Novel Multiphase Assessment for Predicting Left Ventricular Outflow Tract Obstruction Before Transcatheter Mitral Valve Replacement

JACC: CARDIOVASCULAR INTERVENTIONS

VOL.

-, NO. -, 2019

ª 2019 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION PUBLISHED BY ELSEVIER

Novel Multiphase Assessment for Predicting Left Ventricular Outflow Tract Obstruction Before Transcatheter Mitral Valve Replacement Christopher U. Meduri, MD, MPH,a Michael J. Reardon, MD,b D. Scott Lim, MD,c Elliot Howard, PHD,d Gan Dunnington, MD,e David P. Lee, MD,f David Liang, MD,f Robert Gooley, MBBS, PHD,g Daniel O’Hair, MD,h,* Martin K. Ng, MBBS, PHD,i Antony Walton, MD,j Konstantinos Spargias, MD,k Daniel Blackman, MD,l Augustin Coisne, MD, PHD,m David Hildick-Smith, MD,n Marine De Gouy, MS,d Sharla Chenoweth, MS,o Saibal Kar, MD,p,y Patrick M. McCarthy, MD,q Nicolo Piazza, MD,r Atif Qasam, MD,s Randolph P. Martin, MD,t Martin B. Leon, MD,u Michael J. Mack, MD,v David H. Adams, MD,w Vinayak Bapat, MBBS, MS, MCHx,y

ABSTRACT OBJECTIVES This study proposes a physiologic assessment of left ventricular outflow tract obstruction (LVOTO) that accommodates changes in systolic flow and accounts for the dynamic neo–left ventricular outflow tract (LVOT). BACKGROUND Patients considered for transcatheter mitral valve replacement trials often screen-fail because of the perceived risk of LVOTO. In the Intrepid Global Pilot Study, assumed risk of LVOTO was based on computed tomography estimates of the neo-LVOT area computed at end-systole. However, this may overestimate actual risk. METHODS Retrospective analyses were performed for screen-failed patients for potential LVOTO (n ¼ 33) and treated patients (n ¼ 29) with available dynamic computed tomography. A multiphase assessment of the neo-LVOT area was performed and represented as: 1) multiphase average; and 2) early systolic value. Prospective evaluation was performed in 9 patients approved for enrollment with multiphase and early systole methods that would have previously screen-failed with the end-systolic approach. RESULTS Of 166 patients screened for possible inclusion; 32 were screen-failed for nonanatomical reasons. Screen failure for assumed LVOTO risk occurred in 37 of 134 (27.6%) patients. Retrospective analysis indicated a potential enrollment increase of 11 of 33 (33.3%) and 18 of 33 (54.5%) patients using multiphase and early systolic assessment methods. In the prospective cohort, there were no clinical observations of LVOTO 30 days post-procedure, despite assumed risk based on end-systolic estimates. CONCLUSIONS Multiphase, and specifically early systolic, assessment of the neo-LVOT may better determine risk of LVOTO with transcatheter mitral valve replacement compared with end-systolic estimates. This novel approach has the potential to significantly increase patient eligibility, with over one-half of patients previously screen-failed now eligible for treatment. (J Am Coll Cardiol Intv 2019;-:-–-) © 2019 by the American College of Cardiology Foundation.

From the aMarcus Heart Valve Center, Piedmont Heart Institute, Atlanta, Georgia; bDepartment of Cardiothoracic Surgery, Houston Methodist DeBakey Heart and Vascular Center, Houston, Texas; cDivision of Cardiology, Bon Secours, Richmond, Virginia;

d

Coronary and Structural Heart Research and Innovation, Medtronic, Redwood City, California; eDepartment of

Cardiothoracic Surgery, St. Helena Hospital, St. Helena, California; fDepartment of Cardiovascular Medicine, Stanford University, Stanford, California;

g

Monash Cardiovascular Research Centre and MonashHeart, Melbourne, Australia;

h

Department of

Cardiothoracic Surgery, Aurora Health Center, Milwaukee, Wisconsin; iDepartment of Cardiology, Royal Prince Alfred Hospital, Sydney, Australia; jDepartment of Cardiology, The Alfred Hospital, Melbourne, Australia; kDepartment of Transcatheter Heart Valves, Hygeia Hospital, Athens, Greece; lDepartment of Cardiology, Leeds Teaching Hospitals NHS Trust, Leeds, United Kingdom; mHeart Valve Center, Lille University Hospital, Lille, France; nDepartment of Interventional Cardiology, Sussex Cardiac Centre, Brighton and Sussex University Hospitals, United Kingdom; oCoronary and Structural Heart, Department of Biostatistics, Medtronic, Mounds View, Minnesota; pSmidt Heart Institute, Cedars Sinai Medical Center, Los Angeles, California; qBluhm Cardiovascular Institute and Division of Cardiac Surgery, Northwestern University, Chicago, Illinois; rDepartment of Medicine,

ISSN 1936-8798/$36.00

https://doi.org/10.1016/j.jcin.2019.06.015

2

Meduri et al.

