Unsupervised classification of multi-contrast magnetic resonance histology of peripheral arterial disease lesions using a convolutional variational autoencoder with a Gaussian mixture model in latent space: A technical feasibility study.

PURPOSE: To investigate the feasibility of a deep learning algorithm combining variational autoencoder (VAE) and two-dimensional (2D) convolutional neural networks (CNN) for automatically quantifying hard tissue presence and morphology in multi-contrast magnetic resonance (MR) images of peripheral arterial disease (PAD) occlusive lesions. METHODS: Multi-contrast MR images (T2-weighted and ultrashort echo time) were acquired from lesions harvested from six amputated legs with high isotropic spatial resolution (0.078 mm and 0.156 mm, respectively) at 9.4 T. A total of 4014 pseudo-color combined images were generated, with 75% used to train a VAE employing custom 2D CNN layers. A Gaussian mixture model (GMM) was employed to classify the latent space data into four tissue classes: I) concentric calcified (c), II) eccentric calcified (e), III) occluded with hard tissue (h) and IV) occluded with soft tissue (s). Test image probabilities, encoded by the trained VAE were used to evaluate model performance. RESULTS: GMM component classification probabilities ranged from 0.92 to 0.97 for class (c), 1.00 for class (e), 0.82-0.95 for class (h) and 0.56-0.93 for the remaining class (s). Due to the complexity of soft-tissue lesions reflected in the heterogeneity of the pseudo-color images, more GMM components (n=17) were attributed to class (s), compared to the other three (c, e and h) (n=6). CONCLUSION: Combination of 2D CNN VAE and GMM achieves high classification probabilities for hard tissue-containing lesions. Automatic recognition of these classes may aid therapeutic decision-making and identifying uncrossable lesions prior to endovascular intervention.

Peripheral arterial disease treatment planning using noninvasive and invasive imaging methods.

With the growing prevalence and mortality of peripheral arterial disease, preoperative assessment, risk stratification, and determining the correct indication for endovascular and open surgical procedures are essential for therapeutic decision-making. The effectiveness of interventional procedures is significantly influenced by the plaque composition and calcification pattern. Therefore, the identification of patients for whom endovascular treatment is the most appropriate therapeutic solution often remains a challenge. The most commonly used imaging techniques have their own limitations and do not provide findings detailed enough for specific, personalized treatment planning. Using state-of-the-art noninvasive and invasive imaging modalities, it is now possible to obtain a view, not only of the complex vascular anatomy and plaque burden of the lower extremity arterial system, but also of complex plaque structures and various pathologic calcium distribution patterns. In the future, as these latest advancements in diagnostic methods become more widespread, we will be able to obtain more accurate views of the plaque structure and anatomic complexity to guide optimal treatment planning and device selection. We reviewed the implications of the most recent invasive and noninvasive lower extremity imaging techniques and future directions.

Scoping review of atherectomy and intravascular lithotripsy with or without balloon angioplasty in below-the-knee lesions.

Objective: We evaluated how contemporary data on infrapopliteal vessel preparation have been reported to identify knowledge gaps and opportunities for future research. Methods: A literature search was performed on Web of Science, PubMed, and Google Scholar to identify clinical research studies reporting on the outcomes of vessel preparation in below-the-knee lesions between 2006 and 2021. Studies were excluded if they were case reports or case series with a sample size of <10. Results: A total of 15 studies comprising 5450 patients were included in this review, with vessel preparation performed in 2179 cases (40%). Of the 15 studies, 2 were randomized controlled trials, 6 were prospective cohort studies, and 7 were retrospective studies. Only 2 of the 15 studies evaluated intravascular lithotripsy devices, and 6 were noncomparative studies. The mean diameter stenosis treated was 86.7% ± 12.6%, and the lesion length was 71.7 ± 55.3 mm. Large heterogeneity was found in the choice and definitions of end points and lesion characterization. Procedural success ranged between 84% and 90%, and bailout stenting was performed in 0.8% to 15% of cases. Of the five studies comparing procedural success of atherectomy with or without balloon angioplasty to balloon angioplasty alone, only one was in favor of the former (99% vs 90%; P < .001). The remaining studies did not show any statistically significant differences. Similarly, atherectomy had a significantly superior limb salvage rate in only one of seven studies (91% vs 73%; P = .036). In contrast, the seven studies evaluating target lesion revascularization reported conflicting outcomes, with two in favor of atherectomy, two against atherectomy, and three reporting similar outcomes between atherectomy and balloon angioplasty alone. None of the studies evaluating intravascular lithotripsy was comparative. Conclusions: The current body of evidence on vessel preparation in tibial arteries is largely based on observational studies with a large amount of heterogeneity and a number of inconsistencies. Further clinical and experimental studies with more robust study designs are warranted to investigate the comparative efficacy and safety of vessel preparation in calcified tibial arteries.

