Cardiac Imaging • Original Research

Saremi et al.
MDCT of Sinoatrial Node Artery

Cardiac Imaging
Original Research

MDCT of the S-Shaped Sinoatrial

Node Artery
Farhood Saremi1 OBJECTIVE. The purpose of this study was to use 64-MDCT to investigate the anatomic
Stephanie Channual1 characteristics of the S-shaped variant of the sinoatrial node (SAN) artery and to describe the
Amir Abolhoda2 clinical implications of the findings in ablative procedures involving the left atrium.
Swaminatha V. Gurudevan 3 MATERIALS AND METHODS. Coronary CT angiograms of 250 patients (152 men, 98
Jagat Narula3 women; mean age, 60 ± 12 [SD] years) were retrospectively analyzed for identification of the ori-
gin, number, anatomic course, mode of termination, and S-shaped variant of the SAN artery.
Jeffrey C. Milliken2
RESULTS. At least one SAN artery was detected in 244 patients. The S-shaped variant was
American Journal of Roentgenology 2008.190:1569-1575.

Saremi F, Channual S, Abolhoda A, Gurudevan SV,
Narula J, Milliken JC
seen in 35 (14.3%) of these patients. Thirty-four of the variants (30.6% of all left SAN arteries)
arose from the proximal to middle portion of the left circumflex artery (mean distance between
the ostium of the left circumflex artery and the origin of S-shaped variant, 28.7 ± 13.1 mm). The
other variant (0.7% of all right SAN arteries) originated from the distal right coronary artery.
The S-shaped variant was the only artery supplying the SAN in 28 (11.4%) of the patients. In
patients with two arteries supplying the SAN, the right SAN artery and the S-shaped variant of
the left SAN artery were seen together in seven patients. The S-shaped SAN artery (mean dis-
tance from atrial wall, 2.43 ± 0.992 mm) had a predictable proximal course, lying in the poste-
rior aspect in a groove between the orifices of the left superior pulmonary vein and the left
atrial appendage close to the left atrial wall. The terminal segment of the artery approached the
nodal tissue posterior to the superior vena cava in 22 patients, anterior to the vena cava in 10
patients, and through branches surrounding the vena cava in two patients.
CONCLUSION. The S-shaped variation of the SAN artery is common and has a char-
acteristic anatomic course. MDCT can be used to plan surgical and catheter-based left atrial
interventions in which this artery is at risk of injury.

Keywords: ablation, coronary artery, coronary CT
angiography, left atrium, sinoatrial node artery
DOI:10.2214/AJR.07.3127
Received September 8, 2007; accepted after revision
December 25, 2007.
1
Department of Radiological Sciences, University of
California, Irvine, UCI Medical Center, 101 The City Dr.,
Route 140, Orange, CA 92868-3298. Address correspon-
dence to F. Saremi (fsaremi@uci.edu).
2 to the anterior wall of the left atrium [1–7].
Department of Cardiothoracic Surgery, University of
California, Irvine, UCI Medical Center, University of
California, Irvine, Orange, CA.
3
Department of Cardiology, University of California, Irvine,
UCI Medical Center, University of California, Irvine, Orange,
CA.
AJR 2008; 190:1569–1575
0361–803X/08/1906–1569
© American Roentgen Ray Society
R
eports of cadaveric dissections
and angiographic studies of the
human heart describe the S-
shaped sinoatrial node (SAN) ar-
tery, an anatomic variant of the left SAN ar-
tery, as a relatively large vessel arising from
the left circumflex (LCX) artery and coursing
posteriorly between the left atrial appendage
(LAA) and the ostium of the left superior pul-
monary vein (LSPV) and then anteriorly close
The unusual anatomic course and proximity
to the left atrial wall predispose this vessel to
injury during cardiac interventions. Accurate
anatomic imaging of this variant blood vessel
can influence planning for surgical and cathe-
ter-based ablation of the left atrium.
Coronary CT angiography has been found
useful for evaluation of coronary artery steno-
sis, anomalous coronary anatomic features,
and the arterial supply of the cardiac con­
duction system [8–13]. Newer-generation
64-MDCT scanners provide enough temporal
and spatial resolution for evaluation of epicar-
dial coronary arteries and their smaller branch
vessels [14, 15]. We have described our expe-
rience with MDCT for detailing the anatomic
features of the arterial blood supply of the
sinoatrial and atrioventricular nodes [13]. The
purpose of this study was to evaluate our ex-
tended results with 64-MDCT for imaging of
this variant of the SAN artery.
Materials and Methods
A retrospective analysis of ECG-gated MDCT
(Aquilion unit, Toshiba) examinations of the
coronary artery was conducted over 4 months
(February–June 2007) with 265 consecutively
registered subjects, either self-referred (n = 130)
because of personal interest in participating in
research, for which they had given consent, or
referred by a physician because of suspected