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

ABBREVIATIONS AND ACRONYMS CT = computed tomography LVOT = left ventricular outflow tract

LVOTO = left ventricular

MR = mitral regurgitation TEE = transesophageal echo TMVR = transcatheter mitral

TTE = transthoracic echocardiography

ranscatheter mitral valve replace-

been a clinically significant LVOTO in those patients

ment (TMVR) has emerged as a

treated.

less invasive alternative to surgery

To address this, we propose a physiologic assess-

(1–3). However, the complex mitral valve

ment of LVOTO that accounts for the changes in flow

anatomy and concerns for the risk of left

throughout systole, as well as the dynamic nature of

ventricular

outflow tract obstruction

valve replacement

T

-, NO. -, 2019 - 2019:-–-

LVOT Assessment in TMVR

outflow

tract

obstruction

the neo-LVOT. Specifically, we hypothesize that an

(LVOTO) have resulted in high screen-

assessment of both multi- and early systolic phase

failure rates in recent trials (1,2). In TMVR,

would better predict clinically relevant LVOTO, and

various factors have been identified which

that early systolic phase may be the most accurate

contribute to LVOTO, specifically device pro-

predictor, as blood flow out of the ventricle in pa-

trusion into the left ventricle, aortomitral

tients with mitral regurgitation (MR) varies with time

angle, and anterior leaflet displacement (4).

with substantial flow in early systole. If proven, these

Pre-procedural contrast-enhanced dynamic

advances may potentially reduce the screen-failure

computed tomography (CT) has been used

rate for high LVOTO risk.

to identify patients at increased anatomical risk for LVOTO. To date, these methods have been performed

METHODS

using end-systolic measurements (5–7). Within the ongoing Intrepid Global Pilot Study (1),

PATIENTS. Retrospective

cohort. The initial analysis

assessment of LVOTO risk was based on CT estimates

cohort comprised 166 patients screened (screen-failed

of the neo-LVOT area computed at end-systole. This

and screen-accepted) in the Intrepid Global Pilot

end-systolic assessment may represent a worst-case

Study (1). A total of 50 treated patients from the early

approximation of the neo-LVOT area and risk for

feasibility study were considered for the present

LVOTO. Early experience in this trial has demon-

study and the analysis cohorts are summarized in

strated an approximately 40% screen-failure rate

Figure 1. Retrospective analyses were performed for

among patients within the treatable range for annulus

patients screen-failed for LVOTO concerns (n ¼ 33)

size using this method. Furthermore, there has not

and for treated patients (n ¼ 29) with available

Division of Cardiology, McGill University Health Centre, Montreal, Canada; sDivision of Cardiology, University of California, San Francisco, San Francisco, California; tDivision of Cardiology, Emory University, Atlanta, Georgia;

u

Division of Cardiology,

Columbia University Medical Center, New York, New York; vDepartment of Cardiothoracic Surgery, Baylor Scott and White Health, Plano, Texas; wDepartment of Cardiovascular Surgery, Mount Sinai Medical Center, New York, New York; xSt Thomas’ Hospital, London, United Kingdom; and the yDepartment of Surgery, New York Presbyterian/Columbia University Medical Center, New York, New York. *Dr. O’Hair is currently at Boulder Heart, Boulder, Colorado. yDr. Kar is currently at the Center for Advanced Cardiac and Vascular Interventions, Los Angeles, California. The study was supported by Medtronic. Dr. Meduri has served as a consultant for Medtronic and Boston Scientific; and on the advisory board for Boston Scientific. Dr. Reardon has received personal fees from Medtronic and Boston Scientific outside the submitted work. Dr. Lim has received institutional grant support from Medtronic. Dr. Howard, Ms. De Gouy, and Ms. Chenoweth are employees and shareholders of Medtronic. Dr. Dunnington has received research support and consulting honoraria and served on a medical review panel for Medtronic. Dr. Liang has served as consultant for and received consulting honoraria from Medtronic and Twelve. Dr. Gooley has received proctoring fees from Medtronic and Boston Scientific; and consulting fees from Boston Scientific. Dr. O’Hair has served as a consultant to Medtronic. Dr. Ng has served as an advisor to Medtronic and Edwards Lifesciences; and received research grant support from Edwards Lifesciences; and served as a consultant and proctor for Medtronic, HighLife, and Microport. Dr. Walton has served as a proctor and advisor for Medtronic. Dr. Spargias has served as consultant and proctor for Medtronic. Dr. Blackman has served as a consultant and proctor for Medtronic and Boston Scientific. Dr. Hildick-Smith has served as a proctor and advisor for Medtronic, Boston Scientific, and Edwards Lifesciences. Dr. Kar has received consulting honoraria for Abbott Vascular, Boston Scientific, Medtronic, and W.L. Gore. Dr. McCarthy has served as a consultant for and received royalties from Edwards Lifesciences; received consulting honoraria from AtriCare; has received speaker fees from Medtronic; and served on an advisory board for Abbott Vascular. Dr. Piazza has served as a proctor for HighLife, Microport, and Medtronic. Dr. Qasim has received grant support from Abbott Vascular, Siemens, and Medtronic; and served as a consultant for Siemens and Biotelemetry. Dr. Martin has served on the executive committee for Medtronic. Dr. Mack has served on the executive board for the Medtronic APOLLO trial; and served on the Speakers Bureau and as a principal investigator for Medtronic and Edwards Lifesciences. Dr. Adams has served as the national co-principal investigator of the Medtronic APOLLO Pivotal Trial and the Medtronic CoreValve US Pivotal Trial. In addition, The Icahn School of Medicine at Mount Sinai receives royalty payments from Edwards Lifesciences and Medtronic for intellectual property related to development of valve repair rings. Dr. Bapat has served as a consultant for Medtronic, Edwards Lifesciences, 4C, and Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Manuscript received January 30, 2019; revised manuscript received June 3, 2019, accepted June 6, 2019.

