Update on supra-annular sizing of transcatheter aortic valve prostheses in raphe-type bicuspid aortic valve disease according to the LIRA method

Abstract

Introduction

Transcatheter aortic valve replacement (TAVR) indications are expanding towards low surgical risk and younger patients with constant improvements in clinical outcomes.^1–3 In this cohort, the prevalence of bicuspid aortic valve (BAV) anatomy is expected to be higher.^4,5 Furthermore, octogenarians and nonagenarians account for 5–28% of the BAV population⁶ and this cohort are likely to be referred for TAVR due to a higher risk of surgical complications related to age, frailty, and other comorbidities. Transcatheter aortic valve replacement in patients with BAV still represents a challenge due to the peculiar anatomy and lack of consensus concerning the optimal method for multi-detector computed tomography (MDCT) scan sizing and prosthesis selection.⁷ Success rates have improved with the introduction of new-generation transcatheter heart valves (THVs),^8,9 although higher rates of procedural complications and clinical events persist in the BAV cohort,¹⁰ especially in anatomical configurations combining a calcific raphe with excessive leaflet calcification.¹¹ Appropriate THV sizing is essential to optimize the interaction between the prosthesis and individual patient anatomy. Recent evidence has shown that THV anchoring in BAV patients occurs at the level of the raphe, also known as the LIRA (Level of Implantation at the RAphe) plane.¹² We have previously shown that the use of a novel supra-annular sizing method based on measurement of the perimeter at the level of the raphe (LIRA method) was safe and effective in a cohort of 20 BAV patients with severe aortic stenosis.¹³ The goals of this study were to confirm the feasibility and clinical outcomes of LIRA sizing in a larger cohort of patients, assess interobserver reproducibility of this methodology, and analyse the interaction of LIRA-sized prostheses with the surrounding anatomy.

Methods

Study population

The LIRA sizing method was applied prospectively in all consecutive patients with aortic stenosis and raphe-type BAV disease treated with TAVR using a variety of supra-annular self-expandable prostheses at our centre between November 2018 and October 2021. Bicuspid aortic valve was defined as a deformed aortic valve with two functional cusps forming a valve mechanism with <3 zones of parallel apposition. The LIRA sizing method was applied to BAV Types 1 and 2 according to Sievers classification (Type 1—one raphe, Type 2—two raphes).¹⁴ The diagnosis of BAV was confirmed by the local heart team based upon a review of the medical records, multi-detector CT scans, and transthoracic/transoesophageal echocardiography analysis when available. Patients were scheduled for TAVR following multidisciplinary discussion as indicated by current recommendations.¹⁵ Post-procedural CT scans were performed according to local clinical protocols.

MDCT assessment

All subjects underwent electrocardiography-gated cardiac CT scanning before their TAVR procedure. All exams were independently analysed by one radiology technician and two cardiologists (V.R., G.I., and M.M.) using the Osirix programme, version 9 (Pixmeo SARL, Bernex, Switzerland, 2018). Computed tomography scans were analysed in systole (20–40% R-R interval) and the aortic valve virtual basal ring (VBR) defined as the short-axis plane through the nadir of each coronary cusp. The minimal diameter (D_min), maximal diameter (D_max), perimeter, and the area of the aortic valve VBR were determined (Figure 1A).

Figure 1

Prosthesis sizing according to the LIRA method. (A) Virtual basal ring measurements. (B) LIRA plane measurements. (C) LIRA plane level. (D) Transcatheter heart valve at virtual basal ring. (E) Transcatheter heart valve at the LIRA plane. (F) Transcatheter heart valve at coaptation level. LIRA, Level of Implantation at the RAphe.

Open in new tabDownload slide

LIRA plane

The level of the aortic root where prosthesis anchoring was expected to occur (known as the LIRA plane) was identified on the pre-TAVR CT scan. The LIRA plane comprises the most rigid anatomical structures in the aortic root (calcific/fibrotic raphes) that might facilitate THV anchoring during deployment. In BAV Type 1 anatomy, the prosthesis should anchor at the level of the calcific/fibrotic raphe, whereas in BAV Type 2 anatomy presents two raphes and the THV prosthesis should anchor at the level of the major raphe (defined as that with the larger width and length, and greater amount of calcium). In both instances, the LIRA plane was identified as the level cutting the major raphe at its point of maximum protrusion along the aortic root (Figure 1B and C).¹²

