Assessment of left atrial function by magnetic resonance
imaging myocardial feature tracking
-1-
Abstract
Purpose
Left atrium (LA) volumes and function are predictors of cardiovascular events. Since
LA function cannot be assessed from cardiovascular magnetic resonance imaging
(MRI) using the well-established left ventricular tagging techniques, we hypothesized
that adequate feature tracking (FT) applied to conventional cine MRI data could
characterize LA function accurately.
Materials and Methods
We studied 10 young (28±7 years) and 10 elderly (64±6 years) healthy subjects as
well as 20 patients with moderate to severe aortic valve stenosis (AVS; 73±15 years,
effective aortic valve area: 0.67±0.36cm²). MRI cine 2, 3 and 4 chamber views were
analyzed using a newly proposed FT method based on spatial correlation and
endocardial detection resulting in: regional and global longitudinal strain and strain
rate, radial motion fraction and relative velocity for the three LA motion phases
including reservoir, conduit, and LA contraction.
Results
FT reliability was indicated by a good overlap between tracking results and manual
LA endocardial borders, the low error for comparison against theoretical strains
introduced in a synthetic phantom and the good inter-observer reproducibility
(coefficient of variation < 15%). While all LA functional parameters were
significantly impaired in AVS patients (p<0.04), subclinical age-related variations
induced a decreasing trend on all LA parameters but were significant only for radial
conduit
function
parameters
(p<0.03).
-2-
Finally,
LA
functional
parameters
characterized LA alteration in AVS with higher sensitivity than conventional LA
volumetric parameters.
Conclusions
Left atrial FT is feasible on MRI cine images and its addition to conventional analysis
tools might enhance the diagnosis value of MRI in many heart diseases.
Keywords
Left atrial function; feature tracking; magnetic resonance imaging; myocardial strain
and strain rate.
INTRODUCTION
LA function has been associated with atrial fibrillation (1) as well as cardiovascular
morbidity and mortality (2). Left atrium (LA) function can be quantified by both
volumetric and contractile function parameters and its alteration is strongly associated
with left ventricular (LV) performance because of the functional interplay between
both chambers in healthy aging as well as in various pathological conditions.
While LV diastolic and systolic function, whether assessed using global or regional
indices, are widely described in the literature, functional characterization of the LA is
mostly performed through global morphological indices such as diameters, areas and
volumes. Such indices might be insufficient in describing the complex LA function,
which is divided into reservoir, conduit and contraction phase, and which is
influenced by LV contraction, relaxation and filling pressures, pulmonary vein
orientation as well as LA electrical activity. Accordingly, LA function indices such as
strain and strain rate have been proposed using non-invasive imaging modalities such
as echocardiography speckle tracking (3–5) and Color tissue Doppler. Although
-3-
speckle tracking is presently the only available reference for LA strain estimation,
ultrasound beam direction as well as heart motion relative to the probe may influence
measurements (6) and inter-vendor variability, essentially described in the setting of
LV function, need to be further investigated (7).
Cardiac magnetic resonance imaging (MRI) has been reported as the gold standard for
LV mass and volumes as well as systolic function evaluation. However, only few
studies have been dedicated to LA geometry and volumes (8), highlighting the higher
accuracy of MRI measurements (9) . This latter study advised LA characterization in
all recommended routine MRI exams. Furthermore, despite the improvements in
terms of spatial resolution required to capture the thin LA wall, MRI is the only noninvasive modality able to characterize myocardial tissue in both LV and LA (10, 11).
Tissue magnetization strategies such as tagging or strain encoding have been
developed for myocardial strain evaluation from MRI data and have been widely used
for LV regional function characterization (12). However such techniques cannot be
used for LA strain evaluation because of the small LA myocardial wall thickness
which does not allow to obtain sufficient myocardial signal. In this context, our
hypothesis is that the addition of innovative image processing tools for LA strain
estimation from routine cine MRI data would enhance the clinical usefulness of MRI
for LA characterization and allow retrospective use on large existing datasets.
Ultimately, their combination with recent advances in LA electrophysiological
mapping might help in planning resynchronization therapy.
Accordingly, the objectives of our study were to develop a feature tracking algorithm
and to assess its ability to evaluate LA function from conventional cine MRI imaging,
-4-
which is systematically acquired in standard clinical MRI exams. Such development
should account for LA complex geometry and wall motion.
The reliability of the tracking technique was studied in terms of: 1) LA borders
detection, 2) inter-operator variability and 3) its ability to characterize LA function
alterations in healthy aging and in moderate to severe aortic valve stenosis (AVS).
MATERIALS AND METHODS
Synthetic Phantom Data
To test the algorithm developed for feature tracking on cine MRI data in controlled
conditions, we created a dynamic dataset of synthetic images simulating LA
myocardial motion as in the study proposed on the LV by Hor et al. (13). Briefly, LA
shape is simulated by an ellipse fitted on the maximum contraction time of a real LA
image truncated between extremities of the mitral valve. An additional truncated
ellipse is added at the opposite side of LA to simulate the LV and to test mitral valve
annulus tracking. Ellipse motion is generated using physiological longitudinal strain
values (39.5% max strain) and the respective time variations extracted from a
previous study (14). The ellipse has a 2 pixels grey border and was filled with white
to mimic blood pool signal in MRI images. This phantom comprised 60 time phases
and its pixel size was 0.73mm. Moreover, Rician noise has been added to simulate
realistic MRI images. Finally, effect of spatial and temporal resolution on functional
parameters was studied while: 1) varying pixel size between 0.73 and 3.14mm and 2)
decreasing the temporal resolution by a factor of 2, 3, 4, 5 and 6 (by averaging
successive time frames to mimic the blurring effect of low temporal resolution). The
resulted functional parameters of radial and longitudinal strains were compared
against the reference used to generate the initial phantom.
