Download Quantification of left-to-right shunting in adult congenital heart

The British Journal of Radiology, 82 (2009), 386–391
Quantification of left-to-right shunting in adult congenital heart
disease: phase-contrast cine MRI compared with invasive oximetry
1
K DEBL, MD, 2B DJAVIDANI, MD, 1S BUCHNER, MD, 1N HEINICKE,
2
S FEUERBACH, MD, 1G RIEGGER, MD and 1A LUCHNER, MD
MD,
2
F POSCHENRIEDER,
MD,
1
Klinik und Poliklinik für Innere Medizin II and 2Institut für Röntgendiagnostik, Universitätsklinikum Regensburg, Germany
ABSTRACT. Atrial septum defects (ASDs), ventricular septum defects (VSDs) and patent
ductus arteriosus (PDA) are the most common adult congenital heart defects. The degree
of left-to-right shunting as assessed by the ratio of flow in the pulmonary (Qp) and
systemic circulation (Qs) is crucial in the management of these conditions. This study
compared phase-contrast cine magnetic resonance imaging (PC-MRI), a non-invasive
imaging technique, with invasive oximetry for the measurement of shunt volumes during
cardiac catheterisation in adults with left-to-right shunting. Both invasive oximetry and
shunt quantification by PC-MRI (1.5 T scanner; Sonata, Siemens Medical Solutions) were
performed on 21 patients with left-to-right shunting (14 ASD, 5 VSD, 2 PDA) and data on
Qp/Qs ratios and left-to-right shunt fraction compared. Mean Qp/Qs ratios assessed by
PC-MRI and oximetry were 2.10¡0.76 and 1.96¡0.77, respectively (p 5 0.37). Mean
shunt fraction was 46.3¡19.6% when calculated by PC-MRI and 42.3¡20.1% when
obtained by oximetry (p 5 0.12). There was a strong correlation of Qp/Qs ratios and shunt
fraction between both methods (r 5 0.61, p , 0.01 and r 5 0.84, p , 0.0001, respectively).
The two methods had a good agreement according to Bland and Altman plots with a
small but non-significant overestimation of Qp/Qs-ratios and shunt fraction by PC-MRI.
On receiver operating characteristic analysis, the sensitivity and specificity of PC-MRI to
detect an oximetry-derived Qp/Qs ratio of >1.5:1 were 93% and 100% at a PC-MRI
threshold of a Qp/Qs ratio >1.75:1 (area under curve (AUC) 5 0.99). Quantification of
left-to-right shunting can be performed reliably and accurately by PC-MRI and the data
obtained by this method correlate closely to those from invasive oximetry.
Atrial septum defects (ASDs), ventricular septum
defects (VSDs) and patent ductus arteriosus (PDA) are
the most common congenital heart defects in both
children and adults. Quantification of shunt volumes
and of the ratio of flow in the pulmonary (Qp) and
systemic circulation (Qs) and assessment of cardiac
dilatation as a consequence of volume overload are
crucial in the management of patients with left-to-right
shunting [1–3]. Transthoracic and transoesophageal
echocardiography usually allow exact localisation and
sizing of the defect. However, determination of shunt
volumes by Doppler echocardiography has limitations
[4]. Invasive oximetry using Fick’s principle is the
standard of reference for the quantification of left-toright shunting based on measurements of blood oxygenation [5–7]. Interventional closure or corrective surgery
is usually indicated in patients with a Qp/Qs ratio >1.5.
MRI is a non-invasive method that allows an exact
visualisation of cardiac anatomy and an accurate noninvasive sizing of congenital septal defects [8–10].
Additionally, phase-contrast cine MRI (PC-MRI) allows
the comparative quantification of cardiac blood flow in
the pulmonary artery with that in the ascending aorta.