Automatic Classification of Magnetic Resonance Histology of Peripheral Arterial Chronic Total Occlusions Using a Variational Autoencoder: A Feasibility Study.

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.

Experimental Protocol and Phantom Design and Development for Performance Characterization of Conventional Devices for Peripheral Vascular Interventions.

Conventional catheter-based interventions for treating peripheral artery disease suffer high failure and complication rates. The mechanical interactions with the anatomy constrain catheter controllability, while their length and flexibility limit their pushability. Also, the 2D X-ray fluoroscopy guiding these procedures fails to provide sufficient feedback about the device location relative to the anatomy. Our study aims to quantify the performance of conventional non-steerable (NS) and steerable (S) catheters in phantom and ex vivo experiments. In a 10 mm diameter, 30 cm long artery phantom model, with four operators, we evaluated the success rate and crossing time in accessing 1.25 mm target channels, the accessible workspace, and the force delivered through each catheter. For clinical relevance, we evaluated the success rate and crossing time in crossing ex vivo chronic total occlusions. For the S and NS catheters, respectively, users successfully accessed 69 and 31% of the targets, 68 and 45% of the cross-sectional area, and could deliver 14.2 and 10.2 g of mean force. Using a NS catheter, users crossed 0.0 and 9.5% of the fixed and fresh lesions, respectively. Overall, we quantified the limitations of conventional catheters (navigation, reachable workspace, and pushability) for peripheral interventions; this can serve as a basis for comparison with other devices.

To Cross or Not to Cross: Using MRI-Histology to Characterize Dense Collagenous Plaque in Critical Limb Ischemia.

Peripheral artery disease (PAD) is caused by atherosclerotic buildup in the lower extremities, leading to obstruction and inadequate perfusion to the peripheral vasculature. Impenetrable plaques initially treated with percutaneous vascular intervention (PVI) have led to worse secondary bypass outcomes and amputation in patients. In this case report, we discuss the importance of using magnetic resonance imaging (MRI) histology in PVI planning in a patient with critical limb ischemia. PVI attempts to recanalize the limb failed because of an impenetrable occlusion in the popliteal artery that was not identified on routine preoperative imaging. Subsequent bypass occluded multiple times eventually requiring an above-knee amputation. An MRI-histology protocol-using ultrashort echo time (UTE) and T2-weighted (T2W) sequences-that was performed prior to the index PVI identified the occlusion as a dense collagen plaque. Histology analysis of the amputated specimen confirmed the MRI finding. This imaging modality offers a novel approach to characterize plaque composition and morphology, thereby identifying lesions at greatest risk of PVI failure and potentially playing an important role in selecting the right candidates for an endovascular-first approach.

Human Cadaveric Model for Vessel Preparation Device Testing in Calcified Tibial Arteries.

To describe an ex vivo model for vessel preparation device testing in tibial arteries. We performed orbital atherectomy (OA), intravascular lithotripsy (IVL), and plain balloon angioplasty (POBA) on human amputated limbs with evidence of concentric tibial artery calcification. The arterial segments were then harvested for ex vivo processing which included imaging with microCT, decalcification, and histology. The model was tested out in 15 limbs and was successful in 14 but had to be aborted in 1/15 case due to inability to achieve wire access. A total of 22 lesions were treated with OA on 3/22 lesions, IVL on 8/22, and POBA without vessel preparation on the remaining 11/22. Luminal gain was assessed with intravascular ultrasound and histology was able to demonstrate plaque disruption, dissections, and cracks within the calcified lesions. A human cadaveric model using amputated limbs is a feasible, high-fidelity option for evaluating the performance of vessel preparation devices in calcified tibial arteries.