AJR:190, June 2008 1569


coronary artery disease (CAD). Two hundred fifty
patients were included in the study. Fifteen patients
were excluded because of severe artifacts produced
by low contrast-to-noise ratio (n = 4), motion (n =
10), or pacemaker leads (n = 1).
Completed clinical history questionnaires were
reviewed: 16.1% of the patients reported a history of
atrial fibrillation, 8.6% had structural heart disease
such as heart failure or cardiomyopathy, 6.8% had a
history of coronary artery stent placement, 5.4% had
undergone coronary artery bypass graft (CABG),
surgery, and 5.4% had a history of myocardial in­
farction. The mean age of the patients was 60 ± 12
(SD) years (range, 27–86 years). One hundred fifty-
two (60.8%) of the patients were men, and 98
(39.2%) were women. The mean weight of the entire
cohort was 176.3 ± 37.3 lb (80.0 ± 16.9 kg) (range,
100–280 lb [45.4–127.0 kg]). The study was approved
by the institutional review board at our institution.
Informed consent was waived, and the study was in
compliance with HIPAA regulations.
American Journal of Roentgenology 2008.190:1569-1575.
Patient Preparation
ECG was performed and vital signs were
obtained for all patients when they arrived in the
imaging suite. When necessary, patients were given
oral and IV metoprolol to achieve a target heart rate
less than 65 beats/min. Unless contraindicated, a
sublingual nitroglycerin tablet (0.4–0.8 mg) was
given 1 minute before image acquisition. The mean
heart rate before data acquisition was 58 ± 6 beats/
min (range, 40–73 beats/min).
Scan Protocol and Image Reconstruction
The 64-MDCT scanner had an individual
detector width of 0.5 mm. Contrast enhancement
was achieved with a mean dose of 74.91 ± 3.32 mL
(range, 65–92 mL) of iohexol (Omnipaque 350
mg/mL, GE Healthcare) injected at 4–5 mL/s and
followed by an injection of 50 mL of saline
solution at 5 mL/s through an 18-gauge catheter
into an antecubital vein to allow washout of
contrast material from the right side of the heart
and the superior vena cava (SVC). The scan
parameters were collimation, 64 × 0.5 mm; table
feed per rotation, 7.2 mm; gantry rotation time,
400 milliseconds; tube voltage, 120 kVp; tube
current, 400 mA. Start delay was defined with
bolus tracking in the descending aorta at the level
of tracheal bifurcation, and scan start was
automatically initiated 4 seconds after a threshold
of 180 H was reached. Retrospective ECG-gated
volumetric data sets were acquired during a single
breath-hold. The mean scan duration was 9.0 ± 1.3
seconds with a range of 7.4–15 seconds.
Axial slices were reconstructed and syn­chro­nized
to the ECG with a nonsegmented (≤ 65 beats/min) or
segmented (> 65 beats/min) image reconstruction
1570 AJR:190, June 2008
Saremi et al.
algorithm based on the heart rate throughout the
examination. When necessary, R-wave indicators
were manually repositioned to improve the quality
of synchronization. On the basis of a relative delay
strategy, diastolic data sets were reconstructed at
70%, 75%, and 80% of R-R intervals. In case of
persistence of artifacts related to coronary motion at
the atrioventricular groove, a second reconstruction
approach was made, and systolic images were
reconstructed with an absolute delay strategy with a
350- to 400-millisecond delay after the previous R
wave. Data sets reconstructed during the diastolic
phase were used for all patients. Axial slices with a
thickness of 0.5 mm (increment, 0.3 mm) and a
cardiac CT angiography algorithm were used for
evaluating coronary vessels and conduction system
branches. The data set least affected by cardiac
motion was transferred to an off-line 3D workstation
(Vitrea, Vital Images) for further analysis.
CT Data Analysis
A radiologist and a cardiologist with advanced
training in cardiovascular CT interpretation
rendered and evaluated multiplanar reformations of
the contrast-enhanced axial images according to
training guidelines published by a task force of the
American College of Cardiology Foundation and
the American Heart Association [16]. Visualization
techniques such as maximum intensity projection
and 3D reconstruction with tissue sculpting
depended on individual coronary anatomy and
image quality. Overall image quality was
qualitatively evaluated and classified as excellent,
good, or adequate primarily on the basis of common
image-degrading artifacts related to metal, motion,
and background noise. The imaging findings related
to CAD were graded as normal, mild (any vessel
with less than 50% stenosis), moderate (any vessel
with 50–70% stenosis), or severe (any vessel with
greater than 70% stenosis).
Axial images were reviewed. Coronary vessels
were examined to identify the SAN artery. Images
were analyzed for identification of the origin,
number, anatomic course, and anatomic variants of
the arteries to the SAN with specific attention to the
S-shaped variant of the SAN artery. The diameter
for all SAN arteries was measured close to the
origin of the artery. The distance between the
ostium of the LCX artery and the origin of the
S-shaped SAN artery also was measured. The
anatomic course of the S-shaped SAN artery was
assessed, and its closest distance from the left atrial
wall in the groove between the ostium of the LSPV
and the mouth of the LAA was measured. The
diameter of the left S-shaped SAN artery in this
groove also was measured. The mode of termination
of the S-shaped SAN artery in relation to the SVC
was classified as retrocaval (posterior to the SVC),
precaval (anterior to the SVC), or pericaval (through
multiple branches surrounding the SVC). Each set
of images was evaluated for the presence of
additional anatomic findings, including unusual
origin of the S-shaped SAN artery, atrial branches,
and dual blood supply to the SAN.
Statistical Analysis
Statistical analysis was performed with SAS
software (version 9.1.3, SAS Institute). Mean ±
standard deviation was used to report univariate
statistics for all continuous data. In addition, 95%
CI estimates of percentages were reported.
Statistical analysis for bivariate comparisons was
performed. A value of p < 0.05 was considered
statistically significant.
Results
Image Quality and Vessel Visualization
Image quality was classified as excellent
in 76% of the cases, good in 20%, and ade-
quate in 4%. Axial and 3D images were the
optimal views for visualization of the ana-
tomic course of the SAN arteries and their
atrial branches.
S-Shaped Posterior SAN Artery
Among the 250 patients in the study, the
SAN artery was not visualized in six patients.
In the other 244 patients, a single SAN artery
originated from the right coronary artery
(RCA) in 133 (54.5%) of the patients and from
the LCX artery in 99 (40.6%) of the patients
(Table 1). A dual blood supply to the SAN
from the RCA and LCX arteries was seen in
12 (4.9%) of the patients. Among 250 coro-
nary CT angiograms, 46.2% were normal,
30.5% had evidence of mild CAD, 13.3% evi-
dence of moderate CAD, and 10.0% evidence
of severe CAD in any vessel.
An S-shaped SAN artery was seen in 35
(14.3%) of the 244 patients with SAN arteries.
Thirty-four (30.6%) of the 111 left SAN arter-
ies (99 single, 12 part of a dual supply) origi-
nated from the proximal to middle portion of
the LCX artery (mean distance from the LCX
ostium, 28.7 ± 13.1 mm; range, 12.2–65.0
mm). One (0.7%) of the 145 right SAN arteries
(133 single, 12 part of a dual supply) originated
from the distal RCA. Among the 34 patients
with a left-sided S-shaped SAN artery, 67.6%
had normal findings on coronary CT angiogra-
phy, whereas 29.4% had mild CAD and 2.94%
had moderate CAD in the LCX artery. There
was no relation between the dominance of the
coronary system and the origin of the S-shaped
SAN artery. Among 35 S-shaped vessels, 27
were from a right-dominant, seven from a left-
 MDCT of Sinoatrial Node Artery

