Artificial intelligence evaluation of coronary computed tomography angiography for coronary stenosis classification and diagnosis.

BACKGROUND: Ruling out obstructive coronary artery disease (CAD) using coronary computed tomography angiography (CCTA) is time-consuming and challenging. This study developed a deep learning (DL) model to assist in detecting obstructive CAD on CCTA to streamline workflows. METHODS: In total, 2929 DICOM files and 7945 labels were extracted from curved planar reformatted CCTA images. A modified Inception V3 model was adopted. To validate the artificial intelligence (AI) model, two cardiologists labelled and adjudicated the classification of coronary stenosis on CCTA. The model was trained to differentiate the coronary artery into binary stenosis classifications <50% and ≥50% stenosis. Using the quantitative coronary angiography (QCA) consensus results as a reference standard, the performance of the AI model and CCTA radiology readers was compared by calculating Cohen's kappa coefficients at patient and vessel levels. The net reclassification index was used to evaluate the net benefit of the DL model. RESULTS: The diagnostic accuracy of the AI model was 92.3% and 88.4% at the patient and vessel levels, respectively. Compared with CCTA radiology readers, the AI model had a better agreement for binary stenosis classification at both patient and vessel levels (Cohen kappa coefficient: .79 vs. .39 and .77 vs. .40, p < .0001). The AI model also exhibited significantly improved model discrimination and reclassification (Net reclassification index = .350; Z = 4.194; p < .001). CONCLUSIONS: The developed AI model identified obstructive CAD, and the model results correlated well with QCA results. Incorporating the model into the reporting system of CCTA may improve workflows.

Improving detection of obstructive coronary artery disease with an artificial intelligence-enabled electrocardiogram algorithm.

BACKGROUND AND AIMS: To evaluate the risk of coronary artery disease (CAD), the traditional approach involves assessing the patient's symptoms, traditional cardiovascular risk factors (CVRFs), and a 12-lead electrocardiogram (ECG). However, currently, there are no established criteria for interpreting an ECG to diagnose CAD. Therefore, we sought to develop an artificial intelligence (AI)-enabled ECG model to assist in identifying patients with CAD. METHODS: In this study, we included patients who underwent coronary angiography (CAG) at a single center between 2017 and 2019. Preprocedural 12-lead ECG performed within 24 h was obtained. Obstructive CAD was defined as ≥ 50% diameter stenosis. Using age, gender and ECG data, we developed stacking models using both deep learning and machine learning. Then we compared the performance of our models with CVRFs and with cardiologists' ECG interpretation. Additionally, we validated our model on an external cohort from a different hospital. RESULTS: We included 4951 patients with a mean age of 65.5 ± 12.5 years, of whom 67.0% were men. Based on CAG, obstructive CAD was confirmed in 2637 patients (53.2%). Our best AI model demonstrated comparable performance to CVRFs in predicting CAD, with an AUC of 0.70 (95% CI: 0.66-0.75) compared to 0.71 (95% CI: 0.66-0.76). The sensitivity and specificity of the AI model were 0.75 and 0.54, respectively, while those of CVRFs were 0.67 and 0.63. Compared to cardiologists, the AI model showed better performance with an F1 score of 0.68 vs 0.41. The external validation showed generally consistent diagnostic findings, although there was a slightly lower level of agreement observed in the external cohort. Incorporating ECG and CVRFs improved the AUC to 0.72. CONCLUSIONS: Our study suggests that an AI-enabled ECG model can assist in identifying patients with obstructive CAD, with diagnostic performance similar to that of the traditional approach based on CVRFs. This model could serve as a useful clinical tool in an outpatient setting to identify patients who require further diagnostic tests.

Clinical assessment of resting full-cycle ratio and fractional flow reserve for coronary artery disease in a real-world cohort.

