Cardiovascular magnetic resonance-guided right heart catheterization in a conventional CMR environment – predictors of procedure success and duration in pulmonary artery hypertension

CMR-guided RHC was approved by the Royal Free London NHS Foundation Trust New Interventional Procedures Committee in November 2017. Consecutive patients referred for RHC at the National PH Service at the Royal Free London NHS Foundation Trust between November 2017 and October 2018 and who were available to attend within the timeframe of newly designated CMR-RHC sessions at our institution were offered to undergo CMR-RHC. The clinical indication for CMR-RHC was the same as that for conventional X-ray fluoroscopy-guided RHC, specifically the investigation of suspected or known PH. In particular, patients with scleroderma were referred for CMR-RHC either for the investigation of symptomatic dyspnea and/or using the DETECT algorithm, a systematic clinical screening program used for the evidence-based detection of PAH in this condition [8]. Exclusion criteria for CMR-RHC were arrhythmia, patients requiring an additional X-ray fluoroscopy-guided procedure (including coronary angiography or conventional pulmonary angiography), pregnancy and contraindications to CMR. All ethics were approved by the UCL/UCLH Joint Committees on the Ethics of Human Research Committee, and all participants provided written informed consent.

All procedures were performed in a suite equipped with a 1.5 T CMR scanner (Magnetom Aera, Siemens Healthcare, Erlangen, Germany). A 40-in. CMR-compatible monitor (NordicNeuroLab AS, Bergen, Norway) was used to mirror the scanner display for in-room catheter visualization. A CMR-compatible blood pressure monitor (Expression MR400 or Expression IP5, Invivo International, Philips Healthcare, Best, The Netherlands) was used to measure invasive pressures with the transducers attached to a CMR-compatible drip stand (Fig. 1). Patients receiving continuous parenteral prostanoid therapy had their infusion administered using a volumetric infusion pump (Infusomat Space pump, B. Braun Medical Ltd., Sheffield, United Kingdom) placed inside a unit with shielded aluminium housing (SpaceStation MRI, B. Braun Medical Ltd., Sheffield, United Kingdom).

The staff present for the procedure included an interventional cardiologist, imaging cardiologist, CMR radiographer, cardiac physiologist and a second operator in the scanner room for patient nursing and to assist the interventional operator when required.

Patients initially lay on the scanner table to have surface vector-cardiogram (VCG) electrodes attached and a phased array body coil positioned. The scanner table was then undocked and transferred to the preparation area just outside the scanner room door (Additional file 1: Video S1).

Patients undergoing CMR-RHC were offered the option of mild sedation using intravenous benzodiazepines prior to venous sheath insertion. Vascular access was obtained under local anaesthetic using either manual palpation or under ultrasound guidance (Additional file 1: Video S2). The right femoral vein was chosen as the preferred route of access due to our conventional cardiac catheterization laboratory standard operating procedures, with right internal jugular access used in cases in which femoral venous access was not feasible. In order to maintain sterility of the operating field, the sterile drape was folded over onto the patient firstly from the operator side and subsequently from the opposite side of the table (Additional file 1: Video S3). At this time, the side of the scanner and bore closest to the operator were also draped (Fig. 1). The patient was then transferred back into the CMR scanner room and the table was docked (Additional file 1: Video S4). A non-invasive blood pressure cuff (set to automatically measure at 5-min intervals) and a peripheral oxygen saturation finger probe were attached to the patient. The patient is then positioned inside the scanner bore, following which the drapes are unfolded whilst maintaining sterility.

The catheter operator used the patient’s headphones, allowing communication with the scanner operator. The patient used standard CMR ear defenders and had an alarm to activate if they wished to speak to the team during the procedure in addition to regularly being verbally checked upon by a staff member in the scanner room.

Patient demographics are summarized in Table 1. The majority of the patient cohort (32, 64%) were in World Health Organisation (WHO) functional class III. The significant female preponderance of patients (80%) is reflective of the underlying disease processes in our cohort; 30 out of 40 female patients (75%) had connective tissue disease and seven (18%) had idiopathic PH.

Twenty-six (52%) patients already had a known diagnosis of PH, the majority of whom (25, 96%) were already on pulmonary vasodilator therapy. This included 2 patients who were receiving continuous intravenous epoprostenol. Twenty-four patients had suspected PH and the CMR-RHC was their index invasive hemodynamic assessment. In this group, 13 patients had the diagnosis confirmed and 11 patients (10 with scleroderma and 1 with antisynthetase syndrome) had normal PA pressures. Invasive and CMR derived measures are shown in Table 2, demonstrating a wide spectrum of pulmonary hemodynamics and right heart sizes amongst the patient population.

