Deforming Patient-Specific Models of Vascular Anatomies to Represent Stent Implantation via Extended Position Based Dynamics.

PURPOSE: Angioplasty with stent placement is a widely used treatment strategy for patients with stenotic blood vessels. However, it is often challenging to predict the outcomes of this procedure for individual patients. Image-based computational fluid dynamics (CFD) is a powerful technique for making these predictions. To perform CFD analysis of a stented vessel, a virtual model of the vessel must first be created. This model is typically made by manipulating two-dimensional contours of the vessel in its pre-stent state to reflect its post-stent shape. However, improper contour-editing can cause invalid geometric artifacts in the resulting mesh that then distort the subsequent CFD predictions. To address this limitation, we have developed a novel shape-editing method that deforms surface meshes of stenosed vessels to create stented models. METHODS: Our method uses physics-based simulations via Extended Position Based Dynamics to guide these deformations. We embed an inflating stent inside a vessel and apply collision-generated forces to deform the vessel and expand its cross-section. RESULTS: We demonstrate that this technique is feasible and applicable for a wide range of vascular anatomies, while yielding clinically compatible results. We also illustrate the ability to parametrically vary the stented shape and create models allowing CFD analyses. CONCLUSION: Our stenting method will help clinicians predict the hemodynamic results of stenting interventions and adapt treatments to achieve target outcomes for patients. It will also enable generation of synthetic data for data-intensive applications, such as machine learning, to support cardiovascular research endeavors.

SDF4CHD: Generative modeling of cardiac anatomies with congenital heart defects.

Congenital heart disease (CHD) encompasses a spectrum of cardiovascular structural abnormalities, often requiring customized treatment plans for individual patients. Computational modeling and analysis of these unique cardiac anatomies can improve diagnosis and treatment planning and may ultimately lead to improved outcomes. Deep learning (DL) methods have demonstrated the potential to enable efficient treatment planning by automating cardiac segmentation and mesh construction for patients with normal cardiac anatomies. However, CHDs are often rare, making it challenging to acquire sufficiently large patient cohorts for training such DL models. Generative modeling of cardiac anatomies has the potential to fill this gap via the generation of virtual cohorts; however, prior approaches were largely designed for normal anatomies and cannot readily capture the significant topological variations seen in CHD patients. Therefore, we propose a type- and shape-disentangled generative approach suitable to capture the wide spectrum of cardiac anatomies observed in different CHD types and synthesize differently shaped cardiac anatomies that preserve the unique topology for specific CHD types. Our DL approach represents generic whole heart anatomies with CHD type-specific abnormalities implicitly using signed distance fields (SDF) based on CHD type diagnosis. To capture the shape-specific variations, we then learn invertible deformations to morph the learned CHD type-specific anatomies and reconstruct patient-specific shapes. After training with a dataset containing the cardiac anatomies of 67 patients spanning 6 CHD types and 14 combinations of CHD types, our method successfully captures divergent anatomical variations across different types and the meaningful intermediate CHD states across the spectrum of related CHD diagnoses. Additionally, our method demonstrates superior performance in CHD anatomy generation in terms of CHD-type correctness and shape plausibility. It also exhibits comparable generalization performance when reconstructing unseen cardiac anatomies. Moreover, our approach shows potential in augmenting image-segmentation pairs for rarer CHD types to significantly enhance cardiac segmentation accuracy for CHDs. Furthermore, it enables the generation of CHD cardiac meshes for computational simulation, facilitating a systematic examination of the impact of CHDs on cardiac functions.

Digital twinning of cardiac electrophysiology for congenital heart disease.

In recent years, blending mechanistic knowledge with machine learning has had a major impact in digital healthcare. In this work, we introduce a computational pipeline to build certified digital replicas of cardiac electrophysiology in paediatric patients with congenital heart disease. We construct the patient-specific geometry by means of semi-automatic segmentation and meshing tools. We generate a dataset of electrophysiology simulations covering cell-to-organ level model parameters and using rigorous mathematical models based on differential equations. We previously proposed Branched Latent Neural Maps (BLNMs) as an accurate and efficient means to recapitulate complex physical processes in a neural network. Here, we employ BLNMs to encode the parametrized temporal dynamics of in silico 12-lead electrocardiograms (ECGs). BLNMs act as a geometry-specific surrogate model of cardiac function for fast and robust parameter estimation to match clinical ECGs in paediatric patients. Identifiability and trustworthiness of calibrated model parameters are assessed by sensitivity analysis and uncertainty quantification.