Catheter-Directed Interventions for the Treatment of Lower Extremity Deep Vein Thrombosis.

Lower extremity deep vein thrombosis (DVT) leads to significant morbidity including pain, swelling, and difficulty walking in the affected limb. If left untreated, DVT increases the risk of pulmonary embolism (PE), recurrent venous thromboembolism (VTE), and post thrombotic syndrome (PTS). The objective of this review was to identify catheter-directed interventions and their success rates for the treatment of lower extremity DVT. A comprehensive search of current and emerging catheter-directed interventions for lower extremity DVT treatment was conducted in PubMed and Google Scholar. Clinical trials, retrospective and prospective observational studies, and case reports were identified to classify percutaneous mechanical thrombectomy (PMT), catheter-directed thrombolysis (CDT), and pharmacomechanical CDT (PCDT) devices based on their mechanism of action and indication of use. Catheter-directed interventions such as PMT, CDT, and PCDT offer an alternative therapeutic strategy for DVT management, particularly in patients with limb-threatening conditions and absolute contraindications to anticoagulants. Currently, there are limited guidelines for the use of mechanical and pharmacomechanical devices because of the lack of clinical evidence available for their use in treatment. Future studies are required to determine the short and long-term effects of using catheter-directed interventions as well as their effectiveness in treating acute versus subacute and chronic DVT.

From bench to bedside: A narrative review and institutional experience with experimental models for vascular device design and translation.

The past 2 decades have seen a rise in vascular innovations and a rapid evolution in endovascular device technology, with the emergence of atherectomy, intravascular lithotripsy, drug elution technology, thrombectomy devices, and many more. Like all other medical devices, vascular devices undergo a life cycle composed of a concept phase, a planning and design phase, a regulatory process, a launch phase, and a post-market stage. Experimental and preclinical models are required at various stages of the life cycle to aid in the designing, refining, and feasibility testing of novel devices before they are transferred to clinical practice. The experimental testing of these devices relies heavily on the ability to simulate human anatomy and physiology, and to mimic or induce specific disease processes. Computational and benchtop models play very important roles at the early stages of the manufacturing process, and animal and cadaveric models are indispensable for testing the mechanistic performance, safety, and efficacy of novel devices before they are used in clinical trials and regulatory approval is obtained for public use.

Imaging: New Frontiers in Vascular Training.

Advances in medical imaging have redefined the practice of vascular surgery. Current training programs for vascular surgery do not incorporate formal training in vascular imaging other than in duplex ultrasound when a physician is undergoing the vascular interpretation certification process. Yet imaging modalities and techniques have grown exponentially in the adjacent fields of interventional radiology, interventional and diagnostic cardiology, and neuroradiology, so much so that advanced imaging fellowships have been established in these fields. This article reviews the current state of vascular imaging training, identifies gaps in the current training regimen, and proposes an advanced vascular imaging fellowship for the future.

Performance Assessment of a Radiofrequency Powered Guidewire for Crossing Peripheral Arterial Occlusions Based on Lesion Morphology.

Endovascular wires and devices for peripheral arterial disease therapy have evolved greatly, yet failure rates of these procedures remain high. Information on lesion composition may inform device selection to improve the success rates of these procedures. This paper, presents an approach for informed guidewire selection. The objective of this study is to quantitatively assess the performance of a radiofrequency powered guidewire in the crossing of various morphology types of peripheral chronic total occlusions. Samples taken from amputated patient limbs are characterized by magnetic resonance imaging. Using a customized catheter test station, the performance of a radiofrequency powered guidewire in puncturing these lesions is compared to a conventional guidewire, and to itself when not powered. The analysis includes quantitative and statistical comparisons of the puncture forces experienced by the different guidewires in "hard" vs. "soft" lesions as well as qualitative assessment of deflections, buckling and puncture success of the wires. Results indicate that the use of radiofrequency ablation significantly reduces the required puncture force, reduced events of buckling and deflection, and resulted in a significantly higher puncture success rate.

MRI for peripheral artery disease: Introductory physics for vascular physicians.

Magnetic resonance imaging (MRI) has advanced significantly in the past decade and provides a safe and non-invasive method of evaluating peripheral artery disease (PAD), with and without using exogenous contrast agents. MRI offers a promising alternative for imaging patients but the complexity of MRI can make it less accessible for physicians to understand or use. This article provides a brief introduction to the technical principles of MRI for physicians who manage PAD patients. We discuss the basic principles of how MRI works and tailor the discussion to how MRI can evaluate anatomic characteristics of peripheral arterial lesions.

