Adjustment method for mechanical Boston scientific corporation 30 MHz
intravascular ultrasound catheters connected to a Clearview?console
Mechanical 30 MHz IVUS Catheter Adjustment
Nico Bruining1, Ronald Hamers1, Tat-Jin Teo2, Pim J. de Feijter1, Patrick W. Serruys1&
Jos R.T.C. Roelandt1
1Erasmus MC Rotterdam, Rotterdam, Netherlands;
Center, Fremont, California, USA
2Boston Scientific Corporation, IVUS Technology
Received 6 May 2003; accepted in revised form 13 October 2003
Key words: coronary disease, imaging, ultrasonics
Intracoronary ultrasound (ICUS) is often used in studies evaluating new interventional techniques. It is
important that quantitative measurements performed with various ICUS imaging equipment and materials
are comparable. During evaluation of quantitative coronary ultrasound (QCU) software, it appeared that
Boston Scientific Corporation (BSC) 30 MHz catheters connected to a Clearview?ultrasound console
showed smaller dimensions of an in vitro phantom model than expected. In cooperation with the manu-
facturer the cause of this underestimation was determined, which is described in this paper, and the QCU
software was extended with an adjustment. Evaluation was performed by performing in vitro measure-
ments on a phantom model consisting of four highly accurate steel rings (perfect reflectors) with diameters
of 2, 3, 4 and 5 mm. Relative differences (unadjusted) of the phantom were respectively: 15.92, 13.01, 10.10
and 12.23%. After applying the adjustment: )0.96, )1.84, )1.35 and )1.43%. In vivo measurements were
performed on 24 randomly selected ICUS studies. These showed differences for not adjusted vs. adjusted
measurements of lumen-, vessel- and plaque volumes of )10.1 ± 1.5, )6.7 ± 0.9 and )4.4 ± 0.6%. An
off-line adjustment formula was derived and applied on previous numerical QCU output data showing
relative differences for lumen- and vessel volumes of 0.36 ± 0.51 and 0.13 ± 0.31%. 30 MHz BSC
catheters connected to a Clearview?ultrasound console underestimate vessel dimensions. This can retro-
spectively be adjusted within QCU software as well as retrospectively on numerical QCU data using a
Intravascular ultrasound (IVUS) and intracoron-
ory ultrasound (ICUS) allow transmural, highly
detailed tomographic imaging of blood vessels
providing insights into the pathology of vessel wall
disease by defining its geometry and showing ma-
jor components of atherosclerotic plaques. The
advantages of this technology have been described
in many publications [1–3]. The ability of ICUS to
detect and show small amounts of intima hyper-
plasia has made this technique popular to visualize
and quantify results of new interventional tech-
niques such as for example, the revolutionary new
drug-eluting stents. The results of the first drug-
eluting stent studies are showing extremely good
The International Journal of Cardiovascular Imaging 20: 83–91, 2004.
? 2004 Kluwer Academic Publishers. Printed in the Netherlands.
results . However, it is possible that small
amounts of intima will grow into these stents
over time. To monitor this process, ICUS, fol-
lowed by quantitative coronary ultrasound analy-
sis (QCU), is being applied in most of these on-
going studies. The quantification of small amounts
of plaque requires accurate and reliable QCU
Recently, we started a validation project of a
new QCU package CURAD (CURAD, Wijk bij
Duurstede, Netherlands) . Since in our institu-
tion multiple different ICUS catheters and con-
soles are used, it was decided to perform an in
vitro study to validate the whole measurement
‘chain’ from catheter until final analysis. During
this validation, an underestimation of dimensional
measurements was found when using 30 MHz
mechanical rotating ICUS catheters from the
Boston Scientific Corporation (BSC, IVUS Tech-
nology Center, Fremont, CA, USA), formerly
(CVIS), connected to a BSC Clearview?console.
