3D Heart Research Articles

Multipoint detection of structural deformation of pulsating 3D heart model using flexible organic piezoelectric-sensor array

Multipoint detection of the structural deformation of a pulsating 3D heart model was conducted using a flexible piezoelectric sensor array with a poly(vinylidene fluoride-trifluoroethylene) [P(VDF-TrFE)] thin film sealed with biocompatible parylene C. The piezoelectric signals detected from the sensor array attached to the 3D heart model were correlated with an electrocardiogram signal. These piezoelectric signals can be converted into the compressive stress applied to the flexible sensors by the pulsating motion of the 3D heart model. The experimental results mean that the contraction force, generated by the heart muscle (i.e., myocardium), can be directly evaluated. The different output voltages depending on the location were obtained from each sensor, indicating the possibility of a real-time detection of the irregular motion of the heart and the early detection of ischemic heart disease, which leads to the loss of local contractile force in myocardial tissue.

Are 3D printed models acceptable in assessment?

BackgroundThree‐dimensional (3D) printed models are increasingly used in undergraduate anatomy teaching. However, their role and value in anatomy assessment remains under consideration. The aim of this study was to evaluate student and educator perspectives on acceptability of using novel 3D printed heart models for assessment.MethodsWe used printed 3D models of the heart for first‐year medical students, in small group teaching, formative assessment and revision at home. We adopted a mixed methods approach involving questionnaires, then focus groups to collect student and educator views. We used QSR Nvivo to manage thematic analysis of responses, carried out by student and educators, respectively.FindingsOverall, students 89% (n = 75/84) and educators 91% (n = 10/11) found the assessment acceptable. Thematic analysis of focus groups (n = 4 students, n = 5 educators) identified five key perceptions shared across student and educator groups: 3D models are the future, realism is valued, models appear feasible, consistent and provide a potential for a range of applications in assessment.DiscussionThere was agreement between educators and students that the use of 3D heart models was acceptable. Key recognised benefits include accessibility and consistency across settings, made more relevant in the current COVID‐19 pandemic. We recommend integration of 3D models into teaching and assessment for educational alignment and careful selection of anatomy to model. Further research is required to explore the use of models in summative assessments.

FRESH 3D bioprinting a contractile heart tube using human stem cell-derived cardiomyocytes

Here we report the 3D bioprinting of a simplified model of the heart, similar to that observed in embryonic development, where the heart is a linear tube that pumps blood and nutrients to the growing embryo. To this end, we engineered a bioinspired model of the human heart tube using freeform reversible of embedding of suspended hydrogels 3D bioprinting. The 3D bioprinted heart tubes were cellularized using human stem cell-derived cardiomyocytes and cardiac fibroblasts and formed patent, perfusable constructs. Synchronous contractions were achieved ∼3–4 days after fabrication and were maintained for up to a month. Immunofluorescent staining confirmed large, interconnected networks of sarcomeric alpha actinin-positive cardiomyocytes. Electrophysiology was assessed using calcium imaging and demonstrated anisotropic calcium wave propagation along the heart tube with a conduction velocity of ∼5 cm s−1. Contractility and function was demonstrated by tracking the movement of fluorescent beads within the lumen to estimate fluid displacement and bead velocity. These results establish the feasibility of creating a 3D bioprinted human heart tube and serve as an initial step towards engineering more complex heart muscle structures.

Respiration induced changes and their impact on universal and tailored 3D kT-point parallel transmission pulses for 7T cardiac imaging.

Human heart imaging at ultra-high fields is highly challenging because of respiratory motion-induced artefacts and spatially heterogeneous profiles. This work demonstrates that respiration resolved 3D -maps can be used with a dedicated tailored and universal parallel transmission (pTx) pulse design to compensate respiration related changes in subjects performing shallow and deep breathing (SB/DB). Three-dimensional (3D) -maps of the thorax were acquired in 31 subjects under SB and in 15 subjects under SB and DB. Different universal and tailored non-selective pTx pulses were designed from non-respiration resolved (NRR) and respiration resolved (RR) reconstructions of the SB/DB -maps. The performance of all pulses was tested with RR-SB/DB -maps. Respiration-robust tailored and universal pulses were applied in vivo in 5 subjects at 7T in 3D gradient-echo free-breathing scans. All optimized pTx pulses performed well for SB. For DB, however, only the universal and the tailored respiration-robust pulses achieved homogeneous flip angles (FAs) in all subjects and across all respiration states, whereas the tailored respiration-specific pulses resulted in a higher FA variation. The respiration-robust universal pulse resulted in an average coefficient of variation in the FA maps of 12.6% compared to 8.2% achieved by tailored respiration-robust pulses. In vivo measurements at 7T demonstrate the benefits of using respiration-robust pulses for DB. Universal and tailored respiration-robust pTx pulses based on RR -maps are highly preferred to achieve 3D heart FA homogenization at 7T when subjects perform DB, whereas universal and tailored pulses based on NRR -maps are sufficient when subjects perform SB.