-5-
Study Population and Data Acquisition
40 subjects including 20 healthy volunteers free from overt cardiovascular disease (10
young controls: 27.7±6.8 years and 10 elderly controls: 63.7±5.7years) and 20
patients with AVS (73±14.5years) were studied. In this latter group, severity of AVS
was characterized by echocardiography (mean pressure gradient: 45.8±13.9mmHg
and effective orifice area: 0.67±0.36cm²). Subjects with atrial fibrillation at the time
of MRI were not included in the study. MRI exam, including steady state free
precession (SSFP) acquisition on a 1.5T magnet in three long axis views and short
axis views, during breath-holding, to cover the whole left heart, was performed using
the following scan parameters: acquisition matrix = 260×192, TR = 3.7 ms, TE = 1.5
ms, flip angle = 50°, pixel size = 0.74 mm×0.74 mm, slice thickness = 8 mm, views
per segment=12, temporal resolution ranged between 20 and 30ms. The study
protocol was approved by the institutional review board and informed consent was
obtained from all participants.
LA Volumes and Ejection Fraction
End-systolic and end-diastolic LA endocardial borders were manually traced on both
2 and 4 chamber views using the Qmass software (v7.6, Leiden, Netherlands) to
derive LA volumes and ejection fraction (EF). The same software was also used for
the estimation of LV mass and volumes as well as LV EF after manual tracing of LV
endocardial and epicardial borders on a stack of short axis slices at end-systole and
end-diastole.
Feature Tracking Algorithm
First, 2, 3 and 4 chamber SSFP images were filtered using a non-local mean filter
(15) to enhance the contrast between LA blood pool and myocardium. The feature
-6-
tracking algorithm developed using Matlab (Matworks, 8.0 R2012b, USA) is based on
spatial correlation, as commonly done for speckle tracking in echocardiography and
constrained with endocardial contours detection which can be easily performed on the
highly contrasted MRI images. It comprised the following two main steps (Figure 1):
1) user manual initialization of the LA endocardial border on a single phase
corresponding to maximal LA dilatation by positioning N markers which were
interpolated into N*3 markers and corrected manually if required to optimize the
superimposition of the resulting contour on the sharp transition in contrast between
LA blood and myocardium. Indeed, conversely to speckle tracking where an entire
myocardial feature was considered, features on MRI data consists in an interface
between blood and myocardium, in other words the LA endocardial wall. Markers
were carefully positioned to exclude LA appendage and junctions with pulmonary
veins.
2) Tracking of the initial markers was performed on neighboring images towards the
beginning and the end of the cardiac cycle. Two-dimensional correlation was used to
track similar anatomic features on a delimited neighborhood of each marker. Indeed,
rectangular (6x6 pixels) region of interest (ROI) was defined around the current
marker, perpendicular to the LA endocardial border to take into account physiological
knowledge about LA wall motion. Spatial correlation between the defined ROI and
the rectangular regions centered on each of its pixels and defined on the neighboring
image were calculated. This resulted in a rectangular correlation map in which only
pixels comprised in a conic shape defined from the LA center of mass were
considered to avoid erroneous tracking. The highest correlation measure from this
map corresponded to the new position of the marker on the neighboring image. In
addition to correlation, another weighting which takes into account the distance of the
-7-
marker to: a) the approximated endocardial edge stored in a binary image calculated
from maximal contrast variations (edge method), b) the approximated endocardial
border detected on the native image using active contours, based on (16).
The aforementioned processing steps were combined into three different tracking
options: tracking on gradient images calculated from native images (Track 1),
tracking on filtered native images with the final weighting performed using either
edge images (Track 2) or both active contours and edge images (Track 3). For all
tracking options, radial position of each marker relative to the LA center of mass was
temporally filtered by a 3 median filter to smooth the LA endocardial border and
correct for erroneous tracking. Finally, the user defined 3 segments by choosing their
first and last markers on the initial trace.
LA Functional Parameters
Contours resulting from the tracking were used to calculate regional longitudinal
strain and regional radial motion fraction indices on standardized LA segments
(anterior/inferior, antero-Septal/infero-lateral, infero-septal/antero-lateral, see Figure
2.A and B). LA regional function was characterized by longitudinal and radial indices
as suggested in (4) and adapted from the common denomination for the LV
measurements: longitudinal strain and radial motion fraction (Figure 2.A, Additional
file 1 ).
Longitudinal strain (Figure 2.A and C) is defined as the temporal variation of the
length of the studied segment and is calculated as: 𝑆𝑙(𝑡) =
𝐿0 −𝐿𝑡
𝐿0
, with 𝐿0 the initial
length of the segment and 𝐿𝑡 its length at time t.
Radial motion fraction (Figure 2.A and E) corresponded to the radial relative
displacement of the considered segment: 𝑀𝑟(𝑡) =
-8-
𝑅0 −𝑅𝑡
𝑅0
, with 𝑅0 the initial radius, 𝑅𝑡
the radius at time t. For each endocardial marker, radius was calculated relative to the
LA centre of mass and then averaged for all the markers of the studied segment. Of
note, LA centre of mass was calculated for each phase of the cardiac cycle using the
tracked endocardial markers.