LIRA sizing method

The LIRA plane represents a ‘neo-VBR’ where the perimeter traces the internal border of the leaflets, excluding all other structures encountered at this level (fused commissures, heavy calcification, and calcific or fibrotic raphes) (Figure 1B).¹³ Transcatheter heart valves were sized according to manufacturer recommendations based on baseline CT scan perimeters at the VBR and LIRA plane. If there was a discrepancy in measurements between the two planes, the THV was sized according to the smallest perimeter. Transcatheter heart valve oversizing at the VBR and LIRA plane was expressed according to the cover index [100 − (prosthesis diameter − MDCT annular size)/prosthesis diameter]¹⁶ and the ratio between the implanted and expected THV diameter, according to VBR and LIRA plane measurements. We also evaluated the distance between the LIRA plane and VBR, the raphe length and its characteristics, and the ICD at a point 4 mm above the VBR (a previously established reference for THV sizing in BAV disease).¹⁷

Post-procedural MDCT assessment

Where available we evaluated post-TAVR CT scans to determine the interaction of the THV with surrounding anatomy. Specifically, we evaluated the perimeter of the THV at the LIRA plane, its distance from the VBR wall and the leaflets at the LIRA plane, and investigated the level of THV circularity at the VBR, LIRA plane, and coaptation level. Circularity was expressed as the eccentricity index [EI = 1 − (D_min/D_max)].¹⁸ Implantation depth was measured as the mean distance of the ventricular edge of the prostheses from the non-coronary cusp and the left coronary cusp of the native aortic valve.¹⁹

Study endpoints

The primary endpoint of the study was post-procedural device success, defined according to VARC-2 criteria. Secondary endpoints included in-hospital and 30-day outcomes. The severity of paravalvular leak (PVL) was qualitatively assessed and graded using transthoracic echocardiography.²⁰ To assess the reproducibility of measurements, we evaluated interobserver correlation of the LIRA plane perimeter. Correlation of pre- and post-procedural LIRA plane perimeters and THV interaction with the surrounding anatomy was determined by comparison of pre- and post-TAVR CT scans. The study was conducted in accordance with the Declaration of Helsinki.

Statistical analysis

Categorical variables were expressed as frequencies and percentages, and continuous variables as means ± standard deviations or median and interquartile ranges, depending on the presence or absence of normal distribution. A paired t-test was used to test for differences in means with continuous variables, and a two-sided P-value <0.05 was considered statistically significant. The correlation of measurements was analysed using the Pearson correlation coefficient. Statistical analysis was performed using IBM SPSS Statistics version 22.0 (IBM Corporation, New York, NY, USA).

Results

A total of 50 patients were included between November 2018 and October 2021. Mean age was 80 ± 6 years and 70% were male, with a median Society of Thoracic Surgeons (STS) predicted risk of mortality score of 4.3 (3.0–6.5). Baseline characteristics of the included cohort are reported in Table 1. Computed tomography scans identified three different BAV anatomies: 41 patients with Type 1 BAV and a calcific raphe, 6 with Type 1 BAV and a fibrotic raphe, and 3 with Type 2 BAV. The VBR perimeter was significantly larger than the LIRA plane perimeter (81.4 ± 7.1 vs. 73 ± 7.9 mm, P < 0.001), confirming the typical tapered (‘volcano’) aortic root configuration of BAV anatomy. Two commissures were visible at a point 4 mm above the VBR in 38% of cases. Baseline CT scan measurements are reported in Table 2 and confirm high interobserver reproducibility of LIRA plane perimeter measurements (r = 0.980; P < 0.001) (Figure 2).

Figure 2

Interobserver correlation of the perimeter at the LIRA plane. LIRA, Level of Implantation at the RAphe.

Open in new tabDownload slide

Procedural characteristics

Transcatheter aortic valve replacement was performed under conscious sedation and THVs implanted via transfemoral access. Various supra-annular self-expandable THVs were implanted [ACURATE Neo and Neo 2 (Boston Scientific, Marlborough, MA, USA) 56%; Evolut R/Pro (Medtronic, Minneapolis, MN, USA) 44%]. Discrepancies between VBR and LIRA perimeters would have led to different valve size selection in most cases, with LIRA sizing leading to selection of a THV one size smaller in comparison with conventional annular tracing (Figure 3A and B). The THV was therefore sized according to manufacturer recommendations based on the LIRA plane perimeter (diameter prosthesis implanted/diameter prosthesis according to LIRA plane = 1) (DPI/DP LIRA = 1) and significantly downsized according to the VBR perimeter (DPI/DP VBR 0.89; P < 0.001). The median prosthesis size was 26 mm (range 23–27 mm). The need for pre-dilatation was relatively high (88%) with a median balloon size of 20 mm (range 20–23 mm), whereas post-dilatation was applied in 26% of cases (median balloon size 23 mm, range 22–24 mm). Procedural characteristics of the overall population are reported in Table 3—VARC-2 device success was 100% in the overall population with no valve migration, procedural mortality, cases with mean gradient >20 mmHg or moderate-severe PVL (Table 4).

