Automated segmentation of metal stent and bioresorbable vascular scaffold in intravascular optical coherence tomography images using deep learning architectures

Percutaneous Coronary Intervention (PCI) with stent placement is a treatment effective for coronary artery diseases. Intravascular optical coherence tomography (OCT) with high resolution is used clinically to visualize stent deployment and restenosis, facilitating PCI operation and for complication inspection. Automated stent struts segmentation in OCT images is necessary as each pullback of OCT images could contain thousands of stent struts. In this paper, a deep learning framework is proposed and demonstrated for the automated segmentation of two major clinical stent types: metal stents and bioresorbable vascular scaffolds (BVS). U-Net, the current most prominent deep learning network in biomedical segmentation, was implemented for segmentation with cropped input. The architectures of MobileNetV2 and DenseNet121 were also adapted into U-Net for improvement in speed and accuracy. The results suggested that the proposed automated algorithm’s segmentation performance approaches the level of independent human observers and is feasible for both types of stents despite their distinct appearance. U-Net with DenseNet121 encoder (U-Dense) performed best with Dice’s coefficient of 0.86 for BVS segmentation, and precision/recall of 0.92/0.92 for metal stent segmentation under optimal crop window size of 256.

299 3-Year results of STEMI patients treated with a pre-specified BVS implantation strategy: BVS SYTEMI strategy-it long term

Aims Data on Absorb bioresorbable vascular scaffold (BVS) use in patients presenting with ST-segment elevation myocardial infarction (STEMI) are limited. Furthermore, Absorb studies including STEMI patients lacked a prespecified implantation technique to optimize BVS deployment. To assess the 3-year clinical outcomes STEMI patients undergoing primary percutaneous coronary intervention (pPCI) with a pre-specified BVS implantation strategy. Methods and results BVS STEMI STRATEGY-IT (NCT02601781) included 505 STEMI patients undergoing pPCI with BVS following a dedicated implantation protocol. Device-oriented composite endpoint [(DOCE) cardiac death, ischemia-driven target lesion revascularization (ID-TLR) and target vessel myocardial infarction (TV-MI)] and scaffold thrombosis (ScT) at 3-year were evaluated. A subgroup analysis was performed on patients in whom dual antiplatelet therapy (DAPT) was continued 36 months after pPCI (‘longer-term DAPT’), and outcomes of this cohort compared to the remaining population (‘shorter-term DAPT’: <36 months) are also reported. Three-year DOCE and ScT rates were low (2.4% and 1.0%, respectively). In 319 (63.2%) patients DAPT was continued 36 months after pPCI. DOCE rate was significantly lower in the longer- compared to shorter-DAPT group (HR: 0.29; 95% CI: 0.1–0.9; P = 0.03), driven by a lower rate of TV-MI (0% vs. 2.2%, P = 0.018). Significantly lower rate of ScT was observed in longer-DAPT group (0% vs. 2.7%, P = 0.007). Conclusions In STEMI patients undergoing pPCI, a strategy of optimized BVS implantation technique is associated with favourable DOCE and ScT rates at 3 year. DAPT extension up to 36 months further improves long-term clinical outcomes.

Application of in silico Platform for the Development and Optimization of Fully Bioresorbable Vascular Scaffold Designs

Bioresorbable vascular scaffolds (BVS), made either from polymers or from metals, are promising materials for treating coronary artery disease through the processes of percutaneous transluminal coronary angioplasty. Despite the opinion that bioresorbable polymers are more promising for coronary stents, their long-term advantages over metallic alloys have not yet been demonstrated. The development of new polymer-based BVS or optimization of the existing ones requires engineers to perform many very expensive mechanical tests to identify optimal structural geometry and material characteristics. in silico mechanical testing opens the possibility for a fast and low-cost process of analysis of all the mechanical characteristics and also provides the possibility to compare two or more competing designs. In this study, we used a recently introduced material model of poly-l-lactic acid (PLLA) fully bioresorbable vascular scaffold and recently empowered numerical InSilc platform to perform in silico mechanicals tests of two different stent designs with different material and geometrical characteristics. The result of inflation, radial compression, three-point bending, and two-plate crush tests shows that numerical procedures with true experimental constitutive relationships could provide reliable conclusions and a significant contribution to the optimization and design of bioresorbable polymer-based stents.