dominant, and one from a balanced coronary
system. Women were less likely than men to
have the S-shaped variant (relative risk,
0.4596; 95% CI, 0.2177–0.9700; chi-square
value, 4.56; p = 0.0327).
The S-shaped variant was the only artery
supplying the SAN in 28 patients; 27 of these
TABLE 1:  Incidence and Measurements of Arteries Supplying the Sinoatrial
Node and Their Associated Branches (n = 244)
Vessel
Single SAN artery from RCA
Single SAN artery from LCX artery
Dual SAN arteries from both RCA and LCX
Atrial branch
S-shaped SAN artery from RCA
S-shaped SAN artery from LCX artery
Distance between ostium of LCX artery and origin of S-shaped
SAN artery (mm)
American Journal of Roentgenology 2008.190:1569-1575.
variants were from the LCX artery and one left SAN artery were seen together in five
from the RCA. These 28 cases were 11.4% of (41%) of the patients, and the right SAN ar-
all 244 SAN arteries, 27.3% of 99 single tery and the S-shaped variant of the left SAN
SAN arteries arising from the LCX artery, artery were seen together in seven (58.3%) of
and 0.8% of 133 single SAN arteries arising the patients. This dual supply involving the
from the RCA. In the 12 cases of dual blood S-shaped SAN variant was seen in 2.9% of
supply to the SAN, the right SAN artery and all 244 patients with SAN arteries.
The S-shaped variant of the left SAN ar-
tery appeared larger and longer than the usual
SAN artery. Whereas the mean diameters of
n % the regular right and left SAN arteries were
1.58 ± 0.287 mm (range, 0.90–2.2 mm) and
133 54.5
1.40 ± 0.314 mm (range, 0.70–1.9 mm), re-
99 40.6 spectively, the mean diameter of the left S-
12 4.9 shaped SAN artery was 2.12 ± 0.347 mm
15 6.1
(range, 1.3–2.8 mm) close to its origin. The
left S-shaped SAN artery invariably coursed
1 0.7
posteriorly in a groove at the junction between
34 30.6 the orifice of the LSPV and the mouth of the
28.7 ± 13.1a 12.2–65.0b LAA and had a mean diameter of 1.90 ± 0.331
mm (range, 1.4–2.7 mm) within the groove

