Characterization of Late Incomplete Stent Apposition: A
Comparison among Bare-Metal Stents, Intracoronary Radiation
and Siroli

Recent studies have shown that late incomplete stent apposition (LISA) develops following implantation of conventional bare-metal stents (BMS), drug-eluting stents (DES) or adjunctive intracoronary radiation (IR).^1–6 While the clinical relevance of LISA remains somewhat controversial, several reports have suggested that LISA may be associated with adverse cardiac events, including stent thrombosis.^4,7,8 Intravascular ultrasound (IVUS) studies have demonstrated that morphological features of LISA range from subtle detachment of the stent strut(s) from the vessel wall, to prominent aneurysmal changes at the stented segment. In the present study, the morphometric features of LISA accompanying BMS, IR or sirolimus-eluting stents (SES) were compared using serial IVUS images, thereby addressing the underlying mechanisms of LISA that develop with various treatment modalities.

Methods
Patient population. Patients who met the following criteria were selected from the IVUS database of the Cardiovascular Core Analysis Laboratory at Stanford University: (1) implantation of either BMS, SES or IR; (2) native coronary artery lesion; (3) availability of high-quality serial IVUS images at baseline (post intervention) and follow up (mandated 6–8 months after procedure); and (4) confirmation of LISA through comparative serial IVUS interrogation. In all patients, the IVUS follow up was prospectively scheduled at 6 to 8 months, regardless of symptoms, as part of the clinical research protocols. Written, informed consent was obtained from all patients prior to the procedure.

IVUS analysis. All IVUS images were acquired and analyzed using a dedicated software (EchoPlaque, Indec System, Santa Clara, California) as previously described.⁶ ISA was defined as separation of at least 1 stent strut from the intimal surface, with evidence of blood flow behind the stent strut(s). LISA was defined as ISA identified at follow up, but not at baseline. To specify LISA location, the IVUS image of the stent was longitudinally divided into 3 segments: proximal edge (within 5 mm from proximal edge), body and distal edge (within 5 mm from distal edge). For quantitative analyses, lumen area (LA), stent area (SA) and vessel area (VA), using a single still image at the site of greatest stent-lumen separation at follow up, were measured. Plaque area (PA) was defined as VA minus LA. A matched corresponding still image at baseline was selected for comparison based on the length from stent edge(s) or perivascular landmarks.
Both baseline and follow-up still images were then exported to custom-designed software (VAnQuISH, Stanford, California).⁹ With the follow-up LISA image, “LISA arc” and “apposed arc” were defined based on the location of the LISA (Figure 1). Gap area (GA) was calculated by subtracting SAfrom the LA within the LISA arc. LISA and apposed arcs were then reflected to the corresponding baseline images to obtain: (1) LA, SA, VA and PA of the LISA arc; (2) plaque thickness at the middle of each arc; (3) mean plaque thickness of each arc; and (4) wall thickness (WT) ratio: (WT on LISA arc)/(WT on apposed arc). Vessel area change (VA = CA at follow up minus VA at baseline), plaque area change (PA at follow up minus PA at baseline), and contribution of vessel or plaque area change to GA (VA or PA/GA x 100[%]) were also calculated.
Statistical analysis. Data are presented as mean ± standard deviation (SD). Using StatView 5.0 (SAS Institute, Cary, North Carolina), baseline and follow-up data were compared by the paired t-test, and those among BMS, IR or SES were compared by Bonferroni’s correction test. IVUS measurements were also compared between the BMS-alone group and the BMS with prestent directional coronary atherectomy (DCA) group using the unpaired t-test (GA, wall thickness) or two-way analysis of variance (area change at the LISA arc). Categorical data were compared using Fisher’s exact probability test. A p-value < 0.05 was considered statistically significant.

Results
Among a total of 881 cases reviewed (552 BMS, 93 IR, 236 SES), 30 LISA were identified (12 BMS, 6 IR and 12 SES). The incidence of LISA was 2.2% (12/552), 6.5% (6/93) and 5.1% (12/236) for BMS, IR and SES, respectively. In 8 of 12 BMS LISA cases, patients were treated with DCA prior to stenting. In the 6 IR cases, beta-radiotherapy was used in 5 patients, and gamma-radiotherapy was used in 1.