Forty-nine CMR-RHC studies were performed using femoral venous access and one study (patient number 16) was performed from the right internal jugular vein due to the presence of an inferior vena cava filter. The femoral vein was successfully identified and cannulated by manual palpation in 48 patients and using ultrasound guidance in 1 case; ultrasound guidance was used as the default technique for internal jugular venous access. There were no significant vascular complications, with a simple hematoma reported in one patient that was conservatively managed. Sedation was offered to all patients and subsequently used in 17 (34%) patients without complication; the remaining 33 (66%) of patients declined any sedation as per their preference. No patients requested the study to be terminated early.

A full CMR-RHC study was successfully completed in 47 (94%) patients and failed in 3 (6%) patients (patients 7, 47 and 50). In 2 cases (patients 7 and 47) the catheter could not be advanced from the RV to PA and in the other case (patient 50) the catheter could not be advanced from the right atrium (RA) to a stable RV position. All three procedural failures used the femoral venous approach. These patients subsequently underwent conventional X-ray fluoroscopic guided RHC on the same day, with patients 7 and 47 requiring a loop to be formed in the RA and patients 47 and 50 also necessitating the use of a Nitinol guidewire. CMR-conditional guidewires (Nano4Imaging = 1, EPFlex = 5) were used in six patients without complication. In two of these patients, the procedure ultimately failed and in the other four it succeeded. Real-time cine imaging was required in a total of nine (19%) patients due to difficulties in breath-holding cine acquisition, only one of whom had received intravenous sedation. In addition, four patients (8%; cases 6, 15, 17 and 23) had additional clinically indicated tissue characterisation assessments performed, one of whom (subject #6) included late gadolinium enhancement imaging and extracellular volume quantification at 15-min post-contrast administration. The only two catheter-related adverse events comprised asymptomatic hypotension in two patients (cases 3 and 47) that corrected immediately with intravenous crystalloid whilst in the scanner bore.

Metrics that were significantly associated with guidewire use or procedural failure were those that reflected the severity of PAH, right heart dilatation, RV hypertrophy and RV dysfunction (Table 3). All three patients in whom CMR-RHC was unsuccessful were in WHO functional class IV and had severe pre-capillary PAH with significantly dilated and impaired right hearts (Table 4). The composite endpoint of procedural failure or guidewire use was not associated with procedure number, cardiac output or tricuspid or pulmonary regurgitant fraction.

Procedural timings are summarized in Table 5, with a median catheter time of 3.6 (2.0–7.7) minutes (range 0.6–15.1). The mean time to complete the full CMR-RHC protocol (including catheter advancement, great vessel flow assessment and biventricular cine stack) was 31.7 ± 7.8 min (range 19.0–49.8). The median total department time (CMR-RHC plus patient preparation and transfer time) was 60.0 (range 54.0–68.5) minutes (range 45.0–92.0), respectively. There were no significant differences in any of the four timed components of the study with or without sedation use.

Times plotted against procedure number are shown in Fig. 5, with all times becoming significantly shorter with increasing procedural experience (p < 0.05). However, it can be seen that total department time fell more quickly than catheter time or total procedure time. Severe PH, right heart dilatation, RV hypertrophy and RV dysfunction were all associated with longer catheter times (Table 6). Catheter time was not associated with cardiac output or tricuspid or pulmonary regurgitant fraction.

Only three out of 50 RHC cases were unsuccessful in the CMR environment, although in six instances a guidewire was used with two of these still ending in failure. Due to the significant overlap between guidewire usage and failure, we used a composite outcome of guidewire use or failure to investigate hemodynamic and CMR associations. The main measures associated with guidewire use or failure were higher PA pressures and increasing RV dilation, hypertrophy and dysfunction. In fact, all three patients in whom the procedure failed had significantly deranged pulmonary hemodynamics and severely hypertrophied, dilated and impaired right hearts. The relationship between right heart size and failure or guidewire use is unsurprising as catheter manipulation does become more difficult in dilated chambers, particularly if systolic function is reduced. We also found that the catheter can become lodged in the thickened trabeculations of the hypertrophied RV, explaining the association with RV mass. It is likely that higher PA pressures associated with guidewire use or failure are a surrogate for right heart chamber dilatation, rather than representing a direct effect on catheter manipulation.

