Elective percutaneous coronary interventions (PCI) on the unprotected left main (ULM) are increasing in popularity thanks to improved stent design, optimization of PCI techniques and increasing use of intravascular imaging (IVI), and have long-term mortality comparable to surgical revascularization [1]. On the other hand, Urgent PCI of the culprit lesion in the ULM segment still has high in-hospital mortality rates despite immediate reperfusion, mainly due to a large amount of myocardium at risk [2, 3]. Furthermore, the long-term outcomes of these patients after hospital discharge are still not comparable to those of patients undergoing elective ULM PCI, with higher rates of restenosis and thrombosis [4, 5]. To ensure favourable long-term outcomes after drug-eluting stent (DES) implantation, procedural IVI guidance is recommended to optimize stent implantation, particularly stent expansion and good vessel wall opposition, thereby promoting endothelialization and long-term device patency [6]. Angiographically undetectable stent malapposition occurs more frequently in large coronary anatomies, e.g., the left main bifurcation segment, and after PCI with complex two-stent techniques [7, 8, 9, 10, 11]. Despite its recognized benefit, reported rates of imaging during primary PCI-ULM procedures are low, particularly due to the clinical circumstances with impending circulatory collapse that mandate prompt stent implantation and rapid restoration of flow [3]. Consequently, reports on short- and long-term IVI evaluations of angio-guided ULM stenting during primary PCI are lacking. Therefore, we performed long-term optical coherence tomography (OCT) to investigate long-term DES healing in patients with uneventful clinical follow-up, who underwent angio-guided ULM primary PCI.

The present analysis provides OCT insights into the late imaging results of DES implantation during the primary PCI of culprit ULM bifurcation lesions. The key findings are: (1) Regarding the primary endpoint, the strut coverage is not optimal in proximal bifurcation segments; (2) Although contemporary stent optimization procedures were performed, late malapposition and stent under-expansion were common because of the disproportionately larger proximal bifurcation segments compared to the DES used; (3) Strut malapposition distance affects strut endothelization during long-term follow-up; (4) Invasive imaging revealed a high incidence of subclinical de-novo atherosclerosis and restenotic intimal hyperplasia, necessitating late optimization and additional treatment.

PCI of bifurcation lesions is more likely to result in death, myocardial infarction, and repeat revascularization compared to non-bifurcation PCI [17]. During bifurcation PCI, complex stent implantation techniques are often required, in which the stent must be specifically adapted to the underlying anatomy [18, 19]. Early pathological post-mortem analyses indicated that bifurcation stenting is a significant risk factor for ST due to the presence of uncovered and malapposed struts [20]. However, these analyses didn’t identify the extent or pattern of strut coverage for stents deployed at the bifurcation. In our patient population, we demonstrated a persistent suboptimal level of strut coverage in left main bifurcation lesions, particularly in large proximal portions (78.9 and 79.8% for LM and POC, respectively).

Our findings are in line with previous reports that demonstrated, albeit at earlier time points, higher rates of incomplete endothelialization, also limited to the proximal bifurcation segments [10, 12, 21]. In the largest reported serial OCT analysis of 33 patients undergoing elective ULM PCI with DES, Fujino et al. [10] reported relatively high rates of uncovered and malapposed struts (18.7% and 5.3%, respectively) in the LM region at 9 months. Although, the percentage of malapposed struts decreased in follow-up compared to baseline (5.3% vs. 13.9%; p 0.001), it still demonstrates that in larger bifurcation segments, acute, and persistent malapposition represents a frequent and clinically important finding, which impacts strut coverage and proper device endothelization. Our long-term analysis with observed rates of malapposed (11.7%) and uncovered struts (9.7%) in the proximal LM segment, confirms that these adverse findings persist beyond the 1-year period, which highlights the importance of prolonged duration of a dual antiplatelet regimen following ULM PCI in acute coronary syndrome (ACS). Since prior OCT analyses of patients with late ST showed a significant amount of uncovered (46%) and malapposed struts (39–42%), patients with suboptimal DES healing may be more susceptible to late ST and other adverse cardiovascular events [22, 23, 24].

