A physics-informed deep learning framework for modeling of coronary in-stent restenosis

Machine learning (ML) techniques have shown great potential in cardiovascular surgery, including real-time stenosis recognition, detection of stented coronary anomalies, and prediction of in-stent restenosis (ISR). However, estimating neointima evolution poses challenges for ML models due to limitations in manual measurements, variations in image quality, low data availability, and the difficulty of acquiring biological quantities. An effective in silico model is necessary to accurately capture the mechanisms leading to neointimal hyperplasia. Physics-informed neural networks (PINNs), a novel deep learning (DL) method, have emerged as a promising approach that integrates physical laws and measurements into modeling. PINNs have demonstrated success in solving partial differential equations (PDEs) and have been applied in various biological systems. This paper aims to develop a robust multiphysics surrogate model for ISR estimation using the physics-informed DL approach, incorporating biological constraints and drug elution effects. The model seeks to enhance prediction accuracy, provide insights into disease progression factors, and promote ISR diagnosis and treatment planning. A set of coupled advection-reaction-diffusion type PDEs is constructed to track the evolution of the influential factors associated with ISR, such as platelet-derived growth factor (PDGF), the transforming growth factor- β (TGF- β ), the extracellular matrix (ECM), the density of smooth muscle cells (SMC), and the drug concentration. The nature of PINNs allows for the integration of patient-specific data (procedure-related, clinical and genetic, etc.) into the model, improving prediction accuracy and assisting in the optimization of stent implantation parameters to mitigate risks. This research addresses the existing gap in predictive models for ISR using DL and holds the potential to enhance patient outcomes through predictive risk assessment.

Association between periprocedural myocardial injury and neointimal characteristics in patients with in-stent restenosis: an optical coherence tomography study

Background

The relationship between neointimal characteristics of in-stent restenosis (ISR) and periprocedural myocardial injury (PMI) remains unclear. Therefore, this study aimed to investigate the relationship between PMI and neointimal characteristics of ISR by using optical coherence tomography (OCT).

Methods

This was a retrospective study. We enrolled 140 patients diagnosed with ISR with normal baseline high-sensitivity troponin T (hs-cTnT) levels who underwent OCT and subsequent revascularization by means of drug-coated balloon (DCB) or drug-eluting stent (DES) between October 2018 and October 2022 in the Affiliated Hospital of Zunyi Medical University. Based on the 4th universal definition of myocardial infarction, patients whose hs-cTnT were increased five times above the upper reference limit (URL) after percutaneous coronary interventions (PCI) were deemed to PMI. The patients were subdivided into PMI (n=53) and non-PMI (n=87) groups. In the univariable analysis, variables in the baselines, angiography characteristics and OCT findings were analyzed with binary logistic regression. A P value of <0.2 was included in the multivariable model. Multivariable logistic regression analysis was used to identify the independent predictors of PMI.

Results

The prevalence of intra-intimal microvessels in patients with PMI was higher than in those without PMI (58.5% vs. 32.2%, P=0.003). The ratio of intra-stent plaque rupture (PR) was also higher in patients with PMI (60.4% vs. 40.2%, P=0.021). Multivariable logistic regression analysis showed that intra-intimal microvessels [odds ratio (OR): 3.193, 95% confidence interval (CI): 1.280-7.966; P=0.013] and intra-stent PR (OR: 2.124, 95% CI: 1.153-4.732; P=0.035) were independently associated with PMI.

Conclusions

Intra-intimal microvessels and intra-stent PR were independently associated with PMI. Accurate identification and recognition of intra-intimal microvessels and intra-stent PR may be helpful in preventing PMI.

Association of the monocytes to high-density lipoprotein cholesterol ratio with in-stent neoatherosclerosis and plaque vulnerability: An optical coherence tomography study

BACKGROUND

Monocyte-to-high-density lipoprotein cholesterol ratio (MHR) is an independent predictor of atherosclerosis and in-stent restenosis (ISR). However, the association between MHR and the incidence of in-stent neoatherosclerosis (ISNA) remains to be validated.