Distance between left S-shaped SAN artery and atrial wall (mm) 2.43 ± 0.992a 1.0–4.8b (Fig. 1). In this groove, the artery was close
(mean distance, 2.43 ± 0.992 mm; range,
Route of approach of terminal segment of S-shaped variant of left 1.0–4.8 mm) to the left atrial wall. The course
SAN artery to SAN
of the artery continued toward the SVC along
Retrocaval 22 64.7 the transverse sinus anterior to the left atrium
Precaval 10 29.4 and interatrial groove.
Pericaval 2 5.9 The terminal segment of the S-shaped
variant of the left SAN artery approached
Note—SAN = sinoatrial node artery, RCA = right coronary artery, LCX = left circumflex.
aMean ± SD. the SAN by three different routes (Fig. 2):
bRange. posterior to the SVC (retrocaval) in 22 (64.7%)

A B C
Fig. 1—62-year-old man with chest pain. Anatomic course of S-shaped SAN artery.
A–C, Axial CT scan at level of left atrial appendage (LAA) (A) and left lateral 3D CT scans in two different projections (B and C) of heart show that on axial images,
S-shaped posterior sinoatrial node (SAN) artery (large arrows, A–C) can easily be identified where it courses between LAA and left superior pulmonary vein (LSPV).
Artery arises from proximal left circumflex artery and courses along lateral wall of left atrium. It usually gives off branches to atrial wall (small arrows, B) before making
U-turn toward LAA–LSPV groove. In groove, SAN artery is very close to atrial wall (arrow, A) and can be damaged in surgical procedures on LAA or pulmonary vein
isolation procedures. From this point, anatomic course of S-shaped SAN artery is similar to that of left SAN artery, which courses toward superior vena cava along
anterior wall of left atrium. LPV = left pulmonary vein trunk, AA = ascending aorta.

