Content uploaded by Giulia Costa
Author content
All content in this area was uploaded by Giulia Costa on Jan 19, 2024
Content may be subject to copyright.
Contents lists available at ScienceDirect
Journal of Cardiovascular Computed Tomography
journal homepage: www.elsevier.com/locate/jcct
Research paper
Bicuspid aortic valve sizing for transcatheter aortic valve implantation:
Development and validation of an algorithm based on multi-slice computed
tomography
Anna S. Petronio
a
, Marco Angelillis
a,∗
, Ole De Backer
b
, Cristina Giannini
a
, Giulia Costa
a
,
Claudia Fiorina
c
, Fausto Castriota
d
, Francesco Bedogni
e
, Jean C. Laborde
f
, Lars Søndergaard
b
a
Cardiothoracic and Vascular Department, University Hospital Pisa, Italy
b
The Heart Centre, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark
c
Ospedali Civili Brescia, Italy
d
Cardiovascular Department, Humanities Gavazzeni, Bergamo, Italy
e
S.Donato Hospital, S.Donato Milan, Italy
f
St. George Hospital, London, UK
ARTICLE INFO
Keywords:
Aortic stenosis
Bicuspid aortic valve
MSCT
TAVI
ABSTRACT
Background: No indication are available for transcatheter aortic valve implantation (TAVI) sizing in bicuspid
aortic valve (BAV). Aim of the study is to develop and validate a Multi-Slice Computed Tomography (MSCT)-
based algorithm for transcatheter heart valve (THV) sizing in patients with stenotic BAV under evaluation for
TAVI.
Methods: A two steps method was applied: 1)evaluation of a cohort of 19 consecutive patients with type I BAV
stenosis undergoing TAVI through pre and post-procedural MSCT, and development of an algorithm for THV
sizing; 2)validation of the algorithm on a new cohort of 21 patients.
Results: In the first cohort, a high correlation was found between the raphe-level area measured at pre-proce-
dural MSCT and the smallest THV area measured at post-procedural MSCT (p < 0.001). Moreover, reduced
THV expansion was observed among patients with higher calcium burden (p = 0.048). Then, a new algorithm
for TAVI sizing in BAV was develop (CASPER: Calcium Algorithm Sizing for bicusPid Evaluation with Raphe).
This algorithm is based on the reassessment of the perimeter/area derived annulus diameter, according to three
main anatomical features: 1) the ratio between raphe length and annulus diameter; 2)calcium burden; 3)calcium
distribution in relation to the raphe.
The algorithm was then validated in a new cohort of 21 patients, achieving 100% of procedural success and
excellent TAVI performance.
Conclusion: MSCT assessment of raphe length, calcium burden and its distribution is of crucial relevance in the
pre-procedural evaluation of patients with BAV. These anatomical features can be combined in a new and simple
algorithm for TAVI sizing.
1. Introduction
Bicuspid aortic valve (BAV) is a relatively common congenital de-
fect (1%) that frequently leads to aortic valve stenosis, particularly in
the younger, but also in the elderly population.
1–3
Type I morphology,
according to Sievers classification
4
of BAV, is highly prevalent in
Western countries (88%) and is characterized by the presence of a
raphe that represents an incomplete septation of two cusps. During the
last decade, transcatheter aortic valve implantation (TAVI) has been
recognized as a safe and effective treatment for aortic stenosis (AS) in
patients at increased surgical risk.
5,6
However, in large randomized
clinical TAVI trials, patients with severe bicuspid AS have been ex-
cluded. Indeed, TAVI use in BAV stenosis is still considered an off-label
use and it is associated with higher rates of paravalvular leak (PVL),
new permanent pacemaker (PPM) implantation, and annulus rupture
compared to tricuspid AS.
7
These complications may partially be ex-
plained by a suboptimal transcatheter heart valve (THV) size selection.
The introduction of a new generation of repositionable THV devices
https://doi.org/10.1016/j.jcct.2020.01.007
Received 28 September 2019; Received in revised form 1 January 2020; Accepted 23 January 2020
∗
Corresponding author. Cardiothoracic and Vascular Department, Pisa University Hospital, IT, 56121, Italy.
E-mail address: emodinamica@ao-pisa.toscana.it (M. Angelillis).
Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
1934-5925/ © 2020 Published by Elsevier Inc. on behalf of Society of Cardiovascular Computed Tomography
Please cite this article as: Anna S. Petronio, et al., Journal of Cardiovascular Computed Tomography,
https://doi.org/10.1016/j.jcct.2020.01.007
may result in a safer and more effective treatment of BAV, especially if
size selection is based on a dedicated MSCT assessment.
8–10
To date, there is no consensus on BAV sizing and/or THV im-
plantation technique when considering TAVI treatment.
11
This un-
certainness may be due to the different morphology of bicuspid aortic
valves compared to tricuspid ones. Therefore, TAVI in BAV patients are
mostly based on the operator's experience.
Currently, multi-slice computed tomography (MSCT) is the re-
ference standard technique for THV sizing and procedural planning in
tricuspid AS.
12
The use of MSCT has led to a better understanding of
BAV morphological features.
1,4
The virtual basal ring of BAV is usually
oval and has a larger area than the aortic orifice area at the level of cusp
coaptation. Heterogeneous cusp and sinus morphology, long inter-
commissural distance, as well as severe and asymmetric calcifications
have also been described.
1
Based on these differences, operators adopt different sizing techni-
ques that are mostly empirical and difficult to replicate, and that can be
divided in three main categories: annular, suprannular or balloon-sizing
evaluations.
