Addressing data limitations in seizure prediction through transfer learning.

According to the literature, seizure prediction models should be developed following a patient-specific approach. However, seizures are usually very rare events, meaning the number of events that may be used to optimise seizure prediction approaches is limited. To overcome such constraint, we analysed the possibility of using data from patients from an external database to improve patient-specific seizure prediction models. We present seizure prediction models trained using a transfer learning procedure. We trained a deep convolutional autoencoder using electroencephalogram data from 41 patients collected from the EPILEPSIAE database. Then, a bidirectional long short-term memory and a classifier layers were added on the top of the encoder part and were optimised for 24 patients from the Universitätsklinikum Freiburg individually. The encoder was used as a feature extraction module. Therefore, its weights were not changed during the patient-specific training. Experimental results showed that seizure prediction models optimised using pretrained weights present about four times fewer false alarms while maintaining the same ability to predict seizures and achieved more 13% validated patients. Therefore, results evidenced that the optimisation using transfer learning was more stable and faster, saving computational resources. In summary, adopting transfer learning for seizure prediction models represents a significant advancement. It addresses the data limitation seen in the seizure prediction field and offers more efficient and stable training, conserving computational resources. Additionally, despite the compact size, transfer learning allows to easily share data knowledge due to fewer ethical restrictions and lower storage requirements. The convolutional autoencoder developed in this study will be shared with the scientific community, promoting further research.

GlyT1 Inhibition by NFPS Promotes Neuroprotection in Amyloid-ß-Induced Alzheimer's Disease Animal Model.

Alzheimer's disease (AD) is a neurodegenerative disorder characterized by the accumulation of amyloid-ß, leading to N-methyl-D-aspartate (NMDA) receptor-dependent synaptic depression, spine elimination, and memory deficits. Glycine transporter type 1 (GlyT1) modulates glutamatergic neurotransmission via NMDA receptors (NMDAR), presenting a potential alternative therapeutic approach for AD. This study investigates the neuroprotective potential of GlyT1 inhibition in an amyloid-ß-induced AD mouse model. C57BL/6 mice were treated with N-[3-([1,1-Biphenyl]-4-yloxy)-3-(4-fluorophenyl)propyl]-N-methylglycine (NFPS), a GlyT1 inhibitor, 24 h prior to intrahippocampal injection of amyloid-ß. NFPS pretreatment prevented amyloid-ß-induced cognitive deficits in short-term and long-term memory, evidenced by novel object recognition and spatial memory tasks. Moreover, NFPS pretreatment curbed microglial activation, astrocytic reactivity, and subsequent neuronal damage from amyloid-ß injection. An extensive label-free quantitative UPLC-MSE proteomic analysis was performed on the hippocampi of mice treated with NFPS. In proteomics, KEGG enrichment analysis revealed increased in dopaminergic synapse, purine-containing compound biosynthetic process and long-term potentiation, and a reduction in Glucose catabolic process and glycolytic process pathways. The western blot analysis confirmed that NFPS treatment elevated BDNF levels, correlating with enhanced TRKB phosphorylation and mTOR activation. Moreover, NFPS treatment reduced the GluN2B expression after 6 h, which was associated with an increase on CaMKIV and CREB phosphorylation. Collectively, these findings demonstrate that GlyT1 inhibition by NFPS activates diverse neuroprotective pathways, enhancing long-term potentiation signaling and countering amyloid-ß-induced hippocampal damage.

Impact of hyperaemic stenosis resistance on long-term outcomes of stable angina in the ILIAS Registry.