We report our findings of this validation study,
identify the found problem and describe solutions
developed in cooperation with BSC, the manu-
For the in vitro validation study we used two
ultrasound consoles: (1) a BSC Clearview?console
and (2) a Jomed In-Vision Gold console. Two
different catheters were connected to the Clear-
view?console: (1) a 30 MHz mechanical rotating
element  catheter and (2) a 40 MHz mechanical
rotating element catheter, both manufactured by
BSC. To the In-Vision Gold system we connected
a 20 MHz electrical phased array catheter 
In vitro ICUS imaging
A phantom model consisting of four steel rings
with diameters of 2, 3, 4 and 5 mm (deviation
<1%) was constructed for the validation study.
The rings were mounted in a transparent synthetic
housing (Figure 1). A special construction allowed
the outer sheath of the mechanical BSC catheters
to be straightened in the opening of the phantom
and the ultrasound element positioned in the cen-
ter of the steel rings. Unfortunately, such catheter
Figure 1. In this figure the measurement ‘chain’ is presented. At the left side the phantom is visible. Then there are two ultrasound
consoles, (1) a BSC Clearview?and (2) a Jomed In-Vision Gold. One 30 MHz Ultracross?and one AtlantisTM40 MHz catheter were
connected to the Clearview?. To the In-Vision Gold console an Avanar?20 MHz catheter was connected. Image data were stored on
S-VHS video-tape, digitally on a 3D workstation (EchoScan, TomTec GmbH) and for the In-Vision Gold console also on CD-ROM.
On the far right is the analysis station.
guidance could not be applied with the design of
the phased array catheters; those do not have an
To transport the ultrasound waves, we used a
fluids mixture containing 90% degassed water and
10% ethanol as the in vitro replacement for blood.
The room temperature was 22?. This resulted in an
ultrasound propagation speed of approximately
1.548 mm/ls . The ultrasound propagation
speed used to perform calculations in the BSC
Clearview?console was 1.563 mm/ls. The speed
used in the In-Vision Gold system was unknown.
The catheters were inserted into the phantom
and pulled back using a device that pulled the
catheters with a continuous speed of 0.5 mm/s .
The images were recorded for both systems on S-
VHS videotape and simultaneously digitized using
a three-dimensional (3D) workstation (EchoScan,
TomTec GmbH, Unterschleissheim, Germany)
. On the In-Vision Gold system, the images
were also stored on CD-ROM in the DICOM
image format .
For the in vivo measurements, we used 24 ran-
domly selected ICUS image data sets, e.g. pullback
sequences, previously acquired in patients partici-
pating in different interventional studies and were
selected for analysis in this study since they were
imaged with 30 MHz BSC mechanical rotating
element catheters connected to a BSC Clearview?
ultrasound console. All patients, except one, were
imaged post-stent implantation.
In vivo ICUS imaging
All patients received 250 mg aspirin and 10,000 U
heparin IV. If the duration of the entire catheter-
ization procedure exceeded 1 h, the activated
clotting time was measured, and intravenous
heparin was administered to maintain an activated
clotting time of >300 s. The 30 MHz catheter used
was equipped with a 2.9F 15-cm-long sonolucent
distal sheath with a lumen that alternatively
houses the guide-wire (during catheter introduc-
tion) or the ultrasound transducer (during imag-
ing, after retraction of the guide-wire). This design
avoids direct contact of the ICUS imaging element
to the vessel wall. The ICUS transducer was
withdrawn with a pullback device operating with a
continuous speed of 0.5 mm/s.
ICUS analysis protocol
The in vitro and in vivo ICUS image data sets were
analyzed with an off-line semi-automated QCU
software package, CURAD . The pullback
sequences recorded on S-VHS videotapes were
digitized with a frame-grabber. The software
translated them into the DICOM imaging stan-
dard . Calibration of the videotaped images is
performed using the calibrated grid on the images
provided by the manufacturer. It also imported
already digitized data recorded on Magneto-
Optical disk from the 3D EchoScan workstation
or imports any other DICOM image sequence,
such as stored on CD-ROM from the In-Vision
Gold system (Figure 1). The acquired in vitro
ICUS images from the phantom with the 30 MHz
BSC catheter were also analyzed on the 3D
EchoScan workstation and on the Clearview?
console itself (Figure 2).