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019

Meduri et al.

- 2019:-–-

LVOT Assessment in TMVR

F I G U R E 1 Analysis Cohorts for the Present Study

A total of 50 treated patients from the Early Feasibility Study of the Intrepid Global Pilot Study were considered for the retrospective analysis portion, while 9 patients who otherwise would have been screen-failed under end-systolic assessment methods were prospectively analyzed. *One patient was enrolled under multiphase analysis and is therefore counted in the prospective analysis: total enrollment for early feasibility study–treated patients in this analysis was 50. CT ¼ computed tomography; LVOT ¼ left ventricular outflow tract; LVOTO ¼ left ventricular outflow tract obstruction; TTE ¼ transthoracic echocardiography.

contrast-enhanced dynamic CT. The cohort of treated

with a minimum of 10 phases in 10% increments

patients was chosen based on patients’ end-systolic

across the entire RR interval (0% to 90%). All 10

neo-LVOT area predictions being #250 mm 2. A sub-

phases of the 4-dimensional CT reconstructions were

set of these treated patients (n ¼ 20) were used for

exported in DICOM (Digital Imaging and Communi-

paired analyses of CT and echocardiographic assess-

cations in Medicine) format. These imaging recom-

ment at baseline and 30 days post-procedure. The

mendations were used for pre- and post-procedural

retrospective analyses included estimates of pre-

assessments. CT images were imported and analyzed

dicted neo-LVOT areas at end-systole and assessment

with the 3Mensio Structural Heart software version

at multi- and early systolic phases.

9.0 and corresponding Mitral package (Pie Medical

P r o s p e c t i v e v a l i d a t i o n c o h o r t . Between June 2017

Imaging BV, Maastricht, the Netherlands).

and Jun 2018, 9 patients were enrolled in the Intrepid Global Pilot Study using multiphase (n ¼ 6) or early systolic (n ¼ 3) criteria (Figure 1). These patients would have otherwise been screen-failed due to predicted LVOTO risk based on end-systolic CT screening. Baseline and 30-day echocardiography data are reported.

END-SYSTOLIC LVOT ASSESSMENT. End-systole was

chosen as the cardiac phase of aortic valve closure, typically at the 30% or 40% phase of the cardiac cycle, which was determined visually from the multiphase, electrocardiography-gated CT. End-systole was visually confirmed as the cardiac phase in which both the

CONTRAST-ENHANCED DYNAMIC CT ACQUISITION

aortic and mitral valve leaflets were closed, and

AND IMAGE ANALYSIS. All subjects considered for

typically corresponded to the cardiac phase when the

trial

contrast-

left ventricular volume was smallest. A virtual device

enhanced dynamic cardiac CT imaging by the site

implantation was performed in 3Mensio in which the

per

dynamic

uncompressed geometry of the Intrepid device

4-dimensional acquisition with retrospective elec-

(Medtronic, Inc., Redwood City, California) was

trocardiography gating was performed with the

positioned in the mitral position and dimensions

following scan parameters: detector collimation of

determined based on the device size fit to the corre-

0.4

thickness

sponding mitral valve anatomy. The neo-LVOT area

of #0.8 mm, and minimum slice overlap of 0.4 mm. A

was computed at 2 device locations: 1) where the

minimum contrast volume of 80 ml was used, with a

device is largest in diameter; and 2) the most ven-

minimum iodine concentration of 320 mg/ml and

tricular portion of the device (Figure 2). The minimum

minimum flow rate of 6 ml/s. Image reconstruction

value of the 2 measured locations was reported as the

was retrospective, starting at peak of the R-wave (0%)

end-systolic minimum neo-LVOT area.

enrollment imaging

to

0.625

were

submitted

recommendations.

mm,

minimum

for A

slice

3

4

Meduri et al.

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019 - 2019:-–-

LVOT Assessment in TMVR

F I G U R E 2 Methods of Neo-LVOT Area Measurement for the Intrepid TMVR Implant

The neo-LVOT area is measured at 2 locations along the device, specifically at (A) the location where device diameter is largest and (B) the most ventricular, outflow end of the device. Abbreviations as in Figure 1.

MULTIPHASE AND EARLY SYSTOLIC LVOT ASSESSMENT.