The role of 3D printed heart models in immediate and long-term knowledge acquisition in medical education.

The long-term effect of three-dimensional printed heart model (3DPHM) on knowledge acquisition of congenital heart disease (CHD) remains unknown. This prospective cohort study aims to investigate the role of 3DPHM in improving immediate knowledge gain and long-term knowledge retention on CHD among the medical students. Fifty-three second and third year medical students were assigned into two groups to compare their immediate knowledge acquisition and knowledge retention after an education session on anatomy and pathophysiology of CHD. During the 1.5 hour-long education session, both the control (n = 25) and study groups (n = 28) had access to identical teaching materials: digital 3D heart models, 2D diagrams, and medical images, except for 3DPHM which were only used in the study group. The immediate knowledge gain was assessed via an online quiz, whereas the long-term knowledge retention was assessed using another quiz in 6-weeks' time post-intervention. A survey was also conducted to evaluate the participants' learning experience. There is no significant difference in the immediate knowledge acquisition and long-term knowledge retention between the groups (U = 272, p = 0.16 and r = -0.143, p = 0.15 respectively). Majority of the students (96% in control group and 85% in 3DPHM group) responded that the 3DPHM would have/had improved their learning experience. Despite that, there is no significant difference in the self-perceived knowledge improvement between the groups. This study concludes that the 3DPHM do not significantly improve both immediate knowledge acquisition and knowledge retention among the medical students. However, further research with larger sample size, as well as categorizing the type of questions in the quiz, is needed to better assess the role of 3DPHM in different educational components.

Augmented Reality 3D Heart as Learning Media at Midwifery Lab University of 'Aisyiyah Surakarta

Abstract— Utilization of Information Technology (IT) in the process of delivering information has been developed rapidly. Currently, the latest technology used in the delivery of information is Augmented Reality (AR) technology.. Pada teknologi Augmented Reality (AR), pengguna dapat menvisualisasikan objek atau benda bersejarah dalam bentuk 3 dimensi. Augmented Reality (AR) has the advantage of being interactive and real time so Augmented Reality (AR) is widely implemented in the field of education as a medium to introduce historical objects which are cultural heritage or objects related to a science..Augmented Reality (AR) has the advantage of being interactive and real time, so Augmented Reality (AR) is widely implemented in various fields. Based on the explanation of the study, researchers are interested in conducting research using information technology in the health sector, by making Augmented Reality (AR) modeling to facilitate the community and midwifery students in providing learning about human organs, especially the heart which can be used digitally without having to dissect human organs and bringing him a modeling of the heart with the presence of a 3D heart organ provides convenience in learning to the public about the function and structure of the heart digitally.Keywords :3D, Augmented Reality, Heart 3D Design.

In silico screening method for non-responders to cardiac resynchronization therapy in patients with heart failure: a pilot study

