This retrospective observational study was conducted at the Xiangtan Central Hospital from March 2015 to November 2019. A total of 1058 lesions were selected from 961 consecutive patients subjected to IVUS-guided percutaneous coronary intervention (PCI) for de novo lesions. Amongst them, 232 cases possessed 251 CTLs. The exclusion criteria were: (1) non-tapering lesions (n = 541), (2) left main coronary artery lesions (n = 35), (3) ostial lesions (n = 86), (4) chronic total occlusion (CTO) lesions (n = 59), (5) administration with drug-coated balloons (n = 47), and (6) non-satisfactory angiographic or IVUS image quality (n = 39) (Fig. 1). CTLs were defined by IVUS and were based on differences in the proximal and distal references for each lesion of $\ge$1.0 mm, or $\ge$30% [2]. This study was carried out according to the principles of the Helsinki Declaration and was approved by The Ethical Board of Xiangtan Central Hospital. Written informed consent was obtained from patients prior to the study.

Fig. 1.

Schematic of the study flow. PCI, percutaneous coronary intervention; IVUS, intravascular ultrasound; CTLs, coronary tapering lesions; DoCEs, device-oriented clinical endpoints.

Procedure-associated strategies were decided upon by the operator. Second-generation drug-eluting stents were used in all cases. Preintervention IVUS was employed to analyze CTLs prior to balloon dilatation. Once stenting was complete, another IVUS was carried out to verify the results for stent deployment. For ineffective cases with MSA 4.5 mm${}^2$, we performed stent improvement using non-compliance balloons and guidance with IVUS and angiography until an acceptable result was achieved, as determined by the final IVUS and angiogram. All patients continued to receive dual antiplatelet therapy (DAPT) for at least 6 months.

An offline, commercially available software (QAngio® XA, Medis, Leiden, the Netherlands) was employed for quantitative coronary angiography (QCA) of CTLs. QCA analysis included the minimal lumen diameter, percent diameter stenosis, lesion length, reference vessel diameter, calcification, etc. [14]. The three epicardial arteries were divided into left main (LM) (5), distal (3, 4, 8–10, 12, 14, 15), mid (2, 7, 13), and proximal (1, 6, 11) segments, in accordance with the American Heart Association classification [15].

When nitroglycerin (0.1–0.2 mg) was used for intracoronary administration, automated pullback (0.5 mm/s) was employed to obtain the CTL IVUS images (40 MHz OptiCross™, Boston Scientific, Marlborough, MA, USA) for both before and after PCI. Two independent readers who were blinded to patient information evaluated all IVUS images using a frequency domain available offline software (QIvus®, Medis, Leiden, the Netherlands). The CTL IVUS measurements were performed every 1 mm for the administered segment (pre-PCI) and stent, and every 5 mm for the proximal and distal reference segments. We examined the distal and proximal references in the site, reaching the maximal lumen 5 mm distal and proximal to the stented segment. Reference luminal areas were also examined in frames with minimal plaque burden. The calculation for each percent area of stenosis was: ((reference lumen area - minimal lumen area)/reference lumen area) $\times$ 100 pre-PCI. The respective volumes were determined in accordance with the Simpson rule [16]. The percentage plaque volume refers to the total plaque/vessel volume for the pre-procedure IVUS investigation. Pre-PCI IVUS qualitative analysis included: superficial calcium (hyperechoic region with acoustic shadow), calcified nodule (protruding and irregular calcium with intimal surface), and attenuated plaque (noncalcified plaque with echo attenuation). Post-PCI IVUS qualitative analysis included: stent edge dissection (intimal, medial, intramural hematoma, or outside the external elastic membrane (EEM)), stent malapposition (blood speckle behind stent struts not overlaying a side branch), and tissue protrusion (plaque and/or thrombi intrusion through the stent struts into the vessel lumen) on post-PCI IVUS [17].

The indices for stent expansion were specified in advance and are described below (Fig. 2):

(1) MSA was derived from the automatic minimal cross-sectional lumen area within the post-stented lesion [18].

(2) MSA/vessel area at the MSA cross-sectional [19].