(TTE) imaging was performed at screening (baseline)

Neo-LVOT areas were computed throughout the sys-

before TMVR implantation to determine suitability

tolic interval, approximately 10% to 40% of the car-

of the procedure. Post-procedure TTE is reported at

diac interval in 10% phase increments at the 2 device

1-month follow-up data for this analysis. Baseline

locations described previously. The minimum neo-

and post-procedure LVOT peak gradients were ob-

LVOT area at each cardiac phase was then averaged

tained by TTE pulsed- and continuous-wave Doppler

(i.e., temporal average) across the systolic phases and

TTE. Peak LVOT gradient was calculated as the

represented as: 1) a multiphase average; and 2) an

maximum peak from pulsed- or continuous-wave

early systolic estimate (i.e., at the 10% phase) (Central

Doppler

for

each

patient.

Clinically

Illustration). The multiphase average was then used to

LVOTO

has

been

defined

previously

determine patient risk for LVOT obstruction. Treat-

LVOT gradient $30 mm Hg (8). Evaluation of echo-

ment eligibility cutoff values for multiphase average

cardiographic data was performed by a core echo-

and early systolic area were determined by comparing

cardiography laboratory (University of California,

screening area estimates with post-procedure peak

San Francisco, San Francisco, California).

LVOT gradient change from baseline. Area cutoff (1.5

cm 2,

end-systole

and

multiphase

average;

1.6 cm 2, early systole) was defined as an occurrence of peak LVOT gradient at 30 days post-procedure in 2 or more patients of $20 mm Hg, which was considered a conservative estimate of LVOTO risk. To validate the improved risk predictability of the multiphase and early systolic LVOT screening, compared with the previous end-systolic methods, post-procedure CT analysis of observed neo-LVOT area in treated patients was performed using 3Mensio.

significant as

peak

STATISTICAL ANALYSIS. Data are reported as mean

 SD with minimum and maximum values unless otherwise noted. The Wilcoxon signed rank test was performed for paired data. The linear correlation between predicted and observed LVOT area is measured by the Pearson correlation coefficient. All statistical analysis was performed using statistical software SAS version 9.4 (SAS Institute, Cary, North Carolina).

RESULTS

ECHOCARDIOGRAPHIC ANALYSIS. Transesophageal echo-

PATIENT SCREENING AND CLINICALLY RELEVANT

cardiographic and transthoracic echocardiographic

LVOTO. From March 6, 2015, to July 21, 2017, a total

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019

Meduri et al.

- 2019:-–-

LVOT Assessment in TMVR

C ENTR AL I LL U STRA T I O N Multiphase and Early Systolic LVOT Assessment

% of Stroke Volume

A

Systole

100 90 80 70 60 Early systole 50 40 30 30 20 10 0 10% 0%

Diastole 95

100

75 End systole

20%

B

30%

40% 50% 60% 70% % of Cardiac Cycle Period

80%

Early Systole

90%

100%

End Systole Systolic Phases

Meduri, C.U. et al. J Am Coll Cardiol Intv. 2019;-(-):-–-.

(A) The majority of stroke volume is ejected in early systole. (B) Cardiac output through the different phases of systole.

of 166 patients were screened for possible study in-

50 patients were approved for inclusion. All patients,

clusion through end-systolic CT methods. Of these

except 1, had no clinically relevant LVOTO, as defined

patients, 32 were screen-failed for reasons unrelated

as a peak LVOT gradient $30 mm Hg at 30-day follow-

to patient anatomy, including left ventricular func-

up (Table 1).

tion and low surgical risk. The remaining 134 patients are the subject of the present analysis. A summary of

POST-IMPLANT

anatomical reasons for screening exclusion is pro-

LVOTO risk based on neo-LVOT area estimates was

LVOT

vided in Figure 3. Concern for LVOT obstruction was

performed using the following 3 methods: 1) end-

the cause for screen failure in 37 (27.6%) of these

systolic minimum; 2) multiphase average; and 3)

patients, whereas annulus size (too large or too small)

early systolic maximum. Multiphase and early sys-

was cause for screen failure in 47 (35.1%) patients.

tolic methods were retrospectively applied to pa-

Of patients within the anatomical range for annulus

tients

size, the baseline LVOTO exclusion rate using end-

methods. Multiphase average and early systole pre-

systolic criteria was 37 of 87 (42.5%). Finally,

dictors clearly delineated differences across patients

previously

GRADIENTS. Prediction

screened

with

of

end-systolic

5

6

Meduri et al.

