Cerebral thrombo-embolism is a dreaded complication of pulmonary vein isolation (PVI) for atrial fibrillation; its surrogate, silent cerebral embolism (SCE) can be detected by diffusion-weighted brain magnetic resonance imaging (bMRI). Initial investigations have raised a concern that very high-power, short-duration (vHPSD; 90 W/4 s) temperature-controlled PVI with the QDOT Micro catheter may be associated with a higher incidence of SCE compared with low-power long-duration ablation. We aimed to assess the incidence of procedural complications of vHPSD PVI with an emphasis on cerebral safety. We enrolled 328 consecutive patients undergoing their PVI procedure using vHPSD. A subgroup of 61 consecutive patients underwent diffusion-weighted bMRI within 24 h of the procedure, and incidence and predictors of SCE were studied. The mean procedure time and left atrial dwell time for the overall cohort were 69.6 ± 24.1 and 46.5 ± 21.5 min, respectively. First-pass isolation was achieved in 82%. No stroke or transient ischaemic attack occurred. Silent cerebral embolism was identified in 5 of 61 patients (8.2%). Silent cerebral embolism following procedures was significantly associated with lower baseline generator impedance (105.8 vs. 112.6 Ω, P < 0.0001) and with intermittent loss of catheter–tissue contact during ablation (14.1% vs. 6.1%, P < 0.0001). Very high-power, short-duration PVI is a safe technique with an excellent acute success rate. Silent cerebral embolism incidence in our cohort was below the previously reported range, with no clinically overt cerebral complications. Lower baseline generator impedance and loss of contact during ablation may contribute to a higher risk of SCEs.

Filtered back projection (FBP) and adaptive statistical iterative reconstruction (ASIR) are ubiquitously applied in the reconstruction of coronary CT angiography (CCTA) datasets. However, currently no data is available on the impact of a model-based adaptive filter (MBAF2), recently developed for a dedicated cardiac scanner. Our aim was to determine the effect of MBAF2 on subjective and objective image quality parameters of coronary arteries on CCTA. Images of 102 consecutive patients referred for CCTA were evaluated. Four reconstructions of coronary images (FBP, ASIR, MBAF2, ASIR + MBAF2) were co-registered and cross-section were assessed for qualitative (graininess, sharpness, overall image quality) and quantitative [image noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR)] image quality parameters. Image noise and signal were measured in the aortic root and the left main coronary artery, respectively. Graininess, sharpness, and overall image quality was assessed on a 4-point Likert scale. As compared to FBP, ASIR, and MBAF2, ASIR + MBAF2 resulted in reduced image noise [53.1 ± 12.3, 30.6 ± 8.5, 36.3 ± 4.2, 26.3 ± 4.0 Hounsfield units (HU), respectively; p < 0.001], improved SNR (8.4 ± 2.6, 14.1 ± 3.6, 11.8 ± 2.3, 16.3 ± 3.3 HU, respectively; p < 0.001) and CNR (9.4 ± 2.7, 15.9 ± 4.0, 13.3 ± 2.5, 18.3 ± 3.5 HU, respectively; p < 0.001). No difference in sharpness was observed amongst the reconstructions (p = 0.08). Although ASIR + MBAF2 was non-superior to ASIR regarding overall image quality (p = 0.99), it performed better than FBP (p < 0.001) and MBAF2 (p < 0.001) alone. The combination of ASIR and MBAF2 resulted in reduced image noise and improved SNR and CNR. The implementation of MBAF2 in clinical practice may result in improved noise reduction performance and could potentiate radiation dose reduction.