Photorewriting, Time-Resolved Encryption, and Unclonable Anticounterfeiting with Artificial Intelligence Authentication via a Reversible Photoswitchable System.

In the fields of photolithographic patterning, optical anticounterfeiting, and information encryption, reversible photochromic materials with solid-state fluorescence are emerging as a potential class of systems. A design strategy for reversible photochromic materials has been proposed and synthesized through the introduction of photoactive thiophene groups into the molecular backbone of aryl vinyls, compounds with unique aggregation-induced emission properties, and solid-state reversible photocontrollable fluorescence and color-changing properties. This work develops novel photochromic inks, films, and cellulose hydrogels for enhancing the security of information encryption and anticounterfeiting technologies. They achieve rapid and reversible color change under ultraviolet light irradiation. Dependent upon the rate of color change, higher levels of time-resolved security can be achieved. This feature is important for enhancing the confidentiality of encrypted information and the reliability of security labels. Color-changing cellulose hydrogels, inks, and films consisting of three photochromic fluorescent molecules have quick photoactivity, great photoreversibility and photostability, and good processability, making them ideal for time-delayed anticounterfeiting and smart encryption. Furthermore, specialized algorithms are used to construct convolutional neural networks, and image analysis is performed on these systems, thus solving the current problem of the time-consuming information decryption process. This artificial intelligence method offers new opportunities for enhanced data encryption.

A wireless fluorescent sensing device for on-site closed-loop detection of hydrazine levels in the environment.

Given the extensive need for the detection of hydrazine (N2H4) in the biomedical and chemical-pharmaceutical sectors, there is a necessity to devise a fast, sensitive, specific, and portable technique for precisely quantifying hydrazine at environmental levels. In our work, an "OFF-ON" type fluorescent probe namely 2-(4-(10-(naphthalen-2-yl)anthracen-9-yl)phenyl)isoindole-1,3-dione (NAP), which was inspired by the "Gabriel" reaction, was synthesized. The NAP fluorescent cellulose film successfully achieved the detection of hydrazine vapor with a LOD = 0.658 ppm. Compared to previous qualitative methods for detecting hydrazine, this study successfully achieved quantitative identification of hydrazine at low concentrations. In addition, a portable sensor device based on NAP cellulose film was successfully integrated, enabling ultra-sensitive, wireless, remote, and real-time detection of N2H4 vapor. It was determined that the probe (NAP) exhibited excellent detection performance when applied to various environmental samples including distilled water, tap water, creek water, soil and plants. This study introduces a potentially effective approach for detecting hydrazine in real-world settings.

Virtual Shape-Editing of Patient-Specific Vascular Models Using Regularized Kelvinlets.

OBJECTIVE: Cardiovascular diseases, and the interventions performed to treat them, can lead to changes in the shape of patient vasculatures and their hemodynamics. Computational modeling and simulations of patient-specific vascular networks are increasingly used to quantify these hemodynamic changes, but they require modifying the shapes of the models. Existing methods to modify these shapes include editing 2D lumen contours prescribed along vessel centerlines and deforming meshes with geometry-based approaches. However, these methods can require extensive by-hand prescription of the desired shapes and often do not work robustly across a range of vascular anatomies. To overcome these limitations, we develop techniques to modify vascular models using physics-based principles that can automatically generate smooth deformations and readily apply them across different vascular anatomies. METHODS: We adapt Regularized Kelvinlets, analytical solutions to linear elastostatics, to perform elastic shape-editing of vascular models. The Kelvinlets are packaged into three methods that allow us to artificially create aneurysms, stenoses, and tortuosity. RESULTS: Our methods are able to generate such geometric changes across a wide range of vascular anatomies. We demonstrate their capabilities by creating sets of aneurysms, stenoses, and tortuosities with varying shapes and sizes on multiple patient-specific models. CONCLUSION: Our Kelvinlet-based deformers allow us to edit the shape of vascular models, regardless of their anatomical locations, and parametrically vary the size of the geometric changes. SIGNIFICANCE: These methods will enable researchers to more easily perform virtual-surgery-like deformations, computationally explore the impact of vascular shape on patient hemodynamics, and generate synthetic geometries for data-driven research.