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019 - 2019:-–-

LVOT Assessment in TMVR

F I G U R E 3 Reasons for Anatomic Exclusion From the Intrepid Global Pilot Study Using

End-Systolic Risk Assessment

overestimate LVOTO risk compared with multiphase average and early systolic estimates, as reflected by significantly greater multiphase and early systolic estimates (Figure 5). Screening predictions were significantly correlated with post-implant neo-LVOT area measurements by CT using both multiphase and early systole. Pearson correlation coefficients for linear correlation were smallest for the end-systole predictor (r ¼ 0.39, p ¼ 0.09) and increased for multiphase average (r ¼ 0.75, p < 0.001) and early systole (r ¼ 0.72, p < 0.001) predictors. Despite significant correlation, area predictions were significantly less than the actual post-implant CT values (Figure 5). To understand potential contributing factors of these differences, additional analyses were performed to segment the population by MR etiology (primary or mixed vs. secondary) determined by the echo core lab. On average, the differences in endsystolic prediction and actual post-implant LVOT area measurements per CT were greater in patients with secondary MR (n ¼ 15; 163.1  26.1 mm 2 predicted vs. 360.1  81.4 mm 2 actual) compared with

A total of 166 patients were screened for possible study participation. Of these, 32 were screen-failed for various reasons unrelated to patient anatomy. The remaining 134 pa-

those with primary or mixed MR (n ¼ 5; 183.3 

tients are the subject of this analysis and represented here. Thirty-seven (27.6%) were

45.8 mm 2 predicted vs. 250.9  67.3 mm 2 actual)

screen-failed due to concern for LVOTO, which corresponds to approximately 42.5% of

(Online Figure 1).

patients deemed anatomically suitable. MV ¼ mitral valve; other abbreviations as in

Of 37 patients originally screen-failed for LVOTO

Figure 1.

risk, 33 patients had dynamic, 10-phase contrast CT data available for analysis; 4 patients did not have dynamic CT available and were excluded from analin LVOT risk; whereas, these differences were

ysis. Using the multiphase assessment, 11 of 33

reduced when an end-systolic predictor was used,

(33.3%) screen-failed patients would now be eligible

making

for enrollment. An additional 7 of 33 (21.2%) patients

differentiation

of

patient

risk

un-

would be eligible for enrollment based on the early

clear (Figure 4). CT SCREENING AND RETROSPECTIVE ASSESSMENT OF

LVOTO

SCREEN

FAILURE

IMPROVEMENT.

systolic assessment. In total, 18 of 33 (54.5%) patients previously screen-failed for LVOTO risk with dynamic CT available would now be eligible for enrollment

Screen failure rates were compared among the 3

using either the multiphase or early systolic approach

LVOTO risk predictor methods in treated patients as

(Figure 5B). The reduction in screen-failure rate is

well as patients previously screen-failed based

based on a cutoff value of 1.5 cm 2 applied for multi-

on

phase average and early systole neo-LVOT area. This

end-systolic

metrics.

End-systolic

predictors

cutoff value was determined from correlation of predicted neo-LVOT area and post-procedure peak

T A B L E 1 Baseline and 30-Day Echocardiographic Data

Measurement

Left ventricular ejection fraction, % Mitral valve mean gradient, mm Hg

Baseline

LVOT gradient (Discussion). 30 Days

44.0  11.1 [26.0–70.0] (20) 36.2  10.0 [21.0–59.0] (20) 3.1  1.9 [0.7–9.5] (16)

4.0  1.4 [2.3–8.0] (18)

Mitral valve peak gradient, mm Hg

8.1  3.7 [2.4–16.4] (16)

9.8  2.8 [6.2–17.3] (18)

LVOT mean gradient, mm Hg

3.6  2.5 [0.7–11.2] (18)

5.7  5.1 [2.1–23.1] (18)

LVOT peak gradient, mm Hg

6.5  4.3 [1.5–18.3] (18)

10.2  8.2 [3.8–37.0] (18)

Values are mean  SD [range] (n). Pre-implant and 30-day left ventricular outflow tract (LVOT) peak gradients were obtained by transthoracic echocardiographic pulse-wave and continuous wave Doppler transthoracic echocardiographic, with the maximum peak from either chosen for each patient.

PROSPECTIVE CASE EXAMPLES OF PATIENTS WHO PREVIOUSLY WOULD HAVE BEEN SCREEN-FAILED.

In June 2017, the first patient originally screen-failed due the end-systolic LVOT assessment was screenaccepted based on the multiphase average criteria and successfully implanted with the Intrepid device. Since then, 5 additional patients have been approved through the multiphase method and 3 through the early systolic method. Echocardiographic data for these patients are summarized in Figure 6. No

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019

Meduri et al.

- 2019:-–-

LVOT Assessment in TMVR

F I G U R E 4 Correlation of Predicted Neo-LVOT Areas With 30-Day Post-Procedure LVOT Gradient

Correlation of predicted neo-LVOT areas with 30-day post-procedure LVOT gradient (n ¼ 20). Neo-LVOT area prediction using (A) end-systolic, (B) multiphase, and (C) early systolic methods and corresponding cutoff value (150 mm2). Thirty-day gradients could not be read in 2 patients. LVOT peak gradients were obtained by TTE pulsed- and continuous-wave Doppler TTE, with the maximum peak from either chosen for each patient. Abbreviations as in Figure 1.

clinically significant LVOT gradient (defined as peak

compared with those assumed during screening

LVOT gradient $30 mm Hg) was observed in any of

(Online Table 1, Figure 8).

these patients at 30 days. Example CT measurements of LVOT area in 2 treated patients are presented in Figure 7. One patient had a gradient of 37 mm Hg