AJR:190, June 2008 1571


A B
American Journal of Roentgenology 2008.190:1569-1575.
C D
of the patients, anterior to the SVC (precaval)
in 10 (29.4%), and through multiple branches
surrounding the SVC (pericaval) in two
(5.9%) of the patients. The retrocaval mode
of termination was therefore the most com-
mon pattern, and there was a significant dif-
ference between this mode and the precaval
mode of termination (chi-square value, 4.50;
p = 0.0339). At the level of the interatrial
groove, the S-shaped SAN artery typically
sent multiple branches to supply the inter-
atrial Bachmann’s bundle. In the retrocaval
mode of termination, the S-shaped SAN ar-
tery penetrated the interatrial bundle before
reaching the SAN in the lateral aspect of the
superior cavoatrial junction.
An unusual variant of the S-shaped SAN
artery originating from the RCA distal to the
origin of the posterolateral artery was identi-
fied in one (0.4%) of the 244 patients with
SAN arteries. This artery coursed posterior-
ly around the posterior aspect of the coro-
nary sinus and left atrium and then anteriorly
toward the ostium of the left inferior pulmo-
nary vein and then the LSPV (Fig. 3). The
course was 1.2 mm from the left atrial wall
in the groove between the LAA and the
LSPV and along the transverse sinus, termi-
nating precavally to supply the SAN.
An atrial branch from the S-shaped poste-
rior SAN artery was found in 6.0% of all 250
1572 AJR:190, June 2008
Saremi et al.
cases reviewed. The proximal portion of this
branch was commonly seen supplying the
posterior wall of the left atrium (Fig. 1). In
two patients, an atrial branch was seen aris-
ing from the proximal LCX artery, which
had a course similar to that of the S-shaped
artery between the LAA and the LSPV but
did not reach the SAN area. The SAN was
supplied by the right SAN artery in one of
these patients and by the left SAN artery in
the other (Fig. 4).
Discussion
Several previously published reports [1–6,
17, 18] provide anatomic descriptions of the
blood supply to the cardiac conduction sys-
tem. These reports, however, are based solely
on findings at invasive transarterial angiogra-
phy or cadaveric dissection of the human
heart. These descriptions also focus primarily
on the main named epicardial vessels supply-
ing the nodal tissues, such as the left or right
SAN arteries. We have reported [13] our ex-
perience with noninvasive MDCT of the arte-
rial supply to the SAN and the atrioventricu-
lar node. To our knowledge, the current report
is the first of noninvasive anatomic imaging
of a common variant of the left SAN artery,
namely, the S-shaped SAN artery, with
64-MDCT (0.5-mm isotropic resolution) of
living human subjects. Our results are in con-
Fig. 2—Terminal anatomic course of S-shaped
sinoatrial node (SAN) artery.
A–D, 85-year-old man with retrocaval (A and B)
and 61-year-old woman with precaval (C and D)
mode of termination of artery. Sculptured 3D CT
images (A and C) and corresponding axial images
(B and D) at level of superior cavoatrial junction
show S-shaped SAN arteries (short arrows, A and
C). Proximal courses of artery are similar in most
cases. Arising from proximal left circumflex artery,
S-shaped SAN artery turns posteriorly and courses
in groove between left atrial appendage (LAA) and
left superior pulmonary vein (LSPV) orifices (short
arrows, B and D). We found that in most cases distal
artery coursed close to interatrial groove, penetrated
interatrial muscle bundle, and followed its course
behind superior vena cava (SVC) (retrocaval) to reach
SAN area on lateral aspect of cavoatrial junction
(long arrows, A and B). This variant is prone to injury
in superior septal approach to mitral valve repair. In
precaval mode of termination (long arrows, C and D),
artery courses away from interatrial groove to reach
anterior margin of superior cavoatrial junction and
is less susceptible to injury. AA = ascending aorta,
RSPV = right superior pulmonary vein.
cordance with the previous angiographic and
cadaveric findings. We found that the S-
shaped variant of the left SAN artery accounts
for approximately 30% of SAN arteries aris-
ing from the LCX coronary artery and 14.3%
of all SAN arteries. Kyriakidis et al. [19] and
Nerantzis and Avgoustakis [2] have similarly
reported that this variant can be seen in 27%
and 21%, respectively, of left SAN arteries.
These angiographic and cadaveric findings
confirm the accuracy of our observations.
McAlpine [1] extensively described the S-
shaped variant as a posterior SAN artery aris-
ing from the proximal 50 mm of the LCX ar-
tery coursing over the lateral wall of the left
atrium between the orifices of the LAA and
LSPV and then commonly proceeding intra-
murally through the interatrial septal bundle
to the superior cavoatrial junction. This artery
supplies the SAN and surrounding area, in-
cluding a large part of the right atrial free wall
and interatrial septum, the dome of the left
atrium, and, on occasion, a portion of the
atrioventricular nodal area [2, 3]. McAlpine’s
descriptions of the S-shaped variant in cadav-
ers mirror our noninvasive imaging findings.
It is of interest that the S-shaped SAN artery
follows a course similar to that of the left su-
perior cardinal vein of the fetus and the
oblique vein of Marshall in adults. Therefore,
structures such as a persistent left SVC or a
 MDCT of Sinoatrial Node Artery

This pattern of termination is different for

normal left-sided SAN arteries, for which
the precaval route is most common.
Busquet et al. [31] conducted a cadaveric
study on the normal SAN artery and described
its termination as precaval (58%), retrocaval
(36%), or encircling (6%). We [13] reported
that the terminal SAN arteries (right or left)
approached the SAN retrocavally in 47.5% of
the patients, precavally in 42.5%, and peri-
A cavally in 10%. Berdajs et al. [27] analyzed 50
human cadaveric hearts that did not have pre-
vious pathologic alterations and found that the
sinus node artery crossed the superior poste-
rior border of the interatrial septum in 54% of
the hearts. Because retrocaval variants course
in proximity to the interatrial groove [13], the
S-shaped posterior SAN artery may be more
prone than normal SAN arteries to surgical
trauma during superior transseptal incisions.
Intraoperative damage to the SAN artery can
American Journal of Roentgenology 2008.190:1569-1575.