METHODS

This study included 216 patients with acute coronary syndrome who had 220 ISR lesions and had undergone optical coherence tomography (OCT). All eligible patients were divided into three groups according to their MHR tertile level. OCT characteristics were comparatively analyzed between groups of different MHR levels, and univariate and multivariate logistic regression analyses were constructed to assess correlations between MHR level and ISNA as well as in-stent thin-cap fibroatheroma (TCFA). A receiver operating characteristic curve was used to determine the optimal MHR thresholds for predicting ISNA and in-stent TCFA.

RESULTS

The incidence of ISNA (70.3% vs. 61.1% vs. 20.3%, P < 0.001) and in-stent TCFA (40.5% vs. 31.9% vs. 6.8%, P < 0.001) was the highest in the third tertile, followed by the second and first tertiles, respectively. Multivariate analysis revealed that MHR was independently associated with ISNA (odds ratio [OR], 7.212; 95% confidence interval [CI], 1.287-40.416; P = 0.025) and in-stent TCFA (OR, 5.610; 95% CI, 1.743-18.051; P = 0.004) after adjusting for other clinical factors. The area under the curve was 0.745 (95% CI, 0.678-0.811; P < 0.001) for the prediction of ISNA and 0.718 (95% CI, 0.637-0.778; P < 0.001) for the prediction of in-stent TCFA.

CONCLUSION

MHR levels are an independent risk factor for ISNA.

Evaluation of a Rabbit Model of Vascular Stent Healing: Application of Optical Coherence Tomography

Percutaneous coronary intervention (PCI) is a management strategy for symptomatic obstructive coronary artery disease (CAD). Despite advancements, in-stent restenosis (ISR) still imparts a 1-2% annual rate of repeat revascularization-a focus of ongoing translational research. Optical coherence tomography (OCT) provides high resolution virtual histology of stents. Our study evaluates the use of OCT for virtual histological assessment of stent healing in a rabbit aorta model, enabling complete assessment of intraluminal healing throughout the stent. ISR varies based on intra-stent location, stent length, and stent type in a rabbit model-important considerations for translational experimental design. Atherosclerosis leads to more prominent ISR proliferation independent of stent-related factors. The rabbit stent model mirrors clinical observations, while OCT-based virtual histology demonstrates utility for pre-clinical stent assessment. Pre-clinical models should incorporate clinical and stent factors as feasible to maximize translation to clinical practice.

Identification of differentially expressed genes associated with coronary in-stent restenosis by integrated bioinformatics approaches

BACKGROUND

In-stent restenosis (ISR) is a common complication following percutaneous coronary intervention (PCI). Designing ISR-specific therapeutic targets is important for optimizing coronary stenosis therapy. This study aims to find new potential key genes related to ISR and the usefulness of differentially expressed genes (DEGs) as diagnostic biomarkers for ISR.

METHODS

: Microarray datasets of peripheral blood was downloaded from the Gene Expression Omnibus (GEO) public database (GSE46560, n=11). The DEGs between the ISR group and the control were obtained through the limma package analysis of the R language. Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were used to obtain the related signal pathways. STRING database was used to constructed a protein-to-protein interaction network with DEGs. Networkanalyst database was used to predict the target miRNA of the target gene. The area under the curve (AUC) value obtained from receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic effectiveness of the hub gene in distinguishing ISR patients from normal individuals.

RESULTS

A total of 154 DEGs were identified between ISR group and control group, including 57 up-regulated and 97 down-regulated DEGs. GO enrichment analysis showed that genes were enriched in homeostasis, cell morphology and cell connection. KEGG pathway analysis shows that genes are mainly involved in human papillomavirus infection, salmonella infection, human T-cell leukemia virus type 1 infection, axon guidance and MAPK signaling pathway. CLTA, CAT, STAT5A, CD300A, CA1, NCF2, HBQ1, AHSP, SLC4A1 and EPB42 were identified as 10 hub genes. A total of 151 target miRNAs of 7 DEGs were obtained, and 149 mRNA-miRNA pairs were identified. ROC curve indicates that CA1, STAT5A and HBQ1 have high diagnostic value in ISR.