BACKGROUND: The hyperaemic stenosis resistance (HSR) index was introduced to provide a more comprehensive indicator of the haemodynamic severity of a coronary lesion. HSR combines both the pressure drop across a lesion and the flow through it. As such, HSR overcomes the limitations of the more traditional fractional flow reserve (FFR) or coronary flow reserve (CFR) indices. AIMS: We aimed to identify the diagnostic and prognostic value of HSR and evaluate the clinical implications. METHODS: Patients with chronic coronary syndromes (CCS) and obstructive coronary artery disease were selected from the multicentre ILIAS Registry. For this study, only patients with combined Doppler flow and pressure measurements were included. RESULTS: A total of 853 patients with 1,107 vessels were included. HSR more accurately identified the presence of inducible ischaemia compared to FFR and CFR (area under the curve 0.71 vs 0.66 and 0.62, respectively; p<0.005 for both). An abnormal HSR measurement was an independent and important predictor of target vessel failure at 5-year follow-up (hazard ratio 3.80, 95% confidence interval: 2.12-6.73; p<0.005). In vessels deferred from revascularisation, HSR seems to identify more accurately those vessels that may benefit from revascularisation rather than FFR and/or CFR. CONCLUSIONS: The present study affirms the theoretical advantages of the HSR index for the detection of ischaemia-Âinducing coronary lesions in a large CCS population. (Inclusive Invasive Physiological Assessment in Angina Syndromes Registry [ILIAS Registry], ClinicalTrials.gov: NCT04485234).

Concept-drifts adaptation for machine learning EEG epilepsy seizure prediction.

Seizure prediction remains a challenge, with approximately 30% of patients unresponsive to conventional treatments. Addressing this issue is crucial for improving patients' quality of life, as timely intervention can mitigate the impact of seizures. In this research field, it is critical to identify the preictal interval, the transition from regular brain activity to a seizure. While previous studies have explored various Electroencephalogram (EEG) based methodologies for prediction, few have been clinically applicable. Recent studies have underlined the dynamic nature of EEG data, characterised by data changes with time, known as concept drifts, highlighting the need for automated methods to detect and adapt to these changes. In this study, we investigate the effectiveness of automatic concept drift adaptation methods in seizure prediction. Three patient-specific seizure prediction approaches with a 10-minute prediction horizon are compared: a seizure prediction algorithm incorporating a window adjustment method by optimising performance with Support Vector Machines (Backwards-Landmark Window), a seizure prediction algorithm incorporating a data-batch (seizures) selection method using a logistic regression (Seizure-batch Regression), and a seizure prediction algorithm with a dynamic integration of classifiers (Dynamic Weighted Ensemble). These methods incorporate a retraining process after each seizure and use a combination of univariate linear features and SVM classifiers. The Firing Power was used as a post-processing technique to generate alarms before seizures. These methodologies were compared with a control approach based on the typical machine learning pipeline, considering a group of 37 patients with Temporal Lobe Epilepsy from the EPILEPSIAE database. The best-performing approach (Backwards-Landmark Window) achieved results of 0.75 ± 0.33 for sensitivity and 1.03 ± 1.00 for false positive rate per hour. This new strategy performed above chance for 89% of patients with the surrogate predictor, whereas the control approach only validated 46%.

Guide catheter extension use are associated with higher procedural success in chronic total occlusion percutaneous coronary interventions.

BACKGROUND: Guide catheter extensions (GCEs) increase support and facilitate equipment delivery, but aggressive instrumentation may be associated with a higher risk of complications. AIM: Our aim was to assess the impact of GCEs on procedural success and complications in patients submitted to chronic total occlusion (CTO) percutaneous coronary intervention (PCI). METHODS: We analyzed data from the multicenter LATAM CTO Registry. Procedural success was defined as <30% residual stenosis and TIMI 3 distal flow. Major adverse cardiac and cerebrovascular events (MACCE) was defined as the composite of all-cause death, myocardial infarction, target vessel revascularization, and stroke. Propensity score matching (PSM) was used to compare outcomes with and without GCE use. RESULTS: From August 2010 to August 2021, 3049 patients were included. GCEs were used in 438 patients (14.5%). In unadjusted analysis, patients in the GCE group were older and had more comorbidities. The median J-CTO score and its components were higher in the GCE group. After PSM, procedural success was higher with GCE use (87.7% vs. 80.5%, p = 0.007). The incidence of coronary perforation (odds ratio [OR]: 1.46, 95% confidence interval [CI]: 0.78-2.71, p = 0.230), bleeding (OR: 1.99, 95% CI: 0.41-2.41, p = 0.986), in-hospital death (OR: 1.39, 95% CI: 0.54-3.62, p = 0.495) and MACCE (OR: 1.07, 95% CI: 0.52-2.19, p = 0.850) were similar in both groups. CONCLUSION: In a contemporary, multicenter cohort of patients undergoing CTO PCI, GCEs were used in older patients, with more comorbidities and complex anatomy. After PSM, GCE use was associated with higher procedural success, and similar incidence of adverse outcomes.