BackgroundCardiac resynchronization therapy (CRT) is an effective treatment option for patients with heart failure (HF) and left ventricular (LV) dyssynchrony. However, the problem of some patients not responding to CRT remains unresolved. This study aimed to propose a novel in silico method for CRT simulation.MethodsThree-dimensional heart geometry was constructed from computed tomography images. The finite element method was used to elucidate the electric wave propagation in the heart. The electric excitation and mechanical contraction were coupled with vascular hemodynamics by the lumped parameter model. The model parameters for three-dimensional (3D) heart and vascular mechanics were estimated by matching computed variables with measured physiological parameters. CRT effects were simulated in a patient with HF and left bundle branch block (LBBB). LV end-diastolic (LVEDV) and end-systolic volumes (LVESV), LV ejection fraction (LVEF), and CRT responsiveness measured from the in silico simulation model were compared with those from clinical observation. A CRT responder was defined as absolute increase in LVEF ≥ 5% or relative increase in LVEF ≥ 15%.ResultsA 68-year-old female with nonischemic HF and LBBB was retrospectively included. The in silico CRT simulation modeling revealed that changes in LVEDV, LVESV, and LVEF by CRT were from 174 to 173 mL, 116 to 104 mL, and 33 to 40%, respectively. Absolute and relative ΔLVEF were 7% and 18%, respectively, signifying a CRT responder. In clinical observation, echocardiography showed that changes in LVEDV, LVESV, and LVEF by CRT were from 162 to 119 mL, 114 to 69 mL, and 29 to 42%, respectively. Absolute and relative ΔLVESV were 13% and 31%, respectively, also signifying a CRT responder. CRT responsiveness from the in silico CRT simulation model was concordant with that in the clinical observation.ConclusionThis in silico CRT simulation method is a feasible technique to screen for CRT non-responders in patients with HF and LBBB.

Analysis of mechanical bi-leaflet heart valve made of biomaterial-polypropylene

The heart has four major valves, which are responsible for directing blood flow and producing a pulse. Using reference geometry from [1] Krzysztof Jozwik, Artificial Heart Valves Designs, Medical Engineering & Physics, March 2011, we designed a mechanical bi-leaflet heart valve. The main goal of this article is to conduct an examination of a 3D printable mechanical heart valve made of polypropylene (PP) biomaterial and to compare the findings with other 3D printing materials such as Acrylonitrile Butadiene Styrene (ABS).

The Reliability of Automated Three-Dimensional Echocardiography-HeartModelA.I. Versus 2D Echocardiography Simpson Methods in Evaluation of Left Ventricle Volumes and Ejection Fraction in Patients With Left Ventricular Dysfunction.

Background:Two-dimensional echocardiography (2DE) Simpson methods is the most frequently used imaging modality to assess Left ventricular ejection fraction (LVEF). LVEF is an important predictor of morbidity and mortality in a wide range of patients and clinical scenarios. Despite its importance in prognosis and clinical decision making, most echocardiography laboratories currently determine EF primarily by visual estimation, which is highly experience-dependent and sensitive to intra- and inter-observer variability and suboptimal accuracy and repeatability. Over the last decade, 3-dimensional echocardiography (3DE) has become increasingly implemented in clinical practice. The automated 3D HeartModelA.I. tracks every frame over the cardiac cycle using 3D speckle technology. HeartModelA.I. is a fully automated program that simultaneously detects LA and LV endocardial surfaces using an adaptive analytics algorithm that consists of knowledge-based identification of initial global shape and orientation followed by patient-specific adaptation.Objective:The objective of the study was to compare the automated 3D HeartModelA.I echocardiography and 2D Simpson methods echocardiography in evaluation of the left ventricular ejection fraction and left ventricular volumes in patients with left heart dysfunction.Methods:The study prospectively enrolled 165 patients with symptoms of LV dysfunction (ischemic or nonischemic) and New York Heart Association (NYHA) functional class I-III, referred for an echocardiographic study to evaluate the LV volumes and LV ejection fraction (LVEF) during the period from March 2020 to March 2022. Echocardiographic images were acquired by experienced echocardiographers using a commercially available Philips EPIQ machine (Koninklijke Philips Ultrasound, USA) equipped with X5-1 Matrix probe for 2DE and DHM 3DE acquisitions, respectively.Results:2D Simpson methods echocardiography results for estimated LVEF were 38.43 ± 1.70 in patients with NYHA class I-II, 30.53 ± 1.60 in patients with NYHA class III. Using 3D Heart Model, LVEF were 38.23 ± 1.71 in patients with NYHA class I-II and 30.27 ± 1.50 in patients with NYHA class III. The results of 2D Simpson methods echocardiography for estimated LVEDVi in NYHA class I-II and NYHA class III were 99.06 ± 6.36 ml/m2, 121.96 ± 2.93 ml/m2 respectively, LVESVi were 60.91 ± 3.91 ml/m2, 84.74 ± 2.70 ml/m2 respectively, for 3D Heart Model, LVEDVi in NYHA class I-II and NYHA class III were 100.07 ± 6.72, 121.38 ± 3.01 ml/m2 respectively, LVESVi were 61.75 ± 3.94 ml/m2, 84.73 ± 2.33 ml/m2 respectively. 2DE measurement of LV volumes and EF was completed in 6.1 ± 0.8 min. per patient. 3DE HeartModelA.I acquisition and analysis in most patients was completed in <3.2 min., an average time of 2.9 ± 1.3 min. per patient. The result of our study shows that the 3D HeartModelA.I. is a reliable and robust method for LVEF and LV volume analysis, which has similar results to 2D echocardiography performed by experienced sonographers. In this study, we found that 3DE DHM fully automated tool is also significantly faster than 2DE analysis and thus can help overcome the time-consuming nature and its present a strong argument for its incorporation into the clinical workflow. In this study, we found that 3DE DHM fully automated tool is also significantly faster than 2DE analysis and thus can help overcome the time-consuming nature and its present a strong argument for its incorporation into the clinical workflow.Conclusion:3D DHM provides fast and accurate LV volumes and LVEF quantitation, as it avoids geometric assumptions and left ventricular foreshortening, has better reproducibility and has incremental value to predict adverse outcomes in comparison with conventional 2DE. In the future major benefit of AI in echocardiography is expected from improvements in automated analysis and interpretation to reduce workload and improve clinical outcome.