CONCLUSIONS

CA1, STAT5A and HBQ1 are identified as specific related genes for ISR patients. These genes may provide new targets for the early identification and treatment of ISR.

A novel mechanism of inhibiting in-stent restenosis with arsenic trioxide drug-eluting stent: Enhancing contractile phenotype of vascular smooth muscle cells via YAP pathway

Objective

Arsenic trioxide (ATO or As₂O₃) has beneficial effects on suppressing neointimal hyperplasia and restenosis, but the mechanism is still unclear. The goal of this study is to further understand the mechanism of ATO's inhibitory effect on vascular smooth muscle cells (VSMCs).

Methods and results

Through in vitro cell culture and in vivo stent implanting into the carotid arteries of rabbit, a synthetic-to-contractile phenotypic transition was induced and the proliferation of VSMCs was inhibited by ATO. F-actin filaments were clustered and the elasticity modulus was increased within the phenotypic modulation of VSMCs induced by ATO in vitro. Meanwhile, Yes-associated protein (YAP) nuclear translocation was inhibited by ATO both in vivo and in vitro. It was found that ROCK inhibitor or YAP inactivator could partially mask the phenotype modulation of ATO on VSMCs.

Conclusions

The interaction of YAP with the ROCK pathway through ATO seems to mediate the contractile phenotype of VSMCs. This provides an indication of the clinical therapeutic mechanism for the beneficial bioactive effect of ATO-drug eluting stent (AES) on in-stent restenosis (ISR).

Plasma levels of miR-143 and miR-145 are associated with coronary in-stent restenosis within 1 year of follow-up after drug-eluting stent implantation

Background

ISR remains the major adverse outcome after percutaneous coronary intervention (PCI). MicroRNAs have been demonstrated to be associated with coronary plaque and stable in the blood and can be used as biomarkers/predictors. This study aimed to investigate whether circulating microRNAs could predict in-stent restenosis (ISR).

Methods

MicroRNA array was used to detect differently expressed microRNAs between 30 ISR patients and 30 non-ISR patients in the derivation cohort. Quantitative real-time polymerase chain reaction (qRT-PCR) was used to validate the microRNA array results and to detect levels of target microRNAs in the validation cohort. All patients were followed up for at least 1 year, and major adverse cardiac events (MACEs) were recorded. Univariate and multivariate logistic regression analysis were applied to find factors associated with ISR. Receiver operating characteristics (ROC) and Kaplan-Meier survival curves were used to analyze the predictive ability of the microRNA score for ISR.

Results

MicroRNA array and qRT-PCR showed that miR-143, 145, 425, 208, and let-7g were differently expressed between ISR patients and non-ISR patients. Multivariate analysis demonstrated that lower levels of mir-143 (OR =2.36, 95% CI: 1.43–3.67) and mir-145 (OR =2.12, 95% CI: 1.56–3.48) were associated with ISR. MicroRNA scores differed statistically between ISR patients and non-ISR patients (49.18±2.05 vs. 52.10±2.41, P<0.01) and has predictive ability for ISR with an area under the curve (AUC) of 0.8206 (95% CI: 0.7155–0.9256, P<0.01). In the validation cohort, Kaplan-Meier survival curves demonstrated that patients with higher microRNA scores have better prognosis in 1 year of follow-up.

Conclusions

A lower plasma level of mir-143/145 predicts a higher risk of ISR and a worse outcome.