be avoided if preoperative MDCT angiogra-

B C phy accurately depicts the course and mode of
termination of the artery, leading to modifica-
Fig. 3—45-year-old woman with S-shaped sinoatrial node artery (short black arrows, C) arising from right tion of the surgical approach as indicated.
coronary artery (RCA).
A–C, Axial CT images at level of left atrial appendage (LAA) (A) and coronary sinus (B) and 3D posterior CT
Several previous reports [2, 4, 31] have de-
scan (C) of heart show rare variant found in one patient. S-shaped sinoatrial node artery originates from scribed a blood supply to the SAN through the
terminal branches of dominant right coronary artery posterior to coronary sinus (CS) (B and C) and courses common variants of the right and left SAN ar-
along posterolateral wall of left atrium toward groove between LAA and left superior pulmonary vein (LSPV). teries. Kyriakidis et al. [19] reported a dual
IVC = inferior vena cava, LIPV = left inferior pulmonary vein, PLA = posterolateral artery (long black arrows, C),
RA = right atrium. White arrows indicate areas of A and B that correspond to C. supply to the SAN involving the right SAN
artery and the S-shaped variant of the left
recanalized Marshall ligament should be con- benefit of the ablation procedure. Therefore, SAN artery. To our knowledge, our report is
sidered in the differential diagnosis of an en- atrial fibrillation ablation procedures should the first on the MDCT findings of this unusual
hancing structure passing between the LAA be care­fully planned with this anatomic vari- variant of a dual blood supply to the SAN. The
and the LSPV (Fig. 4). In addition, although ant in mind. presence of a dual blood supply to the same
saphenous vein grafts usually lie behind the Imaging information on the mode of ter- anatomic cardiac region may serve as a pro-
LAA in patients who have undergone CABG mination of the SAN artery, the S-shaped tective mechanism against ischemia in pa-
graft surgery, the proximity of the graft to the variant in particular, can be of clinical rele- tients with progressive coronary artery athero-
LSPV–LAA groove can cause the graft to be vance [27]. For example, in the popular sclerosis. For instance, the collateral arcade
mistaken for the S-shaped SAN artery (Fig. 4). transseptal superior dome approach to the between the right SAN artery and the S-
Potential clinical implications of our im- mitral valve apparatus, cardiac surgeons de- shaped left SAN artery arising from the mid-
aging findings in patients undergoing inva- liberately extend the septal incision toward portion of the LCX coronary artery may less-
sive procedures on the left and right atria are the roof the left atrium and the LAA [28]. en coronary ischemia in the LCX territory in
beyond anatomic curiosity. For instance, the The left SAN artery traverses the transverse the case of development of hemodynamically
long, aberrant, yet predictable course of the sinus and reaches the nodal tissue by a retro- significant proximal LCX stenosis. Further-
S-shaped variant of the SAN artery can ex- caval route and becomes susceptible to inju- more, a patient with a dual blood supply would
pose this vessel to injury during epicardial ry. It is clinically well documented that pa- be potentially less vulnerable to ischemia-
or endocardial ablation performed in prox- tients who undergo the superior transseptal mediated sick sinus syndrome.
imity to the posterosuperior left atrial wall surgical approach to mitral valve surgery of- Hutchinson [32] reported that the S-shaped
and the base of the LAA. These procedures ten have postoperative disturbances in sinus posterior SAN artery originated from the ter-
include epicardial pulmonary vein isolation node function, but these changes are tran- minal part of the RCA and coursed in a clock-
and ligation of the LAA as part of any modi- sient [29, 30]. Our study revealed that the wise direction around the walls of the right and
fied Cox maze operation for correction of termination mode of the S-shaped variant of left atria to the SAN in one (2.5%) of 40 hearts.
atrial fibrillation [20–26]. The consequence the SAN artery can differ from that of nor- We found this variant of the S-shaped poste-
of such injury, especially if the vessel is the mal SAN arteries. We observed that the S- rior SAN artery branching from the terminal
sole blood supply to the SAN, can be sinus shaped variant most often (64.7%) has a ret- branches of the RCA in one (0.4%) of 244 pa-
node dysfunction and even junctional escape rocaval termination, followed by precaval tients with SAN arteries. This rare variant also
rhythm, negating the potential therapeutic (29.4%) and pericaval (5.9%) terminations. coursed in a clockwise direction around the