(3) Traditional SEI: MSA/mean proximal and distal reference lumen cross-sectional area.

(4) Minimal volumetric expansion index (MVEI) [13]: (actual lumen area/ideal lumen area $\times$ 100) in the minimal value cross-sectional area through the stented site. The ideal lumen cross-sectional area without plaque was calculated using the mathematical relationship for proximal and distal reference cross-sectional areas and side branch diameter ($>$0.5 mm), as described by Huo et al. [20], and referred to as the H–K model. If the vessel has no intermediate side branch (diameter $>$0.5 mm), the ideal lumen diameter of the uniform tapering vessel was calculated using the linear model [13].

(5) IVUS-XPL standards, calculated by an MSA $>$100% of the distal reference lumen cross-sectional area [21].

(6) ULTIMATE standards, calculated by an MSA $>$5.0 mm${}^2$ or $>$90% of the distal referencelumen cross-sectional area [7].

(7) ILUMIEN IV standards, calculated by an MSA of the proximal site $>$90% of the proximal reference lumen cross-sectional area, and an MSA of the distal site $>$90% of the distal reference lumen cross-sectional area [22].

(8) ILUMIEN III standards, calculated by mean stent expansion: mean stent area (total of stent area/total of stent length)/mean reference lumen cross-sectional area [23].

Fig. 2.

The calculation formula for stent expansion indices. MSA, minimal stent area; IVUS-XPL, Impact of Intravascular Ultrasound Guidance on the Outcomes of Xience Prime Stents in Long Lesions; ULTIMATE, Intravascular Ultrasound Guided Drug Eluting Stents Implantation in “All-Comers” Coronary Lesions; ILUMIEN IV, Observational Study of Optical Coherence Tomography in Patients Undergoing Fractional Flow Reserve and Percutaneous Coronary Intervention IV.

Device-oriented clinical endpoints (DoCEs) included target lesion revascularization (TLR), myocardial infarction (MI) or stent thrombosis associated with the target vessel, and cardiac death [24]. Cardiac death was defined as any death due to cardiac-related causes, procedure-related deaths, and death of unknown cause. MI was reported in accordance with European Society of Cardiology guidelines [25]. TLR refers to an ischemia-driven repeat PCI, or to coronary artery bypass surgery of the target lesion for angiographic target lesion restenosis or ischemia-driven clinical complications. Stent thrombosis refers to either probable or definite stent thrombosis [24]. Periodic clinical follow-up occurred at 6-month intervals through either a telephone interview or a clinical visit. In general, recruited patients were subjected to almost 3-years of clinical follow-ups, and at least one year of follow-ups.

Continuous variables are expressed as mean $\pm$ standard deviation when meeting a normal distribution, whereas, for an abnormal distribution, they are described as the median value with the interquartile. Categorical variables are expressed as numbers (percentages). The Mann–Whitney U test or Student’s t test was employed for the analysis of continuous outcome data, while the chi-square test or Fisher’s exact test was used for the categorical variables. For lesion-specific variables, continuous and categorical variables were analyzed by generalized estimating equations (GEEs) adopted for clarifying the clustering of multiple lesions in the respective patient. For continuous variables, we employed a GEE model that was subjected to normal distributions, with the expression of the least square means (95% confidence interval). For categorical variables, we employed a GEE model that was subjected to logit link and binomial distributions. Analysis with multivariable marginal Cox proportional hazards was performed using a stepwise selection procedure to identify independent stent expansion indices related to the DoCE. Log-rank and Kaplan–Meier tests were used to compare DoCE incidences between the stent expansion indices. Receiver operating characteristic curve analysis was performed to evaluate the stent expansion indices for their ability to predict DoCEs through the use of minimal under-expansion. We used the Youden index to determine the cut-off value. Associations between CTL morphological parameters, PCI parameters, and stent expansion indicators were analyzed using multivariable linear regression. Statistical significance was defined by two-sided p values 0.05. IBM SPSS Statistics 24.0 (IBM-SPSS Statistics, Chicago, IL, USA) was used to perform statistical analyses.

