely reflects a decrease in RV volume and pressure overload, and hence a reduction in interventricular dependence from paradoxical septal motion. Conversely, the LVEF remained unchanged due to the insignificant increase in LVESVI.

Of note, one patient with an implantable cardiac defibrillator was fully studied with CMR with a comprehensive evaluation of RV and RV outflow tract function. The reason for this good result is that the housing of the device was implanted on the right side of the chest, causing artefacts to not influence heart imaging. This CMR was performed because the patient had been enrolled in an IRB-approved prospective study aimed at evaluating the MR safety in patients having pacemakers or implantable cardiac defibrillators.

PPVI implies some possible complications as stent fracture [23] or stent restenosis [23, 24]. In our population, two patients had stent fracture and one had stent restenosis. In the two cases of stent fracture, a surgical implantation of a conduit was performed; in the case of restenosis, a new stent was implanted. We did not find any CMR sign or parameter that could predict the adverse events, the baseline and follow-up results (see Table 3) not being different from those patients who did not have any adverse events. Thus, while endocarditis is the most probable cause of restenosis, the only possible explanation for the fracture is a suboptimal correspondence between the stent geometry and the individual outflow tract anatomy.

This study had some limitations. First, we retrospectively evaluated a heterogeneous sample of patients with congenital heart diseases originally treated with different surgical procedures. However, the series of patients was consecutive and all patients presented the same condition of pulmonary stenosis/insufficiency treated with PPVI. Second, we did not estimate the inter- and intra-observer reproducibility of CMR data. However, CMR is widely considered more reproducible than echocardiography in the evaluation of the right chambers and pulmonary artery in a multiplanar and comprehensive evaluation [15, 21]. Finally, only one type of pulmonary valve stent was involved, as described in the Methods section, so that our conclusion cannot be applied to other pulmonary valve stents.

In conclusion, a CMR four-year follow-up after PPVI demonstrated restored pulmonary conduit function, reduced RV volumes and increased RV and LV function. However, it did not predict valve fracture/restenosis.