Glycine transporter-1 inhibition by NFPS promotes neuroprotection against striatal damage models.

The striatum, an essential component of the brain's motor and reward systems, plays a pivotal role in a wide array of cognitive processes. Its dysfunction is a hallmark of neurodegenerative diseases like Parkinson's disease (PD) and Huntington's disease (HD), leading to profound motor and cognitive deficits. These conditions are often related to excitotoxicity, primarily due to overactivation of NMDA receptors (NMDAR). In the synaptic cleft, glycine transporter type 1 (GlyT1) controls the glycine levels, a NMDAR co-agonist, which modulates NMDAR function. This research explored the neuroprotective potential of NFPS, a GlyT1 inhibitor, in murine models of striatal injury. Employing models of neurotoxicity induced by 6-hydroxydopamine (PD model) and quinolinic acid (HD model), we assessed the effectiveness of NFPS pre-treatment in maintaining the integrity of striatal neurons and averting neuronal degeneration. The results indicated that NFPS pre-treatment conferred significant neuroprotection, reducing neuronal degeneration, protecting dopaminergic neurons, and preserving dendritic spines within the striatum. Additionally, this pre-treatment notably mitigated motor impairments resulting from striatal damage. The study revealed that GlyT1 inhibition led to substantial changes in the ratios of NMDAR subunits GluN2A/GluN1 and GluN2B/GluN1, 24 h after NFPS treatment. These findings underscore the neuroprotective efficacy of GlyT1 inhibition, proposing it as a viable therapeutic strategy for striatum-related damage.

Comparison between epileptic seizure prediction and forecasting based on machine learning.

Epilepsy affects around 1% of the population worldwide. Anti-epileptic drugs are an excellent option for controlling seizure occurrence but do not work for around one-third of patients. Warning devices employing seizure prediction or forecasting algorithms could bring patients new-found comfort and quality of life. These algorithms would attempt to detect a seizure's preictal period, a transitional moment between regular brain activity and the seizure, and relay this information to the user. Over the years, many seizure prediction studies using Electroencephalogram-based methodologies have been developed, triggering an alarm when detecting the preictal period. Recent studies have suggested a shift in view from prediction to forecasting. Seizure forecasting takes a probabilistic approach to the problem in question instead of the crisp approach of seizure prediction. In this field of study, the triggered alarm to symbolize the detection of a preictal period is substituted by a constant risk assessment analysis. The present work aims to explore methodologies capable of seizure forecasting and establish a comparison with seizure prediction results. Using 40 patients from the EPILEPSIAE database, we developed several patient-specific prediction and forecasting algorithms with different classifiers (a Logistic Regression, a 15 Support Vector Machines ensemble, and a 15 Shallow Neural Networks ensemble). Results show an increase of the seizure sensitivity in forecasting relative to prediction of up to 146% and in the number of patients that displayed an improvement over chance of up to 300%. These results suggest that a seizure forecasting methodology may be more suitable for seizure warning devices than a seizure prediction one.

Classifier combination supported by the sleep-wake cycle improves EEG seizure prediction performance.

OBJECTIVE: Seizure prediction is a promising solution to improve the quality of life for drug-resistant patients, which concerns nearly 30% of patients with epilepsy. The present study aimed to ascertain the impact of incorporating sleep-wake information in seizure prediction. METHODS: We developed five patient-specific prediction approaches that use vigilance state information differently: i) using it as an input feature, ii) building a pool of two classifiers, each with different weights to sleep/wake training samples, iii) building a pool of two classifiers, each with only sleep/wake samples, iv) changing the alarm-threshold concerning each sleep/wake state, and v) adjusting the alarm-threshold after a sleep-wake transition. We compared these approaches with a control method that did not integrate sleep-wake information. Our models were tested with data (43 seizures and 482 hours) acquired during presurgical monitoring of 17 patients from the EPILEPSIAE database. As EPILEPSIAE does not contain vigilance state annotations, we developed a sleep-wake classifier using 33 patients diagnosed with nocturnal frontal lobe epilepsy from the CAP Sleep database. RESULTS: Although different patients may require different strategies, our best approach, the pool of weighted predictors, obtained 65% of patients performing above chance level with a surrogate analysis (against 41% in the control method). CONCLUSION: The inclusion of vigilance state information improves seizure prediction. Higher results and testing with longterm recordings from daily-life conditions are necessary to ensure clinical acceptance. SIGNIFICANCE: As automated sleep-wake detection is possible, it would be feasible to incorporate these algorithms into future devices for seizure prediction.