Adjustment of BSC 30 MHz catheter measurement
During the validation of the QCU software, it
became apparent that the 30 MHz BSC catheters
showed considerably different measurement results
than the other two catheter types (results in
Table 1), and from what could be expected theo-
retically. After consulting with the research and
development department of BSC the following
explanation was given for this underestimation:
The catheter configuration of the BSC catheters
contains an outer sheath to prevent direct contact
of the ultrasound element to the vessel wall. For
the Ultracross?30 MHz catheters this sheath has
a wall thickness of 127 lm. This thickness will
delay the ultrasound signal (Figure 3). This issue
was not taken into account with previous software
versions on the BSC ultrasound consoles. The
speed of sound in the sheath material is 2.38 mm/
ls. The temperature dependence of this speed is
negligible for the sheath material used over the
range of temperatures of interest. The transducer
is tilted 5? from the normal to the sheath. The
curvature of the sheath is not taken into account.
The Clearview?instrument assumes a 1.563 mm/
ls propagationspeed. Thetwo-way transit
through the sheath in pulsed-echo mode therefore
introduces a radial adjustment term (RAT) that
can be calculated as
cos 5 ?
The factor of two accounts for the pulsed echo
two-way transit and the numerator accounts for
the difference in transit time due to the sheath and
what the machine assumes. The cosine factor
corrects for the tilt angle.
The adjustment term applies to any radial
measurements, thus also influences area measure-
ments. Any diameter measurement made with
Ultracross?30 MHz catheters on a Clearview?
console with software version 4.11, 4.12 and 4.22
should add 2 ? 87:6 ¼ 175:2 lm to bring their re-
sults in line with AtlantisTM40 MHz catheters or
20 MHz Jomed phased array catheters or all BSC
catheters operated on the BSC GalaxyTMultra-
For an accurate adjustment, the RAT needs to
be applied to the raw contour data. Therefore, it
RAT ¼ 2 ?1:5625
Figure 2. Images of the phantom. Panel A shows a cross-sectional ultrasound image of the 5 mm steel ring acquired with a 30 MHz
Ultracross?catheter. In panel B, a 2D reconstructed longitudinal reconstruction of the 2, 3, 4 and 5 mm steel rings from the phantom
is shown. In panel C, a screen dump of the measurements on the BSC Clearview?ultrasound console itself is presented.
Table 1. In vitro validation.
BSC 30 MHzBSC 40 MHz Jomed 20 MHz
C2 vs. C1
C4 vs. C1
C6 vs. C1
C8 vs. C1
In column 1 (C1) the theoretical areas are given. In C2, C6 and C8 the measured areas are presented. In C4 the area results from the
adjusted quantitative coronary ultrasound software are presented. In C3, C5, C7 and C9 the relative differences between the measured
and the theoretical areas are given.
Figure 3. The increased propagation speed of the ultrasound
signal crossing the catheter sheath was neglected in the BSC
30 MHz Ultracross?software in the Clearview?instrument.
This must be taken into account for accurate radial, and thus
also area measurements. Applying Equation 1 can retrospec-
tively do this.
has been implemented in the core of the QCU
BSC 30 MHz mathematical model based
Besides the integrated QCU solution to adjust for
the underestimation of measurements, as described
above, a mathematical model was developed to be
applied to numerical QCU result data (areas and
volumes), avoiding the necessity to perform a time
consuming complete re-analysis of previously
The model is based on two output parameters
incorporated in the QCU software, namely the
projected maximum and
diameter of each contour . In the model, an
adjustment factor to be applied to the measured
cross-sectional area, is calculated as the ratio of
adjusted and not adjusted areas of a mathematical
ellipse described by the maximum and minimum
projected diameters as its long and short axis
¼ Unadjusted area ?