DISCUSSION

(Figures 5A and 7A), and the early systolic area,

The primary findings of this analysis are that multi-

multiphase average, and end-systolic areas were

phase CT analysis, and more specifically early systolic

similar in magnitude. Figure 7B shows CT scans from a

CT assessment, better predict clinically relevant

patient treated based on the multiphase approach

LVOTO. These advances have the potential to reduce

(this patient was previously screen-failed based on

the screen-failure rate in TMVR clinical trials and

the end-systolic criteria).

improve the accessibility of this therapy to additional

POST-IMPLANT DEVICE GEOMETRY. The post-implant

geometry of the Intrepid device is significantly compressed in the anterior-posterior dimension and ovalized around its perimeter. Although the smallest

patient populations previously screen-failed from TMVR trials. Further, these findings are likely not device-specific and may therefore be generalized to additional TMVR devices.

neo-LVOT area typically occurs at the ventricular end

LVOTO PREVENTION AND TREATMENT. A number of

of the fixation ring, a discrepancy was noted between

approaches have been made to prevent and treat

the post-implant frame dimensions at this location

LVOTO related to TMVR and mitral valve-in-valve or

7

8

Meduri et al.

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019 - 2019:-–-

LVOT Assessment in TMVR

F I G U R E 5 Retrospective Assessment of LVOT Area Predictions

(A) Retrospective assessment demonstrating LVOT area predictions in treated patients with pre-procedural computed tomography data available (n ¼ 20). Two patients with elevated gradients (denoted with arrows) demonstrated a similar LVOT area predictions regardless of the assessment method used. (B) Retrospective assessment of LVOTO risk predictions in patients screen-failed based on end-systolic estimates (n ¼ 33). (C) Retrospective assessment of treated patients (n ¼ 20) with paired baseline predictions and post-procedure measurements of LVOT area. End-systolic predictors (mean  SD) significantly overestimate LVOTO risk compared with multiphase average and early systolic estimates. Screen predictions were significantly less than actual post-implant LVOT area measurements per computed tomography (p < 0.001). In panel C, the box is centered at the median, with upper and lower bounds of the box representing the interquartile range. The diamond indicates the mean of the value. Upper and lower ends of the whiskers represent minimum and maximum values. Abbreviations as in Figure 1.

valve-in-ring or valve-in-mitral annular calcification F I G U R E 6 Prospective Analysis of Patients Approved Through Multiphase and Early

procedures. For those with a predicted high risk of

Phase Methods

neo-LVOT based on late-systolic CT measurement, several techniques have been used to facilitate TMVR. One method has been the use of preemptive ethanol septal ablation in appropriate anatomical patients. Though this has been effective in some patients, it is not without risk, including further reduction of LV function, permanent pacemaker, ventricular septal defects, and delaying treatment an additional 4 to 6 weeks. Another approach has been the novel LAMPOON procedure, which causes an intentional laceration of the anterior mitral leaflet, thereby reducing the risk of LVOTO by preventing the anterior mitral leaflet from covering the open cells of the

Prospective analysis of 7 patients approved through multiphase and early phase methods

transcatheter prosthesis. Despite being a complex

with available 30-day post-procedure TTE data. Peak LVOT gradients are based on TTE.

procedure, it has demonstrated an additional oppor-

Pre-implant and 30-day LVOT peak gradients were obtained by TTE pulsed- and

tunity to expand therapy to patients without further

continuous-wave Doppler TTE, with the maximum peak from either chosen for each patient. Baseline data were not available in 2 patients and therefore assumed to be

options. Unfortunately, this technique is only appli-

0 mm Hg. For all patients, there were no observations of clinically relevant LVOTO

cable to TMVR devices with an open-cell design (i.e.,

defined as a peak gradient of $30 mm Hg (8). Abbreviations as in Figure 1.

SAPIEN valve [Edwards Lifesciences, Irvine, California]). In most new TMVR devices, lacerating the

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019

Meduri et al.

- 2019:-–-

LVOT Assessment in TMVR

anterior mitral leaflet would not reduce the risk of LVOTO because these valves are closed-cell design (9).

F I G U R E 7 Computed Tomography Images of Treated Patients Post-Implant

Other methods focus on addressing LVOTO after it has occurred, including a valve-in-ring case series of 8 patients, 2 of whom experienced LVOTO. In 1 patient, LVOTO was eliminated with repositioning of the transcatheter valve. In the second case, repositioning did not improve the obstruction, so the valve was recaptured and the procedure was abandoned. The patient subsequently underwent alcohol septal ablation (10). Successful bail-out septal alcohol ablation has also been described in the setting of TMVR in a patient with nearly circumferential calcification of the mitral annulus who experienced LVOTO (11). Additionally, there has been the use of a novel perfusion balloon to maintain patency of the LVOT during TMVR valve-in-valve, to prevent LVOTO (12). POST-IMPLANT

DEVICE

GEOMETRY. The

primary

discrepancy between predicted LVOTO risk and actual post-implant LVOT area measurements via CT assessment potentially involves post-implant device geometry. Indeed, the post-implant geometry of the Intrepid device is significantly compressed in the anteroposterior dimension and ovalized around its