Impact of sex on the assessment of the microvascular resistance reserve.

BACKGROUND: The microvascular resistance reserve (MRR) is an innovative index to assess the vasodilatory capacity of the coronary circulation while accounting for the presence of concomitant epicardial disease. The MRR has shown to be a valuable diagnostic and prognostic tool in the general coronary artery disease (CAD) population. However, considering the fundamental aspects of its assessment and the unique hemodynamic characteristics of women, it is crucial to provide additional considerations for evaluating the MRR specifically in women. AIM: The aim of this study was to assess the diagnostic and prognostic applicability of the MRR in women and assess the potential differences across different sexes. METHODS: From the ILIAS Registry, we enrolled all patients with a stable indication for invasive coronary angiography, ensuring complete physiological and follow-up data. We analyzed the diagnostic value by comparing differences between sexes and evaluated the prognostic value of the MRR specifically in women, comparing it to that in men. RESULTS: A total of 1494 patients were included of which 26% were women. The correlation between MRR and CFR was good and similar between women (r = 0.80, p < 0.005) and men (r = 0.81, p < 0.005). The MRR was an independent and important predictor of MACE in both women (HR 0.67, 0.47-0.96, p = 0.027) and men (HR 0.84, 0.74-0.95, p = 0.007). The optimal cut-off value for MRR in women was 2.8 and 3.2 in men. An abnormal MRR similarly predicted MACE at 5-year follow-up in both women and men. CONCLUSION: The MRR seems to be equally applicable in both women and men with stable coronary artery disease.

EEG epilepsy seizure prediction: the post-processing stage as a chronology.

Almost one-third of epileptic patients fail to achieve seizure control through anti-epileptic drug administration. In the scarcity of completely controlling a patient's epilepsy, seizure prediction plays a significant role in clinical management and providing new therapeutic options such as warning or intervention devices. Seizure prediction algorithms aim to identify the preictal period that Electroencephalogram (EEG) signals can capture. However, this period is associated with substantial heterogeneity, varying among patients or even between seizures from the same patient. The present work proposes a patient-specific seizure prediction algorithm using post-processing techniques to explore the existence of a set of chronological events of brain activity that precedes epileptic seizures. The study was conducted with 37 patients with Temporal Lobe Epilepsy (TLE) from the EPILEPSIAE database. The designed methodology combines univariate linear features with a classifier based on Support Vector Machines (SVM) and two post-processing techniques to handle pre-seizure temporality in an easily explainable way, employing knowledge from network theory. In the Chronological Firing Power approach, we considered the preictal as a sequence of three brain activity events separated in time. In the Cumulative Firing Power approach, we assumed the preictal period as a sequence of three overlapping events. These methodologies were compared with a control approach based on the typical machine learning pipeline. We considered a Seizure Prediction horizon (SPH) of 5 mins and analyzed several values for the Seizure Occurrence Period (SOP) duration, between 10 and 55 mins. Our results showed that the Cumulative Firing Power approach may improve the seizure prediction performance. This new strategy performed above chance for 62% of patients, whereas the control approach only validated 49% of its models.

Novel Proline Transporter Inhibitor (LQFM215) Presents Antipsychotic Effect in Ketamine Model of Schizophrenia.