Magnetic resonance imaging characteristics of lesions relate to the difficulty of peripheral arterial endovascular procedures.

OBJECTIVE: Limitations with current peripheral arterial imaging modalities make selection of patients for percutaneous vascular interventions difficult. The purpose of this study was to determine whether a novel preprocedural magnetic resonance imaging (MRI) method can identify lesions that would be more challenging to cross during percutaneous vascular intervention. METHODS: Fourteen patients with peripheral arterial disease underwent MRI before their intervention. A novel steady-state free precession flow-independent magnetic resonance (MR) angiogram was used to locate lesions, and an ultrashort echo time image was used to characterize hard lesion components including calcium and dense collagen. Lesions were characterized as hard if ≥50% of the lumen was occluded with calcium or collagen (as determined by MR image characteristics) in the hardest cross section within the lesion. The primary outcome was the time it took to cross a guidewire through the target lesion. The secondary outcome was the need for stenting. RESULTS: Of 14 lesions, 8 (57%) were defined as hard and 6 (43%) were soft on the basis of MR image characteristics. Hard lesions took significantly longer to cross than soft lesions (average, 14 minutes 49 seconds vs 2 minutes 17 seconds; P = .003). Hard lesions also required stenting more often than soft lesions (Fisher exact test, P = .008). Of 14 lesions, 2 (14%) could not be crossed with a guidewire, and both lesions were hard. MR images also detected occult patencies and noncalcified hard lesions that could not be seen on X-ray angiography. CONCLUSIONS: MRI can be used to determine which peripheral arterial lesions are more difficult to cross with a guidewire. Future work will determine whether MRI lesion characterization can predict long-term endovascular outcomes to aid in procedure planning.

Perfusion measures for symptom severity and differential outcome of revascularization in limb ischemia: Preliminary results with arterial spin labeling reactive hyperemia.

BACKGROUND: Previously, a theoretical model based on microvascular physiology was established to facilitate the interpretation of calf perfusion dynamics recorded by arterial spin labeling (ASL). PURPOSE: To investigate the clinical relevance of novel perfusion indices by comparing them to the symptoms, response to revascularization, and short-term functional outcome in patients with peripheral arterial disease (PAD). STUDY TYPE: Prospective cohort study. POPULATION: Nineteen patients with PAD. FIELD STRENGTH/SEQUENCE: Pulsed ASL at 3T. ASSESSMENT: The mid-calf reactive hyperemia induced by 2 minutes of arterial occlusion was recorded in PAD patients. The perfusion responses were characterized by the peak, time-to-peak, and physiological model-derived indices including the baseline perfusion fr , arterial resistance Ra , and compliance Ca , and sensitivity gATP and response time τATP of downstream microvasculature to metabolic stress. These indices were compared to the disease severity and outcome within 6 months after revascularization assessed by self-reported symptoms and the ankle-brachial index. Disease severity was categorized as asymptomatic, claudication, or critical limb ischemia. The outcome was categorized as symptom resolved or limited improvement. STATISTICAL TESTS: Severity and outcome groups were compared using Mann-Whitney and Kruskal-Wallis tests with Holm-Sidak adjustments. RESULTS: The peak perfusion decreased and model arterial resistance increased progressively with increasing severity of limb ischemia (P = 0.0402 and 0.0413, respectively). Eleven patients had a successful endovascular procedure, including six patients who had symptoms resolved, four patients who had remaining leg pain, and one patient lost to follow-up. The subjects with limited improvement had significantly lower preintervention microvascular sensitivity gATP than those with symptoms resolved (8.72 ± 1.46 vs. 4.93 ± 0.91, P = 0.0466). DATA CONCLUSION: ASL reactive hyperemia reflects multiple aspects of the pathophysiology. Measures of macrovascular arterial disease are related to the manifested symptom severity, whereas preintervention gATP associated with microvascular dysfunction is related to prognosis following revascularization. LEVEL OF EVIDENCE: 1 Technical Efficacy: Stage 4 J. Magn. Reson. Imaging 2018;47:1578-1588.