Computed tomography and actual neo-LVOT area are shown. In panel A, the actual

perimeter. The smallest neo-LVOT area typically oc-

neo-LVOT average area was 160 mm2, with a 30-day peak gradient of 37 mm Hg. In

curs at the ventricular end of the fixation ring. How-

panel B, the actual neo-LVOT average area was 226 mm2 and the 30-day peak gradient

ever, it was noted that the AP diameter at this

was 8 mm Hg.

location is markedly different than the dimensions assumed during screening. These findings demonstrated that device compression occurred, as oversizing is used for implant size selection. This observed device compression is important, as oversizing is used for TMVR implant size selection; however, as device technology and sizing continues to improve, it is likely that this challenge can be overcome to more precisely predict LVOTO. Prior research has established cutoff values of 1.7 to 1.9 cm2 with

risk and screen-failed patients otherwise eligible from clinical trials for TMVR. Therefore, we retrospectively performed CT measurements in patients that were both screen-accepted (i.e., treated) and screen-failed, noting that for all of these patients only the endsystolic criteria were used for the treatment decision. We used a novel multiphase approach and applied it to CT measurements from these patients,

other TMVR devices (13,14), and those neo-LVOT predictions were well correlated with post-implant

F I G U R E 8 Post-Implant Frame Geometry

measurements. In contrast, for the Intrepid device, we have shown that our predictions underestimate the actual neo-LVOT area observed post-implant, as demonstrated by predicted neo-LVOT areas that were less than those observed post-implant. This necessitates improved CT methods that can accurately predict neo-LVOT area. IMPROVEMENTS IN LVOT SCREENING WITH MULTIPHASE ASSESSMENT. In the present analysis cohort from the

Intrepid Global Pilot Study, 42.5% of patients deemed anatomically suitable were screen-failed for concerns

These post-procedure computed tomography measurements of frame geometry demonstrate the discrepancy between predicted LVOTO risk and actual post-implant

of LVOTO. Current methods of assessing risk for

LVOT area measurements per computed tomography (Online Table 1). AP ¼ anterior-

LVOTO are based on CT measurements at late systole.

posterior; CC ¼ commissure-commissure; other abbreviations as in Figure 1.

As such, these methods overestimated the LVOTO

9

10

Meduri et al.

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019 - 2019:-–-

LVOT Assessment in TMVR

based on our hypothesis that this approach would

therapy is limited due to a high anatomical screen fail

better predict the risk of LVOTO compared with static

rate. A significant portion of these screen fails come

end-systolic estimates. We proposed that a multi-

from LVOTO, which will remain an issue based on

phase CT analysis, including early systolic phase,

current screening assessment of late-systolic phase.

would provide a more accurate assessment of risk and

The data reported here show that the use of a physi-

thus lead to a higher proportion of patients eligible

ologic assessment of LVOTO, based on early systolic

for TMVR. The area cutoff for the multiphase and

neo-LVOT, allows for a better estimation of post-

early systolic criteria was based on correlations of

implant LVOT area. Through its use, we have found

predicted neo-LVOT area and observed peak LVOT

that 54.5% of patients previously screen-failed for

gradient at 30 days post-procedure. We used a con-

LVOTO would now be eligible for treatment. Further

servative estimate of peak LVOT gradient $20 mm Hg

prospective validation, particularly in larger cohorts

to derive cutoff values of 1.5 cm2 for multiphase

and with multiple devices, will be needed to validate

average and 1.6 cm 2 for early systole assessments.

this method. However, we believe that there is po-

The end-systolic cutoff was set to the multiphase

tential for this screening method to become the new

average cutoff of 1.5 cm 2. However, as prediction ac-

standard of care for predicting risk of LVOTO in pa-

curacy improves, cutoff values will likely evolve from

tients being evaluated for TMVR. Finally, we believe

our current estimate to those previously published

that as more device sizes become available, the

(i.e., 1.5 to 1.9 cm 2) (13,14).

combined impact of early systolic prediction of

Our findings demonstrate that one-third of previ-

LVOTO and increased annular size coverage will

ously screen-failed patients would now be eligible for

result in substantially more patients screened being

enrollment based on multiphase alone. However, as

anatomically suitable for TMVR, allowing for a much

the greatest flow occurs in early systole (15), we

broader treatment population.

evaluated use of a single measurement during early

ACKNOWLEDGEMENT Jessica

systole, hypothesizing that this would be more

CMPP, an employee of Medtronic, provided medical

physiological and would correlate well or better than

writing support under the direction of the lead

the multiphase approach. Using the early systolic

authors.

Dries-Devlin,

PhD,

method, an additional 21% of previously screen-failed patients would be eligible for enrollment. In total, 18

ADDRESS

of 33 (54.5%) of patients previously screen-failed for

Bapat, CUMC/Milstein Hospital Building, 177 Fort

LVOTO risk with dynamic CT available would now be

Washington Avenue, Floor: MHB 7-435, New York,

eligible for enrollment using either the multiphase or

New York 10032. E-mail: [email protected].

early systolic criteria. The proportion of eligible

edu. Twitter: @chrismeduri.