Recent randomized controlled trials have demonstrated a notable prevalence of immediate technical failures in percutaneous vascular interventions (PVIs) for complex arterial lesions associated with chronic limb-threatening ischemia. Current imaging modalities present inherent limitations in identifying these lesions, making it challenging to determine the most suitable candidates for PVI. We present a novel preprocedural magnetic resonance imaging (MRI) histology protocol for identifying lesions that might present a higher rate of immediate and midterm PVI failure. We enrolled 22 patients (13 females, average age 65.8 ± 9.72 years) scheduled for PVI were prospectively and underwent 3T MRI using ultrashort echo time and steady-state free precession contrasts to characterize target lesions before PVI. Lesions were scored as hard if >50% of the lumen was occluded by hard components (calcium/dense collagen) on MRI in the hardest cross-section. Two readers evaluated MRI datasets. Trans-Atlantic Inter-Society Consensus Document on Management of Peripheral Arterial Disease (TASC)/Global Limb Anatomic Staging System (GLASS)/Wound, Ischemia and Foot infection scoring was performed based on intraprocedural angiograms and chart review. The relationship between MRI scoring, TASC/GLASS scoring, and procedural outcomes was investigated using univariate analysis. Midterm follow-up (revascularization and amputation rates) was recorded at 3 and 6 months after the intervention. Our cohort of 22 patients yielded 40 target lesions. Five lesions were excluded (two for nondiagnostic image quality; three PVIs were ultimately diagnostic only). Six lesions (17%) were scored as hard. MRI-scored hard lesions had a higher proportion of immediate technical failure (hard vs soft 83% [5/6] vs 3% [1/29]; P < .001). Hard vs soft MRI scoring was the only factor significantly associated with immediate PVI technical success (P < .001), as opposed to TASC/GLASS scoring. Both at 3 months and 6 months after PVI, the reintervention rate was significantly higher among those lesions which were scored hard on MRI (3 months hard, 80% vs soft, 16% [P =.011]; 6 months hard, 80% vs soft, 27%; P = .047). MRI histology could be a valuable tool for optimizing PVI patient selection and treatment strategies.

The novel approach of our study consists in adapting and in evaluating a custom-made variational autoencoder (VAE) using two-dimensional (2D) convolutional neural networks (CNNs) on magnetic resonance imaging (MRI) images for differentiate soft vs. hard plaque components in peripheral arterial disease (PAD). Five amputated lower extremities were imaged at a clinical ultra-high field 7 Tesla MRI. Ultrashort echo time (UTE), T1-weighted (T1w) and T2-weighted (T2w) datasets were acquired. Multiplanar reconstruction (MPR) images were obtained from one lesion per limb. Images were aligned to each other and pseudo-color red-green-blue images were created. Four areas in latent space were defined corresponding to the sorted images reconstructed by the VAE. Images were classified from their position in latent space and scored using tissue score (TS) as following: (1) lumen patent, TS:0; (2) partially patent, TS:1; (3) mostly occluded with soft tissue, TS:3; (4) mostly occluded with hard tissue, TS:5. Average and relative percentage of TS was calculated per lesion defined as the sum of the tissue score for each image divided by the total number of images. In total, 2390 MPR reconstructed images were included in the analysis. Relative percentage of average tissue score varied from only patent (lesion #1) to presence of all four classes. Lesions #2, #3 and #5 were classified to contain tissues except mostly occluded with hard tissue while lesion #4 contained all (ranges (I): 0.2–100%, (II): 46.3–75.9%, (III): 18–33.5%, (IV): 20%). Training the VAE was successful as images with soft/hard tissues in PAD lesions were satisfactory separated in latent space. Using VAE may assist in rapid classification of MRI histology images acquired in a clinical setup for facilitating endovascular procedures.

Recent years have brought increasing development of artificial intelligence (AI)-based approaches aimed at optimizing clinical decision-making in vascular medicine. Models increasingly facilitate diagnostic interpretation, treatment planning, and prognosis prediction, particularly in complex cases of peripheral artery disease (PAD). In parallel, a major effort to integrate deep learning methods and imaging modalities has led to new opportunities for using anatomical data to enhance diagnostic capabilities. Among these modalities, magnetic resonance imaging (MRI) is gaining popularity as a noninvasive, nonionizing imaging tool with excellent soft tissue contrast and plaque characterization capabilities. Unlike conventional imaging, MRI can provide information on vascular composition—eg, soft tissue, fibrosis, and calcification—without exposing PAD patients to ionizing radiation and thus is an especially desirable modality for serial monitoring and long-term follow-up of PAD. In this review, we present an overview of the current state of AI integration with MRI in vascular imaging, with a particular focus on PAD. We also introduce several contributions from our research group, with our uniquely designed MRI protocol that incorporates ultrashort echo time and multicontrast sequences to improve vascular tissue characterization in PAD. These high-resolution datasets have served as the foundation for the development of deep learning-based classification systems, including both supervised and unsupervised models. Specifically, we describe how variational autoencoders and hybrid scoring systems have been used to phenotype PAD lesions and predict procedural factors such as guidewire crossability. Our work represents a step toward more objective, reproducible, and clinically interpretable tools that bridge high-fidelity imaging with scalable, AI-driven analysis.