LinFlo-Net: A Two-Stage Deep Learning Method to Generate Simulation Ready Meshes of the Heart.

We present a deep learning model to automatically generate computer models of the human heart from patient imaging data with an emphasis on its capability to generate thin-walled cardiac structures. Our method works by deforming a template mesh to fit the cardiac structures to the given image. Compared with prior deep learning methods that adopted this approach, our framework is designed to minimize mesh self-penetration, which typically arises when deforming surface meshes separated by small distances. We achieve this by using a two-stage diffeomorphic deformation process along with a novel loss function derived from the kinematics of motion that penalizes surface contact and interpenetration. Our model demonstrates comparable accuracy with state-of-the-art methods while additionally producing meshes free of self-intersections. The resultant meshes are readily usable in physics based simulation, minimizing the need for postprocessing and cleanup.

Digital twinning of cardiac electrophysiology for congenital heart disease.

In recent years, blending mechanistic knowledge with machine learning has had a major impact in digital healthcare. In this work, we introduce a computational pipeline to build certified digital replicas of cardiac electrophysiology in pediatric patients with congenital heart disease. We construct the patient-specific geometry by means of semi-automatic segmentation and meshing tools. We generate a dataset of electrophysiology simulations covering cell-to-organ level model parameters and utilizing rigorous mathematical models based on differential equations. We previously proposed Branched Latent Neural Maps (BLNMs) as an accurate and efficient means to recapitulate complex physical processes in a neural network. Here, we employ BLNMs to encode the parametrized temporal dynamics of in silico 12-lead electrocardiograms (ECGs). BLNMs act as a geometry-specific surrogate model of cardiac function for fast and robust parameter estimation to match clinical ECGs in pediatric patients. Identifiability and trustworthiness of calibrated model parameters are assessed by sensitivity analysis and uncertainty quantification.

SDF4CHD: Generative Modeling of Cardiac Anatomies with Congenital Heart Defects.

Congenital heart disease (CHD) encompasses a spectrum of cardiovascular structural abnormalities, often requiring customized treatment plans for individual patients. Computational modeling and analysis of these unique cardiac anatomies can improve diagnosis and treatment planning and may ultimately lead to improved outcomes. Deep learning (DL) methods have demonstrated the potential to enable efficient treatment planning by automating cardiac segmentation and mesh construction for patients with normal cardiac anatomies. However, CHDs are often rare, making it challenging to acquire sufficiently large patient cohorts for training such DL models. Generative modeling of cardiac anatomies has the potential to fill this gap via the generation of virtual cohorts; however, prior approaches were largely designed for normal anatomies and cannot readily capture the significant topological variations seen in CHD patients. Therefore, we propose a type- and shape-disentangled generative approach suitable to capture the wide spectrum of cardiac anatomies observed in different CHD types and synthesize differently shaped cardiac anatomies that preserve the unique topology for specific CHD types. Our DL approach represents generic whole heart anatomies with CHD type-specific abnormalities implicitly using signed distance fields (SDF) based on CHD type diagnosis, which conveniently captures divergent anatomical variations across different types and represents meaningful intermediate CHD states. To capture the shape-specific variations, we then learn invertible deformations to morph the learned CHD type-specific anatomies and reconstruct patient-specific shapes. Our approach has the potential to augment the image-segmentation pairs for rarer CHD types for cardiac segmentation and generate cohorts of CHD cardiac meshes for computational simulation.

A Conjugated Microporous Polymer/Wood Aerogel with Physical Adsorption, Chemical Degradation and Antibacterial Self-Cleaning Triple Sewage Treatment Functions.