The aim of this study was to analyse the principal anatomical
parameters, to understand their influence on the difference between the
suprannular space and the annular one and consequently develop an
algorithm to optimize BAV sizing using MSCT in patients undergoing
TAVI.
For this purpose, we divided our project in two-steps:
1. Analysis and comparison of pre and post-procedural MSCT in a
series of TAVI implants using both mechanical expandable and self
expandable valve, in order to correlate valve anatomical features
with prosthesis expansion and develop a sizing algorithm based on
the more significant parameters
2. Validation of the algorithm in a new consecutive prospective cohort
of patients.
2. Methods
2.1. First step
A prospective and consecutive series of patients with BAV were
collected in a database. All selected patients had to have a type I con-
figuration according to the Sievers classification.
4
All operators were
experienced in BAV treatment with self-expanding and mechanically
expanding THV devices and work in high volume centres (more than
100 TAVI/year). Sizing was routinely done according to the operator's
experience with no predetermined method. All patients underwent
MSCT assessment of the aortic valve before and after the procedure.
MSCT measurements were analysed to evaluate which anatomical
features could influence a different sizing in respect to the annulus one
and according to the results, a new sizing algorithm was elaborated.
For the first step, the Lotus valve (Boston Scientific, Natick, MA) was
chosen for its characteristic of being nestled by the cusps and for
symmetrical shape of the prosthesis, allowing an easy analysis of the
aortic valve anatomy after THV deployment. Moreover, in order to
validate results also with other TAVI prosthesis, we collect data also
from patients treated with Corevalve Evolut R (Medtronic Inc,
Minneapolis).
2.2. Second step
In a consecutive cohort of patients with type I BAV, MSCT evalua-
tion and valve sizing for TAVI was upstream decided using the pre-
viously derived algorithm. A self-expanding valve (Evolut R/Medtronic
Inc, Minneapolis) was used in all patients. The self-expanding device
was chosen for its suprannular structure and also because the me-
chanical THV was not available anymore.
Clinical outcomes and adverse events, principally those strictly
correlated to valve expansion such as paravalvular leak (PVL), mean
gradients and new permanent pacemaker implantation (PPM), were
collected at discharge and were defined according to the Valve
Academic Research Consortium criteria 2 (VARC 2)
13
The study protocol was approved by the Istitutional Medical Ethics
Committee and complied with the Declaration of Helsinki.
14
All study
subjects provided written informed consent to the collection of the data
and study.
2.3. MSCT assessment
All baseline and post-procedural MSCT data were evaluated both by
each TAVI center and by an independent Core Lab, which was blinded
to the valve size selected by operators. Scans were performed using a
high definition 64-row scanner (Discovery CT 750HD, General Electric
Milwaukee, WI®). All patients gave written informed consent. All MSCT
scans were acquired with retrospective ECG-gating, and image re-
construction was performed every 10% of the R-R interval. Tube cur-
rent and voltage were modulated according to patients’BMI in order to
mitigate beam artifacts from the prosthesis frame (BMI < 22 kg/m2:
80 KeV; BMI ≥22 kg/m2: 120–140 KeV). Contrast media injection:
40–60 ml at a flow rate of 4–5 ml/s followed by 20–30 ml saline bolus.
According international indications, we performed measurement in a
systolic phase (30–40% R-R)”both for pre- and post-procedural scans.
Scan slice thickness was sub-millimeter. An iterative reconstruction
algorithm (ASIRTM) was applied to improve signal-to-noise ratio while
reducing radiation dose. The effective dose was estimated from the
product of the dose-length product (DLP) and a conversion coefficient
for the chest (effective dose [mSv] = DLP [mGy*cm] * 0.014 [mSv/
mGy*cm]). The 3Mensio Structural Heart software version 8.0 (Pie
Medical Imaging, Maastricht, The Netherlands) was used for all MSCT
analysis.
2.4. Pre-procedural MSCT
All measurements were performed in a plane obtained in a multi-
plane reconstruction, after aligning the aortic annulus perpendicular to
the axis of the aorta. Aortic valve dimensions (both perimeter and area)
were evaluated at different standardized levels: the aortic annulus as
determined by the virtual basal ring, linking the hinge points of the
aortic cusps; the left ventricular outflow tract (LVOT) 4 mm below the
annulus; the sinuses of Valsalva (SOV); the sinus-tubular junction; the
height of coronary arteries.
15
The inter-commissural distance (i.e. the
diameter obtained at the cusps opening height) was measured at a plane
approximately 4–6 mm above from the virtual ring. The measurement
of raphe length was performed at the level of best visualization (i.e. the
level where the raphe is displayed in all its length), irrespective from
the plane height. Furthermore, the ratio between raphe length and
area/perimeter-derived aortic annulus diameter was calculated. TAVI
sizing was based on area/perimeter measurement according the type of
valve used (perimeter for self-expandable valve, area for mechanical
expandable). We collect both data in our analyses, and, according to
IFU, we used the are-derived diameter for the Lotus valve and the
Abbreviations
AS aortic stenosis
BAV bicuspid aortic valve
LVOT left ventricular outflow tract
MSCT multi-slice computed tomography
PM pacemaker
PVL paravalvular leak
TAVI transcatheter aortic valve implantation
THV transcatheter heart valve
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
2
perimeter-derived diameter for the Evolut valve”.
2.5. Valve calcium analysis
The absolute calcium value was obtained by adjusting thresholds in
the 3Mensio program in order to compensate overestimation due to
scan protocols and contrast enhancement. A threshold of 450
Hounsfield Units (HU) was used when the density measurement in as-
cending aorta was lower than 300 HU, while a threshold of 850 HU was
used when the density measurement in ascending aorta was higher than
300 HU, as previously described
16,17
(Supplementary appendix, Fig. 1).
