The principal finding of this study is that, after deployment of BES in coronary arteries of real-world patients, the stent length measured by IVUS is significantly different from the M-L. Both measurements of stent length represented by the longitudinal distance between the proximal and the distal stent edges (E-E length) and stent borders (A-A length) are significantly longer than the manufacturer-stated length. The reason for this may be the elongation of the stent during high-pressure balloon inflation. It was seen after stent deployment under high pressure, and consequently, stent elongation would occur [11]. Of note, 88 stents (92.6%) were deployed under high pressure (14 atm or more) in our trial. In addition to manufacturer-stated length, other factors were important in determining longer stent lengths such as lesion length, deployment of the stent in the left anterior descending vessel, and vessel size. More recently, the interaction between the stent and the vessel wall has been illustrated as a complex interaction that presumably leads to stent elongation. The stent length has been determined by the plaque composition and eccentricity at the lesion site [12]. In contrary to BES, serial IVUS examination of self-expandable stents revealed a significant stent shortening during long-term follow-up [13], which might need stent oversizing at the time of implantation [14].
There may be a concern regarding the withdrawal speed of the IVUS catheter during examination especially in long lesion, tortuous vessel, small and diffusely diseased vessel. These factors may interfere with the automatic pullback speed rendering it more slowly, and consequently, the stent length will be longer. However, in our study, the median lesion length was short, 9.09 mm, most of the lesion types were not complex [types A and B; 87 lesions (91.6%)]; the mean reference vessel was not small, 3.13 mm; the smallest stent diameter, 2.5 mm, was used in only 14 lesions (14.74%); and the long stents, ≥ 20 mm, were used in 32 lesions (33.68%); thus, all these factors made a smooth non-interrupted IVUS catheter withdrawal. Of note, the stent length was longer in 79 stents (83.2%) of the cases. Moreover, another concern regarding the longitudinal movement of the IVUS catheter with each heartbeat is IVUS catheter positions which may change during cardiac cycle. Theoretically, the problem of longitudinal movement can be minimized with EKG gating and measurements made only at end-diastole, however, Kaple et al. [15] found similar results on measuring stent length by using ECG-gated IVUS with standard greyscale IVUS. Also, it was not superior to the standard IVUS.
The axial resolution of IVUS frame is important to identify stent struts and hence stent length in this study. The axial and lateral resolutions depend on the frequency of the ultrasound beam. In our study, the frequency of the IVUS catheter is 40 MHz which gives a satisfactory stent imaging. The commercially available IVUS catheters use 20–40 MHz providing 70–200 μm and 200–400 μm lateral resolution and 5–10 mm imaging depth. Optical coherence tomography (OCT) uses back-scattered infrared light to generate high spatial resolution, 10–30 μm, but shallow penetration depth. Ultra-high frequency IVUS at 80 MHz gives higher axial resolution thus improving stent visualization and a comparable penetration depth to OCT, 2 mm. New IVUS catheter design, multi-frequency IVUS, can use both low and high frequencies (35/90 MHz, 35/120 MHz, and 35/150 MHz) in order to get both advantages of high penetration and high axial resolution [16].
The longitudinal stent architecture is not always straight as in vitro analysis, but it may be curved and even angulated in either end-to-end or side-to-side directions depending on the affected vessel, the lesion site, and the lesion characters. Also, the position of the IVUS catheter and the ultrasound beam may not be perpendicular to the short axis of the stent struts, i.e., the IVUS catheter may not pass through the center or may change with the cardiac cycle. To overcome these problems, the stent length was measured in two ways. The start point and the end point were identified at two levels: (1) the start/end of viewing of stent strut, i.e., the E-E frame, where any strut could be seen at even one quadrant was considered the start/end of E-E length; (2) the frame where stent struts could be seen at two or more IVUS quadrants, i.e., A-A frame, was considered the start/end of A-A length. Thus, not only one length was used to measure stent length but two measurements. The stent edges would lie definitely at either E-E frame or A-A frame. Interestingly, both E-E stent length and A-A stent length were longer than the manufacturer stent length, M-L length. Moreover, the same method was used before in Dvir’s trial in 2014 [9]. Of note, the E-E length was significantly longer than the A-A length; consequently, the distance between the A-A length and the E-E length added a significant length to the stent. Although that distance might have a concern regarding the angel between the IVUS catheter and the stent struts (i.e., the ultrasound beam was not perpendicular to the stent), it was functionally covered part of the vessel and should not be left. Nevertheless, the A-A length was significantly longer than both manufacturer-stated lengths, the label stent length and the calculated foreshortened stent length too. Thus, whatever the method used to measure the stent length, it was significantly longer than the manufacturer-stated length data. Tanaka et al. [17] measured the stent length by using a similar pullback device (CVIS) in 45 patients. In contrary to our findings, the length was significantly shorter. However, the stent length was measured between the first complete circumferential appearance of the stent struts and the disappearance of complete circumferential visualization of the stent struts. The complete circumferential appearance of stent struts means the struts should be seen in the four quadrants of IVUS image while the measurements of our study were between the stent struts seen in one quadrant (E-E length) and in two or more quadrants (A-A length), i.e., not arbitrary seen in the four quadrants. We frequently found the complete circumferential stent struts lied proximal and distal to the first and the last stent struts at the distal and the proximal stent edges, respectively. This made the measurements shorter because it excluded part of the stent seen in two and three IVUS quadrants at both edges. Moreover, we do not know exactly the shape of the circumference of the stent at both stent edges and whether or not it is perpendicular to the vessel wall. Also, we do not know the plane of the ultrasound beam in relation to both the vessel wall and the stent. Besides, the patients’ number in our study was much more than the previous study, 90 patients versus 45 patients. Nevertheless, the correlations between the known M-L and the stent lengths measured in this study were r = 0.89 for the E-E length (p < 0.001) and r = 0.91 (p < 0.001) for the A-A length. Tanaka et al. obtained nearly a similar correlation, r = 0.92.
Recently, stent longitudinal integrity has got a lot of interest [3, 5]. The term LSD was used to describe some problems affecting it. Stent elongation was considered as one form of the LSD. Although the stent length in this study was longer than the known stent length, LSD was not considered because the incidence of LSD was generally uncommon, 1.1% [4, 18], and the incidence of stent elongation was rare too, 0.19% [3]. It was caused mainly by mechanical factors as occurring during retrieval of a protection wire [19].
In addition to M-L, three important factors were found to determine the stent length: the lesion length, the affected vessel, and the minimal vessel diameter. All these factors correlated directly with the difference in stent length except the minimal vessel diameter that had an inverse correlation. These factors may explain the difference in the stent length in nearly 60% of cases. Other factors such as balloon elongation may be present.
IVUS examination remains a crucial tool for studying the longitudinal stent architecture so that the researchers compare the stent length by using the recent investigative tools like optical coherence tomography with IVUS measurements [20]. Also, the stent length proved to affect the outcomes of percutaneous coronary intervention. It was associated with increased major adverse cardiac events (a composite of death, myocardial infarction, and target vessel revascularization). On the other hand, the use of IVUS during coronary stenting with long stents reduced the risk of these events significantly (hazard ratio, 0.47 and 0.57 for stent lengths measured from 23 to 32 mm and more than 32 mm, respectively) [21].
Study limitations
Several stent types were included in this study. Third-generation DES was not included in the study. The number of used stents with 4.0 mm stent diameter was only four stents that rendered the comparison between the M-L and E-E length insignificant.