If acute stent malapposition is not corrected, it decreases over time as a result of neointimal proliferation, and the distance of the strut to the vessel wall [25]. Serial OCT evaluations showed that endothelialization progresses rapidly in the early period after DES implantation reaching more than 90% at three months [26]. In our later analysis (average 1580 days), the time interval to follow-up OCT did not correlate with the percentage of struts covered (p = 0.343). Since this was not an imaging analysis with serial assessments of the same devices but only a very late observational evaluation, we could not investigate the relationship between strut coverage and the post-PCI time interval.

In our analysis, significantly malapposed struts $>$400 µm were less likely to be covered by proliferating intimal hyperplasia compared to 400 µm malapposed struts (15.4 $\pm$ 21.6 vs. 24.8 $\pm$ 23.9%; p = 0.011). Previous reports have shown that stent endothelialization is significantly influenced by the distance of the malapposed struts from the vessel wall: the greater the distance of acute malapposition, the greater the likelihood of persistent malapposition during follow-up and delayed healing [27]. In contrast, struts closer to the vessel wall, although still malapposed, may undergo endothelialization to a greater extent over time (Fig. 4) [28, 29]. Therefore, the risk of late ST is significantly increased in struts with severe malposition ($\ge$400 µm) because they often have incomplete, delayed or absent endothelialization [30]. Therefore, to promote long-term healing and reduce the risk of late complications, our results, along with prior reports, emphasize the importance of achieving optimal acute stent apposition, or the smallest achievable malapposition distance, during the initial procedure. Notably, neither on OCT nor during follow-up, was there any evidence of recent subclinical thrombi formation or clinically apparent ST, even though larger portions of the ULM had a high rate of uncovered and malpositioned stent struts. Two patients showed evidence of organized thrombi on overhanging, uncovered struts (one case with large metallic neocarina after TAP stent implantation). However, our small sample size precludes drawing definitive conclusions about this clinically important but relatively rare adverse event.

All baseline primary PCI procedures were performed by experienced operators in a high-volume primary PCI center (around 800 primary PCIs and 100 ULM PCIs annually during the study period). POT was performed with angiographically, 1:1 sized, non-compliant or semi-compliant balloons at high pressures. However, high rates of late malapposition were still noted in the LM and POC regions due to the disproportionately larger proximal reference size, highlighting the importance of balloon sizing according to IVI measurements and achieving a larger stent area, in contrast to the results of angiographic balloon sizing reported in this study. We found that larger reference area to mean stent area mismatch, was associated with the percentage of total covered struts (R² = 0.268, p 0.001), significantly malapposed $>$400 µm (R² = 0.189, p = 0.028), and significantly malapposed $>$400 µm and uncovered struts (R² = 0.176, p = 0.005), across bifurcation segments. The importance of IVI device sizing, as suggested by consensus documents, was reported in the clinical setting in the prospective LEMON study [31]. OCT-guided PCIs of ULM lesions were found to be feasible and provided overall positive study results, with stent expansion reaching 94% for the proximal LM segment (compared to 63% in our population) and a low significant malapposition rate in 18% of cases. However, whether OCT guidance can improve the acute and long-term outcomes of ULM PCI compared with angio-guided treatment remains to be elucidated. Currently, the advantage of OCT-guided bifurcation PCI was recently demonstrated in the randomized, outcomes-oriented OCTOBER trial, which demonstrated a lower incidence of adverse events in the OCT-guided group at 2 years [32]. Patients undergoing PCI for a ULM lesion represented a significant proportion of the population studied and contributed to the overall positive outcomes, although subgroup analysis showed no beneficial effect for this type of lesion.