We then stratified the patients according to the quartiles of aortic
valve calcium score, as previously described,
18,19
obtaining four classes:
class 0 (from 0 to < 300 mm
3
), class I (from 300 to <600 mm
3
), class
II (from 600 to 1000 mm
3
), and class III (> 1000 mm
3
).
2.6. Post-procedural MSCT
Post implant measurements of THV dimensions (perimeter and
area) at inflow level, raphe level, and at stent waist level (defined as the
smallest dimension observed at any plane of the implanted prosthesis)
were obtained both for Lotus than for Evolut valve. Eccentricity index
and valve expansion index was calculated only in Lotus valve, given
that its symmetrical shape. Eccentricity was defined as 1-(minimal
diameter/maximal diameter) and was calculate both at inflow level that
at narrowest level. Expansion index was defined as the ratio between
the effective measured area of the valve and the nominal prosthesis
valve area of the prosthesis, and was calculated both at the inflow level
(generally corresponding to the annulus) and at the narrowest level
(generally corresponding with the raphe plane). Under-expansion was
defined as a ratio between the MSCT area and the nominal area ≤90%.
For Evolut valve, since its asymmetrical shape, measurement of ex-
pansion index at raphe level was not applicable, considering that
nominal area of prosthesis changes depending the height.
The raphe length (defined as the longest measurable dimension of
the structure) after deployment was measured at the same level of pre-
procedural CT analyses, and the ratio between the pre and post-pro-
cedural length was calculated for the assessment of raphe shortening for
both valve type.
2.7. Statistical analyses
Continuous variables were presented as median and interquartile
range (IQR) and compared using the ANOVA test. Inter-group com-
parison was performed with Bonferroni correction. Categorical vari-
ables were presented as frequency and percentage. Measurements cor-
relations were analysed using Pearson correlation coefficient. A
p < 0.05 was considered as statistically significant. Statistical analyses
were performed with the SPSS (IBM SPSS version 23.0, New York).
3. Results
3.1. First step
Patient characteristics. During the inclusion period, 19 patients
with a BAV type 1 were enrolled in our study. Mean age was
75.4 ± 8.6 years, and 13 patients (68.4%) were males. Mean Society
of Thoracic Surgeons (STS) score was 3.4 ± 1.7%. All patients were for
TAVI after Heart Team evaluation. All THVs were implanted according
to the operator intentions, trying to obtain an implant as high as pos-
sible (as commonly recommended in BAV compared to tricuspid ones).
Clinical outcomes are presented in Table 1. Procedural success was
achieved in 100% of patients. No in-hospital deaths occurred. Ten pa-
tients (52.7%) had no PVL, while 9 patients (47.3%) had only trivial or
mild PVL at discharge, with no cases of moderate or severe PVL.
Average mean residual gradient was 10.5 ± mmHg. Three patients
Table 1
Procedural data and outcome at discharge in cohort 1.
PT-001 PT-002 PT-003 PT-004 PT-005 PT-006 PT-007 PT-008 PT-009 PT-010 PT-011 PT-012 PT-013 PT-014 PT-015 PT-016 PT-017 PT-018 PT-019 Mean ± SD % (N)
Post dilatation 1 0 0 1 0 1 0 0 1 0 1 0 0 1 1 1 1 1 0 52.6% (10)
PVL* 0 1 0 0 0 1 0 0 0 0 1 1 1 1 1 1 1 0 0 47.3% (9)
Mean *Gradient (mmHg) 13 13 15 10 8 12 14 14 12 10 9 14 9 8 7 8 7 7 11 10.5 ± 2.7
PPM *implanted 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 1 0 0 0 15%(3)
PVL: paravalvular leak assessed with echo evaluation; PPM: permanent pacemaker * evaluated at discharge.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
3
Table 2
Baseline MSCT assessment in cohort 1.