AJR:190, June 2008 1573


A B
American Journal of Roentgenology 2008.190:1569-1575.
C D
posterior wall of the left atrium. We also iden-
tified multiple posterior atrial branches arising
from the S-shaped variant of the left SAN ar-
tery in 6% of the patients. In addition, we
found two atrial branches arising directly from
the LCX artery and having an anatomic course
similar to that of the S-shaped variant. These
atrial branches should not be confused with
the S-shaped SAN artery because they do not
extend to the SAN area (Fig. 4).
Our study had the expected limitations of a
retrospective observational review, and our data
were primarily collected from patients with
structurally normal hearts. The latter limitation
may make our findings difficult to generalize to
patients with cardiac anomalies or pathologic
heart conditions. Nevertheless, our results are in
accordance with those of historical angiographic
studies and cadaveric dissections of the human
heart, emphasizing the validity of our observa-
tions. In addition, the number of patients includ-
ed in our study was larger than in past studies of
SAN arteries [3, 7, 27, 31, 32], which may allow
greater generalization of our results.
The S-shaped variant of the SAN artery
as imaged with 64-MDCT is a common vari-
ant of the left SAN artery in patients under-
1574 AJR:190, June 2008
Saremi et al.
going noninvasive evaluation for CAD.
MDCT is a useful imaging adjunct for pre-
cise description of the course and mode of
termination of the S-shaped variant of the
SAN artery. This imaging technique can be
used to avoid injury to this vessel during sur-
gical or catheter-based procedures on the
atrial chambers of the human heart.
References
1. McAlpine WA. Heart and coronary arteries: an
anatomic atlas for clinical diagnosis, radiologic
investigation, and surgical treatment. New York,
NY: Springer-Verlag, 1975:151–162
2. Nerantzis C, Avgoustakis D. An S-shaped atrial
artery supplying the sinus node area: an anatomic
study. Chest 1980; 78:274–278
3. Tanaka S, Lee HY, Mizukami S, Nakatani T,
Chung IH. Posterior sinus node artery and acces-
sory atrioventricular node artery arising by a
common origin: a case report. Clin Anat 1998;
11:106–111
4. Kyriakidis M, Vyssoulis G, Barbetseas J, Toutou-
zas P. A clinical angiographic study of the arterial
blood supply to the sinus node. Chest 1988;
94:1054–1057
5. Anderson KR, Ho SY, Anderson RH. Location
Fig. 4—Findings to be differentiated from S-shaped
sinoatrial node artery. S-shaped sinoatrial node
artery invariably courses in groove between left
atrial appendage and left superior pulmonary vein
and is best localized on axial images.
A, 65-year-old man with a history of coronary artery
disease and chest pain. CT scan shows independent
small atrial branch arising from left circumflex artery
(not shown) and coursing in groove between left
atrial appendage and left superior pulmonary vein
and behind left atrium (white arrows) but not reaching
sinoatrial node area. True sinoatrial node artery is
left-sided artery arising from left circumflex artery
(black arrows).
B, 56-year-old woman with recanalized ligament
of Marshall. CT scan shows ligament of Marshall
(arrow) is recanalized because left brachiocephalic
vein is occluded (not shown).
C, 61-year-old man with persistent left superior
vena cava. CT scan shows superior vena cava
(arrow) partially filled by collateral vessels. Cardiac
chambers are poorly filled with contrast material.
D, 72-year-old man who has undergone coronary
artery bypass graft surgery. CT scan shows typical
course of saphenous vein graft to obtuse marginal
artery. Graft usually courses behind left atrial
appendage but with enough distance from groove
(white arrow) and should not be mistaken for
S-shaped sinoatrial node artery on axial images.
S-shaped variant of sinoatrial node artery (black
arrows) is evident.
and vascular supply of sinus node in human heart.
Br Heart J 1979; 41:28–32
6. Sow ML, Ndoye JM, Lo EA. The artery of the
sinuatrial node: anatomic considerations based on
45 injection-dissections of the heart. Surg Radiol
Anat 1996; 18:103–109
7. Ortale JR, Paganoti Cde F, Marchiori GF. Ana-
tomic variations in the human sinuatrial nodal
artery. Clinics 2006; 61:551–558
8. Achenbach S, Giesler T, Ropers D, et al. Detection
of coronary artery stenoses by contrast-enhanced,
retrospectively electrocardiographically-gated,
multislice helical CT. Circulation 2001;
103:2535–2538
9. Vogl TJ, Abolmaali ND, Diebold T, et al. Tech-
niques for the detection of coronary atherosclero-
sis: multidetector row CT coronary angiography.
Radiology 2002; 223:212–220
10. Ropers D, Baum U, Pohle K, et al. Detection of
coronary artery stenoses with thin-slice multide-
tector row helical CT and multiplanar reconstruc-
tion. Circulation 2003; 107:664–666
11. Duran C, Kantarci M, Durur Subasi I, et al. Re-
markable anatomic anomalies of coronary arter-
ies and their clinical importance: a MDCT angio-
graphic study. J Comput Assist Tomogr 2006;
30:939–948