Figure 3

(A) Virtual basal ring and LIRA sizing of ACURATE Neo and Neo2 prostheses. (B) Virtual basal ring and LIRA sizing of Evolut R/Pro prostheses. NA: outside the range for specific prosthesis sizing charts and therefore not available. LIRA, Level of Implantation at the RAphe.

Open in new tabDownload slide

Echocardiographic data and in-hospital and 30-day outcomes

Procedural, in-hospital, and 30-day outcomes are reported in Table 4. The mean THV gradient pre-discharge was 8.2 ± 3.4 mmHg—no patients had a mean gradient >20 mmHg or moderate-severe PVL. There were no strokes or mortality, and the incidence of in-hospital pacemaker implantation was 10%.

Post-procedural MDCT assessment

A total of 10 post-procedural CT scans were available for analysis and demonstrated a strong correlation between the pre-TAVR LIRA plane perimeter and post-TAVR prosthesis perimeter at the same level (r = 0.981; P < 0.001) (Figure 4A). Transcatheter heart valve anchoring and sealing take place at the LIRA plane (mean maximal horizontal distance of 0 mm between the prosthesis cage and the leaflets), whereas at the level of the VBR there is partial interaction between the THV and the aortic annular walls (mean maximal horizontal distance of 4.5 ± 2.3 mm between the prosthesis cage and annular plane walls). The supra-annular prosthesis is elliptical at the VBR and LIRA planes, whereas a greater degree of circularity is maintained at coaptation level (mean eccentricity index 0.05 ± 0.05) (Figure 1D–F). The implantation level of the prosthesis was high (mean prosthesis implantation depth 3.0 ± 3.3 mm), with two cases where the ventricular edges of the prostheses frames were higher than VBR levels. Interactions between THV prostheses and the aortic root in bicuspid anatomy with raphe are reported in Table 5.

Figure 4

(A) Correlation between pre-transcatheter aortic valve replacement perimeter measurement at the LIRA plane and post-transcatheter aortic valve replacement prosthesis perimeter at the same level. (B) Prediction of the perimeter occupied by a transcatheter heart valve using CT LIRA plane sizing. Columns A and B show the LIRA plane of the bicuspid aortic valve and the perimeter traced according to the LIRA method. Column C shows the perimeter of the implanted transcatheter heart valve. Column D shows the overlap of the transcatheter heart valve perimeter on the LIRA plane, demonstrating that each valve interacts differently with the surrounding anatomy. Column E demonstrates the final shapes of the implanted valves—note significant and unpredictable variation in final ellipsoid (not elliptical) shape. LIRA, Level of Implantation at the RAphe.

Open in new tabDownload slide

Discussion

Our study demonstrates that (i) supra-annular self-expandable THV sizing using the LIRA method is safe with a high device success rate in a cohort of 50 patients with raphe-type BAV disease, (ii) the aortic root generally has a tapered (‘volcano’) configuration in raphe-type BAV disease and LIRA sizing mostly leads to selection of a THV one size smaller compared to traditional annular tracing, (iii) LIRA plane perimeter measurement is highly reproducible, (iv) two commissures are not always visible at a point 4 mm above the VBR ( < 38% of cases), (v) the LIRA plane perimeter determined from a pre-procedural CT scans can predict the perimeter occupied by the supra-annular self-expandable THV, (vi) supra-annular self-expandable prostheses maintain adequate circularity at coaptation level in raphe-type BAV disease.