þ4 ? RAT2
d ? D
1 þ 2 ? RAT ?
d þ D
d ? D
where d is the projected minimum diameter and D
the projected maximum diameter and RAT ¼
87.6 lm (Figure 4A and C). The last term,
to the final adjusted area (<0.5%) and was there-
fore neglected. The applied formula is thus:
d ?DÞ, is a very small contributing component
¼ Unadjusted area ?
1 þ 2 ? RAT ?
d þ D
d ? D
This formula (Equation 2b) was also applied for
adjustment of the total measured volume (mean
area times segment length) of the analyzed seg-
ment, with d and D in this case taken as the mean
projected diameters of the analyzed segment. This
was done to investigate if a time-consuming re-
calculation per frame could be avoided.
There are other QCU software packages that do
not calculate the projected contour diameters in
their analyses. Therefore, we investigated if a cir-
cular model could also be applied for retrospective
adjustment to area measurements:
Adjusted area ¼ Unadjusted area ?
1 þ2 ? RAT
Where d is the mean contour diameter in a single
cross-sectional frame as derived from the mea-
sured area d ¼
, RAT ¼ 87.6 lm (Figure 4B
Quantitative data are presented as mean ±
standard deviation (SD). All results are plotted in
diagrams according to Bland and Altman .
Figure 4. In panels A and C, the elliptical model based on the
projected contour diameters is presented. The ellipse is defined
by the projected minimum contour diameter (d) and the pro-
jected maximum contour diameter (D) (Equation 2). In panels
B and D, the circular model is presented. From the detected
contour, the enclosed area is calculated. This area defines the
circle to which the radial adjustment factor can be applied
In vitro validation
In Table 1, the results of the in vitro phantom
measurements, additional adjustment and recal-
culations are presented. Area measurements for
the BSC 40 MHz (C6 and C7) and the Jomed
20 MHz (C8 and C9) were within a maximum
relative range of 2% difference. However, the BSC
30 MHz catheter showed a relative difference up
to 15.92% (C2 and C3). After applying the RAT
within the QCU software, the recalculated results
of the 30 MHz catheters came into the same
accuracy range as compared to the other two
catheters (C4 and C5).
In vivo measurements
In Table 2, all in vivo results are presented. The
relative differences between the actual and the ad-
justed volumetric results were quite large (C1–C3).
Applying the mathematical models, operating
retrospectively on the numerical output, the rela-
tive differences found were for area adjustment per
frame and subsequent recalculation of the volumes
for the lumen 0.31 ± 0.53% (C7). It was for
adjustment on the total volume 0.36 ± 0.51%
(C5) and finallyfor
0.66 ± 0.52% (C9). Similar results were found for
the total vessel and plaque measurements. Bland–
Altman results for these measurements are pre-
sented in results plot 1 (Figure 5).
The present study demonstrates that measure-
ments performed with BSC 30 MHz catheters
connected to a Clearview?console underestimate
true dimensions. Such findings have been previ-
ously reported [14, 15]. With the explanation of the
cause of the problem as supplied by BSC, the
manufacturer, an adjustment was integrated into
the QCU analysis software.
After verification that this method of adjust-
ment brought dimensional measurements for BSC
30 MHz catheters well into range with the other
investigated catheters and consoles, the adjusted
results could be used to test a mathematical model
working retrospectively on previously calculated
numerical QCU data. From the derived mathe-
matical models (Equations 2b and 3) it was sur-
prising to see that of all three different methods of
recalculating the volumes, even the worst relative
difference was, on average, well below 1% in
comparison to adjusted re-analysis within the
QCU software. Even the simple circular model,
resulting in a relative difference of 0.66 ± 0.52%
for lumen and 0.26 ± 0.32% for the total vessel
volumes, appeared to be a valid method for retro-
Table 2. In vivo measurements.