A total of 232 consecutive patients with 251 CTLs were assessed. Of these, 17 patients (7.3% of all patients) with 17 lesions (6.8% of all lesions) had 2-year follow-ups for the DoCEs. The average follow-up was 729 days (interquartile range: 705–733 days). As shown in Table 1, no significant differences were observed in any of the clinical characteristics between DoCE(–) and DoCE(+) patient groups. The procedural and angiographic findings were also compared between the two groups. Again, no significant differences were observed between patients who did or did not suffer DoCEs (Table 2).

The final overall IVUS MSA after PCI was examined as 5.9 $\pm$ 1.6 mm${}^2$. Table 3 shows the results for lesions with IVUS prior to PCI. There were no significant differences in lesions at the MSA site between patients with or without DoCEs. However, the MVEI calculated by the linear model or the H–K model had significantly fewer lesions in the DoCE(+) patients compared to the DoCE(–) patients. Among the different stent expansion indices, only MVEI (hazard ratio [HR], 1.91; 95% CI 1.16–3.96; p = 0.028) was significantly associated with an increased risk of DoCEs in the multivariable analysis (Table 4). Higher balloon inflation pressures were negatively related to the volumetric expansion index (HR: 0.28; 95% CI: 0.10–0.47; p = 0.01) (Table 5). Receiver operating characteristic analysis revealed that the optimal MVEI cut-off value for predicting a DoCE was 62.2% (area under the curve [AUC]: 0.71; 95% CI: 0.65–0.77) (Fig. 3). The Kaplan–Meier analysis of the DoCE after the two-year follow-up, in relation to the MVEI is shown in Fig. 4. A significant difference was observed in the incidence rate of 2-year DoCEs between patients with MVEI 62.2% and MVEI $\ge$62.2% (10.9% vs 3.5%; p 0.011), as shown in Table 6. Summaries of representative cases for MVEI are shown in Fig. 5.

Fig. 3.

Receiver operating characteristic curve analysis. AUC, area under curve.

Fig. 4.

Two-year Kaplan–Meier curves for DoCEs. DoCEs, device-oriented clinical endpoints; CI, confidence interval; MVEI, minimal volumetric expansion index.

Fig. 5.

Representative IVUS images with different patterns of MVEI. (A) Both MSA and MVEI are large. (B) MSA is large and MVEI is small. (C) MSA is small and MVEI is large. (D) Both MSA and MVEI are small. MVEI, minimal volumetric expansion index; IVUS, intravascular ultrasound; MSA, minimal stent area.

CTLs with reference lumen cross-sectional area mismatching remain difficult to treat by interventional cardiologists, with no optimal interventional strategy confirmed as yet. The remodeling of vessels at the reference site of the target lesion is the primary treatment for CTLs [26]. Although self-expandable stents and tapered stents are employed to revascularize CTLs [27, 28], they have yet to be extensively employed. Moreover, in contrast to existing balloons and symmetrical stents, their efficacy has also yet to be confirmed. Due to the effect of the symmetrical design, and without considering strategies involving stents and balloons, PCI, in terms of tapering lesions through the application of self-expandable stents and tapered stents that are symmetrical, is subject to dissection and overstretching risks within the distal segment, or to incomplete stent apposition and thrombus formation at proximal sites of the tapering lesion. The above sub-optimal conditions for tapering lesions are likely to trigger common PCI complications, such as in-stent restenosis and stent thrombosis [3]. The present study highlights the use of the volumetric expansion index under the guidance of IVUS to assess whether tapering stent expansion is important for reducing adverse clinical outcomes. In addition, the volumetric analysis algorithm could also optimize the success of stent implantations with symmetrical devices.