Transcatheter aortic valve replacement in patients with raphe-type BAV stenosis poses several anatomical and technical challenges that may be associated with suboptimal procedural and clinical outcomes. Importantly, BAV anatomy has been excluded from randomized clinical trials evaluating the efficacy of TAVR in the treatment of severe aortic stenosis.^1–3 Despite the introduction of new-generation THVs, device success in this setting is limited by a higher likelihood of significant PVL with self-expanding prostheses and annular rupture with balloon-expandable prostheses.¹⁰ These issues require tailored solutions as TAVR indications expand to younger patients with lower surgical risk, and a higher prevalence of BAV anatomy.⁴

Appropriate THV sizing is essential to optimize the interaction between the prosthesis and individual patient anatomy. In particular, the narrowest portion of the aortic root in BAV patients seems to be located at the level of the raphe instead of the VBR. The consequent tapered (‘volcano’) configuration of the aortic root leads to discrepant measurements at the annular and supra-annular plane. Various methods of supra-annular sizing have been proposed for THV sizing in BAV disease as alternatives to conventional annular tracing.⁷ Liu et al.²¹ described a sizing method using balloon pre-dilatation based on the so-called ‘waist sign’—sequential pre-dilatations were performed beginning with a smaller size until the ‘waist sign’ appeared with less than mild regurgitation during contrast injection. Weir-McCall et al.²² used the distance between the innermost aspect of the two commissures (the intercommisural distance, ICD) as the diameter of a circle to calculate perimeter and area. Alternatively, the ICD at a point 4 mm above the VBR may be used as the standardized supra-annular measurement, based upon the fact that the device landing zone usually lies 4 mm below the aortic annulus in patients with a tri-leaflet aortic valve.¹⁷ As an analogy, the landing zone in BAV patients could theoretically lie 4 mm above the aortic annulus given the location of the constraint points. The size of the prosthesis was therefore chosen based on the minimum value between the mean perimeter-derived diameter of the annulus and the ICD 4 mm above the VBR—according to these measurements, annulus-based sizing was applicable in 88% of patients. Furthermore, Petronio et al.²³ proposed the CASPER algorithm for prosthesis sizing that takes account of both the annular and supra-annular plane—starting from the annulus-derived diameter, several millimetres (up to 2.5) are subtracted to take account of the presence of heavy calcium burden, raphe length, and calcium distribution. We recently identified a supra-annular plane at raphe-level that predicts THV anchoring in raphe-type BAV disease—the LIRA (Level of Implantation at the RAphe) plane.¹² Delineation of the internal border of the leaflets (excluding all other structures encountered at this level—fused commissures, heavy calcification, and calcific or fibrotic raphes) allows derivation of the perimeter at this plane (Figure 1B and C) and THV sizing with optimal procedural outcome.¹³ Importantly, none of these methods of supra-annular sizing have been applied in case series exceeding 20 patients. This study, therefore, represents the first substantial real-world experience using a supra-annular sizing method (the LIRA sizing method) in the anatomical setting of raphe-type BAV disease.

Therefore, among all closed curves in the plane of fixed perimeter, the one with the largest area is the circle (the so-called ‘Steiner symmetrization’).^24,Figure 5 illustrates the geometrical assumptions underpinning the LIRA tracing method. In the left-hand panel, a lunar eclipse is shown, where the moon is superimposed upon the sun. The moon and sun are two figures with different shapes and areas but the same perimeter. However, if we turn the small curve of the moon by 180° (right-hand panel), we have two identical figures. This is equivalent to the perimeter traced according to the LIRA method and the perimeter of the implanted THV—two different figures with different areas but the same perimeter. Figure 4A and B and Table 5 demonstrate that the perimeter derived using the LIRA method can accurately predict the actual space occupied by the THV—in Columns B and C, the device perimeter following implantation is almost identical to the LIRA plane perimeter derived from the pre-procedural CT scan. Furthermore, Columns C and D emphasize the difficulty of predicting the final THV shape (e.g. circular²² or hypothetical ellipse) from the pre-procedural CT scan, since this depends on interaction with the surrounding individual patient anatomy (Column E). Specifically, the presence of the raphe inhibits adequate expansion of the prosthesis at this level, independent of its characteristics (fibrotic or calcific).

Figure 5

Geometrical assumptions underpinning the LIRA method. LIRA, Level of Implantation at the RAphe.