Received 30 August 2007
Revised 24 February 2008
Accepted 6 May 2008
DOI: 10.1259/bjr/18500608
’ 2009 The British Institute of
Radiology
Very good agreement of shunt quantifications by PCMRI and invasive oximetry has been shown in children
with congenital heart disease [11–16]. Data on shunt
quantification by PC-MRI in adults are, however, limited
and predictive values in comparison with oximetry are
lacking. The aim of our study was therefore to compare
shunt volumes assessed by PC-MRI and invasive
oximetry in adults with left-to-right shunting.
Methods
Patients
The study included 21 consecutive adult patients with
suspected or known left-to-right shunt admitted to our
institution for shunt quantification. Invasive oximetry
and shunt quantification by PC-MRI were performed in
all patients within 1 week. All patients gave written
informed consent. PC-MRI and catheterisation studies
were well tolerated in all patients.
Magnetic resonance imaging studies
Address correspondence to: K Debl, Klinik und Poliklinik für Innere
Medizin II, Universitätsklinikum, F.J.-Strauss-Allee 11, 93042
Regensburg, Germany. E-mail: [email protected]
K Debl and B Djavidani contributed equally as first author.
386
Magnetic resonance imaging studies were performed
on a 1.5 T scanner (Sonata, Siemens Medical Solutions).
For the assessment of cardiac function and localisation of
The British Journal of Radiology, May 2009
Quantification of shunting in adult congenital heart disease by MRI
(a)
(b)
(c)
Figure 1. (a) 4-chamber view of a patient with an atrial septum defect II (SSFP sequences); (b) phase-contrast image
perpendicular to the ascending aorta; (c) phase-contrast image perpendicular to the main pulmonary artery.
the left-to-right shunt, cine images were acquired in
multiple short axis and long axis views by fast imaging
with steady-state precession (SSFP, slice thickness 8 mm,
echo time 1.53 ms, pixel bandwidth 1085 Hz, repetition
time 3.14 ms, matrix 256*202). In addition, SSFP
sequences were performed to generate scout images of
the ascending aorta and pulmonary trunk in three
planes. Flow volumes in the ascending aorta and
pulmonary trunk were determined in the isocentre of
the magnet using through plane phase-contrast MRI (PCMRI) in a retrospective gating technique during normal
respiration (flash 2D; slice thickness 5 mm, echo time
3.2 ms, pixel bandwidth 391 Hz, repetition time 41 ms,
matrix 256*192), as described previously [17]. Using the
SSFP scout images, PC-MRI was performed orthogonally
to the ascending aorta and the main pulmonary artery.
Slice position in the ascending aorta was approximately
2.0–3.0 cm above the aortic valve distal to the coronary
arteries at the level of the right pulmonary artery. Slice
position in the pulmonary trunk was approximately 1.5–
2.0 cm above the pulmonary valve but proximal to the
bifurcation. Using this approach, the influence of the
angle deviation is usually below 15 ˚ and therefore
negligible. To optimise the accuracy of PC-MRI, the
velocity encoding was set as low as possible while
avoiding aliasing artefacts. Usually, the velocity encoding was 150 cm s21 for the measurement of the aortic
and pulmonary flow. To achieve flow results as
physiological as possible, PC-MRI was performed during
normal respiration. Data analyses were performed using
ARGUS Flow software (Sonata, Siemens Medical
Solutions). For the calculation of pulmonary (Qp) and
systemic (Qs) flow volumes, the cross-sectional areas of
the pulmonary trunk and ascending aorta were drawn
manually for each time frame on the magnitude images
and transferred to the corresponding phase image
(Figure 1). Shunt volumes for left-to-right shunts in
ASD and VSD were calculated by the following formula:
Left-to-right shunt ð%Þ~ðQp{QsÞ=Qp
ð1Þ
and in the case of PDA as:
(Qs{Qp)=Qs
The British Journal of Radiology, May 2009
ð2Þ
Oximetry
Calculation of Qp and Qs values was performed as
described by Grossman [18]. Blood samples in the upper
and lower vena cavae, right atrium and ventricle,
pulmonary arteries and a systemic artery were taken to
determine the oxygen saturation. The shunt volume was
calculated by the difference of mixed venous and
pulmonary oxygen saturation. The invasive calculation
of the blood flow for the systemic and pulmonary
circulation was determined by Fick’s principle:
Qp l min{1 ~Oxygen uptake ml min{1
=PVcontent ml l{1 PAcontent ml l{1
ð3Þ
Qs l min{1 ~Oxygen uptake ml min{1
=SAcontent ml l{1 MVcontent ml l{1
ð4Þ
where PVcontent is the systemic arterial oxygen content,
PAcontent is the oxygen content in the pulmonary artery,
SAcontent is the systemic arterial oxygen content,
MVcontent is the mixed venous oxygen content calculated
as (3 6 superior vena cava oxygen saturation + inferior
vena cava oxygen saturation)/4.