PT-001 PT-002 PT-003 PT-004 PT-005 PT-006 PT-007 PT-008 PT-009 PT-010 PT-011 PT-012 PT-013 PT-014 PT-015 PT-016 PT-017 PT-018 PT-019
ANULUS EVALUATION PRE-
IMPLANT
Area (mm
2
) 510 620 525 724 610 441 491 527 456 563 412 465 461 407 372 685 539 496 652
Perimeter (mm) 82.8 89.2 82.1 97.3 92.2 78.2 79.9 82.6 76 85 73.4 80 78.2 72.8 70.3 94.3 84.1 80 92.5
Ø Derived Area (mm) 25.5 28.1 25.9 30.4 27.9 23.7 25 25.9 24.1 26.8 22.9 24.3 24.2 22.8 21.8 29.6 26.2 25.1 28.7
Ø derived perimeter (mm) 26.3 28.4 26.1 31 29.3 24.9 25.4 26.3 24.2 27.1 23.4 25.6 24.9 23.2 22.4 30 26.8 25.5 28.9
SOV EVALUATION PRE-
IMPLANT
Area SOV (mm
2
) 894.2 1131.2 1257.6 1184.4 1178.4 866.7 888 971.6 952 1022.9 920.5 864 803.4 824.7 664.1 1049 921 865 956
Perimeter SOV (mm) 110 124 133 125 123 108 109 116 112 116 110 108 101 103 93 120 108 115 118
Intercommisural diameter
(mm)
28.6 30 31.2 32.4 31.5 27.3 26.2 26.7 28.2 31 25.2 26 32.5 23.7 29.4 28 24.5 29.5 27.6
Ø SOV right 31.5 35.8 37.4 35.3 36.1 31.4 31.8 32.8 31.5 33.5 30.3 29 31 30.6 28 34 34.1 32.9 33.5
Ø SOV NC 36.5 29.4 37.6 38.4 35.1 30.9 33.8 32.7 37 36.7 36.5 33 31.9 31.2 29.8 39 33.4 36.3 37.8
Ø SOV left 31.9 35.9 38.8 40.2 40.3 34.5 31.3 35 32 35 33 31 29.9 31.4 27.2 35.7 32.3 36.1 35.2
Raphe length (mm) 18.1 16.9 17.5 18.2 15.3 12.6 10.3 13.7 13.9 11 13.6 13 16 10 10.4 11.5 7 14.7 13
Raphe percentage of Ø
annulus
71 60 68 60 55 53 41 53 58 41 59 53 66 44 48 38.8 26 57.6 45
Total calcium score
(mm3)
993 1591 1739 450.4 594 1994 1321 725 1172 1389 1711 447 1263 1711 754 1615 1560 1090 865
Calcium score rating 2 3 3 1 1 3 3 2 3 3 3 1 3 3 2 3 3 3 2
Size valve according
annulus
27 OR 27 OR OR 25 25 27 27 27 25 27 27 25 23 34 34 29 34
Size implanted valve 25 27 23 27 27 23 25 25 25 27 25 25 25 23 23 34 29 26 34
Downsizing yes yes yes yes yes yes no yes yes no no yes yes yes no no yes yes no
Type of valve implanted Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Evolut Evolut Evolut Evolut
SOV=Sinus of Valsalva. Ø = diameter.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
4
(15%) required new PM implantation. Post-dilation was performed in 9
patients (47.3%). No valve embolization or malposition was reported.
MSCT measurements and THV size selection. The THV sizes were
in all cases selected by the treating physician based on pre-procedural
MSCT. In 13 cases, the operators chose to implant a smaller size THV
than the one recommended by the official annulus-based measurements
as indicated in instructions for use (IFUs). In one patient, the valve was
undersized by two sizes, and a 23 mm prosthesis was implanted in an
aortic annulus supposed to fit a 27 mm Lotus valve. Although annulus
measurements were larger than the range covered by the Lotus valve,
three patients received a 27 mm valve prosthesis.
Core-Lab evaluations of pre-procedural MSCT are shown in Table 2.
The ratio between raphe length and area/perimeter-derived aortic an-
nulus diameter was greater than 50% (range 26–71%) in 12 (63%) of
the patients.
As regard to the calcium burden classification, no patients had a
class 0, whereas 3 patients had a class I (mild), 3 patients had a class II
(moderate), and 9 patients had a class III (severe).
Core-Lab analyses of post-procedural MSCT scans are reported in
Table 3. In 13/19 patients (68.4%) the THV was undersized respect to
the IFUs’indications, and the THV expansion index (calculated only in
Lotus valve) ranged from 79% to 103% at inflow level, and from 62% to
98% at waist level (93.2 ± 6.3% vs 82.8 ± 9.8%, respectively;
p = 0.005) (Table 3). In 9/15 Lotus valve patients, the smallest THV
dimension (i.e the “waist”level) was at the raphe level, while in 3/4
Evolut valve patients, though the asymmetric shape of the valve, an
evident indentation was observed at raphe level (supplementary ap-
pendix Fig. 2). A high correlation between the valve area measured at
raphe level on pre-procedural MSCT and at THV waist level on post-
procedural MSCT was observed (p < 0.001) (Fig. 1). Moreover,
comparing pre- and post-implant raphe lengths, a foreshortening from
27.2% to 86% was observed.
Higher calcium scores were significantly correlated with reduced
raphe length foreshortening (r Pearson = -0.59, p = 0.02) (Fig. 2A). A
significantly smaller reduction of the raphe length (p < 0.001) and a
less pronounced THV expansion (p = 0.048) were observed in BAVs
with class II and III calcium scores as compared to those with class I
(Fig. 2B).
In order to assess measurements reproducibility, we also evaluated
inter-observer correlation of the principal parameters: mean diameter
of SOV, raphe length, inter-commissural diameter at the baseline MSCT,
and prosthesis area at the waist level at post-procedural MSCT. All
measurements showed a high reproducibility, except for the inter-
commissural diameter (Fig. 3).
3.2. Proposed algorithm
In the cohort 1, 68% (13/19) of THV were undersized in respect to
the annulus measurement. Nevertheless, prosthesis expansion at the
inflow level was satisfactory, while it was observed to be suboptimal at
the waist level, often corresponding to the raphe level and to the point
of maximum calcium burden. Therefore, to optimize sizing for BAV
stenosis and decide when and how to undersize the prosthesis in respect
to the annulus dimensions and IFU choice, we considered an algorithm
that takes into account the two parameters with a significant correla-
tion to valve expansion and highly reproducible: calcium score and
raphe length. In particular, the latter demonstrated a high inter-ob-
server reproducibility, and, both parameters showed to have a key role
in THV under expansion. It was also observed that the waist level of the
implanted prosthesis was not only due to the raphe, but in some pa-
tients, predominantly dependent to the high calcium burden
(Supplementary appendix, Figs. 2–3).
The CASPER algorithm (CASPER: Calcium Algorithm Sizing for
bicusPid Evaluation with Raphe) is described in Fig. 4. First, perimeter
and area annulus measurement must be calculated according standard
practice, and derived mean diameter must be obtained by perimeter or
Table 3
Post-implantation MSCT assessment in cohort 1.