Time derivatives of longitudinal and radial indices were calculated resulting in
longitudinal strain rate and radial relative velocity curves.
For longitudinal measurements, the following functional indices were derived from
strain and strain rate curves (Figure 2.C-D): reservoir strain (SlR: first strain peak), left
atrial contraction strain (SlA: second strain peak) and the conduit strain (SlE:
difference between R and A peaks), systolic strain rate (SRlS’: positive strain rate
peak), early LV filling strain rate (SRlE’: first negative strain rate peak) and LA
contraction strain rate (SRlA’: second negative strain rate peak).
Similarly, the following indices were extracted from radial motion fraction and
relative velocity curves (Figure 2.E-F): reservoir motion fraction (MrR: first motion
fraction peak), left atrial contraction motion fraction (MrA: second motion fraction
peak) and the conduit motion fraction (MrE: difference between R and A peaks),
systolic relative velocity (VrS’: positive relative velocity peak), early LV filling
relative velocity (VrE’: first negative relative velocity peak) and LA contraction
relative velocity (VrA’: second negative relative velocity peak). Finally, global indices
were calculated similarly when considering the three LA segments of each view.
Feature Tracking Algorithm Evaluation
On phantom data, quality of tracking and performance of the three aforementioned
tracking options were assessed by the comparison of the estimated functional indices
against those initially introduced in the model.
-9-
For MRI data, evaluation in terms of contours overlap between tracking and manual
Qmass contours was performed, using the standard measurements commonly used for
segmentation algorithms evaluation such as Dice, Jaccard, mean and maximal
distances (17). Such analysis was performed on 10 subjects (3 views, all phases of the
cardiac cycle).
These two steps enabled the selection of the best tracking option. Finally, since
initialization of LA markers is the only manual intervention, inter operator variability
regarding such initialization was investigated on the 2,3 and 4 chamber views of 10
subjects in terms of contour overlap and differences in LA functional indices.
Statistical Analysis
All continuous variables are given as mean ± standard deviation. Bland and Altman
analysis was performed for comparisons between repeated measurements and mean
bias and limits of agreement (mean±1.96*standard deviation) were provided. Boxplot
graphs were used to illustrate LA functional parameters variation between young and
elderly controls as well as AVS patients. Differences between young controls and
elderly control as well as elderly controls and AVS patients were tested using the nonparametric Wilcoxon test and a p-value <0.05 indicated statistical significance.
Associations between continuous variables were studied using linear regression and
Pearson correlation coefficients were provided. For inter-observer variability, a
coefficient of variation was calculated as the standard deviation of the differences
between two series of measurements divided by the mean of the measurements. The
ability of the newly proposed LA functional parameters as well as conventional
volumetric parameters to detect LA functional alteration underlying AVS, in terms of
sensitivity, specificity, negative and positive predictive values (NPV and PPV) as well
as accuracy, was evaluated using a receiver operating characteristic (ROC) analysis
- 10 -
and optimal thresholds were defined. Statistical analysis was performed using the R
software.
RESULTS
Feature Tracking Algorithm Evaluation
Table 1 summarizes absolute differences and percentage of difference between results
of the three tracking options and reference longitudinal strain as well as radial motion
fraction parameters, initially applied to generate the synthetic phantom. Percentage of
differences was <12% for all algorithms and all parameters and results obtained using
the algorithm Track 3 were slightly better (<8%).
Figure 3 summarizes percentages of differences between reference longitudinal strain
and radial motion fraction and strains calculated on dataset with lower spatial (Figure
3.A) and temporal resolution (Figure 3.B). Such differences increased with pixel size
and crossed 20% over a resolution of 1.8 mm for both longitudinal strain and radial
motion fraction (Figure 3.A). Regarding the effect of temporal resolution, differences
with reference were under 20% for a reduction in temporal resolution with a factor of
3 for both longitudinal strain and radial motion fraction (Figure 3.B).
MRI Data: LA Contours Comparison and Inter-operator Variability
LA contours obtained by feature tracking algorithms were compared against manually
traced contours in 10 subjects (Table 2), in terms of Dice, Jaccard as well as mean and
maximal distances. While similar results were obtained in terms of Dice and Jaccard
for the three tracking options, slightly lower mean and maximal distances were found
for the algorithm using edge and active contours weighting (Track 3).
- 11 -
Regarding results on phantom and LA contours overlap with the manual reference on MRI
data, quantitative LA parameters obtained when using Track 3 will be further studied in the
remaining parts of the paper.
Inter-Observer variability analysis resulted in coefficients of variation of 4.3%, 7%
and 10.2% for the three phases for longitudinal strains, of 15.1%, 12.2% and 9.8% for
longitudinal strain rate parameters, of 2.9%, 8.3% and 9% for the three phases of
radial motion fraction and of 7.4%, 8.6% and 10.2% for radial relative velocities.
Patients Characteristics
Table 3 summarizes the young and elderly controls as well as AVS patients
characteristics including basic characteristics, LV and LA Qmass derived indices.
While LV parameters, except LVEF, were significantly different between young and
elderly controls, LA parameters were not significantly different between the two
groups except LAEF. Regarding AVS patients, they had preserved LVEF and
significantly different LA parameters (p<0.05) when compared to elderly controls.