12. Stratznig P, Groell R, Rienmueller R. Detection of
the sinus node artery using electron beam CT of the
heart. Int J Cardiovasc Imaging 2002; 18:457–461
13. Saremi F, Abolhoda A, Ashikyan O, et al. Arterial
supply to sinuatrial and atrioventricular nodes:
imaging with multidetector CT. Radiology 2008;
246:99–107
14. Leschka S, Alkadhi H, Plass A, et al. Accuracy of
MSCT coronary angiography with 64-slice tech-
nology: first experience. Eur Heart J 2005;
26:1482–1487
15. Raff GL, Gallagher MJ, O’Neill WW, Goldstein
JA. Diagnostic accuracy of noninvasive coronary
angiography using 64-slice spiral computed to-
mography. J Am Coll Cardiol 2005; 46:552–557
16. Budoff MJ, Achenbach S, Fayad Z, et al. Task
force 12: training in advanced cardiovascular im-
aging (CT)—endorsed by the American Society
of Nuclear Cardiology, Society for Cardiovascu-
lar Angiography and Interventions, Society of
Atherosclerosis Imaging and Prevention, and So-
American Journal of Roentgenology 2008.190:1569-1575.
ciety of Cardiovascular Computed Tomography. J
Am Coll Cardiol 2006; 47:915–920
17. Sow ML, Ndoye JM, Lo EA. The artery of the
atrioventricular node: an anatomic study based on
38 injection-dissections. Surg Radiol Anat 2000;
22:93–96
18. Anderson KR, Murphy JG. The atrio-ventricular
node artery in the human heart. Angiology 1983;
AJR:190, June 2008 1575
MDCT of Sinoatrial Node Artery
34:711–716
19. Kyriakidis MK, Kourouklis CB, Papaioannou JT,
Christakos SG, Spanos GP, Avgoustakis DG. Si-
nus node coronary arteries studied with angiogra-
phy. Am J Cardiol 1983; 51:749–750
20. Rigatelli G, Tranquillo M, Stefano P. Abnormally
huge S-shaped sinus node artery: a possible ca-
veat for mitral valve surgery. Int J Cardiol 2007;
116:401–402
21. Haissaguerre M, Jais P, Shah DC, et al. Electro-
physiological end point for catheter ablation of atrial
fibrillation initiated from multiple pulmonary ve-
nous foci. Circulation 2000; 101:1409–1417
22. Khaykin Y, Marrouche NF, Saliba W, et al. Pul-
monary vein antrum isolation for treatment of
atrial fibrillation in patients with valvular heart
disease or prior open heart surgery. Heart Rhythm
2004: 1:33–39
23. Sueda T, Nagata H, Orihashi K, et al. Efficacy of
a simple left atrial procedure for chronic atrial fi-
brillation in mitral valve operations. Ann Thorac
Surg 1997; 63:1070–1075
24. Wang JZ, Du RY, Ding HX, et al. Limited poste-
rior left atrial linear radiofrequency ablation for
patients with chronic atrial fibrillation undergoing
rheumatic valvular heart surgery. Chin Med J
(Engl) 2004; 117:758–760
25. Cox JL, Schuessler RB, D’Agostino HJ Jr, et al. The
surgical treatment of atrial fibrillation. Part III. De-
velopment of a definitive surgical procedure. J Tho-
rac Cardiovasc Surg 1991; 101: 569–583
26. Cox JL, Jaquaiss RD, Schuessler RB, Boineau JP.
Modification of the maze procedure for atrial flut-
ter and atrial fibrillation. Part II. Surgical tech-
nique of the maze III procedures. J Thorac Car-
diovasc Surg 1995; 110:485–495
27. Berdajs D, Patonay L, Turina MI. The clinical
anatomy of the sinus node artery. Ann Thorac
Surg 2003; 76:732–735
28. Gaudino M, Alessandrini F, Glieca F, et al. Con-
ventional left atrial versus superior septal ap-
proach for mitral valve replacement. Ann Thorac
Surg 1997; 63:1123–1127
29. Misawa Y, Fuse K, Kawahito K, Saito T, Konishi
H. Conduction disturbances after superior septal
approach for mitral valve repair. Ann Thorac Surg
1999; 68:1262–1264
30. Garcia-Villarreal OA, Gonzalez-Oviedo R, Ro-
driguez-Gonzalez H, Martinez-Chapa HD. Supe-
rior septal approach for mitral valve surgery: a
word of caution. Eur J Cardiothorac Surg 2003;
24:862–867
31. Busquet J, Fontan F, Anderson RH, Ho SY, Da-
vies MJ. The surgical significance of the atrial
branches of the coronary arteries. Int J Cardiol
1984; 6:223–236
32. Hutchinson MC. A study of the atrial arteries in
man. J Anat 1978; 125:39–54