The glutamatergic hypothesis of schizophrenia suggests a correlation between NMDA receptor hypofunction and negative psychotic symptoms. It has been observed that the expression of the proline transporter (PROT) in the central nervous system (CNS) is associated with glutamatergic neurotransmission, as L-proline has the capacity to activate and modulate AMPA and NMDA receptors. In this study, we aimed to investigate whether inhibition of proline transporters could enhance glutamatergic neurotransmission and potentially exhibit antipsychotic effects in an experimental schizophrenia model. Using molecular dynamics analysis in silico, we validated an innovative PROT inhibitor, LQFM215. We quantified the cytotoxicity of LQFM215 in the Lund human mesencephalic cell line (LUHMES). Subsequently, we employed the ketamine-induced psychosis model to evaluate the antipsychotic potential of the inhibitor, employing behavioral tests including open-field, three-chamber interaction, and prepulse inhibition (PPI). Our results demonstrate that LQFM215, at pharmacologically active concentrations, exhibited negligible neurotoxicity when astrocytes were co-cultured with neurons. In the ketamine-induced psychosis model, LQFM215 effectively reduced hyperlocomotion and enhanced social interaction in a three-chamber social approach task across all administered doses. Moreover, the compound successfully prevented the ketamine-induced disruption of sensorimotor gating in the PPI test at all tested doses. Overall, these findings suggest that PROT inhibition could serve as a potential therapeutic target for managing symptoms of schizophrenia model.

Prognostic Implications of Individual and Combinations of Resting and Hyperemic Coronary Pressure and Flow Parameters.

Background: Coronary pressure- and flow-derived parameters have prognostic value. Objectives: This study aims to investigate the individual and combined prognostic relevance of pressure and flow parameters reflecting resting and hyperemic conditions. Methods: A total of 1,971 vessels deferred from revascularization after invasive pressure and flow assessment were included from the international multicenter registry. Abnormal resting pressure and flow were defined as distal coronary pressure/aortic pressure ≤0.92 and high resting flow (1/resting mean transit time >2.4 or resting average peak flow >22.7 cm/s), and abnormal hyperemic pressure and flow as fractional flow reserve ≤0.80 and low hyperemic flow (1/hyperemic mean transit time <2.2 or hyperemic average peak flow <25.0 cm/s), respectively. The clinical endpoint was target vessel failure (TVF), myocardial infarction (MI), or cardiac death at 5 years. Results: The mean % diameter stenosis was 46.8% ± 16.5%. Abnormal pressure and flow were independent predictors of TVF and cardiac death/MI (all P < 0.05). The risk of 5-year TVF or MI/cardiac death increased proportionally with neither, either, and both abnormal resting pressure and flow, and abnormal hyperemic pressure and flow (all P for trend < 0.001). Abnormal resting pressure and flow were associated with a higher rate of TVF or MI/cardiac death in vessels with normal fractional flow reserve; this association was similar for abnormal hyperemic pressure and flow in vessels with normal resting distal coronary pressure/aortic pressure (all P < 0.05). Conclusions: Abnormal resting and hyperemic pressure and flow were independent prognostic predictors. The abnormal flow had an additive prognostic value for pressure in both resting and hyperemic conditions with complementary prognostic between resting and hyperemic parameters.

Changes in microvascular resistance following percutaneous coronary intervention - From the ILIAS global registry.

BACKGROUND: Microvascular resistance (MR) has prognostic value in acute and chronic coronary syndromes following percutaneous coronary intervention (PCI), however anatomic and physiologic determinants of the relative changes of MR and its association to target vessel failure (TVF) has not been investigated previously. This study aims to evaluate the association between changes in MR and TVF. METHODS: This is a sub-study of the Inclusive Invasive Physiological Assessment in Angina Syndromes (ILIAS) registry which is a global multi-centre initiative pooling lesion-level coronary pressure and flow data. RESULTS: Paired pre-post PCI haemodynamic data were available in n = 295 vessels out of n = 828 PCI treated patients and of these paired data on MR was present in n = 155 vessels. Vessels were divided according to increase vs. decrease % in microvascular resistance following PCI (ΔMR % ≤ 0 vs. ΔMR > 0%). Decreased microvascular resistance ΔMR % ≤ 0 occurred in vessels with lower pre-PCI fractional flow reserve (0.67 ± 0.15 vs. 0.72 ± 0.09 p = 0.051), coronary flow reserve (1.9 ± 0.8 vs. 2.6 ± 1.8 p < 0.0001) and higher hyperemic microvascular resistance (2.76 ± 1.3 vs. 1.62 ± 0.74 p = 0.001) and index of microvascular resistance (24.4 IQ (13.8) vs. 15. 8 IQ (13.2) p = 0.004). There was no difference in angiographic parameters between ΔMR % ≤ 0 vs. ΔMR > 0%. In a cox regression model ΔMR % > 0 was associated with increased rate of TVF (hazard ratio 95% CI 3.6 [1.2; 10.3] p = 0.018). CONCLUSION: Increased MR post-PCI was associated with lesions of less severe hemodynamic influence at baseline and higher rates of TVF at follow-up.