LA Functional Parameters
Table 4 reports global LA functional parameters obtained for young and elderly
controls as well as AVS patients. Longitudinal strain, strain rate, radial motion
fraction and relative velocity were reduced in elderly controls as compared to young
controls. Such reduction reached statistical significance only for radial motion
fraction and relative velocity corresponding to LA conduit phase (p<0.03).
Regarding comparisons between AVS patients and elderly controls, all LA functional
parameters were significantly lower in AVS patients (p<0.04).
- 12 -
Anatomical Variation of LA Functional Parameters
Distribution of regional LA functional parameters in controls is illustrated in Figure 4.
Analysis of regional variations in terms of longitudinal strains indicated higher
reservoir and conduit phases values for lateral and inferior segments than septal and
anterior segments for each view (p<0.03). This anatomical trend was not significant
for the radial measurements.
Association between LA Volumetric and Functional Parameters
Associations between LA functional parameters and LAEF or indexed LA enddiastolic volumes are presented in Figure 5 for the three studied groups. LA
longitudinal strains resulted in higher correlation with LAEF than with LA indexed
end-diastolic volumes. The highest correlations were found for associations between
reservoir phase longitudinal strain and LAEF. Similar results were obtained when
considering LA radial motion fractions (for associations with LAEF: r=0.82 for MrR,
r=0.74 for MrA, p<0.001; for association with LA indexed end-diastolic volumes:
r=0.54 for MrR, r=0.53 for MrA, p<0.001).
Detection of LA Functional Alterations
Table 5 reports results of receiver operating characteristic (ROC) analysis performed,
in a patient basis, to determine the ability of LA volumetric and functional parameters
to characterize LA functional alterations present in the AVS patients, as compared to
controls. Functional parameters characterized LA alterations with overall higher
values of accuracy and area under the ROC curve (AUC), as compared to volumetric
parameters including LAEF. As highlighted above, LA strain and radial motion
fraction corresponding to the reservoir and conduit phases characterized LA
alterations in AVS with higher performances than those corresponding to the LA
- 13 -
contraction phase. Furthermore, strain rate and relative velocity parameters
corresponding to reservoir and conduit phases (S’ and E’) were those providing the
highest sensitivity, specificity, accuracy and AUC. ROC curves corresponding to
conventional volumetric parameters and to a selection of highly accurate LA
functional parameters (global strains and motion fraction) are illustrated in Figure 6.
DISCUSSION
In this study, a new LA feature tracking algorithm based on spatial correlation
constrained by both edge detection and active contours was described and
successfully applied on cine SSFP MRI data for the evaluation of both regional and
global longitudinal and radial LA functional parameters. Its reliability was indicated
by the good inter-observer reproducibility, the good overlap between tracking and
manual LA endocardial borders on all the phases of the cardiac cycle, and the low
error for comparison against theoretical strains introduced in a synthetic phantom.
Moreover, LA functional parameters obtained in young controls, elderly controls and
AVS patients indicated the ability to our method to characterize LA alterations related
to age and to the presence of AVS. Of note, LA functional parameters were able to
detect LA alterations in AVS with higher accuracy than conventional LA volumetric
parameters. Finally, consistency of our indices was indicated by their significant
association with LAEF.
LA functional and volumetric evaluation has been associated with cardiovascular
morbidity and mortality (2). While most of the studies reported in the literature
regarding LA function and strain were performed using echocardiography, only two
studies were performed using MRI (18, 19). Indeed, the reference standard for
myocardial strain in MRI is based on analysis of tagging images, which is not as
- 14 -
practical for LA evaluation as for LV evaluation because of the thin LA wall. The
DENSE sequences, which have been recently applied on the aortic wall or right
ventricle (20, 21), might overcome this drawback but are hampered by such thin
walls. The aforementioned two LA strain MRI studies were performed using feature
tracking technique, previously described for LV strain evaluation (13). Similar to our
method, this technique has the original feature of being applied on LA cine SSFP data
which are routinely acquired during MRI exam. In our study, LA endocardial
contours tracking was compared against a reference manual tracing using quantitative
measurements. Moreover, additional radial parameters such as radial motion fraction
and radial relative velocities were proposed and their ability to characterize aging and
AVS was shown. On the technical point of view, while our method used spatial correlation as
similarity measure, constraints based on active contours were used rather than optical flow.
This choice can be justified by the fact that optical flow has been reported to increase
calculation time and assumes that brightness gradients are presents in the processed images,
which might induce erroneous tracking because of flow-related heterogeneities within the LA
cavity at the edge of LA borders (22). Finally, coefficients of variation obtained when using
our method are in the same range than those previously reported (18). In addition, regional
variations of LA functional parameters estimated by the proposed feature tracking technique
were in agreement with physiological knowledge and with results previously described in
echocardiography mentioning higher longitudinal strain values for lateral and inferior wall as
compared to the remaining segments (3). Finally, both longitudinal and radial functional
parameters were significantly associated with alteration in LAEF and LA end-diastolic
volume (Figure 5). Despite these significant associations, correlations with end-
diastolic volumes were lower. This finding might be due to the fact that the alteration
in LA functional parameters occurs prior to LA dilation in AVS patients.