PT-001 PT-002 PT-003 PT-004 PT-005 PT-006 PT-007 PT-008 PT-009 PT-010 PT-011 PT-012 PT-013 PT-014 PT-015 PT-016 PT-017 PT-018 PT-019
Raphe level Valve area (mm
2
) 301.6 415.7 336.6 560.7 456 325.7 361.8 396.8 440.3 508.8 402.8 426 380.9 349.9 343.5 370 352 313 345
Valve perimeter (mm) 62.8 73.6 66.3 84 75.9 64.6 67.8 71.3 74.7 80.9 71.3 74 70.5 66.9 66.3 69.5 66.5 63.9 65.9
Raphe length (mm) 9.4 8.7 11.7 4.6 4.5 6.7 7.5 5.9 5.8 4.6 3.5 4.3 7 5.2 4.8 10 5 9.6 6
Raphe shortening (%) 48.1 48.5 33.2 74.7 70.6 46.8 27.2 59.1 58.3 58.2 73.2 66.9 56.3 48 53.8 86 71.4 65 54.5
At Waist level Valve area (mm
2
)299.4 389.9 334.6 545.1 ––351.4 –––––367.8 –––339 ––
Valve perimeter (mm) 63 71.5 66.2 82.9 ––67.3 –––––69.7 –––64.2 ––
At Inflow level Valve area (mm
2
) 474 557.3 395.8 554.9 508.2 405.1 505.7 392.3 455 528.3 468.1 455.7 390.2 403.4 388.7 480 405 356 354
Valve Perimeter (mm) 76 84.5 71.2 83.6 80 71.6 80.1 71 76 82.9 77.1 76.9 72 71.9 70.2 78.6 74.2 68.7 66.8
Expansion index inflow level(%) 96.6 97.3 95.3 96.9 88.8 97.5 103 79.9 92.7 92.3 95.4 92.8 79.5 97.1 93.6 NA NA NA NA
Expansion index at narrowest level(%) 61.5 72.6 81.1 98.0 92.9 78.4 73.7 80.9 89.7 88.9 97.0 86.8 74.8 84.3 82.7 NA NA NA NA
Eccentricity index at inflow leve (%)l 8 10 19 4 7 12 4 4 0 17 12 19 24 15 16 NA NA NA NA
Eccentricity index at waist level (%) 24 24 28 4 4 17 14 24 4 12 4 12 28 12 28 NA NA NA NA
PVL 0100010000111111100
Mean Gradient 13 13 15 10 8 12 14 14 12 10 9 14 98787711
In bold are indicated the narrowest measure of the valve area.
- Empty space indicate that measure at waist level correspond with measure at raphe level.
- NA = not applicable.
- Eccentricity index is calculated as 1-(Min diameter/Max diameter). Eccentricity is defined > 10%.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
5
by area according the type valve. The first step to consider in the flow
chart is the calcium score volume. In patients with high calcium score
(> 300 mm
3
), a 1 mm must be subtract from the area/perimeter de-
rived mean diameter at the annulus level, due to the “calcium impact”.
The second step is consider the raphe length, which can be more or less
than 50% of area/perimeter derived mean diameter at the annulus
level. At this point, four (2 × 2) combinations are possible, due to the
combination of calcium volume and raphe length. If raphe length is
more than 50% of area/perimeter derived mean diameter at the an-
nulus level, a 1 mm must be subtract from mean derived diameter. The
third step is the calcium distribution. Further 0.5 mm can be subtracted
from the mean derived diameter when a high burden of calcium is
distributed predominantly on the raphe site. The final diameter ob-
tained after these three step is used to choose the valve size. Table 4
shows possible sizing according to three methodologies currently used
with different kinds of THV: annulus sizing, inter-commissural sizing
and the proposed algorithm. An example of algorithm application is
further reported in the supplementary appendix Fig. 4.
3.3. Second step
Validation of algorithm: The proposed algorithm was tested in a
new cohort of 21 consecutive patients with BAV type 1 all treated with
self-expanding prosthesis. In 15 cases (71.4%) the algorithm led to the
implantation of an undersized THV respect to the annulus dimension.
Results are reported in Table 5. All cases met the procedural success
criteria. Thirteen patients (61.9%) had no PVL, while 8 patients
(38.1%) had only trivial or mild PVL at discharge, with no cases of
moderate or severe PVL, Mean post-procedural gradient was
8.8 ± 2.6 mmHg. Post-dilation was performed in 10 patients (47.6%).
The rate of PPM implantation was low (14.2%).
4. Discussion
This study performed an analysis of bicuspid aortic valve anato-
mical aspects and their modifications relationship after prosthesis im-
plant, and their correlation with valve expansion.
We observed three principal aspects:
1. In BAVs, the anchoring of THV does not occur at annulus level, as
usually happens in tricuspid valves, but often corresponds to the
raphe level. This level was the narrowest in the majority of patients
in cohort 1. In BAVs, the anatomical structures, in particular the
raphe, generate an asymmetric valve orifice, which is located at a
supra-annular plan and can be smaller than the virtual aortic an-
nular ring.
2. The raphe length and the calcium burden together with its dis-
tribution affect leaflets expansion, which in turn influences the
correct prosthesis sizing.
3. The raphe length and calcium burden should be considered in BAV
sizing for a more accurate dimension evaluation.
Over the past few years, the concept of THV undersizing, especially
in highly calcified BAV has been a topic of discussion. The degree of
undersizing or “correct”sizing remains a challenge in the absence of a
dedicated sizing algorithm.