FOR

CORRESPONDENCE:

Dr. Vinayak

TMVR patients may increase by more than one-half, as evidenced by our series. Importantly, to validate our hypothesis that early

PERSPECTIVES

systolic phase prediction of neo-LVOT would safely predict risk of LVOTO, we prospectively treated 9

WHAT IS KNOWN? The complex anatomy of the

patients that previously would have failed screening

mitral valve and concerns for the risk of LVOTO have

for LVOTO based on end-systolic assessment. Each of

resulted in high screen-failure rates in recent TMVR

these patients was eligible based on early systolic

trials.

measurement. Among these patients, there were no signs of clinically relevant LVOTO on 30-day echo.

WHAT IS NEW? Use of a physiologic assessment of the neo-LVOT more accurately predicts clinically

STUDY LIMITATIONS. It should be noted that our

experience here reflects a small sample size. Further, the TMVR field in general, and specifically that of

relevant obstruction of the LVOT. This resulted in more than 50% of previously screen-failed patients from a pilot study of TMVR to be eligible for treatment.

predicting LVOTO risk, is new and there is still much to be learned from ongoing trials and case experience. As the field advances, we predict further refinement around the ideal cutoff value that should be used.

CONCLUSIONS TMVR has emerged as an exciting potential therapy for patients with MR, but the generalizability of the

WHAT IS NEXT? Additional prospective validation of this assessment is needed across valves. As more TMVR device sizes become available, the combined impact of this assessment method and increased annular size coverage should greatly increase the TMVR treatment population.

JACC: CARDIOVASCULAR INTERVENTIONS VOL.

-, NO. -, 2019

- 2019:-–-

Meduri et al. LVOT Assessment in TMVR

REFERENCES 1. Bapat V, Rajagopal V, Meduri C, et al. Early experience with new transcatheter mitral valve replacement. J Am Coll Cardiol 2018;71:12–21. 2. Muller DWM, Farivar RS, Jansz P, et al. Transcatheter mitral valve replacement for patients with symptomatic mitral regurgitation: a global feasibility trial. J Am Coll Cardiol 2017;69:381–91. 3. Guerrero M, Urena M, Himbert D, et al. 1-year outcomes of transcatheter mitral valve replacement in patients with severe mitral annular calcification. J Am Coll Cardiol 2018;71:1841–53.

transcatheter mitral valve replacement. J Am Heart Assoc 2017;6:e007353.

mitral valve-in-valve replacement. Catheter Cardiovasc Interv 2018;92:601–6.

8. Elliott PM, Gimeno JR, Tome MT, et al. Left ventricular outflow tract obstruction and sudden death risk in patients with hypertrophic cardiomyopathy. Eur Heart J 2006;27:1933–41.

13. Wang DD, Eng MH, Greenbaum AB, et al. Validating a prediction modeling tool for left ventricular outflow tract (LVOT) obstruction after transcatheter mitral valve replacement (TMVR). Catheter Cardiovasc Interv 2018;92: 379–87.

9. Babaliaros VC, Greenbaum AB, Khan JM, et al. Intentional percutaneous laceration of the anterior mitral leaflet to prevent outflow obstruction during transcatheter mitral valve replacement: first-

14. Yoon SH, Bleiziffer S, Latib A, et al. Predictors of left ventricular outflow tract obstruction after

in-human experience. J Am Coll Cardiol Intv 2017;10:798–809.

transcatheter mitral valve replacement. J Am Coll Cardiol Intv 2019;12:182–93.

10. Latib A, Ruparelia N, Bijuklic K, et al. First-inman transcatheter mitral valve-in-ring implantation with a repositionable and retrievable aortic valve prosthesis. EuroIntervention 2016;11: 1148–52.

15. Uretsky S, Chaudhry FA, Gillam L, et al. A novel technique to quantify the instantaneous mitral regurgitant rate. J Cardiovasc Magn Reson 2013; 15:74.

6. Wang DD, Eng M, Greenbaum A, et al. Pre-

11. Deharo P, Urena M, Himbert D, et al. Bail-out alcohol septal ablation for left ventricular outflow

dicting LVOT obstruction after TMVR. J Am Coll Cardiol Img 2016;9:1349–52.

tract obstruction after transcatheter mitral valve replacement. J Am Coll Cardiol Intv 2016;9:e73–6.

KEY WORDS left ventricular outflow tract obstruction, LVOT, mitral valve, neo-LVOT, TMVR

7. Murphy DJ, Ge Y, Don CW, et al. Use of cardiac computerized tomography to predict neo-left ventricular outflow tract obstruction before

12. Herrmann HC, Szeto WY, Litt H, Vernick W. Novel use of perfusion balloon inflation to avoid outflow tract obstruction during transcatheter

4. Bapat V. Mitral valve-in-ring: the good, the bad, and the ugly. EuroIntervention 2016;11: 1092–4. 5. Blanke P, Naoum C, Dvir D, et al. Predicting LVOT obstruction in transcatheter mitral valve implantation: concept of the neo-LVOT. J Am Coll Cardiol Img 2017;10:482–5.

A PP END IX For a supplemental table and figure, please see the online version of this paper.

11