- 15 -
Although an increasing trend in LA indexed volumes with age was found between our
two age groups, it did not reach statistical significance. These findings are consistent
with those previously found in echocardiographic and MRI studies (3, 8) and volume
values were in the same range than those provided by Maceira et al. (8). In line with
findings reported by Sun et al. (3), we had a significant decrease in LAEF in the
elderly as compared with young controls. Regarding LA functional parameters, a
decreasing trend was observed between young and elderly controls for all parameters
and such decrease was statistically significant only for radial motion and velocity
parameters corresponding to the conduit phase. This might be explained by an
alteration of LA compliance and the known alteration in LV relaxation with aging
which induces a subclinical increase in LV filling pressures. Indeed, the LA conduit
phase, which is a passive phase in terms of myocardial electrical activation as
opposed to active LA dilation and contraction, could be interpreted as a surrogate of
LA compliance.
Regarding AVS patients, they had significantly dilated LA and reduced LAEF as
compared with elderly controls, as previously reported (23, 24). This LA dilation
resulted in substantial and significant decrease in all LA functional parameters
indicating the alteration of LA reservoir, conduit and contraction phases. Such
alterations were expected since moderate to severe AVS induces LV hypertrophy and
subsequent diastolic dysfunction which was shown to be associated with LA
alterations (5, 25). Importantly, LA functional parameters, especially longitudinal
strain rate and radial motion fraction and velocity corresponding to the conduit phase,
were able to detect LA alterations in AVS with the highest accuracy as compared to
the remaining LA functional parameters as well as conventional indices including
LAEF. Such findings were not surprising since substantial alteration in myocardial
- 16 -
tissue relaxation, which might expand to LA myocardium, was expected in moderate
to severe AVS. Indeed, diastolic function studies performed either using
echocardiography or MRI highlighted a marked decrease in peak longitudinal
myocardial annulus velocity (E’) in AVS patients (26, 27). In addition to the LA
compliance property, such results highlight the interplay between LA and LV
functions. Indeed, this might be explained by the fact that LA reservoir phase might
be related to the descent of the mitral annulus during LV systole while LA conduit
phase might be related to both LA tissue mechanical properties and LA to LV
pressure gradient at the opening of the mitral valve.
Limitations
The first limitation of our study is the population size which was relatively small.
However, despite the small samples, significant differences were found between the
young and elderly individuals and LA functional parameters were able to characterize
LA functional alterations in AVS patients. In addition, the primary goal of our study
was to describe and assess the reliability of a feature tracking technique and this was
achieved by contours overlap with a manual reference, reproducibility and
consistency of functional parameters. However, further studies should be performed
on larger populations to fully investigate the added clinical value of LA feature
tracking. Another limitation is the temporal resolution of our MRI data, although it
was equivalent to the previous feature tracking MRI studies (18). Improvement of
temporal resolution might enhance: 1) feature tracking quality in reducing temporal
distances and increasing feature similarity between successive images 2) reliability of
longitudinal strain rate and radial relative velocity and thus provide a better
characterization of alterations in rapid LA emptying and LA contraction. However,
our experiment on the phantom data indicated that up to 30 phases by cardiac cycle
- 17 -
the error on strain measurements remains reasonable. Finally, the use of area-length
method for LA volume quantification on 2 and 4 chamber views may be questionable
since the best method for LA volume quantification is based on summing consecutive
disks covering the LA.
Despite these limitations, contours obtained by our feature tracking technique had an
excellent overlap with manual LA reference contours and LA functional parameters
resulted in higher performance than conventional volumetric parameters in
characterizing LA involvement in AVS patients.
In conclusion, conventional cine MRI SSFP images combined with a fast and
reproducible feature tracking technique resulted in accurate LA functional parameters,
which fully described the complex tri-phasic atrial myocardial function in its
longitudinal and radial components. Such parameters were able to characterize
subclinical age related variations in LA function as well as LA functional impairments
secondary to AVS. Since MRI SSFP images used in this study are systematically
acquired during MRI exam, the addition of the proposed feature tracking tool might
enhance the diagnostic value of routine cardiac MRI.
- 18 -
REFERENCES
1. Candan O, Ozdemir N, Aung SM, et al.: Left Atrial Longitudinal Strain Parameters
Predict Postoperative Persistent Atrial Fibrillation Following Mitral Valve Surgery: A
Speckle Tracking Echocardiography Study. Echocardiography 2013; 30:1061–1068.
2. Posina K, McLaughlin J, Rhee P, et al.: Relationship of phasic left atrial volume
and emptying function to left ventricular filling pressure: a cardiovascular magnetic
resonance study. J Cardiovasc Magn Reson 2013; 15:99.
3. Sun JP, Yang Y, Guo R, et al.: Left atrial regional phasic strain, strain rate and
velocity by speckle-tracking echocardiography: normal values and effects of aging in
a large group of normal subjects. Int J Cardiol 2013; 168:3473–3479.
4. Geyer H, Caracciolo G, Abe H, et al.: Assessment of Myocardial Mechanics Using
Speckle Tracking Echocardiography: Fundamentals and Clinical Applications. J Am
Soc Echocardiogr 2010; 23:351–369.
5. D’Ascenzi F, Cameli M, Henein M, et al.: Left atrial remodelling in patients
undergoing transcatheter aortic valve implantation: a speckle-tracking prospective,
longitudinal study. Int J Cardiovasc Imaging 2013; 29:1717–1724.
6. Sühling M, Jansen C, Arigovindan M, et al.: Multiscale motion mapping: a novel
computer vision technique for quantitative, objective echocardiographic motion
measurement independent of Doppler: first clinical description and validation.
Circulation 2004; 110:3093–3099.