Again, shunt volumes for left-to-right shunts were
calculated by the following formula:
Left-to-right shunt ð%Þ~ðQp{QsÞ=Qp
ð5Þ
Statistical analysis
The agreement between the methods of shunt quantification was assessed by univariate regression analysis
and by the Bland–Altman method. Differences in mean
values between two groups were analyzed by Student’s
t-test. The x2 test was performed to compare frequencies
between groups. Receiver operating characteristic (ROC)
analysis was carried out to determine the predictive
ability of PC-MRI to detect left-to-right shunting as
defined by oximetry. A level of significance of below 0.05
387
K Debl, B Djavidani, S Buchner et al
was defined as statistically significant. SPSS version
12.0.1 was used for statistical analysis.
100%, respectively, at a PC-MRI threshold of a Qp/Qs
ratio >1.75:1 (AUC 5 0.99) (Table 2).
Results
Discussion
Patient characteristics are depicted in Table 1. Of the 21
patients, 14 had ASD (9 ASD II, 3 ASD I, 2 sinus venosus
defect), 5 VSD and 2 PDA. No significant differences in
age, gender, rhythm and mean pulmonary artery pressure
were observed in patients with Qp/Qs ratios , 1.5 as
compared to patients with Qp/Qs ratios >1.5.
Our study demonstrates that the quantification of leftto-right shunting in adults by PC-MRI is feasible and
reliable. There is a good correlation to the data derived
by invasive oximetry, with a small but non-significant
overestimation of Qp/Qs ratios and shunt fractions by
PC-MRI. Furthermore, PC-MRI obtained excellent predictive values for the detection of significant left-to right
shunting requiring shunt closure (Table 2).
Comparison of shunt quantification by oximetry
and MRI
Quantification of left-to-right shunting by PC-MRI
compared to oximetry
Mean Qp/Qs ratios ¡ 2 SD assessed by PC-MRI and
cardiac catheterisation were 2.10¡0.76 and 1.96¡0.77,
respectively (p 5 0.37). Mean shunt fraction was
46.3¡19.6% when assessed by PC-MRI and
42.3¡20.1% at cardiac catheterisation (p 5 0.12). Mean
differences of Qp/Qs ratios and shunt fraction between
the two assessment methods were 0.14% and 4.0%,
respectively. Both Qp/Qs ratios and shunt fraction were
strongly correlated between the two methods (r 5 0.61,
p , 0.01 and r 5 0.84, p , 0.0001, respectively). The
Bland–Altman plots also showed good agreement
between the two methods with a small but nonsignificant overestimation of Qp/Qs ratios and shunt
fraction by PC-MRI. Figures 2 and 4 show that in two
patients the Qp/Qs ratios obtained by PC-MRI and
cardiac catheterisation fell at opposite sides of the cut-off
value for clinical decision making of 1.5. PC-MRI would
have suggested defect closure in these two cases,
whereas oximetry would have led to a conservative
strategy (Figures 2–5).