Assessment of superior cavo-pulmonary anastomoses (Glenn shunt) by cardiac magnetic resonance imaging in comparison with multi-slice computed tomography

BackgroundMulti-slice computed tomography (MSCT) angiography is the gold standard imaging modality to evaluate the patency of Glenn shunt and the presence of veno–veno collaterals. The goal of this study is to evaluate the ability of two cardiac magnetic resonance imaging (MRI) techniques to assess the patency of Glenn shunt and the presence of veno–veno collaterals compared to MSCT angiography.ResultsPatients with Glenn shunt had MSCT angiography and cardiac MRI using two techniques: TWIST (Time-resolved angiography With Stochastic Trajectories) and the three-dimensional (3D) post-contrast whole heart techniques. MSCT angiography and cardiac MRI images were post-processed for quantitative and qualitative assessment of Glenn shunt and veno–veno collaterals. Our study included 29 patients (17 male, 59%) with Glenn shunt, the median age was 22 years (range 3–36 years). 3D post-contrast whole heart images give similar results compared to MSCT angiography results in the evaluation of Glenn shunt and veno–veno collaterals, 100% agreement in Glenn shunt visualization and agreement was 86.2% in the detection of veno–veno collaterals with a perfect agreement (kappa = 1) as regards their proximal connection to superior vena cava (SVC). While TWIST showed lower agreement compared to MSCT angiography results, 87.5% agreement in Glenn shunt visualization and agreement was 68.9% in the detection of veno–veno collaterals with poor agreement (kappa = 0.266) as regards their proximal connection to SVC.Conclusions3D post-contrast whole heart MRI images have similar results as MSCT angiography in the evaluation of superior cavo-pulmonary anastomosis and can be a good and safer alternative to MSCT angiography.

Hemodynamic characteristics expose the atherosclerotic severity in coronary main arteries: One-dimensional and three-dimensional approaches

Computed tomography coronary angiography image-based noninvasive virtual fractional flow reserve (vFFR) is a promising clinical practice to determine the physiological condition of coronary lesions. vFFR is the key factor in the diagnosis of coronary plaque. The purpose of this study is to detect the coronary main arteries lesion severity by using 1D (one-dimensional) hemodynamic factors compared to 3D (three-dimensional) heart flow computational models. The results provide the hemodynamic properties of the intraluminal condition by evaluating the vFFR. The computational burden of 3D hemodynamic simulations is one of the important drawbacks in most clinical cases. At first, we have established comparative results between vFFR3D (vFFR in 3D) and published results in the maximal hyperemic flow conditions. Then, we have employed statistical analysis including Pearson correlation test, Bland–Altman test, and computation time test for vFFR3D compared with the vFFR1D (vFFR in 1D) results. We have observed that the vFFR3D and vFFR1D results do not significantly differ as a function of stenosis length, type (concentric or eccentric), or location in the coronary artery. Pearson's product moment was found as r=0.9661,p&amp;lt;0.0001 illustrating a strong correlation between vFFR1D and vFFR3D. In both 3D and 1D cases, the results suggest that proximal stenosis is more severe compared to a distal one, even if they provide the same reduction in lumen (80% or 90% area of stenosis). The 1D inexpensive computational results vFFR1D can be used to predict the severity of atherosclerotic plaque in clinical procedures.