7. Gayat E, Ahmad H, Weinert L, Lang RM, Mor-Avi V: Reproducibility and InterVendor Variability of Left Ventricular Deformation Measurements by Three-
- 19 -
Dimensional Speckle-Tracking Echocardiography. J Am Soc Echocardiogr 2011;
24:878–885.
8. Maceira AM, Cosin-Sales J, Roughton M, Prasad SK, Pennell DJ: Reference left
atrial dimensions and volumes by steady state free precession cardiovascular magnetic
resonance. J Cardiovasc Magn Reson 2010; 12:65.
9. To ACY, Flamm SD, Marwick TH, Klein AL: Clinical utility of multimodality LA
imaging: assessment of size, function, and structure. JACC Cardiovasc Imaging 2011;
4:788–798.
10. Oakes RS, Badger TJ, Kholmovski EG, et al.: Detection and Quantification of
Left Atrial Structural Remodeling With Delayed-Enhancement Magnetic Resonance
Imaging in Patients With Atrial Fibrillation. Circulation 2009; 119:1758–1767.
11. Kuppahally SS, Akoum N, Burgon NS, et al.: Left atrial strain and strain rate in
patients with paroxysmal and persistent atrial fibrillation: relationship to left atrial
structural remodeling detected by delayed-enhancement MRI. Circ Cardiovasc
Imaging 2010; 3:231–239.
12. Garot J, Bluemke DA, Osman NF, et al.: Fast Determination of Regional
Myocardial Strain Fields From Tagged Cardiac Images Using Harmonic Phase MRI.
Circulation 2000; 101:981–988.
13. Hor KN, Baumann R, Pedrizzetti G, et al.: Magnetic Resonance Derived
Myocardial Strain Assessment Using Feature Tracking. J Vis Exp 2011.
14. Hoit BD: Left atrial size and function: role in prognosis. J Am Coll Cardiol 2014;
63:493–505.
- 20 -
15. Darbon J, Cunha A, Chan TF, Osher S, Jensen GJ: Fast nonlocal filtering applied
to electron cryomicroscopy. In 5th IEEE Int Symp Biomed Imaging Nano Macro 2008
ISBI 2008; 2008:1331–1334.
16. Chan TF, Vese LA: Active contours without edges. IEEE Trans Image Process
Publ IEEE Signal Process Soc 2001; 10:266–277.
17. Lynch M, Ghita O, Whelan PF: Segmentation of the Left Ventricle of the Heart in
3-D+t MRI Data Using an Optimized Nonrigid Temporal Model. IEEE Trans Med
Imaging 2008; 27:195–203.
18. Kowallick JT, Kutty S, Edelmann F, et al.: Quantification of left atrial strain and
strain rate using Cardiovascular Magnetic Resonance myocardial feature tracking: a
feasibility study. J Cardiovasc Magn Reson 2014; 16:60.
19. Habibi M, Chahal H, Opdahl A, et al.: Association of CMR-measured LA
function with heart failure development: results from the MESA study. JACC
Cardiovasc Imaging 2014; 7:570–579.
20. Auger DA, Zhong X, Epstein FH, Spottiswoode BS: Mapping right ventricular
myocardial mechanics using 3D cine DENSE cardiovascular magnetic resonance. J
Cardiovasc Magn Reson 2012; 14:4.
21. Haraldsson H, Hope M, Acevedo-Bolton G, et al.: Feasibility of asymmetric
stretch assessment in the ascending aortic wall with DENSE cardiovascular magnetic
resonance. J Cardiovasc Magn Reson 2014; 16:6.
22. Xavier M, Lalande A, Walker PM, Brunotte F, Legrand L: An adapted optical
flow algorithm for robust quantification of cardiac wall motion from standard cine-
- 21 -
MR examinations. IEEE Trans Inf Technol Biomed Publ IEEE Eng Med Biol Soc
2012; 16:859–868.
23. O’Connor K, Magne J, Rosca M, Piérard LA, Lancellotti P: Left atrial function
and remodelling in aortic stenosis. Eur J Echocardiogr 2011; 12:299–305.
24. Farhad H, Neilan T, Abbasi S, et al.: Left atrial passive function after aortic valve
replacement in aortic stenosis. J Cardiovasc Magn Reson 2014; 16(Suppl 1):P237.
25. Lisi M, Henein MY, Cameli M, et al.: Severity of aortic stenosis predicts early
post-operative normalization of left atrial size and function detected by myocardial
strain. Int J Cardiol 2013; 167:1450–1455.
26. Nagueh SF, Appleton CP, Gillebert TC, et al.: Recommendations for the
evaluation of left ventricular diastolic function by echocardiography. J Am Soc
Echocardiogr Off Publ Am Soc Echocardiogr 2009; 22:107–133.
27. Bollache E, Redheuil A, Clément-Guinaudeau S, et al.: Automated left ventricular
diastolic function evaluation from phase-contrast cardiovascular magnetic resonance
and comparison with Doppler echocardiography. J Cardiovasc Magn Reson Off J Soc
Cardiovasc Magn Reson 2010; 12:63.
- 22 -
Tables
Table 1 - Feature tracking algorithms evaluation
Absolute and percentage differences between functional parameters estimated using
the three feature tracking algorithm options and reference strain initially used to
generate the synthetic phantom. Track 1: tracking on gradient images; Track2:
tracking on filtered native images using edge images; Track 3: tracking on filtered
native images using both active contours and edge images. Functional parameters
were: SlR: reservoir longitudinal strain, SlE: conduit longitudinal strain, SlA: LA
contraction longitudinal strain, MrR : reservoir radial motion fraction, MrE : conduit
radial motion fraction, MrA: LA contraction radial motion fraction.