The current study is among the first to compare PCMRI and oximetry by cardiac catheterisation in freebreathing adults who have congenital heart disease. It
confirms previous studies that have already shown the
feasibility of this technique and a good correlation with
invasive oximetry in children. In the management of
patients with left-to-right shunting, quantification of the
ratio of flow in the pulmonary (Qp) and systemic
circulation (Qs) and assessment of cardiac dilatation as
a consequence of volume overload are crucial [1–3].
Invasive oximetry is the gold standard for the quantification of left-to right shunting but has several limitations.
Oxygenation of venous and arterial blood is especially
influenced by several factors such as sedation, agitation
and ventilation disorders. Furthermore, the true mixed
venous saturation is not measured directly but is calculated
by sampling from the superior and inferior vena cavae.
Additionally, there is a large intra-individual variability of
Qp/Qs ratios assessed by oximetry [6]. Despite these
limitations, oximetry remains the standard criterion against
which new non-invasive methods must be judged.
A potential limitation of PC-MRI-based calculation of
the shunt volume calculation may result from the
assessment of systemic flow in the ascending aorta distal
to the coronary ostia, which would miss coronary blood
flow. This technical limitation may result in a small
underestimation of Qs values and consequently lead to a
slight overestimation of Qp/Qs ratios and left-to-right
shunt fraction, respectively. In consequence, underestimation of Qs is probably responsible for a small
overestimation of shunt volumes by PC-MRI compared
to oximetry. In this respect, our study is in agreement
with a recent study [11] that also suggests that
quantification of left-to-right shunting by PC-MRI may
overestimate shunt volumes as compared to invasive
oximetry. Accordingly, ROC analysis determined that
detection of Qp/Qs ratios of >1.5:1 by oximetry was best
achieved at a PC-MRI threshold of >1.75:1.
In our study, PC-MRI derived Qp/Qs ratios pointed to
shunt closure in two patients, whereas oximetry-based
Qp/Qs ratios would have favoured a conservative
strategy. Interestingly, the Qp/Qs ratios assessed by
oximetry were only slightly below the cut-off value of 1.5
in these two patients. Qp/Qs ratio is, however, only one
criterion in clinical decision making in left-to-right
Patients
Predictive values of PC-MRI for severity of left-toright shunting
As assessed by ROC analysis, the sensitivity and
specificity of PC-MRI in detecting Qp/Qs ratios of >1.5:1
(as determined by invasive oximetry) were 93% and
Table 1. Patient characteristics
Age
Gender (% male)
SR (%)
Mean PAP (mm Hg)
Cardiac shunts
ASD (n)
VSD (n)
PDA (n)
Qp/Qs ,1.5 (n56)
Qp/Qs >1.5 (n515)
47.0¡14.0
66
50
19.0¡7.2
50.9¡16.9
53
80
25.4¡12.7
2
4
0
11
1
2
Qp/Qs, pulmonary-to-systemic flow ratio assessed by oximetry; SR, sinus rhythm; PAP, pulmonary artery pressure,
derived at right heart catheterisation; ASD, atrial septum
defect; VSD, ventricular septum defect; PDA, patent ductus
arteriosus.
388
The British Journal of Radiology, May 2009
Quantification of shunting in adult congenital heart disease by MRI
Figure 2. Pulmonary-to-systemic
flow ratios (Qp/Qs) determined by
phase-contrast cine MRI (PC-MRI)
and oximetry in 21 patients with
left-to-right shunting. Only two
patients are in the right lower
quadrant in which a PC-MRI derived
Qp/Qs ratio of >1.5 would have
suggested shunt closure, whereas
data from oximetry would not have.
shunting in adults. Although there are limitations in
estimating Qp/Qs ratios with any method, high levels of
shunting are usually associated with adverse long-term
outcomes. Especially in ASD, further signs such as right
heart dilatation and an ASD minimum diameter
.10 mm must also be considered in the management
of these patients [1]. In VSD, left ventricular volume
overload or progressive aortic valve disease are usually
additional indications for shunt closure [2]. Finally, PDA
closure is clearly indicated for symptomatic adults with
significant left-to-right shunting and for asymptomatic
patients with significant left-to-right shunting and
evidence of left heart enlargement [3].