Conjugated microporous polymers (CMPs) have important applications in the fields of optoelectronics and sewage treatment due to their high specific surface area, broad visible absorption, processability and simple synthesis process. Biocompatibility, recycling, mass production and solar photodegradation are particularly important in wastewater treatment. Here, A CMP with a high specific surface area and a hierarchical pore structure (CPOP) was constructed based on 4,4',4″-Tris(carbazol-9-yl)-triphenylamine (3CZ-TPA). Furthermore, a CMP-loaded wood aerogel (CPOP/wood aerogel) with physical adsorption, chemical degradation, bacterial inhibition and self-cleaning properties was prepared by in situ polymerization and used for wastewater treatment. The obtained CPOP/wood aerogel is highly biocompatible and easy to recycle. In addition, the inherent broad visible light absorption property of CPOP endows it with promising photocatalytic properties. Subsequently, we investigated the photocatalytic mechanism of CPOP, and the results showed that it was mainly affected by peroxyl radicals, which implied and confirmed its microbial self-cleaning for secondary cleaning of water pollutants. The reported studies on CPOP/wood aerogel provide a new direction for water purification materials with excellent adsorption, degradation and antibacterial properties.

Case report: A combined immunotherapy strategy as a promising therapy for MSI-H colorectal carcinomas with multiple HPD risk factors.

Approximately 5% of advanced colorectal carcinomas (CRCs) and 12-15% of early CRCs are microsatellite instability-high (MSI-H) or mismatch repair-deficient (dMMR) tumors. Nowadays, PD-L1 inhibitors or combined CTLA4 inhibitors are the major strategies for advanced or metastatic MSI-H colorectal cancer, but some people still show drug resistance or progression. Combined immunotherapy has been shown to expand the benefit population in non-small-cell lung carcinoma (NSCLC), hepatocellular carcinoma (HCC), and other tumors while reducing the incidence of hyper-progression disease (HPD). Nevertheless, advanced CRC with MSI-H remains rare. In this article, we describe a case of an elder patient with MSI-H advanced CRC carrying MDM4 amplification and DNMT3A co-mutation who responded to sintilimab plus bevacizumab and chemotherapy as the first-line treatment without obvious immune-related toxicity. Our case provides a new treatment option for MSI-H CRC with multiple risk factors of HPD and highlights the importance of predictive biomarkers in personalized immunotherapy.

Learning Whole Heart Mesh Generation From Patient Images for Computational Simulations.

Patient-specific cardiac modeling combines geometries of the heart derived from medical images and biophysical simulations to predict various aspects of cardiac function. However, generating simulation-suitable models of the heart from patient image data often requires complicated procedures and significant human effort. We present a fast and automated deep-learning method to construct simulation-suitable models of the heart from medical images. The approach constructs meshes from 3D patient images by learning to deform a small set of deformation handles on a whole heart template. For both 3D CT and MR data, this method achieves promising accuracy for whole heart reconstruction, consistently outperforming prior methods in constructing simulation-suitable meshes of the heart. When evaluated on time-series CT data, this method produced more anatomically and temporally consistent geometries than prior methods, and was able to produce geometries that better satisfy modeling requirements for cardiac flow simulations. Our source code and pretrained networks are available at https://github.com/fkong7/HeartDeformNets.

Controllable synthesis of conjugated microporous polymer films for ultrasensitive detection of chemical warfare agents.

Nerve agents, one of the most toxic chemical warfare agents, seriously threaten human life and public security. The high toxicity of nerve agents makes the development of fluorescence sensors with suitable limit of detection challenging. Here, we propose a sensor design based on a conjugated microporous polymer film for the detection of diethyl chlorophosphate, a substitute of Sarin, with low detection limit of 2.5 ppt. This is due to the synergy of the susceptible on-off effect of hybridization and de-hybridization of hybrid local and charge transfer (HLCT) materials and the microporous structure of CMP films facilitating the inward diffusion of DCP vapors, and the extended π-conjugated structure. This strategy provides a new idea for the future development of gas sensors. In addition, a portable sensor is successfully integrated based on TCzP-CMP films that enables wireless, remote, ultrasensitive, and real-time detection of DCP vapors.

A deep-learning approach for direct whole-heart mesh reconstruction.