AI-CHD

3D heart modeling and AI bring new cardiac surgery to remote and less-developed regions.

Echocardiography-computed tomography fusion imaging: a new approach for congenital heart disease

Abstract Introduction Diagnosis, management and surgical decision making in children and adults with congenital heart disease are based on echocardiography. One of the most recent developments in cardiac imaging is “fusion” of different imaging modalities. Our objective was to evaluate the feasibility of imaging fusion between computed tomography (CT) and 3D trans-thoracic echocardiography (TTE) in children and adults with congenital heart disease. Methods Fourteen patients with congenital heart disease who had a CT and a 3D TTE (9.5 years [2.7–15.7], 57% male, BSA 0.9 m2 [0.6–1.7]) were prospectively included for their usual monitoring in our center. Congenital heart disease was classified as simple (n=4, 29%), moderate (n=4, 29%) or complex (n=6, 42%). A full-volume 3D heart acquisition on three cycles was performed using a VividTM E95 GE. Image fusion between CT and TTE was done using General Electrics (GE) Fusion Software. Steps required for images importation, for fusion process (alignment, landmarks, superposition), navigation and analysis were described. Results Fusion imaging between 3D TTE and CT was successful for all patients and superposition was good with a median scored at 8.5/10. Median total time required to complete the fusion process was 735 seconds [628–1163], without significant difference according to the degree of complexity of the cardiopathy. Landmarks were significantly placed differently from the reference landmarks in complex congenital heart disease. For complex cardiopathy, navigation into the heart at the same time with multiple modalities demonstrate its interest, for diagnosis or for planning surgery. Conclusion We have investigated the feasibility and the accuracy of imaging fusion between 3D TTE and CT in a varied population of patients with congenital heart disease, from children to adult. These advances allow simultaneous visualization of intracardiac structures and may help in understanding the complex anatomy of congenital heart disease, without limitations in terms of age, weight or type of congenital heart disease. Funding Acknowledgement Type of funding sources: None. Validation of fusion imagingNavigation in fusion imaging

Asymptomatic Giant Left Atrial Appendage Cavernous Haemangiona Misdiagnosed as a Mediastinal Mass

A 50-year-old female was admitted because a chest computed tomography (CT) incidentally noted a mediastinal mass. Physical examination revealed a regular heart rate of 84 beats/minute and blood pressure of 124/88 mmHg. Moreover, no cardiac murmurs were auscultated. Preoperative contrast-enhanced chest CT demonstrated an 8.2 cm × 6.0 cm non-enhancing mass in the left anterior mediastinum (Figure 1A). No abnormalities were illustrated in the 3D heart vasculature reconstruction (Figure 1B) and echocardiogram (Figure 1C).

Real-Time Interactive Simulations of Complex Ionic Cardiac Cell Models in 2D and 3D Heart Structures with GPUs on Personal Computers.

Cardiac modeling in heart structures for the study of arrhythmia mechanisms requires the use of software that runs on supercomputers. Therefore, computational studies are limited to groups with access to computer clusters and personnel with high-performance computing experience. We present how to use and implement WebGL programs via a custom-written library to run and visualize simulations of the most complex ionic models in 2D and 3D, in real time, interactively using the multi-core GPU of a single computer. We use Abubu.js, a library we developed for solving partial differential equations such as those describing crystal growth and fluid flow, along with a newly implemented visualization algorithm, to simulate complex ionic cell models. By combining this library with JavaScript, we allow direct real-time interactions with simulations. We implemented: 1) modification of any model parameters and equations at any time, with direct access to the code while it runs, 2) electrode stimulation anywhere in the 2D/3D tissue with a mouse click, 3) saving the solution of the system at any time to re-initiate the dynamics from saved initial conditions, and 4) rotation/visualization of 3D structures at any angle. As examples of this modeling platform, we implemented a phenomenological cell model and the human ventricular OVVR model (41 variables). In 2D we illustrate the dynamics in an annulus, disk, and square tissue; in 2D and 3D porcine ventricles, we show the initiation of functional/anatomical reentry, spiral wave dynamics in different regimes, initiation of early afterdepolarizations (EADs), and the effects of model parameter variations in real time. We present the first simulations of complex models in anatomical structures with enhanced visualization and extended interactivity that run on a single PC, without software downloads, and as fast as in real-time even for 3D full ventricles.