Clinical relevance of MRI in left-to-right shunt
quantification
With respect to the clinical work-up of patients with
left-to-right shunting, transthoracic and transoesopha-
Figure 3. The Bland and Altman
plot depicts the mean of the pulmonary-to-systemic flow ratios (Qp/
Qs) obtained by phase-contrast cine
MRI (PC-MRI) and oximetry and the
difference between these Qp/Qs
ratios for each of 21 patients with
left-to-right shunting. The mean
difference (solid line) and ¡ 2 SD
from this difference (broken lines)
across all patients are shown.
The British Journal of Radiology, May 2009
389
K Debl, B Djavidani, S Buchner et al
Figure 4. Left-to-right shunt fractions determined by phase-contrast
cine MRI (PC-MRI) and oximetry in
21 patients with left-to-right shunting. Only two patients are in the
right lower quadrant where a PCMRI-derived shunt fraction suggests
shunt closure whereas oximetry
does not.
geal echocardiography are often adequate to determine
the defect size. However, transoesophageal echocardiography is semi-invasive and might have limitations in
visualisation of adjacent structures and anomalous
venous return. Determination of shunt volumes by
Doppler echocardiography also has limitations [4].
Invasive oximetry using Fick’s principle is the standard
of reference for the quantification of left-to-right shunting based on measurements of blood oxygenation, but
also has several limitations as discussed above.
PC-MRI could overcome these problems of defect sizing
and quantification of shunt volumes. Interestingly, ASD
sizing by MRI correlates well with assessments made by
transoesophageal echocardiography and provides additional information on defect shapes, rim distances and
pulmonary and venous anomalies that are often associated
with left-to-right shunting in congenital defects [8–10].
Furthermore, PC-MRI allows the reliable quantification of
left-to-right shunting. MRI can therefore provide anatomical and functional information in a single examination.
Figure 5. The Bland and Altman
plot depicts the mean left-to-right
shunt fraction derived by phasecontrast cine MRI (PC-MRI) and
oximetry and the difference of leftto-right shunt fractions by PC-MRI
and oximetry for each of 21 patients
with left-to-right shunting. The
mean difference (solid line) and ¡
2 SD from this difference (broken
lines) across all patients are shown.
390
The British Journal of Radiology, May 2009
Quantification of shunting in adult congenital heart disease by MRI
Table 2. Predictive values
CATH-Qp/Qs >1.5:1
CATH-shunt >33.3%
Cases /Total (n)
PC-MRI cut-off
ROC area (95% CI)
Sensitivity (%)/specificity (%)
15/21
15/21
1.75:1
43%
0.99 (0.97–1.02)
0.99 (0.97–1.02)
93/100
93/100
CI, confidence interval; ROC, receiver operating characteristic; PC-MRI, phase-contrast cine MRI.
In summary, quantification of left-to-right shunting
can be performed reliably and accurately by PC-MRI and
provides results that correlate closely with those
obtained by invasive oximetry, although there is a small
overestimation of the degree of shunting. In the clinical
management of patients with left-to-right shunting, MRI
can provide anatomical and functional information in a
single examination and is a useful technique for the
assessment of adult congenital heart disease.
References
1. Webb G, Gatzoulis M. Atrial septal defects in the adult:
recent
progress
and
overview.
Circulation
2006;114:1645–53.
2. Minette MS, Sahn DJ. Ventricular septal defects. Circulation
2006;114:2190–7.
3. Schneider DJ, Moore JW. Patent ductus arteriosus.
Circulation 2006;114:1873–82.