Automated construction of surface geometries of cardiac structures from volumetric medical images is important for a number of clinical applications. While deep-learning-based approaches have demonstrated promising reconstruction precision, these approaches have mostly focused on voxel-wise segmentation followed by surface reconstruction and post-processing techniques. However, such approaches suffer from a number of limitations including disconnected regions or incorrect surface topology due to erroneous segmentation and stair-case artifacts due to limited segmentation resolution. We propose a novel deep-learning-based approach that directly predicts whole heart surface meshes from volumetric CT and MR image data. Our approach leverages a graph convolutional neural network to predict deformation on mesh vertices from a pre-defined mesh template to reconstruct multiple anatomical structures in a 3D image volume. Our method demonstrated promising performance of generating whole heart reconstructions with as good or better accuracy than prior deep-learning-based methods on both CT and MR data. Furthermore, by deforming a template mesh, our method can generate whole heart geometries with better anatomical consistency and produce high-resolution geometries from lower resolution input image data. Our method was also able to produce temporally-consistent surface mesh predictions for heart motion from CT or MR cine sequences, and therefore can potentially be applied for efficiently constructing 4D whole heart dynamics. Our code and pre-trained networks are available at https://github.com/fkong7/MeshDeformNet.

Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Segmentation: The M&Ms Challenge.

The emergence of deep learning has considerably advanced the state-of-the-art in cardiac magnetic resonance (CMR) segmentation. Many techniques have been proposed over the last few years, bringing the accuracy of automated segmentation close to human performance. However, these models have been all too often trained and validated using cardiac imaging samples from single clinical centres or homogeneous imaging protocols. This has prevented the development and validation of models that are generalizable across different clinical centres, imaging conditions or scanner vendors. To promote further research and scientific benchmarking in the field of generalizable deep learning for cardiac segmentation, this paper presents the results of the Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Segmentation (M&Ms) Challenge, which was recently organized as part of the MICCAI 2020 Conference. A total of 14 teams submitted different solutions to the problem, combining various baseline models, data augmentation strategies, and domain adaptation techniques. The obtained results indicate the importance of intensity-driven data augmentation, as well as the need for further research to improve generalizability towards unseen scanner vendors or new imaging protocols. Furthermore, we present a new resource of 375 heterogeneous CMR datasets acquired by using four different scanner vendors in six hospitals and three different countries (Spain, Canada and Germany), which we provide as open-access for the community to enable future research in the field.

Automating Model Generation for Image-Based Cardiac Flow Simulation.

Computational fluid dynamics (CFD) modeling of left ventricle (LV) flow combined with patient medical imaging data has shown great potential in obtaining patient-specific hemodynamics information for functional assessment of the heart. A typical model construction pipeline usually starts with segmentation of the LV by manual delineation followed by mesh generation and registration techniques using separate software tools. However, such approaches usually require significant time and human efforts in the model generation process, limiting large-scale analysis. In this study, we propose an approach toward fully automating the model generation process for CFD simulation of LV flow to significantly reduce LV CFD model generation time. Our modeling framework leverages a novel combination of techniques including deep-learning based segmentation, geometry processing, and image registration to reliably reconstruct CFD-suitable LV models with little-to-no user intervention.1 We utilized an ensemble of two-dimensional (2D) convolutional neural networks (CNNs) for automatic segmentation of cardiac structures from three-dimensional (3D) patient images and our segmentation approach outperformed recent state-of-the-art segmentation techniques when evaluated on benchmark data containing both magnetic resonance (MR) and computed tomography(CT) cardiac scans. We demonstrate that through a combination of segmentation and geometry processing, we were able to robustly create CFD-suitable LV meshes from segmentations for 78 out of 80 test cases. Although the focus on this study is on image-to-mesh generation, we demonstrate the feasibility of this framework in supporting LV hemodynamics modeling by performing CFD simulations from two representative time-resolved patient-specific image datasets.

Finite element analysis of MitraClip procedure on a patient-specific model with functional mitral regurgitation.