Reduction of Lipid-Core Burden Index in Nonculprit Lesions at Follow-Up after ST-Elevation Myocardial Infarction: A Randomized Study of Bioresorbable Vascular Scaffold versus Optimal Medical Therapy

Background. Non-flow-limiting nonculprit lesions (NCL) that contain a large lipid-rich necrotic core (nonculprit lipid-rich plaques (NC-LRP)) are most likely to cause recurrent acute coronary syndrome after ST-elevation myocardial infarction (STEMI). Near-infrared spectroscopy (NIRS) detects LRPs using the maximum 4 mm lipid-core burden index (maxLCBI4 mm). Few data are available regarding NIRS-guided therapy of these NC-LRPs, which are a potential target for preventive stenting. Bioresorbable vascular scaffold (BVS) provides local drug delivery and could facilitate plaque passivation after resorption. This study sought to assess the safety of BVS implantation in NC-LRPs and its efficacy in reducing maxLCBI4 mm at 2-year follow-up after STEMI. Methods and Results. In total, 33 non-flow-limiting NCLs from 29 STEMI patients were included in this study. Of these, 15 were LRPs and were randomly assigned to either the BVS + optimal medical therapy (OMT) arm (group 1; N = 7) or the OMT arm (group 2; N = 8). At baseline, there were no differences in plaque characteristics between groups (fractional flow reserve: 0.85 ± 0.04 vs. 0.89 ± 0.06; diameter stenosis (DS): 43.4 ± 8 vs. 40.1 ± 10.7%; plaque burden 54.98 ± 5.8 vs. 49.76 ± 8.31%; and maxLCBI4 mm 402 [348; 564] vs. 373 [298; 516]; p = N S for all comparisons between groups 1 and 2, respectively). Seven BVSs were implanted 3 ± 1 days after STEMI in six patients, without complications. At angiographic follow-up (712 [657; 740] days), a significant and similar reduction of maxLCBI4 mm was observed in both groups, with a median change of 306 [257; 377] in group 1 vs. 300 [278; 346] in group 2 p = 0.44 . DS was significantly lower in group 1 vs. group 2 (19.8 ± 7 vs. 41.7 ± 13%, p = 0.003 ), while plaque burden remained unchanged in both groups. Overall survival was 100%, target lesion failure was 13%, and stent thrombosis was 0%. Conclusions. BVS + OMT and OMT appear as similarly safe and effective in reducing maxLCBI4mm in NC-LRPs at 2-year follow-up after STEMI.

From the real device to the digital twin: A coupled experimental-numerical strategy to investigate a novel bioresorbable vascular scaffold

The purpose of this work is to propose a workflow that couples experimental and computational activities aimed at developing a credible digital twin of a commercial coronary bioresorbable vascular scaffold when direct access to data about material mechanical properties is not possible. Such a situation is be faced when the manufacturer is not involved in the study, thus directly investigating the actual device is the only source of information available. The object of the work is the Fantom® Encore polymeric stent (REVA Medical) made of Tyrocore™. Four devices were purchased and used in mechanical tests that are easily reproducible in any mechanical laboratory, i.e. free expansion and uniaxial tension testing, the latter performed with protocols that emphasized the rate-dependent properties of the polymer. Given the complexity of the mechanical behaviour observed experimentally, it was chosen to use the Parallel Rehological Framework material model, already used in the literature to describe the behaviour of other polymers, such as PLLA. Calibration of the material model was based on simulations that replicate the tensile test performed on the device. Given the high number of material parameters, a plan of simulations was done to find the most suitable set, varying each parameter value in a feasible range and considering a single repetitive unit of the stent, neglecting residual stresses generated by crimping and expansion. This strategy resulted in a significant reduction of computational cost. The performance of the set of parameters thus identified was finally evaluated considering the whole delivery system, by comparing the experimental results with the data collected simulating free expansion and uniaxial tension testing. Moreover, radial force testing was numerically performed and compared with literature data. The obtained results demonstrated the effectiveness of the digital twin development pipeline, a path applicable to any commercial device whose geometric structure is based on repetitive units.