Development of artificial intelligence tools for invasive Doppler-based coronary microvascular assessment.

Aims: Coronary flow reserve (CFR) assessment has proven clinical utility, but Doppler-based methods are sensitive to noise and operator bias, limiting their clinical applicability. The objective of the study is to expand the adoption of invasive Doppler CFR, through the development of artificial intelligence (AI) algorithms to automatically quantify coronary Doppler quality and track flow velocity. Methods and results: A neural network was trained on images extracted from coronary Doppler flow recordings to score signal quality and derive values for coronary flow velocity and CFR. The outputs were independently validated against expert consensus. Artificial intelligence successfully quantified Doppler signal quality, with high agreement with expert consensus (Spearman's rho: 0.94), and within individual experts. Artificial intelligence automatically tracked flow velocity with superior numerical agreement against experts, when compared with the current console algorithm [AI flow vs. expert flow bias -1.68 cm/s, 95% confidence interval (CI) -2.13 to -1.23 cm/s, P < 0.001 with limits of agreement (LOA) -4.03 to 0.68 cm/s; console flow vs. expert flow bias -2.63 cm/s, 95% CI -3.74 to -1.52, P < 0.001, 95% LOA -8.45 to -3.19 cm/s]. Artificial intelligence yielded more precise CFR values [median absolute difference (MAD) against expert CFR: 4.0% for AI and 7.4% for console]. Artificial intelligence tracked lower-quality Doppler signals with lower variability (MAD against expert CFR 8.3% for AI and 16.7% for console). Conclusion: An AI-based system, trained by experts and independently validated, could assign a quality score to Doppler traces and derive coronary flow velocity and CFR. By making Doppler CFR more automated, precise, and operator-independent, AI could expand the clinical applicability of coronary microvascular assessment.

Sex differences in prevalence and outcomes of the different endotypes of chronic coronary syndrome in symptomatic patients undergoing invasive coronary angiography: Insights from the global ILIAS invasive coronary physiology registry.

BACKGROUND AND AIMS: The management of chronic coronary syndrome (CCS) is informed by studies predominantly including men. This study investigated the relationship between patients sex and different endotypes of CCS, including sex-specific clinical outcomes. METHODS: In patients with CCS undergoing coronary angiography, invasive Fractional Flow Reserve (FFR) and Coronary Flow Reserve (CFR) were measured. Patients were stratified into groups: 1) obstructive coronary artery disease (oCAD) (FFR≤0.80, no revascularization), 2) undergoing revascularization, 3) non-obstructive coronary artery disease with coronary microvascular dysfunction (CMD) (FFR>0.80, CFR≤2.5), and 4) non-obstructive coronary artery disease without CMD (FFR>0.80 and CFR>2.5). RESULTS: 1836 patients (2335 vessels) were included, comprising 1359 (74.0%) men and 477 (26.0%) women. oCAD was present in 14.1% and was significantly less prevalent in women than in men (10.3% vs 15.5%, respectively p < 0.01). Revascularization was present in 30.9% and was similarly prevalent in women and men (28.2% vs. 31.9%, respectively p = 0.13). CMD was present in 24.2% and was significantly more prevalent in women than men (28.6% vs 22.6%, respectively p < 0.01). Normal invasive measurements were found in 564 patients (33.0% women vs 30.0% men, p = 0.23). Male sex was associated with an increased risk of target vessel failure compared to women (HR.1.89, 95% CI 1.12-3.18, p = 0.018), regardless of CCS-endotype. CONCLUSIONS: Sex differences exist in the prevalence and outcomes of different endotypes of CCS in symptomatic patients undergoing invasive coronary angiography. In particular, oCAD (and subsequent revascularization) were more prevalent in men. Conversely, CMD was more prevalent in women. Overall, men experienced a worse cardiovascular outcome compared to women, independent of any specific CCS endotype.

Fractional flow reserve use in coronary artery revascularization: A systematic review and meta-analysis.