Absolute
long.
Strain
diff. Absolute radial motion fraction
(Percentage)
diff. (Percentage)
Methods
SlR
SlE
SlA
MrR
MrE
MrA
Track 1
0.1 (0.2)
1.8 (7.6)
1.9 (11.9)
1.8 (4.2)
2.0 (7.9)
0.2 (1.3)
Track 2
1.5 (3.7)
0.4 (1.8)
1.9 (11.9)
2.3 (5.7)
2.0 (8.2)
0.3 (1.8)
Track 3
1.3 (3.1)
0 (0.1)
1.3 (8.0)
1.5 (3.5)
0.9 (3.6)
0.6 (3.5)
- 23 -
Table 2 - Comparison of feature tracking algorithms against manual contours
Comparison of LA endocardial contours extracted by the three tracking algorithms vs.
manual Qmass contours for each images of the cardiac cycle. Track 1: tracking on
gradient images; Track2: tracking on filtered native images using edge images; Track
3: tracking on filtered native images using both active contours and edge images (see
methods section for more details).
Mean Distance Max
Dice
Distance
Jaccard
(mm)
(mm)
Methods
Mean±SD
Mean±SD
Mean±SD
Mean±SD
Track 1
0.91±0.04
0.84±0.02
1.60±0.63
5.59±2.30
Track 2
0.91±0.04
0.84±0.02
1.54±0.53
5.75±2.52
Track 3
0.91±0.03
0.84±0.02
1.50±0.50
5.52±2.32
- 24 -
Table 3 - Study population: basic characteristics including LV and LA volumes and
ejection fraction
LV stands for left ventricular, LA for left atrial, EDV for end diastolic volume, ESV
for end systolic volume, EF for ejection fraction. Statistical differences between
young controls and elderly controls are noted by *, and between elderly controls and
AVS patients by x, with *** or
xxx
for p<0.001,** or
xx
for p<0.01, and *or
p<0.05.
Young
Elderly
Patients
controls
controls
with AVS
(n=10)
(n=10)
(n=20)
Age (years)
27.7±6.8
63.7±5.7***
73±14.5xx
Male (%)
60%
60%
65%
Body Surface Area (m²)
1.92±0.19
1.75±0.16*
1.76±0.21
Body Mass Index (kg/m²)
26.1±3.8
24.7±5.2*
25.1±3.7
Indexed LV Mass (g/m²)
65.3±9.6
57.3±10.4*
103.1±31.5xxx
LV EDV (mL)
156.9±33.5
119.9±32.9**
145.7±81.8
LV ESV (mL)
59.2±15.2
45.4±14.7**
67.7±75.5
Indexed LV EDV (mL/m²)
81.9±14.9
68.1±16.1***
82.4±39
Indexed LV ESV (mL/m²)
31.1±7.7
28.9±7.9**
37.9±38.3
LV EF (%)
62.4±4.9
62.28±5.12
60.3±19.7
LA EDV (mL)
77.0±17.0
75.4±29.1
116.7±53.8x
LA ESV (mL)
30.0±8.5
35.1±14.5
78.7±57.6xx
Indexed LA EDV (mL/m²)
40.8±10.3
43±16.2
67±34.2x
Indexed LA ESV (mL/m²)
16.1±6.0
20±7.7
45.4±35.9xx
LA EF (%)
60.9±7.8
53.1±7.7*
37.2±16.8 x
- 25 -
x
for
Table 4 - Distribution of LA functional indices in controls and AVS
Longitudinal strains: reservoir SlR, conduit SlE and LA contraction SlA; longitudinal
strain rates: SRS' , SRE', SRA'. Radial motion fraction: reservoir MrR, conduit MrE and
LA contraction MrA; radial relative velocities: VrS’, VrE’, VrA’. Statistical differences
between young controls and elderly controls are noted by *, and between elderly
controls and AVS patients by x, with *** or
xxx
for p<0.001,** or xx for p<0.01, and *
or x for p<0.05.
Elderly
Patients with
Young controls
controls
AVS
(n=10)
(n=10)
(n=20)
SlR
24.6±6.4
22.6±4.9
12.4±6.5xxx
SlE
10.4±4.4
8.7±3.1
2.7±1.8 xxx
SlA
14.2±6.5
14.0±4.1
9.8±5.9 x
SlR /SlA
1.9±0.5
1.7±0.3
1.4±0.4x
SRS'
1.6±0.6
1.5±0.6
0.7±0.5xx
SRE'
-1.9±0.9
-1.3±0.5
-0.4±0.2xx
SRA'
-2.2±1.1
-1.9±0.9
-1.2±0.9x
MrR
30.6±7.4
26.8±5.5
13.8±7.4xxx
MrE
14.7±3.6
11.6±2.7*
3.3±2.3xxx
MrA
15.8±5.7
15.3±4.1
10.4±6.9x
MrR /MrA
2.0±1.2
1.8±0.2
1.5±0.8 xx
Longitudinal strain (%)
Longitudinal strain rate (%/s)
Radial motion fraction (%)
- 26 -
Radial relative velocity (%/s)
VrS'
1.9±0.7
1.6±0.4
0.8±0.7xxx
VrE'
-2.2±0.7
-1.4±0.6**
-0.5±0.3xxx
VrA'
-2.4±1.2
-2.1±0.8
-1.2±1.0xxx
- 27 -
Table 5 - Ability of volumetric and functional LA parameters to differentiate patients
with AVS from all controls by using functional LA characterization.