Coronary Magnetic Resonance Angiography in Chronic Coronary Syndromes.

Cardiovascular disease is the leading cause of mortality worldwide, with atherosclerotic coronary artery disease (CAD) accounting for the majority of cases. X-ray coronary angiography and computed tomography coronary angiography (CCTA) are the imaging modalities of choice for the assessment of CAD. However, the use of ionising radiation and iodinated contrast agents remain drawbacks. There is therefore a clinical need for an alternative modality for the early identification and longitudinal monitoring of CAD without these associated drawbacks. Coronary magnetic resonance angiography (CMRA) could be a potential alternative for the detection and monitoring of coronary arterial stenosis, without exposing patients to ionising radiation or iodinated contrast agents. Further advantages include its versatility, excellent soft tissue characterisation and suitability for repeat imaging. Despite the early promise of CMRA, widespread clinical utilisation remains limited due to long and unpredictable scan times, onerous scan planning, lower spatial resolution, as well as motion related image quality degradation. The past decade has brought about a resurgence in CMRA technology, with significant leaps in image acceleration, respiratory and cardiac motion estimation and advanced motion corrected or motion-resolved image reconstruction. With the advent of artificial intelligence, great advances are also seen in deep learning-based motion estimation, undersampled and super-resolution reconstruction promising further improvements of CMRA. This has enabled high spatial resolution (1 mm isotropic), 3D whole heart CMRA in a clinically feasible and reliable acquisition time of under 10 min. Furthermore, latest super-resolution image reconstruction approaches which are currently under evaluation promise acquisitions as short as 1 min. In this review, we will explore the recent technological advances that are designed to bring CMRA closer to clinical reality.

Combining patient-specific, digital 3D models with tele-education for adolescents with CHD.

Adolescents with CHD require transition to specialised adult-centred care. Previous studies have shown that adolescents' knowledge of their medical condition is correlated with transition readiness. Three-dimensional printed models of CHD have been used to educate medical trainees and patients, although no studies have focused on adolescents with CHD. This study investigates the feasibility of combining patient-specific, digital 3D heart models with tele-education interventions to improve the medical knowledge of adolescents with CHD. Adolescent patients with CHD, aged between 13 and 18 years old, were enrolled and scheduled for a tele-education session. Patient-specific digital 3D heart models were created using images from clinically indicated cardiac magnetic resonance studies. The tele-education session was performed using commercially available, web-conferencing software (Zoom, Zoom Video Communications Inc.) and a customised software (Cardiac Review 3D, Indicated Inc.) incorporating an interactive display of the digital 3D heart model. Medical knowledge was assessed using pre- and post-session questionnaires that were scored by independent reviewers. Twenty-two adolescents completed the study. The average age of patients was 16 years old (standard deviation 1.5 years) and 56% of patients identified as female. Patients had a variety of cardiac defects, including tetralogy of Fallot, transposition of great arteries, and coarctation of aorta. Post-intervention, adolescents' medical knowledge of their cardiac defects and cardiac surgeries improved compared to pre-intervention (p < 0.01). Combining patient-specific, digital 3D heart models with tele-education sessions can improve adolescents' medical knowledge and may assist with transition to adult-centred care.

Artificial Intelligence Based Framework to Quantify the Cardiomyocyte Structural Integrity in Heart Slices.