Fractional flow reserve (FFR)-guided percutaneous coronary intervention (PCI) is recommended in revascularization guidelines for intermediate lesions. However, recent studies comparing FFR-guided PCI with non-physiology-guided revascularization have reported conflicting results. PubMed and Embase were searched for studies comparing FFR-guided PCI with non-physiology-guided revascularization strategies (angiography-guided, intracoronary imaging-guided, coronary artery bypass grafting). Data were pooled by meta-analysis using random-effects model. 26 studies enrolling 78,897 patients were included. FFR-guided PCI as compared to non-physiology-guided coronary revascularization had lower risk of all-cause mortality (odds ratio [OR] 0.79 95% confidence interval [CI] 0.64-0.99, I2 = 53%) and myocardial infarction (MI) (OR 0.74 95% CI 0.59-0.93, I2 = 44.7%). However, no differences between groups were found in terms of major adverse cardiac events (MACEs) (OR 0.86 95% CI 0.72-1.03, I2 = 72.3%) and repeat revascularization (OR 1 95% CI 0.82-1.20, I2 = 43.2%). Among patients with coronary artery disease (CAD), FFR-guided PCI as compared to non-physiology-guided revascularization was associated with a lower risk of all-cause mortality and MI.

Microvascular resistance reserve: diagnostic and prognostic performance in the ILIAS registry.

AIMS: The microvascular resistance reserve (MRR) was introduced as a means to characterize the vasodilator reserve capacity of the coronary microcirculation while accounting for the influence of concomitant epicardial disease and the impact of administration of potent vasodilators on aortic pressure. This study aimed to evaluate the diagnostic and prognostic performance of MRR. METHODS AND RESULTS: A total of 1481 patients with stable symptoms and a clinical indication for coronary angiography were included from the global ILIAS Registry. MRR was derived as a function of the coronary flow reserve (CFR) divided by the fractional flow reserve (FFR) and corrected for driving pressure. The median MRR was 2.97 [Q1-Q3: 2.32-3.86] and the overall relationship between MRR and CFR was good [correlation coefficient (Rs) = 0.88, P < 0.005]. The difference between CFR and MRR increased with decreasing FFR [coefficient of determination (R2) = 0.34; Coef.-2.88, 95% confidence interval (CI): -3.05--2.73; P < 0.005]. MRR was independently associated with major adverse cardiac events (MACE) at 5-year follow-up [hazard ratio (HR) 0.78; 95% CI 0.63-0.95; P = 0.024] and with target vessel failure (TVF) at 5-year follow-up (HR 0.83; 95% CI 0.76-0.97; P = 0.047). The optimal cut-off value of MRR was 3.0. Based on this cut-off value, only abnormal MRR was significantly associated with MACE and TVF at 5-year follow-up in vessels with functionally significant epicardial disease (FFR <0.75). CONCLUSION: MRR seems a robust indicator of the microvascular vasodilator reserve capacity. Moreover, in line with its theoretical background, this study suggests a diagnostic advantage of MRR over other indices of vasodilatory capacity in patients with hemodynamically significant epicardial coronary artery disease.

Removing artefacts and periodically retraining improve performance of neural network-based seizure prediction models.

The development of seizure prediction models is often based on long-term scalp electroencephalograms (EEGs) since they capture brain electrical activity, are non-invasive, and come at a relatively low-cost. However, they suffer from major shortcomings. First, long-term EEG is usually highly contaminated with artefacts. Second, changes in the EEG signal over long intervals, known as concept drift, are often neglected. We evaluate the influence of these problems on deep neural networks using EEG time series and on shallow neural networks using widely-used EEG features. Our patient-specific prediction models were tested in 1577 hours of continuous EEG, containing 91 seizures from 41 patients with temporal lobe epilepsy who were undergoing pre-surgical monitoring. Our results showed that cleaning EEG data, using a previously developed artefact removal method based on deep convolutional neural networks, improved prediction performance. We also found that retraining the models over time reduced false predictions. Furthermore, the results show that although deep neural networks processing EEG time series are less susceptible to false alarms, they may need more data to surpass feature-based methods. These findings highlight the importance of robust data denoising and periodic adaptation of seizure prediction models.