Open in new tabDownload slide

Supra-annular self-expanding THVs adapt to the LIRA plane perimeter space available in this anatomical setting without modifying its value, and increase the area interacting with the surrounding anatomy. The elliptical shape of the prosthesis at the VBR and LIRA plane is balanced by greater degree of circularity at coaptation level (Table 5, Figure 1D–F). Preserved symmetry of the THV leaflets at coaptation level might account for the low residual transvalvular gradient (mean gradient <20 mmHg) and may determine future valve durability.²⁵

In this study, the perimeter of the LIRA plane was significantly smaller than that of the VBR, conferring the typical tapered (‘volcano’) configuration of the aortic root in raphe-type BAV disease. Consequently, implanted THVs were at least one size smaller than would have been selected according to conventional annular tracing in most cases (Figure 3A and B). Tchetche et al.¹⁷ previously reported that the ICD 4 mm above the VBR plane provided optimal THV sizing in relation to the mean perimeter-derived annular diameter—accordingly, most aortic root configurations were ‘flare’ or ‘tubular’, leading to a traditional annular based sizing strategy in most cases. The ICD is a unidimensional measurement between commissures with inherent limitations, whereas the perimeter is bidimensional. Figure 6A shows that the same mean diameter can be derived from the mean of the maximal and minimal diameter of different closed curves (circle, ellipse, or irregular shaped curve). However, the inverse derivation cannot be applied since the only figure that can be derived with certainty from a mean diameter (in terms of area and perimeter) is a circle (where maximum and minimum diameter are identical). Therefore, the ICD can only derive a circle, leading to excessive area enlargement as demonstrated by the Steiner symmetrization.²⁴ Furthermore, Figure 6Bshows that ICD may not correlate with perimeter-based diameters at the LIRA plane (same ICD diameter with three different perimeters and three different THV sizes). Consequently, considering the ICD as a reference of supra-annular measurement might overestimate the actual space available, leading to the misconception of a high incidence of ‘flare’ or ‘tube’ aortic root configuration in patients with BAV anatomy and risk of aortic rupture if a balloon-expandable THV is selected.

Figure 6

(A) Direct and inverse derivation of the mean diameter from different closed curves. (B) Inter-commissural distance and perimeter at the LIRA plane. (A–C) Three different patients with similar inter-commissural distances but different perimeters at the LIRA plane. ICD, intercommissural distance; LIRA, Level of Implantation at the RAphe.

Open in new tabDownload slide

The ICD 4 mm above the VBR is based upon the fact that the landing zone is usually 4 mm below the aortic annulus in patients with a tri-leaflet aortic valve. As an analogy, the landing zone in BAV patients could theoretically run 4 mm above the aortic annulus given the location of the constraint points.¹⁷ However, the LIRA plane cannot be arbitrarily located at a fixed distance above the VBR since its location differs in each patient.¹² In addition, measurement of ICD 4 mm above the VBR may be problematic since this plane might not be high enough to identify two commissures in Types 1 and 2 anatomy. Indeed, three commissures were visualized in 62% of our patient cohort.

In this study, we report the feasibility of LIRA sizing in 50 prospective and consecutive TAVR procedures using various supra-annular self-expanding prostheses in a range of raphe-type BAV anatomies. Pre-dilatation was performed frequently (88%) with a median balloon size of 20 mm, whereas post-dilatation was performed in only 26% of cases. We propose that this strategy of aggressive pre-dilatation (20 mm balloon) in relation to the LIRA plane (LIRA perimeter-derived diameter 23 mm) enhanced optimal expansion during THV deployment and reduced the need for post-dilatation to obtain an adequate result Table 3. Procedural outcomes were optimal (100% VARC-2 device success) with no valve migration, adequate prosthesis expansion, trivial/mild residual PVL, and low transprosthetic gradient, and maintained at 30 days. Furthermore, the incidence of in-hospital pacemaker implantation was low (10%), reflecting high THV implantation in relation to the level of the LIRA plane and the low prosthesis interaction with the membranous septum Table 4 and Table 5.²⁶

Limitations

Our study has several limitations. First, this was a single-centre non-randomized study. Second, imaging assessment for the diagnosis of BAV and evaluation of echocardiographic outcomes was not performed by a core laboratory. Third, our proposed sizing algorithm (LIRA method) requires validation in a larger prospective multicentre registry.

Conclusions

This study confirms that supra-annular sizing using the LIRA method in raphe-type BAV patients is safe in a large cohort of patients and is associated with high device success using various supra-annular self-expandable THVs. Further, larger studies are needed to confirm these findings.

Acknowledgement

This study has been awarded with the prize “Premio Vincenzo Masini” 2022 from the Italian Association of Hospital Cardiologists (ANMCO).

Funding

None.

Conflict of interest: M.M. serves as a proctor for Boston Scientific. B.P. has received unrestricted research and educational grants from Edwards Lifesciences, lecture fees from Edwards Lifesciences, Medtronic, and Abbott, and consultancy fees from Anteris and Microport. The other authors have nothing to disclose.

Data availability

The data underlying this article are available in the article and in its online supplementary material.

References

Author notes