In this analysis, we only considered BAV type I, as it is the most
frequent BAV subtype and considered one of the most challenging. A
clear observation in the first group was that the majority of THV (13/
19, 68%) were undersized with respect to the IFUs’indications; how-
ever, the prostheses were well expanded at inflow level but not at the
waist level, which was often at raphe level. This observation suggests
that the raphe is the anatomical structure that limits prosthesis ex-
pansion; so, it should be included in the sizing process. The operators
that adopt undersizing methods according to a suprannular measure-
ment consider as crucial the inter-commissural (IC)
20
diameter obtained
Fig. 1. Correlation between valve area measured at raphe level and at waist level.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
6
at the cusps opening height. However, this method has never been fully
described in literature. We observed that the measurement of the IC
distance is not easily reproducible, most likely because of the lack of a
standardized method. Moreover recently a study comparing annulus
sizing with IC one has underlined a high number of improper size se-
lection.
21
The direct measure of the perimeter or area at IC level de-
pends on many factors (mid systole phase of MSCT, height of the
measurement etc.), therefore it can be difficult to reproduce and mostly
empiric. Recently a THV size selection based on balloon sizing has been
proposed.
20,22
This technique relies on cusp imprint during balloon
expansion, and is not determined by sizing but only by the operator skill
and experience; moreover, it sometimes requires multiple dilations with
increasing diameter of the balloon.
The raphe is the anatomical entity typical of this subtype of BAV,
and together with calcium, reduces prosthesis expansion. These features
increase the risk of a residual gradient when an oversized THV is used
or of an important PVL if the size in not correct and in the worst hy-
pothesis an annulus rupture.
Size selection based on the impact of the raphe allows a correct
adaption of the device. In some patients of the first cohort, we observed
adifferent THV size choice than the one that would have been selected
considering only calcium score or raphe length. The importance of the
calcium distribution together with the other two parameters can be
observed in pt. 3# (supplementary appendix, Fig. 5) of the first cohort,
where calcium score was high and the distribution was mainly at the
raphe side the THV implantation. In this patient, the valve implanted
was two size smaller than the one dictated by annulus measurements,
with no PVL at the end of the procedure. For these reasons, a third step
was added to the sizing flowchart subtracting an additional 0.5 mm
when the calcium is prevalent at the raphe site.
Adifferent situation was observed in two other patients (#11 and
#13), where the calcium burden was high but mainly located at the
opposite side of the raphe. In addition in these patients, the IC distance
would have suggested a larger prosthesis than the one indicated by the
algorithm (Supplementary appendix Fig. 5).
Importantly, this method pertains both to the annulus derived dia-
meter dimension, which is the standard IFU MSCT measurement, and to
the other principal anatomical features peculiar in BAV, resulting in a
practical and easy-to-use algorithm for BAV sizing for TAVI.
Finally, in the second cohort of 21 patients in whom the algorithm
was validated and all THV were chosen with the algorithm; an under-
sized THV in respect to the annulus measurement was selected in 71.4%
of cases. Nevertheless, we were able to obtain a 100% procedural
success rate without valve malposition or embolization. It is also im-
portant to underline that none of the patients showed a high trans-
prosthetic gradient and the mean gradient was similar to the one
usually obtained in tricuspid valves.
23
Moreover, differently than in
other experiences, a small number of new PPM and absence of mod-
erate PVL was seen.
24
5. Study limitations
This study was conducted in a small cohort of patients, though the
total number of both cohorts is acceptable considering the low per-
centage of BAV in Western countries. Nevertheless, our cohort has a
higher number of patients compared to previous reported studies about
Fig. 2. A) Pearson correlation between absolute value of calcium score and raphe shortening; B)Expansion index and raphe shortening correlation according to
calcium score class.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
7
BAVs sizing 20.
We are aware that the THVs implanted in the observational cohort
predominately are different from those implanted in the validation one,
but many reasons can be adduced. The mechanical expanding valve was
chosen for its adaption to the aortic valve anatomy and the purpose of
observing an uniform cohort of treated patients, but unfortunately it
has been temporally removed from the market. Moreover, the Lotus
valve with its symmetrical shape allow reproducible measurement in
the post-procedural CT, which will be not possible with an asymme-
trical shape valve. However, we then evaluated other 4 patients treated
with Evolut R valve. Also in these 4 cases we observed an analogues
behavior of the raphe, which lead to the more constrain section of the
prosthesis. We further believe that the algorithm that has been vali-
dated with a self-expanding THV can be adopted also with other valve
types, but needs further validation. Lastly, the method has been tested
only in type I BAV, since this morphology is not only the most frequent
Fig. 3. Inter-observer correlation of: A) pre-procedural raphe length, B) post-implant valve area at the narrowest level, C) inter-commissural length, D) mean
diameter of SOV.
Fig. 4. Algorithm flow chart.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
8
Table 4
Sizing evaluation in cohort 1 according multiple methods.