A potential strategy for in-stent restenosis: Inhibition of migration and proliferation of vascular smooth muscle cells by Cu ion

The in-stent restenosis (ISR) often happens after the implantation of metal stents, including both bare metal stents (BMSs) and drug-eluting stents (DESs). Drug release from DESs could reduce significantly the occurrence of ISR but also suppress the revascularization and cause thrombosis. In this study, the effect of Cu ion in a range of 0 to 500 μM on the migration and proliferation of rat aortic smooth muscle cells (RASMCs) was investigated by a series of in vitro experiments including wound-healing assay, cell viability assay and flow cytometric analysis. It has been found that the critical concentration of Cu ion should be at least 250 μM in order to significantly inhibit the migration of RASMCs and the proliferation of RASMCs were impeded by every dose of Cu ion used in this study. In addition, the protein level of caspase-3 was upregulated by 250 μM and 500 μM Cu²⁺ exposure, which might be the main reason for RASMCs apoptosis. Thus, it is proposed that ISR might be prevented by the constant release of Cu ion.

Efficacy and safety of drug-eluting stent for the intracranial atherosclerotic disease: A systematic review and meta-analysis

Drug-eluting stent (DES) is a potential endovascular treatment for patients with symptomatic intracranial atherosclerotic disease (sICAD). However, evidence regarding the treatment of ICAD with DES is lacking. We systematically searched Pubmed, Embase, Cochrane database (before 2017-12-21) for literature reporting the application of DES in the treatment of sICAD. The main outcomes were as follows: the incidence of any stroke or death within 30 days (perioperative complications), ischemic stroke in the territory of the qualifying artery beyond 30 days (long-term complications), in-stent restenosis rate (ISR) and symptomatic ISR during follow-up. Those studies with mean stenosis rate greater than 70% and less than 70% were defined as severe and moderate stenosis group, respectively. The random effect model was used to pool the data. Of 518 articles, 13 studies were eligible and included in our analysis (N = 336 patients with 364 lesions). After the implantation of DES, perioperative complications (mortality = 0) occurred in 6.0% (95%CI 2.0%-11.9%), long-term complications occurred in 2.2% (95%CI 0.7%-4.5%), ISR rate was 4.1% (95%CI 1.6%-7.7%) and the symptomatic ISR rate was only 0.5% (95%CI 0-2.2%). In addition, subgroup analysis showed that the perioperative complication rate in severe stenosis group [10.6% (95%CI 6.5%-15.7%)] was significantly (p < 0.01) higher than that in moderate stenosis group [1.0% (95%CI 0.3%-3.5%)]. In summary, endovascular DES implantation is a relatively safe and effective method compared with stents or medical management group in SAMMPRIS and VISSIT trials. However, a higher preoperative stenosis rate may imply a higher risk of perioperative complications. Further studies are needed.

Association of high level of hs-CRP with in-stent restenosis: A case-control study

BACKGROUND

In-stent restenosis (ISR) is one adverse outcome of coronary stent implantation. Although using drug-eluting stents has reduced the rate of ISR, it remains a major problem. Here, we have investigated the relationship between several patient characteristics including serum high sensitive C-reactive protein (hs-CRP) and ISR.

METHODS

This was a case-control study comprising 104 individuals with ISR and 202 patients without. Baseline characteristics were collected using a questionnaire. Fasting blood glucose (FBG), serum triglycerides (TG), total cholesterol (TC), high density lipoprotein cholesterol (HDL-C) and serum high sensitivity C-reactive protein (hs-CRP) were measured using commercial kits on an auto-analyzer. Data were analyzed using SPSS software and a p value ≤ 0.05 was considered significant.

RESULTS

Diabetes mellitus (p < 0.001), stent type (p = 0.005), serum hs-CRP (p = 0.006), FBG (p = 0.038) and serum TG (p = 0.039) were significantly associated with ISR. The association between hs-CRP and ISR remained significant after adjustment for stent type and DM. For patients with a serum hs-CRP <2.64 mg/dL, ISR was only associated with diabetes mellitus (p = 0.016); while for individuals with a serum hs-CRP ≥2.64 mg/dL, ISR was also associated with the presence of diabetes mellitus, serum triglycerides and stent type.

CONCLUSION

Higher levels of serum hs-CRP were significantly associated with the occurrence of ISR.