4. Boehrer JD, Lange RA, Willard JE, Grayburn PA, Hillis LD.
Advantages and limitations of methods to detect, localize,
and quantitate intracardiac left-to-right shunting. Am Heart
J 1992;124:448–55.
5. Daniel WC, Lange RA, Willard JE, Landau C, Hillis LD.
Oximetric versus indicator dilution techniques for quantitating intracardiac left-to-right shunting in adults. Am J
Cardiol 1995;75:199–200.
6. Cigarroa RG, Lange RA, Hillis LD. Oximetric quantitation
of intracardiac left-to-right shunting: limitations of the Qp/
Qs ratio. Am J Cardiol 1989;64:246–7.
7. Antman EM, Marsh JD, Green LH, Grossman W. Blood
oxygen measurements in the assessment of intracardiac left
to right shunts: a critical appraisal of methodology. Am J
Cardiol 1980;46:265–71.
8. Sechtem U, Pflugfelder P, Cassidy MC, Holt W, Wolfe C,
Higgins CB. Ventricular septal defect: visualization of shunt
flow and determination of shunt size by cine MR imaging.
Am J Roentgenol 1987;149:689–92.
9. Piaw CS, Kiam OT, Rapaee A, Khoon LC, Bang LH, Ling
CW et al. Use of non-invasive phase contrast magnetic
resonance imaging for estimation of atrial septal defect size
The British Journal of Radiology, May 2009
10.
11.
12.
13.
14.
15.
16.
17.
18.
and morphology: a comparison with transesophageal echo.
Cardiovasc Intervent Radiol 2006;29:230–4.
Beerbaum P, Korperich H, Esdorn H, Blanz U, Barth P,
Hartmann J et al. Atrial septal defects in pediatric patients:
noninvasive sizing with cardiovascular MR imaging.
Radiology 2003;228:361–9.
Esmaeili A, Hohn R, Koch A, Vogl TJ, Hofstetter R,
Abolmaali N. Assessment of shunt volumes in children
with ventricular septal defects: comparative quantification
of MR flow measurements and invasive oximetry. Clin Res
Cardiol 2006;95:523–30.
Beerbaum P, Korperich H, Barth P, Esdorn H, Gieseke J,
Meyer H. Noninvasive quantification of left–to-right shunt
in pediatric patients: phase–contrast cine magnetic resonance imaging compared with invasive oximetry.
Circulation 2001;103:2476–82.
Beerbaum P, Korperich H, Gieseke J, Barth P, Peuster M,
Meyer H. Rapid left-to-right shunt quantification in
children by phase–contrast magnetic resonance imaging
combined with sensitivity encoding (SENSE). Circulation
2003;108:1355–61.
Hundley WG, Li HF, Lange RA, Pfeifer DP, Meshack BM,
Willard JE et al. Assessment of left-to-right intracardiac
shunting by velocity-encoded, phase–difference magnetic
resonance imaging. A comparison with oximetric and
indicator dilution techniques. Circulation 1995;91:2955–60.
Kalden P, Kreitner KF, Voigtlander T, Roberts H, Roberts T,
Krummenauer F et al. Flow quantification of intracardiac
shunt volumes using MR phase contrast technique in the
breath holding phase. Rofo 1998;169:378–82.
Petersen SE, Voigtlander T, Kreitner KF, Kalden P,
Wittlinger T, Scharhag J et al. Quantification of shunt
volumes in congenital heart diseases using a breath-hold
MR phase contrast technique — comparison with oximetry.
Int J Cardiovasc Imaging 2002;18:53–60.
Lotz J, Meier C, Leppert A, Galanski M. Cardiovascular
flow measurement with phase–contrast MR imaging: basic
facts and implementation. Radiographics 2002;22:651–71.
Grossman W. Shunt detection and measurement. In:
Grossman W, Baim DS, eds. Cardiac catheterization,
angiography and intervention. 4th ed. Philadelphia, PA:
Lea & Febiger; 1991:166–181.
391