Finally, five patients (33.3%) underwent ad hoc OCT-guided ULM reintervention due to subclinical DES failure. According to the collected data, DES is effective in preventing early and mid-term restenosis, but are unable to absolutely prevent proliferative neointimal hyperplasia and the development of neo-atherosclerosis, especially if not properly expanded or adequately sized [33, 34]. Specifically, for the ULM PCI, Kang et al. [34] showed that stent underexpansion detected by IVI, results in lower event-free survival at 9 months. The minimal stent area (MSA) cutoffs that best predicted restenosis was 5.0 mm² for ostial left circumflex, 6.3 mm² for ostial left anterior descending (LAD), 7.2 mm² for the POC, and 8.2 mm² for the LM segment. Although measured MSA in our population was above the specified criteria (10.3 mm² for LM and 7.1 mm² for LAD), neoatheroslerosis was detected in 33.3% of cases at long-term follow-up. According to the published data, neoatherosclerosis is one of the leading underlying reasons for very-late ST, reported to be present in 22–31% of late events [24, 35]. It is well known that coronary bifurcation lesions have a higher risk of developing in-stent restenosis than non-bifurcation lesions [17]. This is partially due to distinct flow patterns and areas of low shear stress [36]. Finally, there is insufficient evidence regarding the optimal treatment strategy for stent failure in ULM, as surgical revascularization is usually considered due to the unique challenges associated with a failed PCI in ULM lesions [37]. Further research is necessary to confirm our findings.

There are several limitations of this study. The main limitation is the observational design and small sample size, therefore, no definitive conclusions regarding the impact of the observed OCT findings on clinical outcomes is possible. This is a pilot study, and findings should be considered hypothesis-generating and need to be investigated in future larger studies. However, this study illustrates for the first time a very late vascular response after angio-guided implantation of DES during ULM primary PCI using high-resolution imaging and unintended stent deformation, which resulted in reinterventions in 33.3% of patients. Since IVI was not performed at baseline, it is not possible to compare very late follow-up imaging findings with the results of the index procedure. In addition, incomplete blood clearance at the LM ostium impaired adequate OCT assessment of the ostial segment. The current study population included only subjects who did not experience any clinical events and TLR during the follow-up, and excluded patients without severe renal and heart failure, which reduces the generalizability of our results.

Long-term OCT evaluation of coronary stents implanted during primary PCI for ULM lesions demonstrates high rates of incomplete strut coverage, late malapposition, de novo atherosclerosis, and restenotic neointimal hyperplasia. The unfavourable OCT results of angio-guided DES implantations advocate for attention and greater implementation of image optimization strategies during primary PCI. Further research is needed to validate these results and improve the percutaneous treatment of ULM lesions during primary PCI.

DES, drug-eluting stents; DKC, double kissing crush; CABG, coronary artery bypass graft; CKF, chronic kidney failure; EF, ejection fraction; FMC, first medical contact; HF, heart failure; IVI, intravascular imaging; KBI, kissing balloon inflation; LAD, left anterior descending; LM, left main; MADS, Main, Across, Distal, Side; MB, main branch; MSA, minimal stent area; NIH, neointimal hyperplasia; Non-TLR, non-target lesion revascularization; OCT, optical coherence tomography; PCI, percutaneous coronary intervention; POC, polygon of confluence; POT, proximal optimization technique; SB, side branch; SD, standard deviation; ST, stent thrombosis; STEMI, ST-elevation myocardial infarction; TAP, T-And Protrusion; TLR, target lesion revascularization; ULM, unprotected left main.

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conceptualization, methodology, article writing ZM and GS; Project supervision – GS; ZM, DJ analysed and performed OCT analysis. ZM, VV, SS, BB, DO, MTom, MD, MTes, DMil, SA, VD, MZ, SJ, DJ, DMla, GS performed either baseline primary PCI or were involved in follow-up invasive angiography or OCT recordings. All authors were involved in drafting or in critical revision of the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The research protocol was approved by the Council of Scientific Field of Medical Sciences, of the Faculty of Medicine, University of Belgrade (61206-4744/2-21), and the Research Board of the Department of cardiology of the University Clinical Center of Serbia (review No: 1883), and all of the participants provided signed informed consent.

This research was funded and supported by: Ministry of Education, Science and Technological Development of Republic of Serbia, Research project No. 175082; Science Fund of the Republic of Serbia, grant No. 7558, SINERGY-ACUTE project; Serbian Academy of Sciences and Arts, project F-42.

The authors declare no conflict of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/j.rcm2512445.