Background: There are few reports published on the comparison of the resting full-cycle ratio (RFR) and fractional flow reserve (FFR) on the assessment of the severity of coronary stenosis. We aimed to investigate the diagnostic accuracy of RFR for detection of functionally significant coronary lesions. Methods: This was an observational, retrospective, single-center study. We evaluated both RFR and FFR for 277 coronary lesions of 235 patients who underwent coronary angiography. Patients presenting with chronic coronary syndrome, unstable angina, or non-ST-elevation myocardial infarction were included. Results: The mean FFR and RFR values were 0.84 ± 0.08 and 0.90 ± 0.08, respectively. RFR significantly correlated with FFR (r = 0.727, P < 0.001). The agreement rate between the FFR and RFR was 79.8% (221/277). The diagnostic performance of RFR vs. FFR was accuracy 79.8%, sensitivity 70.4%, specificity 83.7%, positive predictive value 64.0%, and negative predictive value 87.2%. The discriminative power of RFR to identify lesions with FFR ≤ 0.80 was acceptable when the RFR value was within the gray zone [0.86 ≤ RFR ≤ 0.93; AUC: 0.72 (95% CI:0.63-0.81)], while it was excellent when the RFR value was out of the gray zone [RFR > 0.93 or < 0.86; AUC: 0.94 (95% CI:0.88-0.99)]. Conclusion: RFR was significantly correlated with FFR in the assessment of intermediate coronary stenosis. An RFR-FFR hybrid approach increases the diagnostic accuracy of RFR in the detection of functionally significant lesions.

Machine learning of treadmill exercise test to improve selection for testing for coronary artery disease.

BACKGROUND AND AIMS: The high false-positive rate of the treadmill exercise test (TET) may lead to unnecessary invasive coronary angiography. We aimed to develop a machine learning-based algorithm to improve the diagnostic performance of TET. METHODS: Study included 2325 patients who underwent TET and subsequent coronary angiography within one-year interval. The mean age was 58.7 (48.1-69.3) years, 1731 (74.5%) were male, 1858 (79.9%) had positive TET result, and 812 (34.9%) had obstructive coronary artery disease (≥70% stenosis in at least one vessel). The study population were randomly divided into training (70%) and testing (30%) groups for algorithm development. A total of 93 features, including exercise performance, hemodynamics and ST-segment changes were extracted from the TET results. Clinical features included comorbidity, smoking, height, weight, and Framingham risk score. Support vector machine, logistic regression, random forest, k-nearest neighbor and extreme gradient boosting machine learning algorithms were used to build the predictive models. The performance of each model was compared with that of conventional TET. RESULTS: Four of the five models exhibited comparable diagnostic performance and were better than conventional TET. The random forest algorithm had an area under the curve (AUC) of 0.73. When used with clinical features, the AUC improved to 0.74. The major advantage of the algorithm is the reduction of the false-positive rate compared with conventional TET (55% vs. 76.3%, respectively), while maintaining comparable sensitivity (85%). CONCLUSIONS: Using the information obtained from conventional TET, a more accurate diagnosis can be made by incorporating an artificial intelligence-based model.

Vessel fractional flow reserve in assessment of non-culprit lesions in ST elevation myocardial infarction.

OBJECTIVES: We sought to evaluate the physiology of non-culprit lesions by using vessel fractional flow reserve (vFFR) among patients with ST elevation myocardial infarction (STEMI) and multivessel disease (MVD). METHODS: From January 2017 to December 2019, 354 patients with STEMI in the Taipei Veterans General Hospital Acute Myocardial Infarction Registry were screened. Patients who underwent successful primary percutaneous coronary intervention (PCI) for culprit lesions, with at least one non-culprit lesion with stenosis of ≥50%, were eligible. vFFR was computed retrospectively. RESULTS: A total of 156 patients with 217 non-culprit lesions were eligible for this study. Aortic root pressure and two good angiograms were available for 139 non-culprit lesions for vFFR analysis. Based on the vFFR analysis, 59 non-culprit lesions (43.2%) had a vFFR value >0.80, and PCI was deferred in 45 lesions (76.3%). Meanwhile, 80 non-culprit lesions (56.8%) had a vFFR value ≤0.80; however, PCI was only performed in 31 lesions (38.7%) (p=0.142). The incidence of vessel-oriented composite endpoint was numerically higher in non-culprit lesions with vFFR ≤0.80 than those with vFFR >0.80 (6.3% vs 1.7%, HR: 3.59, 95% CI: 0.42 to 30.8, p=0.243). CONCLUSION: Functional incomplete revascularisation is common among patients with STEMI and MVD. The adoption of vFFR to assess non-culprit lesions may reclassify the coronary revascularisation strategy that is usually guided by angiography only in this acute setting.

