nar
app
Ta
Conclusions: Low intra and interobserver variability can be expected when analyzing OCT data for stent apposition and tissue coverage. This
supports the validity of OCT as a clinical and research tool in the setting of intracoronary stent imaging.
using a 1310 nm near-infrared light source that detects cially available OCT system which uses a proprietary
software for offline analysis of acquired images. We sought
to assess the inter- and intra-observer variability of the
current OCT system for stent strut apposition immediately
International Journal of Cardiology
IJCA-11494; No of Pages 6
ARTICLE IN PRESS⁎ Corresponding author. Department of Cardiology, Royal Brompton© 2008 Published by Elsevier Ireland Ltd.
Keywords: Optical coherence tomography; Strut apposition; Tissue coverage; Reproducibility
Optical coherence tomography (OCT) is rapidly achiev-
ing widespread diffusion in catheterisation laboratories
worldwide. Its advantage over intravascular ultrasound
(IVUS) is its superior resolution (15 μm) and fewer
artifacts. The current technology permits image acquisition
backscattering of light from superficial microstructures
within the coronary wall. OCT has been used as an intra-
coronary imaging tool for the assessment of stent
apposition, neointimal tissue coverage and plaque char-
acterization [1–10]. To date, there is only one commer-Received 29 July 2008; accepted 27 November 2008
Abstract
Background: The introduction of optical coherence tomography (OCT) as an intracoronary imaging modality has allowed accurate
assessment of strut apposition and neointimal tissue coverage. This study set out to assess the inter and intraobserver variability of
measurements of acute stent apposition and strut tissue coverage using OCT.
Methods: Thirty patients were studied (14 immediately after stent implantation and 16 during follow-up angiography [mean of 4.7±
2.8 months]) using OCT (LightLab, Westford, Massachusetts, US). Data analysis was performed by 2 experienced observers. Struts were
classified as “embedded", “protruding" or “malapposed" to the vessel well and recorded as percentage of total struts. Intimal coverage at
follow-up was measured as the thickness of tissue covering each strut expressed in μm. Intra and interobserver variability was assessed by
Bland-Altman plots and by calculation of the intraclass correlation coefficient (ICC).
Results: An average of 3967 struts was examined by each observer and, overall, 53.7% of struts was embedded, 36.4% protruding and 9.9%
malapposed. Low intraobserver variability for all measures of strut apposition was found, with repeatability coefficients that ranged between
5.1% and 9.3% and ICC exceeding 95% in all cases. Interobserver variability was also low (repeatability coefficients 6.6–10.8 and
ICCN91.3%). Mean intimal thickness in the follow-up group was 172.5 µm. Bland-Altman plots demonstrated a low intraobserver and
interobserver variability for intimal thickness, with repeatability coefficients 26.7 μm and 24.1 μm, respectively and ICC exceeding 98.6%
for both.Francesca Del Furia, Carlo Di Mario
Department of Cardiology, Royal Brompton Hospital, London, United KingdomQuantitative analysis of intracoro
measurements of stent strut
Peter Barlis ⁎, Konstantinos Dimopoulos, JunHospital, Sydney Street, London SW3 6NP, United Kingdom. Tel.: +44 20
7352 8616; fax: +61 3 9470 8311.
E-mail address: peter.barlis@gmail.com (P. Barlis).
0167-5273/$ - see front matter © 2008 Published by Elsevier Ireland Ltd.
doi:10.1016/j.ijcard.2008.11.204
Please cite this article as: Barlis P, et al, Quantitative analysis of intracoronary o
tissue coverage, Int J Cardiol (2009), doi:10.1016/j.ijcard.2008.11.204y optical coherence tomography
osition and tissue coverage
nigawa, Ewa Dzielicka, Giuseppe Ferrante,
xx (2009) xxx–xxx
www.elsevier.com/locate/ijcardfollowing stent deployment and for tissue coverage at
follow-up.
ptical coherence tomography measurements of stent strut apposition and

1. Methods
A total of 30 patients were prospectively included in this
study. OCTwas performed immediately after stent deployment
in 14 patients and on follow-up angiography in a separate
group of 16 patients, using the LightLab system (Westford,
Massachusetts, US). Blood clearance was achieved using a
proximal occlusion balloon (Helios, Goodman, Japan) with
intra-coronary flush of lactated Ringer's solution through the
end-hole of the balloon catheter (flow rate 0.6–0.9 ml/s)
during simultaneous image acquisition at 1.0 mm/s using the
dedicated ImageWire (LightLab Imaging, Westford, Massa-
chusetts, US) at 15.6 frames/second.
1.1. OCT data analysis
Offline OCT images were analyzed by 2 experienced
observers. For intraobserver variability, measurements were
repeated 3 months after the initial assessment by the same
observer, who was blinded to the previous results. For inter-
observer variability, measurements were carried out inde-
pendently by the two observers, blinded to each other's
analysis.
1.2. Acute stent strut apposition
cross section, the distance between the endoluminal border
of the strut and the intima was measured. As OCT can only
image the endoluminal strut border, distances were adjusted
based on the thickness of the strut, including the polymer, for
each of the stent types used [9]. Struts were classified as
“embedded” when buried into the vessel wall, “protruding”
when in contact with the vessel wall but protruding into the
lumen and “malapposed” when no contact between the
intima and the strut was detected (Fig. 1).
1.3. Strut tissue coverage
Tissue coverage was assessed by OCT at a minimum of
3 months post-implantation of different stents types. The
thickness of the tissue overlying each stent strut wasmeasured.
1.4. Statistical analysis
Statistical analyses and generation of plots were per-
formed using R version 2.6.1 (The R Foundation for
Statistical Computing, Vienna, Austria). To assess intra and
inter-observer agreement, Bland-Altman plots were produced
and mean difference and 95% limits of agreement with 95%
confidence intervals were calculated. The repeatability
coefficient (within which 95% of all differences are included)
was calculated as twice the standard deviation of the
2 P. Barlis et al. / International Journal of Cardiology xx (2009) xxx–xxx
ARTICLE IN PRESSAnalysis of contiguous cross-sections within the stented
segment was performed at 1 mm intervals. In each selectedFig. 1. Optical coherence tomography classification of stent
Please cite this article as: Barlis P, et al, Quantitative analysis of intracoronary o
tissue coverage, Int J Cardiol (2009), doi:10.1016/j.ijcard.2008.11.204differences between measurements on the same segment, as
described by Bland and Altman [11]. Intra-class correlationapposition into embedded, protruding or malapposed.
ptical coherence tomography measurements of stent strut apposition and