Carotid endarterectomy for treatment of carotid in-stent restenosis: long-term follow-up results and surgery experiences from one single centre

Objective

Few studies have reported the surgical treatment of carotid in-stent restenosis (ISR), more data and longer follow-up are needed. We describe the surgical treatment of ISR by standard carotid endarterectomy (CEA) with stent removal, including long-term follow-up in 10 patients from our centre.

Methods

Ten patients from our centre who underwent CEA with stent removal for ISR were retrospectively analysed, including nine symptomatic and one asymptomatic ISR of at least 70% with mean age 67.3, the median time between carotid artery stenting and CEA was 17 months (range, 2–54 months).

Results

Standard CEA with stent removal was performed in all 10 patients without much technical difficulty (9 male and 1 female, mean age 67.3). Two cases were performed in hybrid operation room. There were a total of three complications that happened in three patients (30%) respectively. An asymptomatic dissecting aneurysm was formed on the petrous internal carotid artery in one patient who was followed up without intervention. In the second case, dissection occurred in the arterial wall distal to the site of the stent after stent removal revealed by intraoperative angiography, and another stent was implanted. The patient sustained temporary hypoglossal nerve dysfunction postoperatively. The third patient suffered cerebral hyperperfusion with complete recovery when discharged. No neurological complications occurred in other seven patients. After follow-up of 25 months (range, 11–54 months), one patient died of rectal cancer without ischaemic attack and restenosis 4 years postoperation; in one patient occurred recurrent symptomatic restenosis (90%) 1 year later; all other patients remained asymptomatic and without recurrent restenosis (>50%) by follow-up carotid ultrasound or CT angiography.

Conclusion

It seems that CEA with stent removal is a reasonable choice, by experienced hand, for symptomatic ISR with higher but acceptable complications. The indication of stent removal for asymptomatic ISR needs further observation.

IVUS-Guided Implantation of Bioresorbable Vascular Scaffolds for Very Late Paclitaxel Stent Thrombosis

Bioresorbable vascular scaffold (BVS) implantation has been shown to be safe in patients with stable coronary disease, and effective against the thrombotic lesion and the in-stent restenosis (ISR) of the drug-eluting stent (DES). BVSs have the advantages of a snow racket concept, positive vessel remodeling, and better conformability compared with DES in acute coronary syndrome (ACS). We report on a young patient with ST-elevation myocardial infarction (STEMI) who presented to our emergency department arising from very late stent thrombosis (VLST) of a 2.5 × 28 mm paclitaxel-eluting stent (Coroflex^® Please) three years after its implantation. After the patient was treated with balloon dilation, intravascular ultrasound (IVUS) revealed a short segment of a guide wire outside the DES mesh. Two BVSs were implanted to prevent a DES recoil. Post-scaffold-implantation IVUS showed adequately expanded strut of BVSs. Six months later, optical coherence tomography (OCT) revealed that some segments of the scaffold had been absorbed and that there was no in-scaffold restenosis. The patient had not complained about angina during the out-patient clinic follow-up. This is the first report of successful BVS implantation for a STEMI patient attributable to DES VLST.

Understanding and managing in-stent restenosis: a review of clinical data, from pathogenesis to treatment

The lumen diameter reduction after percutaneous coronary intervention (PCI) is well known as "restenosis". This phenomenon is due to vessel remodeling/recoil in case of no-stent strategy or, in case of stent employ, "neointimal proliferation" that consists in an excessive tissue proliferation in the luminal surface of the stent otherwise by a further new-occurring atherosclerotic process called "neoatherosclerosis". The exact incidence of in-stent restenosis (ISR) is not easy to determine caused by different clinical, angiographic and operative factors. In the pre-stent era the occurrence of restenosis ranged between 32-55% of all angioplasties, and drop to successively 17-41% in the bare metal stents (BMS) era. The advent of drug-eluting stent (DES), especially 2nd generation, and drug-coated balloon (DCB) further reduce restenosis rate until <10%. We here review the main characteristics of this common complication of coronary interventions, from its pathogenesis to the most appropriate treatment strategy.