Current research suggests that PCI following IVUS guidance can optimize stent expansion by providing accurate lesion assessment, pre-stenting preparation, and post-dilation improvement. MSA is capable of estimating absolute stent expansion and is known to be a critical predictor of future stent failures in IVUS and OCT research. The optimal MSA cut-off values for the prediction of adverse clinical outcomes are reported to be 4.5 to 5.5 mm${}^2$ [10, 29, 30, 31]. Data from a randomized trial with 804 patients who received a long ($\ge$28 mm length) drug-eluting stent implant used IVUS to identify an MSA of 5.0 mm${}^2$ as a threshold for predicting future clinical outcomes [9]. However, MSA is mainly determined by the reference cross-sectional lumen diameter of the target lesion, although this varies depending on the distribution of the three main epicardial coronary arteries and lesions. The concept of “bigger is better” does not apply to all lesions. For example, a lesion at the distal end of the right coronary artery (e.g., 2.5 mm diameter) cannot easily achieve an MSA $>$5.0 mm${}^2$ after a stent is implanted and with optimized post-dilation. The final overall MSA here was 5.9 $\pm$ 1.6 mm${}^2$, thereby suggesting a favorable IVUS-guided stent expansion. This is likely the reason why MSA was unable to predict further stent failures in the current cohort. The above result demonstrates that the independent use of MSA has a limited application for individual cases.

A uniform standard for comparing the minimal luminal area in the intravascular imaging-guided stent has yet to be established for the proximal or distal reference luminal area, or the mean luminal area. Meneveau et al. [31], reported that an optimal cut-off value for stent expansion $>$79.4% and a minimal luminal area $>$5.44 mm${}^2$ could predict a final fractional flow reserve (FFR) $>$0.90. In addition, the recent expert consensus document [8] recommended a value $>$80% to improve clinical outcomes, in terms of the MSA/mean reference luminal area. However, for small vessels, an MSA/mean reference luminal area of $>$80% is not feasible [32]. Furthermore, pooled data from the ADAPTDES (dual antiplatelet therapy evaluation that involves drug-eluting stents) study found that neither MSA nor traditional expansion indices affected the two-year DoCEs [19]. Instead, the IVUS-directed MSA/vessel area ratio at the MSA section was significantly associated with adverse two-year clinical outcomes when the value was 38.9%. Nakamura et al. [13] investigated the relationship between the incidence of DoCEs at 1-year and the H–K model-derived minimal index for volumetric stent expansion and post-stent FFR. Consistent with the present study, the authors reported that a volumetric analysis model that considers vessel tapering is a better predictor of final FFR and clinical events. In contrast, a different cut-off value was reported for OCT-derived volumetric parameters (74.0% in Katsura et al. [12] and 62.2% in the present cohort). One explanation may be that the rate of post-dilatation in the relevant segment of the stent was higher in our study than in Katsura et al. [12] (100% vs. 80.5% ). This may reduce the frequency of stents under expansion and increase the expansion volume of the stents. In summary, the current research indicates that vessel tapering or vessel remodeling in the volumetric stent expansion index is an important criterion for post-stent improvement and for the prediction of adverse clinical events.

The choice of a symmetrical stent for achieving the favorable conformation of a tapering lesion is very challenging in real-world practice [3]. The use of post-stenting improvements with a larger-sized balloon or a greater pressure of inflation helps to address tapering lesions but causes higher rates of stent failure [33, 34, 35]. The primary clinical endpoint was higher in the present study (7.3% of patients) than in previous clinical PCI studies (2.9%–3.9% [13, 19, 36]). All cases in our cohort were accepted post-dilatation and with a higher in-stent balloon inflation pressure (average $>$18 atm), which reduced both the stent under-expansion and the severe incomplete stent apposition. An animal study [37] reported that adventitial myofibroblasts are a vital feature of atherosclerosis in coronary arteries. Following damage to the vessel wall, proliferating cells synthesize growth factors and migrate into the vascular intima. A serial IVUS observational study [38] found that vascular morphology and vascular stretching are altered after stent implantation. The total vascular area post-PCI was correlated with in-stent neointimal proliferation rather than with the lumen or plaque area. Therefore, a stretch or injury to the adventitia rather than the intimal section is important for neointimal growth. Appropriate stretching of the total vascular area plays a significant role in preventing late in-stent neointimal hyperplastic growth. The incidence of adverse clinical events was higher in our study, suggesting that excessive in-stent expansion may lead to vascular injury and to an increased risk of adverse events. Furthermore, the minimal total vascular expansion area should be achieved through post-dilatation.