LISA location did not differ between the three groups (Table 1). Compared to baseline, the LA, VA and PA of IR, and the LA and VA of SES were significantly increased at follow up. The LA, SA and VA of the SES were smaller than those of the BMS at both baseline and follow up.
Analysis of the LISA arc showed that LA at follow up increased significantly in all three groups compared to baseline (Table 2). No difference was observed in GA among groups. VA of the IR and SES significantly increased at follow up, while there was no significant change in PA. In contrast, the BMS group showed no increase in VA, whereas a significant reduction in PA was seen. These area changes and GA were not significantly different between patients treated with BMSalone and those treated with BMS and prestent DCA.

Wall thickness at the apposed arc was thinner in BMS as compared to SES and IR (Table 3). In contrast, that of the LISA arc was thinner for SES as compared to BMS. Similarly, the WT ratio of SES was smaller than that of BMS. The mid-WT ratio of IR was also smaller than BMS. These findings were consistently observed irrespective of prestent DCA, and the measurements were not significantly different between BMS patients who underwent prestent DCA and those who did not.
The contribution of VA change to the GA was significantly greater in SES than in BMS, whereas that of PA change was conversely greater for BMS compared to SES or IR (Figure 2).

Discussion
The present study demonstrates different morphometric features of LISA among patients receiving BMS, IR and SES. For IR and SES cases, vessel enlargement was the predominant factor for LISA, whereas plaque regression mainly contributed to LISA in BMS cases. Several IVUS studies have documented the morphology of LISA;^1,5,10,11 however, the present study is the first systematic one comparing serial IVUS images of LISA accompanying BMS, IR or SES.
Vessel enlargement observed with SES or IR may reflect pathological processes characteristic to each treatment. Regarding SES, their implantation has been reported to induce poor endothelialization and/or inflammatory processes,with or without eosinophilic infiltration.^12,13 Such pathological processes may translate to vessel enlargement and resultant LISA. In fact, persistent inflammation has been documented at the resected coronary wall of SES-induced coronary aneurysms, an extensive phenotype of LISA.^14,15 In addition, IR reportedly causes radiation-induced apoptosis and activation of proteinase such as matrix metalloproteinase,^2,16,17 potentially triggering vessel enlargement.
In BMS cases, by contrast, PA decrease predominantly contributed to LISA formation. One possible explanation is the absorption of persistent thrombus formed at the time of intervention. In fact, 8 of 12 patients in the BMS group underwent DCA before stenting, and such debulking procedures are known to induce thrombus formation.¹⁸ Another possibility is the net regression of overall plaque volume, however, clinical characteristics including changes in lipid profile or the use of statins were not available in the present study.
In the matched post-procedure images, the vessel wall of the LISA arc was thinner in SES compared to BMS. This is in agreement with our previous report, demonstrating that LISA with SES was mainly observed on the relatively diseasefree side of the vessel wall.⁶ Since SES intensely inhibit neointimal growth, delaying early reparative processes may cause greater vessel enlargement at sites along the thin vessel wall. Also, the sirolimus and/or drug-eluting polymer may impede the physiological function of the vasa vasorum. Since the vasa vasorum provides the vessel wall with nutrients,¹⁹ undesirable effect(s) of drug/polymer on the vasa vasorum may translate to vessel expansion.
LISA associated with SES may share a common pathogenesis with late stent thrombosis (LST), a critical adverse event that develops with SES or other DES.^20,21 This possibility is supported by the following observations: (1) a recent autopsy study reported ISA as one of the pathological risk factors for DES-associated late thrombosis;12 (2) some of the pathological features of LST, such as poor endothelialization and persistent vessel wall inflammation, are similarly observed in distinct cases with LISA;^14,15 and (3) several reports have documented patients with SES who developed LST together with or following LISA.^4,7,8,22 It is still controversial whether LISA triggers any adverse clinical event including LST; however, the possible clinical impact of LISA, particularly cases with vessel enlargement, needs to be further addressed.

Study Limitations
There are several limitations in this study. The present study is based on a retrospective analysis of a relatively small number of patients, raising the possibility of patient selection bias. Only SES cases were included, therefore, these results may not be applicable to other DES. Also, a significant portion of BMS patients underwent prestent atherectomy, therefore, these results need to be interpreted cautiously. Studies including a larger number of patients without prestent DCA may be necessary to further address the mechanisms of LISA accompanying BMS. Lastly, preprocedural images were not available, therefore, the assessment of pretreatment plaque characteristics was precluded.

Conclusion
The present IVUS study showed morphometrical diversity of LISA depending on the interventional treatment for occlusive coronary artery disease. Plaque reduction mainly contributes to LISA after BMS implantation, whereas vessel expansion predominates in IR and SES implantations. These findings suggest different underlying mechanisms of LISA between BMS and newer-generation treatments such as IR and SES implantation.