C1 vs. C2
C4 vs. C2
C6 vs. C2
C8 vs. C2
In column 1 (C1) the measured volumes are presented. In C2 the results from the adjustment quantitative coronary ultrasound (QCU)
software are given. In C3 the relative differences between the actual and the adjusted results are given. In C4 the results of the
mathematical model applied to the total measured volume of the coronary segment are presented. C5 shows the relative differences
between the model output and the adjusted results. In C6 the mathematical model is applied to each individual frame followed by
recalculation of the volumes, relative differences are shown in C7. In C8 the volumes are recalculated assuming a circular model for
each individual frame. C9 shows the accompanying relative differences against the model QCU output.
Figure 5. The results plot. In the top two graphs (A and B) the Bland–Altman analysis of the luminal volumetric measurements is
presented. At the left side is the before and after adjustment within the QCU software. At the right side are the relative difference
results of the two mathematical models compared to the QCU adjustment. In panels C and D the results of the vessel volumetric
measurements are presented in a similar fashion as for the lumen results. In panel E, finally the plaque results are shown.
BSC states a diameter accuracy claim of
±300 lm or ±10%, whichever is greater and that
the underestimation found is within this accuracy
claim. The newly developed 40 MHz catheters on
the Clearview?console, as well as catheters on the
GalaxyTMconsole utilize an algorithm that further
It is important to note that the adjustment is not
cancelled in the plaque calculation. This results in
a mean underestimation of plaque volume of 4.4%
(Table 2). The adjusted plaque volume is the dif-
ference of the adjusted vessel and lumen volumes.
In some baseline and follow-up studies we
encountered that patients have been imaged using
various catheters at different times. For example,
at baseline with a 30 MHz catheter and at follow-
up with a 40 MHz catheter, or vice versa. From the
analysis provided in this paper, comparison be-
tween these two data sets would not be appropri-
ate as the plaque volume; for example, from the
30 MHz catheter would yield a 4.4% underesti-
mation. We recommend that the adjustment factor
provided in this paper be applied according to the
methods described in this paper. Some previously
published results such as the choice of balloon size
based on QCU measurements , clinical deci-
sion making of stenting in small vessel using ICUS
based measurements [17, 18] or comparison papers
of QCU vs. QCA [12, 19], may benefit from
applying the adjustment factor retrospectively.
Potential sources of error and study limitations
The liquid solution used with the reported room
temperature did not exactly match the ultrasound
propagation speed implemented by BSC, although
the difference is rather small, 1.548 vs. 1.563 mm/
ls , leading to an area measurement inaccuracy
of about 1.5%. The steel rings used in the phan-
tom cause ‘hard’ reflections, which can lead to a
‘blurred’ effect on the ICUS images causing inac-
curate measurements. Other researchers suggested
using soft reflectors . However, many of these
materials are more temperature dependent, which
can cause shrinkage or enlargement when put in a
30 MHz mechanical element catheters from BSC
connected to BSC Clearview?ultrasound consoles
underestimate true dimensions. This can be solved
prospectively within QCU software as well retro-
spectively with a simple mathematical model on
already analyzed ICUS image data sets.
1. Mintz GS, Douek P, Pichard AD, et al. Target lesion cal-
cification in coronary artery disease: an intravascular
ultrasound study. J Am Coll Cardiol 1992; 20(5): 1149–
2. Fitzgerald PJ, St. Goar FG, Connolly AJ, et al. Intravas-
cular ultrasound imaging of coronary arteries. Is three lay-
ers the norm? Circulation 1992; 86(1): 154–158.
3. Nissen SE, Yock P. Intravascular ultrasound: novel patho-
physiological insights and current clinical applications.
Circulation 2001; 103(4): 604–616.