PT-001 PT-002 PT-003 PT-004 PT-005 PT-006 PT-007 PT-008 PT-009 PT-010 PT-011 PT-012 PT-013 PT-014 PT-015 PT-016 PT-017 PT-018 PT-019
Area/perimeter Derived diameter (mm) 25.5 28.1 25.9 30.4 27.9 23.7 25 25.9 24.1 26.8 22.9 24.3 22.2 22.8 21.8 685 539 496 652
Raphe percentage of Ø 71 60 68 60 55 53 41 53 58 41 59 53 66 44 48 38.8 26 57.6 45
Calcium score rating 2 3 3 1 1 3 3 2 3 3 3 1 3 3 2 3332
Intercommisural diameter 28.6 30 31.2 32.4 31.5 27.3 26.2 26.7 28.2 31 25.2 26.3 32.5 23.7 29.4 28 25 25.5 31
Ø Algorithm calculated (mm) 23.5 26.1 23.9 29.4 26.9 21.7 24 23.9 22.1 25.8 20.9 23.3 20.2 21.8 20.8 29 25.8 23.5 27.9
Size valve according intercommisural diameter OR OR OR OR OR 27 27 27 OR OR 25 27 OR 23 OR 34 29 29 OR
Size valve according derived annulus 27 OR 27 OR OR 25 25 27 25 27 23 25 23 23 23 34 34 29 34
Size valve according algorithm 25 27 25 OR 27 23 25 25 23 27 23 25 23 23 23 34 29 26 34
Size implanted valve 25 27 23 27 27 23 25 25 25 27 25 25 25 23 23 34 29 26 34
Type of valve implanted Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Lotus Evolut Evolut Evolut Evolut
Ø: diameter; OR: out of range.
Table 5
Validation of algorithm in cohort 2.
PT-001 PT-002 PT-003 PT-004 PT-005 PT-006 PT-007 PT-008 PT-009 PT-010 PT-011 PT-012 PT-013 PT-014 PT-015 PT-016 PT-017 PT-018 PT-019 PT-020 PT-021
Perimeter annulus 101 87 98 78.7 78 72.5 88 75.4 79.2 77.5 85.2 86.2 84.1 71.7 82 68 68 69 69.9 77 72
Area Annulus 803 582 761 500 478 406 583 431 492 462 536 574 539 408 508 358 356 373 378 466 397.8
Ø Derived Area (mm) 32 27.2 31.1 25.1 24.7 22.7 27.3 23.4 25 24.3 26.1 27 26.2 22.4 25.5 21.1 21.3 21.8 22 24.4 22.5
Ø Derived perimeter (mm) 32.1 27.7 31.2 25 24.9 23.1 28 24 25.2 24.7 27.1 27.4 26.8 22.5 26.4 21.7 21.7 21.9 22.1 24.6 22.9
Raphe length 20 16.5 18.5 15.8 8.6 11.1 14.5 10 12 12 15 12.2 7 14 10 7.8 11 11.2 11 12.6 11.7
Raphe percentage of Ø 62.3 57.9 59.2 61.9 34.5 48 51.7 41.6 47.6 48.5 55.3 44.5 26.1 63 39.6 36.4 50.6 51.1 50 51 52
Calcium score rating 1215 2336 1881 1090 759 837 442 984 365 393 1057 1181 1560 1652 2820 789 327 4463 2426 769 492
Asymmetrical calcium distribution 000000000011011011101
Intercommisural diameter 30 27 32.5 26.4 25.1 23.7 28 22.7 23 22.5 30 28.7 28 22 28.2 21.7 23.8 23.6 22.1 24.3 22.1
Ø Algorithm calculated (mm) 30 25.7 29.1 23 23.9 22.1 26 23 24.2 23.7 24.6 25.9 25.8 20 24.9 20.7 19.7 19.9 20.1 22.6 20.9
Size valve according intercommisural
diameter
34 34 OR 34 29 29 34 26 26 26 OR 34 34 26 34 26 29 29 26 29 26
Size valve according derived annulus OR 34 OR 29 29 29 34 29 29 29 34 34 34 26 34 26 26 26 26 29 26
Size valve according algorithm 34 29 34 26 29 26 29 26 29 29 29 29 29 23 29 26 23 23 26 26 26
Size implanted valve 34 29 34 26 29 26 29 26 29 29 29 29 29 23 29 26 23 23 26 26 26
POSTDILATATION 010100000011110101011
PVL 100001010101100000101
PPM implanted 000010000000000000011
Mean Gradient 10 95713959127131410871178966
Ø: diameter; OR: out of range. PVL: paravalvular leak.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
9
but also the most challenging; nevertheless, we are convinced that it
could be translated to the other BAV types after a wider experience.
6. Conclusions
Our results confirmed that in BAV there is frequently a discrepancy
between annulus measurements and the size of the THV finally im-
planted. Through analyses of pre- and post-procedural MSCT, we were
able to identify which are the main parameters that can provide THV
valve sizing with good hemodynamic results. However, this study needs
to be evaluated in a larger population of BAV patients with different
morphological varieties; moreover, our method has to be compared to
other sizing techniques and tested with different THV types.
Declaration of competing interest
Lars Søndergaard has received proctor and speaker fees, as well as
institutional research grants from Boston Scientific, Natick, MA and
Medtronic, Minneapolis. Francesco Bedogni, Fausto Castriota and Anna
Sonia Petronio have received proctor fees from Boston Scientific,
Natick, MA and Medtronic, Minneapolis.
Acknowledgments
The authors would like to express their gratitude to Rene Ceuleers
for his great support and contribute in MSCT data analysis and imaging
elaboration.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.jcct.2020.01.007.
References
1. Michelena HI, Della Corte A, Prakash SK, Milewicz DM, Evangelista A, Enriquez-
Sarano M. Bicuspid aortic valve aortopathy in adults: incidence, etiology, and clinical
significance. Int J Cardiol. 2015;201:400–407.
2. Roberts WC, Ko JM. Frequency by decades of unicuspid, bicuspid, and tricuspid
aortic valves in adults having isolated aortic valve replacement for aortic stenosis,
with or without associated aortic regurgitation. Circulation. 2005;111(7):920–925.