Drug induced cardiac toxicity is a disruption of the functionality of cardiomyocytes which is highly correlated to the organization of the subcellular structures. We can analyze cellular structures by utilizing microscopy imaging data. However, conventional image analysis methods might miss structural deteriorations that are difficult to perceive. Here, we propose an image-based deep learning pipeline for the automated quantification of drug induced structural deteriorations using a 3D heart slice culture model. In our deep learning pipeline, we quantify the induced structural deterioration from three anticancer drugs (doxorubicin, sunitinib, and herceptin) with known adverse cardiac effects. The proposed deep learning framework is composed of three convolutional neural networks that process three different image sizes. The results of the three networks are combined to produce a classification map that shows the locations of the structural deteriorations in the input cardiac image. The result of our technique is the capability of producing classification maps that accurately detect drug induced structural deterioration on the pixel level. This technology could be widely applied to perform unbiased quantification of the structural effect of the cardiotoxins on heart slices.

Spectral augmentation for heart chambers segmentation on conventional contrasted and unenhanced CT scans: an in-depth study.

Recently, machine learning has outperformed established tools for automated segmentation in medical imaging. However, segmentation of cardiac chambers still proves challenging due to the variety of contrast agent injection protocols used in clinical practice, inducing disparities of contrast between cavities. Hence, training a generalist network requires large training datasets representative of these protocols. Furthermore, segmentation on unenhanced CT scans is further hindered by the challenge of obtaining ground truths from these images. Newly available spectral CT scanners allow innovative image reconstructions such as virtual non-contrast (VNC) imaging, mimicking non-contrasted conventional CT studies from a contrasted scan. Recent publications have demonstrated that networks can be trained using VNC to segment contrasted and unenhanced conventional CT scans to reduce annotated data requirements and the need for annotations on unenhanced scans. We propose an extensive evaluation of this statement. We undertake multiple trainings of a 3D multi-label heart segmentation network with (HU-VNC) and without (HUonly) VNC as augmentation, using decreasing training dataset sizes (114, 76, 57, 38, 29, 19 patients). At each step, both networks are tested on a multi-vendor, multi-centric dataset of 122 patients, including different protocols: pulmonary embolism (PE), chest-abdomen-pelvis (CAP), heart CT angiography (CTA) and true non-contrast scans (TNC). An in-depth comparison of resulting Dice coefficients and distance metrics is performed for the networks trained on the largest dataset. HU-VNC-trained on 57 patients significantly outperforms HUonly trained on 114 regarding CAP and TNC scans (mean Dice coefficients of 0.881/0.835 and 0.882/0.416, respectively). When trained on the largest dataset, significant improvements in all labels are noted for TNC and CAP scans (mean Dice coefficient of 0.882/0.416 and 0.891/0.835, respectively). Adding VNC images as training augmentation allows the network to perform on unenhanced scans and improves segmentations on other imaging protocols, while using a reduced training dataset.

B-PO05-048 INSERTION OF CORONARY SINUS LEAD UTILIZING INTRACARDIAC ECHOCARDIOGRAM, 3D ELECTRO-ANATOMICAL MAPPING, 3D PRINTED HEART MODEL AND FLUOROSCOPY IN A PATIENT WITH MARFAN SYNDROME AND PECTUS EXCAVATUM

A 30-year-old woman with a history of pectus excavatum and Marfan syndrome s/p Bentall procedure 7 years prior developed complete heart block after aortic mechanical prosthesis and underwent PPM implant with one RA and two RV leads (RV apex and RVOT) in lieu of failed traditional CRT. She developed a reduced ejection fraction to 30% and was referred to our institution for CRT-D upgrade. Describe a case of CS lead insertion in a patient with complex anatomy. N/A CRT-P system and two RV leads were extracted using laser lead extraction. A CS lead could not be placed despite an extensive traditional approach. A cardiac CT was performed in order to better define CS anatomy, a 3D heart model was printed for procedural planning and CT images were merged with Carto 3D electro-anatomical system. Reviewing the model and images above, a significant cranial displacement of the CS ostium was noted with an acute angulation of the proximal CS. We proceeded with 3D anatomical mapping, ICE, and fluoroscopic guided insertion of a decapolar® catheter in CS via right femoral vein which was then exchanged for a glide wire in order to mechanically straighten the angulation in proximal CS. Then, using a long sheath from the left subclavian approach, the ostium of CS was engaged. The sheath was cautiously advanced under fluoroscopy and Carto mapping using a Bard quadripolar catheter. CS lead was inserted in a posterolateral branch with a recorded RV-LV separation of 150 msec. Multi-modality imaging and mapping can help insert CS leads in patients with prior unsuccessful attempts at cardiac synchronization therapy.