4. Sousa JE, Costa MA, Abizaid AC, et al. Sustained sup-
pression of neointimal proliferation by sirolimus-eluting
stents: one-year angiographic and intravascular ultrasound
follow-up. Circulation 2001; 104(17): 2007–2011.
5. Hamers R, Bruining N, Knook M, Sabate M, Roelandt
JRTC. A novel approach to quantitative analysis of intra-
vascular ultrasound images. In: Computers in Cardiology.
Rotterdam: IEEE Computer Society Press, 2001; 589–592.
6. Roelandt JR, di Mario C, Pandian NG, et al. Three-
dimensional reconstruction of intracoronary ultrasound
images. Rationale, approaches, problems, and directions.
Circulation 1994; 90(2): 1044–1055.
7. Nissen SE, Grines CL, Gurley JC, et al. Application of
a new phased-array ultrasound imaging catheter in the
assessment of vascular dimensions. In vivo comparison to
cineangiography. Circulation 1990; 81(2): 660–666.
8. Martin K, Spinks D. Measurement of the speed of sound in
ethanol/water mixtures. Ultrasound Med Biol 2001; 27(2):
9. von Birgelen C, Mintz GS, de Feyter PJ, et al. Recon-
struction and quantification with three-dimensional intra-
coronary ultrasound. An update on techniques, challenges,
and future directions. Eur Heart J 1997; 18(7): 1056–1067.
10. Bruining N, von Birgelen C, de Feyter PJ, et al. ECG-gated
versus nongated three-dimensional intracoronary ultra-
sound analysis: implications for volumetric measurements.
Cathet Cardiovasc Diagn 1998; 43(3): 254–260.
11. Thomas JD. The DICOM image formatting standard: its Download full-text
role in echocardiography and angiography. Int J Card
Imaging 1998; 14(Suppl 1): 1–6.
12. Bruining N, Sabate M, de Feyter PJ, et al. Quantitative
measurements of in-stent restenosis: a comparison between
quantitative coronary ultrasound and quantitative coronary
13. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement.
Lancet 1986; 1(8476): 307–310.
14. Stahr P, Rupprecht HJ, Voigtlander T, et al. Importance of
calibration for diameter and area determination by intra-
vascular ultrasound. Int J Card Imaging 1996; 12(4): 221–
15. Fort S, Freeman NA, Johnston P, Cohen EA, Foster FS. In
vitro and in vivo comparison of three different intravascular
ultrasound catheter designs. Catheter Cardiovasc Interv
2001; 52(3): 382–392.
16. Dussaillant GR, Mintz GS, Pichard AD, et al. Small stent
size and intimal hyperplasia contribute to restenosis: a
volumetric intravascular ultrasound analysis. J Am Coll
Cardiol 1995; 26(3): 720–724.
17. Buchwald AB, Werner GS, Moller K, Unterberg C.
Expansion of Wiktor stents by oversizing versus high-pres-
sure dilatation: a randomized, intracoronary ultrasound-
controlled study. Am Heart J 1997; 133(2): 190–196.
18. Okabe T, Asakura Y, Ishikawa S, Asakura K, Mitamura H,
Ogawa S. Determining appropriate small vessels for stenting
by intravascular ultrasound. J Invasive Cardiol 2000;
19. Briguori C, Tobis J, Nishida T, et al. Discrepancy between
angiography and intravascular ultrasound when analysing
small coronary arteries. Eur Heart J 2002; 23(3): 247–254.
20. von Birgelen C, Di Mario C, Li W, et al. Morphometric
analysis in three-dimensional intracoronary ultrasound: an
in vitro and in vivo study performed with a novel system for
the contour detection of lumen and plaque. Am Heart J
1996; 132(3): 516–527.
Address for correspondence: N. Bruining, Cardiology/H-553, Dr
Molewaterplein 40, 3015 GD Rotterdam, The Netherlands
Tel: +31-10-4633934; Fax: +31-10-4634444