Agreement Between Invasive Wire-Based and Angiography-Based Vessel Fractional Flow Reserve Assessment on Intermediate Coronary Stenoses.

Background: Angiography-based functional assessment of coronary stenoses emerges as a novel approach to assess coronary physiology. We sought to investigate the agreement between invasive coronary wire-based fractional flow reserve (FFR), resting full-cycle ratio (RFR), and angiography-based vessel FFR (vFFR) for the functional assessment of coronary stenoses in patients with coronary artery disease. Materials and Methods: Between Jan 01, 2018, and Dec 31, 2020, 298 patients with 385 intermediate lesions received invasive coronary wire-based functional assessment (FFR, RFR or both) at a single tertiary medical center. Coronary lesions involving ostium or left main artery were excluded. vFFR analysis was performed retrospectively based on aortic root pressure and two angiographic projections. Results: In total, 236 patients with 291 lesions were eligible for vFFR analysis. FFR and RFR were performed in 258 and 162 lesions, respectively. The mean FFR, RFR and vFFR value were 0.84 ± 0.08, 0.90 ± 0.09, and 0.83 ± 0.10. vFFR was significantly correlated with FFR (r = 0.708, P < 0.001) and RFR (r = 0.673, P < 0.001). The diagnostic performance of vFFR vs. FFR was accuracy 81.8%, sensitivity 77.4%, specificity 83.9%, positive predictive value 69.9%, and negative predictive value 88.5%. The discriminative power of vFFR for FFR ≤ 0.80 or RFR ≤ 0.89 was excellent. Area under the receiver operating characteristic curve (AUC) was 0.87 (95% CI:0.83-0.92) for FFR and 0.80 (95% CI:0.73-0.88) for RFR. Conclusion: Angiography-based vFFR has a substantial agreement with invasive wire-based FFR and RFR in patients with intermediate coronary stenoses. vFFR can be utilized to assess coronary physiology without a pressure wire in a post hoc manner.

Risk of ischemic stroke in patients with end-stage renal disease receiving peritoneal dialysis with new-onset atrial fibrillation.

BACKGROUND: The clinical effect of atrial fibrillation (AF)-related ischemic cardiovascular events in patients undergoing peritoneal dialysis (PD) remains uncertain. This study aimed to investigate the risk of ischemic events in patients undergoing PD with new-onset AF compared with that in patients without AF and ascertain the association between the CHA2DS2-VASc score and risk of ischemic stroke. METHODS: This nationwide, population-based cohort study used data from Taiwan's National Health Insurance Research Database from 1998 to 2011 for patients receiving PD with or without new-onset AF. The clinical endpoints included ischemic stroke, all-cause death, and in-hospital cardiovascular death. RESULTS: Patients undergoing PD with new-onset AF (N = 505) had significantly higher risks of ischemic stroke (adjusted hazard ratio [aHR], 1.98; 95% CI, 1.40-2.80), all-cause death (aHR, 1.61; 95% CI, 1.40-1.85), and in-hospital cardiovascular death (aHR, 1.82; 95% CI, 1.50.2.21) compared with those in patients undergoing PD without AF. After considering in-hospital death as a competing risk, AF remained associated with an increased risk of ischemic stroke (hazard ratio [HR], 1.67; 95% CI, 1.17-2.37). The CHA2DS2-VASc score was associated with the risk of ischemic stroke (HR, 1.28; 95% CI, 1.12-1.46). CONCLUSION: The risks of ischemic stroke, all-cause death, and in-hospital cardiovascular death were significantly higher in patients undergoing PD with AF than those in patients without AF. The CHA2DS2-VASc score remained associated with the risk of ischemic stroke in patients undergoing PD with AF.

Deficiency in fibroblast PPARß/δ reduces nonmelanoma skin cancers in mice.