Results of the receiver operating characteristic (ROC) analysis with sensitivity,
specificity, negative predictive values (NPV) and positive predictive values (PPV),
accuracy, area under the curve (AUC) and threshold. Indexed LA EDV.: indexed LA
end-diastolic volume, indexed LA ESV.: indexed LA end-systolic volume, LAEF: LA
ejection fraction. Longitudinal strains: reservoir SlR, conduit SlE and LA contraction
SlA; longitudinal strain rates: SRS’' , SRE', SRA'. Radial motion fraction: reservoir MrR,
conduit MrE and LA contraction MrA; radial relative velocities: VrS’, VrE’, VrA’. ROC
Sensitivity
Specificity
NPV
PPV
Accuracy
AUC
Threshold
curves corresponding to LA parameters highlighted in bold are illustrated in Figure 6.
Indexed LA EDV (mL/m²)
0.50
0.90
0.66
0.83
0.71
0.74
59.3
Indexed LA ESV (mL/m²)
0.75
0.81
0.77
0.79
0.79
0.83
24.3
LA EF (%)
0.90
0.62
0.87
0.69
0.76
0.79
56.9
SlR (%)
0.80
0.95
0.83
0.94
0.73
0.90
16.5
SlE (%)
0.95
0.86
0.95
0.86
0.90
0.94
4.5
SlA (%)
0.55
0.86
0.67
0.79
0.73
0.71
9.2
SRlS' (%/s)
0.90
0.81
0.89
0.82
0.73
0.86
1.19
SRlE' (%/s)
0.95
0.95
0.95
0.95
0.95
0.99
-0.74
LA volumes and EF
Longitudinal strains
Longitudinal strain rates
- 28 -
SRlA' (%/s)
0.65
0.95
0.74
0.93
0.80
0.78
-1.06
MrR (%)
0.90
0.86
0.90
0.86
0.68
0.94
22.7
MrE (%)
1.0
1.0
1.0
1.0
1.0
1.0
7.5
MrA (%)
0.65
0.86
0.72
0.81
0.65
0.76
10.4
VrS' (%/s)
0.75
0.95
0.80
0.94
0.77
0.89
0.99
VrE' (%/s)
0.95
0.95
0.95
0.95
0.95
0.98
-0.85
VrA' (%/s)
0.65
0.95
0.74
0.93
0.80
0.80
-0.98
Radial motion fractions
Radial relative velocities
- 29 -
Figure Legends
Figure 1 - Description of the feature tracking algorithm on MRI data and an example of
the synthetic phantom along with the normal longitudinal strain curve used for its
generation
(A) Spatial correlation between 2 ROI weighted by the distance map to LA edge
and/or active contour resulting in a final map. (B) Ellipsoidal phantom with realistic
size fitted on real LA data, and the maximum dilatation and minimum contraction
were obtained from typical normal longitudinal strain curve (C) derived from
literature (max 39.5%). The phantom had a pixel size of 0.73mm and 60 frames.
Figure 2 - Global and regional LA functional parameters
(A) 2 chamber view restricted to the LA, (B) schematic views of LA orientation and
segments, (C) longitudinal strain curves, (D) longitudinal strain rate curves , (E) radial
motion fraction , (F) radial relative velocities for a control subject. Long.:
longitudinal, LAA: left atrium appendage, Ant.: Anterior segment, Roof: LA roof
segment, Inf: Inferior segment.
Figure 3 – Effect of spatial and temporal resolution on functional parameters
Percentage of difference (Diff.) between reference strain and motion fraction used to
generate the phantom and strains measurements obtains from dataset with lower
spatial resolution (A) and temporal resolution (B).
Figure 4 - Box-plots of regional longitudinal strains in controls for reservoir (SlrR),
conduit (SlrE), and LA contraction (SlrA) phases
Long. : longitudinal, Ant: antero, Lat: Lateral, Inf: infero, Sept: septal (see Figure 2
for LA segments definition). Statistical significance is highlighted by * for p<0.05, **
for p<0.01 and *** for p<0.001.
- 30 -
Figure 5
- Comparison between longitudinal strain and LA ejection fraction
andindexed end-diastolic volume.
Linear regression between longitudinal strain and LA ejection fraction for reservoir
(A) and LA contraction phases (C) for the three study groups. Linear regression
between longitudinal strain and LA indexed end-diastolic volume for reservoir (B)
and LA contraction phases (D) for the three study groups.
Figure 6 - Ability of LA volumetric and functional parameters to characterize LA
alterations in AVS
Receiver operating characteristic (ROC) curves: (A) reservoir longitudinal strain
(SlR) and motion fraction (MrR), systolic longitudinal strain rate (SRlS’) and relative
velocity (VrS’); (B) conduit longitudinal strain (SlE) and radial motion fraction
(MrE): (C) LA contraction longitudinal strain rate (SRlA’) and relative velocity
(VrA’). LA EDVi: indexed LA end-diastolic volume, LA ESVi: indexed LA endsystolic volume, LAEF: LA ejection fraction. Of note, LA volumetric indices were
repeated in each frame to enable their direct comparison with LA functional
parameters.
Additional files
Additional file 1 – Illustration of the LA tracking
Cine data of 3 chamber view along with the endocardial tracking and the resulting
relative velocity and radial motion fraction curves.
- 31 -