3. Roberts WC, Janning KG, Ko JM, Filardo G, Matter GJ. Frequency of congenitally
bicuspid aortic valves in patients >/=80 years of age undergoing aortic valve re-
placement for aortic stenosis (with or without aortic regurgitation) and implications
for transcatheter aortic valve implantation. Am J Cardiol. 2012;109(11):1632–1636.
4. Sievers HH, Schmidtke C. A classification system for the bicuspid aortic valve from
304 surgical specimens. J Thorac Cardiovasc Surg. 2007;133(5):1226–1233.
5. Smith CR, Leon MB, Mack MJ, et al. Transcatheter versus surgical aortic-valve
replacement in high-risk patients. N Engl J Med. 2011;364(23):2187–2198.
6. Reardon MJ, Van Mieghem NM, Popma JJ, et al. Surgical or transcatheter aortic-
valve replacement in intermediate-risk patients. N Engl J Med.
2017;376(14):1321–1331. https://doi.org/10.1056/NEJMoa1700456.
7. Yoon SH, Bleiziffer S, De Backer O, et al. Procedural and clinical outcomes in
transcatheter aortic valve replacement for bicuspid versus tricuspid aortic valve
stenosis. J Am Coll Cardiol. 2017;69(21):2579–2589. https://doi.org/10.1016/j.jacc.
2017.03.017.
8. Tchetche D, Van Mieghem NM. New-generation TAVI devices: description and spe-
cifications. EuroIntervention. 2014;10 Suppl:U90–u100.
9. Yoon SH, Lefevre T, Ahn JM, et al. Transcatheter aortic valve replacement with early-
and new-generation devices in bicuspid aortic valve stenosis. J Am Coll Cardiol.
2016;68(11):1195–1205.
10. Watanabe Y, Chevalier B, Hayashida K, et al. Comparison of multislice computed
tomography findings between bicuspid and tricuspid aortic valves before and after
transcatheter aortic valve implantation. Cathet Cardiovasc Interv.
2015;86(2):323–330.
11. Yousef A, Simard T, Pourdjabbar A, et al. Performance of transcatheter aortic valve
implantation in patients with bicuspid aortic valve: systematic review. Int J Cardiol.
2014;176(2):562–564.
12. Al-NajafiS, Sanchez F, Lerakis S. The crucial role of cardiac imaging in transcatheter
aortic valve replacement (TAVR): pre- and post-procedural assessment. Curr Treat
Options Cardiovasc Med. 2016;18(12):70.
13. Kappetein AP, Head SJ, Généreux P, et al. Updated standardized endpoint definitions
for transcatheter aortic valve implantation: the Valve Academic Research
Consortium-2 consensus document. J Am Coll Cardiol. 2012;60(15):1438–1454.
14. World Medical Association Declaration of Helsinki: ethical principles for medical
research involving human subjects. Jama. 2013;310(20):2191–2194.
15. Philip F, Faza NN, Schoenhagen P, et al. Aortic annulus and root characteristics in
severe aortic stenosis due to bicuspid aortic valve and tricuspid aortic valves: im-
plications for transcatheter aortic valve therapies. Cathet Cardiovasc Interv : Off.J.
Soc. Cardiac Angiography Interv. 2015;86(2):E88–E98.
16. Jilaihawi H, Makkar RR, Kashif M, et al. A revised methodology for aortic-valvar
complex calcium quantification for transcatheter aortic valve implantation. Eur Heart
J Cardiovasc Imaging. 2014;15(12):1324–1332.
17. Hur J, Kim YJ, Lee HJ, et al. Left atrial appendage thrombi in stroke patients: de-
tection with two-phase cardiac CT angiography versus transesophageal echocardio-
graphy. Radiology. 2009;251(3):683–690.
18. Bellasi A, Ferramosca E, Muntner P, et al. Correlation of simple imaging tests and
coronary artery calcium measured by computed tomography in hemodialysis pa-
tients. Kidney Int. 2006;70(9):1623–1628.
19. Podlesnikar T, Prihadi EA, van Rosendael PJ, et al. Influence of the quantity of aortic
valve calcium on the agreement between automated 3-dimensional transesophageal
echocardiography and multidetector row computed tomography for aortic annulus
sizing. Am J Cardiol. 2018;121(1):86–93.
20. Liu X, He Y, Zhu Q, et al. Supra-annular structure assessment for self-expanding
transcatheter heart valve size selection in patients with bicuspid aortic valve. Cathet
Cardiovasc Interv. 2018;91(5):986–994. https://doi.org/10.1002/ccd.27467.
21. Kim WK, Renker M, Rolf A, et al. Annular versus supra-annular sizing for TAVI in
bicuspid aortic valve stenosis. EuroIntervention. 2019;15(3):e231–e238.
22. Shivaraju A, Thilo C, Ott I, et al. Tools and Techniques - clinical: Fluoroscopic balloon
sizing of the aortic annulus before transcatheter aortic valve replacement (TAVR) -
follow the "right cusp rule. EuroIntervention. 2015;11(7):840–842.
23. Falk V, Wohrle J, Hildick-Smith D, et al. Safety and efficacy of a repositionable and
fully retrievable aortic valve used in routine clinical practice: the RESPOND study.
Eur Heart J. 2017;38(45):3359–3366.
24. Mylotte D, Lefevre T, Sondergaard L, et al. Transcatheter aortic valve replacement in
bicuspid aortic valve disease. J Am Coll Cardiol. 2014;64(22):2330–2339.
A.S. Petronio, et al. Journal of Cardiovascular Computed Tomography xxx (xxxx) xxx–xxx
10