In the conventional cardiac catheterization laboratory, these obstacles are overcome by using a metallic guidewire, creating catheter loops in the RA and/or by using internal jugular venous access. We did have access to CMR-conditional guidewires, with a successful outcome in four cases. However, these devices lack the mechanical properties of their metallic equivalents and currently only have discrete markers towards the guidewire tip for visualization. This is in comparison to metallic guidewires that can be visualized along their whole length using X-ray fluoroscopy. Furthermore, only the catheter balloon is visible in CMR whilst the catheter shaft is not seen. We believe that these combined limitations in guidewire and catheter shaft visualization preclude the safe formation of catheter loops using CMR in a conventional CMR scanner environment in which immediate X-ray fluoroscopic back-up is unavailable. Improvements in guidewire technology may overcome these limitations, and recent work has also focused on improving catheter shaft visibility using nitinol-based guidewires in the CMR environment under low-specific absorption rate (SAR) CMR protocols [10]. Overall, our early experience suggests that procedural failure using femoral venous access is uncommon even using standard equipment and was confined to WHO functional class 4 patients with PH. Alternative approaches could be considered for this group of patients attending for CMR-RHC, including the use of different CMR-safe catheters (pre-shaped balloon-tipped or non-balloon-tipped, with or without CMR-conditional guidewires) and internal jugular venous access.

One of the main arguments against CMR-RHC has been long procedure times. However, we demonstrate that catheterization and total procedure times were relatively short. Each patient only occupied the department for approximately 1 hour, allowing 3–4 cases to be performed in a single half-day session. Furthermore, CMR-RHC not only provides measures of PA pressure and PVR, but also of biventricular size and function. Therefore, we believe that CMR-RHC is a feasible and time efficient method of physiologically assessing patients with PH. The catheter time modestly improved with procedure number, reflecting a learning curve with advancing a catheter under CMR-guidance. However, total procedure and department times fell by an even greater amount with increasing procedural experience. This suggests that the main improvements come from improved workflow rather than faster catheterization. Study times could be further improved with the use of accelerated imaging for the acquisition of great vessel flows and cine images. However, the aim of this study was to demonstrate the feasibility of CMR-RHC using standard commercially available hardware and software in order to disseminate the technique without the need for non-standard equipment and sequences.

To our knowledge, this is the first study to investigate the association between catheterization time and hemodynamic parameters. As with procedure failure, the main associations were higher PA pressures and increasing RV dilation, hypertrophy and dysfunction. The reasons for these associations are similar to previously described for failure and guidewire usage. This suggests that it might be useful to ensure a mix of disease severity when performing multiple cases in the same session.

Combining hemodynamic and CMR data provides a comprehensive assessment of patients under investigation for PAH in a single study. However, these data can routinely be obtained at separate appointments. Our CMR-RHC procedure times are faster than the total duration of the component investigations performed individually. This confers potential advantages that include cost savings and faster clinical workflows, benefitting both patients and the institution. The times reported in this study included the learning curve that is inherent in setting up a clinical CMR-RHC service. Analyses of the impact of an established CMR-RHC program on cost and patient experience would be worthwhile in evaluating the clinical utility of this combined approach in a tertiary PAH referral center.

Patients routinely had CMR-RHC performed via the femoral approach as per our standard institution protocol. However, the buoyancy and natural upward-pointing arc of the air-filled balloon catheter via the internal jugular venous approach could be advantageous in more technically challenging cases, and merits further study. Furthermore, the use of alternative commercially available CMR-safe curved-tip catheters may have offered greater potential for procedural success in these cases [7]. Whilst the air-filled balloon was adequately visualized in our cohort of patients, previous studies have shown that gadolinium-filled balloons were more consistently conspicuous especially when using a saturation-preparation sequence [4]. Air-filled balloons have the advantages of greater buoyancy and more favourable inflation characteristics, but the enhanced visibility of gadolinium-filled balloons may be advantageous in cases of more complex anatomy that were not encountered in our cohort. Similarly, dedicated interventional CMR workstations that offer simultaneous real-time visualization of multiple scan planes may be beneficial in such complex cases. Our set-up utilized the scanner headphones to communicate between catheter and scanner operators, with the patient receiving regular verbal checks by nursing staff and also holding the emergency alarm to signal if they needed to communicate. This approach worked well for our short CMR-RHC studies, but commercially available systems that enable communication amongst circulating staff and with patients in the CMR environment provide an alternative solution [4, 7]. These systems may be of particular use in complex interventional CMR, but at an additional cost to purchase. We did not record equivalent time parameters for patients undergoing RHC under X-ray-fluoroscopic guidance as a comparator group. However, our CMR-RHC procedure times are nevertheless of a similar order of magnitude to a conventional cardiac catheterization laboratory. Our patient cohort was also not large enough to perform multivariable analysis of the predictors of guidewire usage and procedural failure. However, the factors identified in this study are likely related to one another, and, as such, provide a signal as to the scenarios in which CMR-RHC may be more challenging.