The incidence of nonmelanoma skin cancer (NMSC) has been increasing worldwide. Most studies have highlighted the importance of cancer-associated fibroblasts (CAFs) in NMSC progression. However much less is known about the communication between normal fibroblasts and epithelia; disruption of this communication affects tumor initiation and the latency period in the emergence of tumors. Delineating the mechanism that mediates this epithelial-mesenchymal communication in NMSC could identify more effective targeted therapies. The nuclear receptor PPARß/δ in fibroblasts has been shown to modulate adjacent epithelial cell behavior, however, its role in skin tumorigenesis remains unknown. Using chemically induced skin carcinogenesis, we showed that FSPCre-Pparb/dex4 mice, whose Pparb/d gene was selectively deleted in fibroblasts, had delayed emergence and reduced tumor burden compared with control mice (Pparb/dfl/fl). However, FSPCre-Pparb/dex4-derived tumors showed increased proliferation, with no difference in differentiation, suggesting delayed tumor initiation. Network analysis revealed a link between dermal Pparb/d and TGF-ß1 with epidermal NRF2 and Nox4. In vitro investigations showed that PPARß/δ deficiency in fibroblasts increased epidermal Nox4-derived H2O2 production, which triggered an NRF2-mediated antioxidant response. We further showed that H2O2 upregulated NRF2 mRNA via the B-Raf-MEK1/2 pathway. The enhanced NRF2 response altered the activities of PTEN, Src, and AKT. In vivo, we detected the differential phosphorylation profiles of B-Raf, MEK1/2, PTEN, Src, and AKT in the vehicle-treated and chemically treated epidermis of FSPCre-Pparb/dex4 mice compared with that in Pparb/dfl/fl mice, prior to the first appearance of tumors in Pparb/dfl/fl. Our study revealed a role for fibroblast PPARß/δ in the epithelial-mesenchymal communication involved in cellular redox homeostasis.

Characterisation of serum progesterone and progesterone-induced blocking factor (PIBF) levels across trimesters in healthy pregnant women.

Progesterone-induced blocking factor (PIBF), which plays an important role in maintaining healthy pregnancies, has shown great promise as a prognostic biomarker for threatened miscarriage. To better characterise the physiological trends of progesterone and PIBF, we analysed serum progesterone and PIBF concentrations in healthy non-pregnant and pregnant women across trimesters. We saw increasing concentrations of progesterone and PIBF in pregnant women with advancing trimesters. The serum progesterone and PIBF percentiles across gestational age in healthy pregnancies can be used as a guide for the formulation of reference ranges. We also demonstrated a significant positive correlation between progesterone and PIBF levels. This study demonstrates increasing progesterone and PIBF concentrations in later trimesters and underscores the importance of progesterone and PIBF in healthy pregnancies. Characterisation of progesterone and PIBF across gestational age in healthy pregnant women may help to prognosticate pregnancy viability and support further research into the importance of progesterone and PIBF in the maintenance of healthy pregnancies.

Controlled-release nanoencapsulating microcapsules to combat inflammatory diseases.

The World Health Organization (WHO) has reported that globally 235 million people suffer from chronic and other inflammatory diseases. The short half-lives of nonsteroidal anti-inflammatory drugs (NSAIDs) and their notoriety in causing gastrointestinal discomforts, warrants these drugs to be released in a controlled and sustained manner. Although polymeric particles have been widely used for drug delivery, there are few reports that showcase their ability in encapsulating and sustaining the release of NSAIDs. In this paper, polymeric nanoencapsulating microcapsules loaded with NSAIDs were fabricated using solid/water/oil/water emulsion solvent evaporation method. Two NSAIDs, ibuprofen and naproxen, were first pre-loaded into nanoparticles and then encapsulated into a larger hollow microcapsule that contained the third NSAID, celecoxib. A high encapsulation efficiency (%) of these NSAIDs was achieved and a sustained release (up to 30 days) of these drugs in phosphate-buffered saline was observed. Then, a gastrointestinal drug - cimetidine (CIM) - was co-loaded with the NSAIDs. This floating delivery system exhibited excellent buoyancy (~88% up to 24 h) in simulated gastric fluid. It also allowed a sequential release of the drugs, whereby an immediate release of CIM followed by NSAIDs was observed. Drug release of the NSAIDs observed Fickian diffusion mechanism, whereas CIM observed non-Fickian diffusion. Therefore, this delivery system is a promising platform to control the delivery of NSAIDs to combat inflammatory diseases, thereby protecting against possible gastrointestinal side effects that may arise from the overuse of NSAIDs.