Mitral valve (MV) repair with the MitraClip device has been shown to reduce mitral regurgitation severity and improve clinical outcomes in symptomatic patients at high surgical risk. MitraClip was recently approved in the US for the treatment of functional mitral regurgitation (FMR), which significantly expands the number of patients that can be treated with this device. This study aims to quantify the morphologic changes and evaluate the biomechanical interaction between the MitraClip device and the mitral apparatus of a real patient case with FMR using computational modeling. MitraClip procedures using a central and a lateral clip were simulated in a validated MV-left ventricle finite element (FE) model with severe MR. The patient-specific model integrated detailed geometries of the left ventricle, mitral leaflets and chordae, incorporated age- and gender-matched nonlinear hyperelastic human material properties, and accounted for chordae tethering forces. Central and lateral positioning gave similar biomechanical outcomes resulting in an improved but incomplete MV coaptation. Antero-posterior distance, annulus area, valve opening orifice area, and regurgitant orifice area decreased by up to 26%, 19%, 48% and 63% when compared to the pre-clip model, respectively. Anterior and posterior leaflet peak stresses increased by up to 64% and 62% after clip placement, respectively, and were located at the region of clip grasp. Similarly, anterior and posterior leaflet peak strains increased by up to 20% and 10%, respectively. FE modeling, as used here, can be a powerful tool to examine the complex MitraClip-host biomechanical interaction.

Finite element analysis of annuloplasty and papillary muscle relocation on a patient-specific mitral regurgitation model.

OBJECTIVES: Functional mitral regurgitation (FMR) is a significant complication of left ventricle (LV) dysfunction associated with poor prognosis and commonly treated by undersized ring annuloplasty. This study aimed to quantitatively simulate the treatment outcomes and mitral valve (MV) biomechanics following ring annulopalsty and papillary muscle relocation (PMR) procedures for a FMR patient. METHODS: We utilized a validated finite element model of the left heart for a patient with severe FMR and LV dilation from our previous study and simulated virtual ring annuloplasty procedures with various sizes of Edwards Classic and GeoForm annuloplasty rings. The model included detailed geometries of the left ventricle, mitral valve, and chordae tendineae, and incorporated age- and gender- matched nonlinear, anisotropic hyperelastic tissue material properties, and simulated chordal tethering at diastole due to LV dilation. RESULTS: Ring annuloplasty with either the Classic or GeoForm ring improved leaflet coaptation and increased the total leaflet closing force while increased posterior mitral leaflet (PML) stresses and strains. Classic rings resulted in larger coaptation forces and areas compared to GeoForm rings. The PMR procedure further improved the leaflet coaptation, decreased the PML stress and strain for both ring shapes and all sizes in this patient model. CONCLUSIONS: This study demonstrated that a rigorously developed patient-specific computational model can provide useful insights into annuloplasty repair techniques for the treatment of FMR patients and could potentially serve as a tool to assist in pre-operative planning for MV repair surgical or interventional procedures.

Finite Element Analysis of Tricuspid Valve Deformation from Multi-slice Computed Tomography Images.

Despite the growing clinical interest in the tricuspid valve (TV), there is an incomplete understanding of TV biomechanics which is important in normal TV function and successful TV repair techniques. Computational models with patient-specific human TV geometries can provide a quantitative understanding of TV biomechanic. Therefore, this study aimed to develop finite element (FE) models of human TVs from multi-slice computed tomography (MSCT) images to investigate chordal forces and leaflet stresses and strains. Three FE models were constructed for human subjects with healthy TVs from MSCT images and incorporated detailed leaflet geometries, realistic nonlinear anisotropic hyperelastic material properties of human TV, and physiological boundary conditions tracked from MSCT images. TV closure from diastole to systole was simulated. Chordal lengths were iteratively adjusted until the simulated TV geometries were in good agreement with the "true" geometries reconstructed from MSCT images at systole. Larger chordal forces were found on the strut (or basal) chords than on the rough zone chords and the total forces applied on the anterior papillary muscles by the strut chords were higher than those on the posterior or septal papillary muscles. At peak systolic pressure, the average maximum stress on the middle sections of the leaflets ranged from 30 to 90 kPa, while the average maximum principal strain values ranged from 0.16 to 0.30. The results from healthy TVs can serve as baseline biomechanical metrics of TV mechanics and may be used to inform TV repair device design. The computational approach developed could be one step towards developing computational models that may support pre-operative planning in complex TV repair procedures in the future.