If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The customized vascular graft offers the potential to simplify the surgical procedure, optimize physiological function, and reduce morbidity and mortality. This experiment evaluated the feasibility of a flow dynamic–optimized branched tissue engineered vascular graft (TEVG) customized based on medical imaging and manufactured by 3-dimensional (3D) printing for a porcine model.
We acquired magnetic resonance angiography and 4-dimensional flow data for the native anatomy of the pigs (n = 2) to design a custom-made branched vascular graft of the pulmonary bifurcation. An optimal shape of the branched vascular graft was designed using a computer-aided design system informed by computational flow dynamics analysis. We manufactured and implanted the graft for pulmonary artery (PA) reconstruction in the porcine model. The graft was explanted at 4 weeks after implantation for further evaluation.
The custom-made branched PA graft had a wall shear stress and pressure drop (PD) from the main PA to the branch PA comparable to the native vessel. At the end point, magnetic resonance imaging revealed comparable left/right pulmonary blood flow balance. PD from main PA to branch between before and after the graft implantation was unchanged. Immunohistochemistry showed evidence of endothelization and smooth muscle layer formation without calcification of the graft.
Our animal model demonstrates the feasibility of designing and implanting image-guided, 3D-printed, customized grafts. These grafts can be designed to optimize both anatomic fit and hemodynamic properties. This study demonstrates the tremendous potential structural and physiological advantages of customized TEVGs in cardiac surgery.
Our integrated virtual surgical planning and flow dynamic profile optimization prove the feasibility of providing an optimized anatomy and hemodynamic parameter of the 3D printed, customized vascular graft.
Each patient with congenital heart disease has a unique vascular defect and hemodynamic profile. Therefore, there is a need for a “best fit” conduit to provide the optimal long-term outcome. This study demonstrates that our integrated approach combining virtual surgical planning, 3-dimensional printing, and electrospinning to create a customized vascular graft can provide an optimized anatomic and hemodynamic result in vivo.
Owing to the anatomic complexity and uniqueness of each vascular defect and the individual physiology of the circulatory system, congenital vascular reconstructive surgery can be very challenging. There is an unmet clinical need for individually customized grafts in congenital vascular repair. Pediatric vascular reconstruction carries an increased risk of graft complications, such as rapid graft dysfunction, anastomotic stricture formation, size mismatch–related geometric disruption, and pulmonary artery (PA) obstruction owing to the patients' rapid growth rate.
Mismatched vasculature calibers, lack of growth potential, suboptimal biocompatibility, and risk of thromboembolic events make the use of synthetic implants for congenital vascular reconstructive an unsettling choice.
In the last decade, 3D printing has been used in the manufacture of tissue engineered vascular grafts (TEVGs). This technology offers the potential to develop a customized biodegradable scaffold to promote cellular proliferation and maturation, leading to the formation of a physiologically functional blood vessel. For congenital vascular repair, the customized TEVG provides an opportunity for high-fidelity anatomic reconstruction of vasculature with growth potential. This can significantly improve the clinical outcomes. Our previous study in an ovine model validated the concept of 3D-printed customized TEVGs using a simple straight vascular graft. Our results showed that the TEVGs were biocompatible with adequate neotissue formation and had mechanical properties comparable to those of the native tissue with a 6-month follow-up after inferior vena cava (IVC) transposition surgery.
The unique aspects to consider for cardiovascular surgery include a dynamic fluid profile of the circulatory system and the physiological compliance of the vasculature. Thus, the long-term surgical outcomes of the grafts are based not only on anatomic specificity, but also on the growth potential of the TEVG to match the size of the growing patient. The TEVG must maintain the fluid dynamic profile of the circulation system to achieve a successful long-term clinical outcome. Our group previously introduced a novel preoperative virtual surgical planning strategy in silico for the Fontan procedure.
This study established that preoperative virtual surgical planning can be used to optimize a conduit design that improves the hemodynamic profile after surgery.
Although a feasibility analysis was performed for a simple straight TEVG in our previous study, a curvilinear or branched TEVG is needed to address cases involving complex anatomy for clinical use. Thus, in this study, we built on previous work by creating a branched and curved TEVG to evaluate the potential to design and implant complex- shaped TEVGs in vivo (Video 1). As a preclinical experiment, the goal of this study was to investigate the feasibility to fabricate patient-specific TEVGs with computational fluid dynamics (CFD) optimization for reconstruction of the central PA (Figure 1). Central PA reconstructive surgery plays an important role in many surgical procedures performed to repair congenital cardiac defects in neonates and young infants.
We then evaluated the short-term outcomes of the conduit optimized by CFD hemodynamic analysis in a porcine model for a 1-month trial period. (Figure 2)
Materials and Methods
Preoperative Imaging, 3D Model Design, and Scaffold Fabrication
Cardiac magnetic resonance imaging (MRI) was performed 4 to 5 weeks before implantation, along with magnetic resonance angiography (MRA) and phase-contrast flow analysis using both 2D and 4D flow acquisitions. Following MRI, MRA was used as a roadmap to build a 3D digital model of the central PA reconstruction, using commercially available image segmentation software (Mimics; Materialise, Leuven, Belgium). Both automatic thresholding and manual methods were used to identify the blood pool of the right ventricular outflow tract and branch PAs in each MRA slice. Based on species-specific growth curves, a small increase in dimension (both length of the outflow tract and diameter of the PAs) was added to the blood pool segmentation to account for growth of the animal during the 30 days between imaging and surgery. This segmentation was converted into a 3D digital model, which was then exported using the stereolithography (STL) file format. This STL file was converted into a computer-aided design (CAD) file, smoothed, and hollowed using computer-aided design software (3-matic; Materialise and SolidWorks; Dassault Systemes, Waltham, Mass).
After validation of the design, a biodegradable version of each conduit was manufactured using a 3D electrospinning technique. For the electrospinning, a stainless-steel mandrel in the shape of the optimized graft was 3D-printed by exporting the STL file of the graft to an external printing house (Protolabs, Rosemount, Minn). The mandrel was designed such that the electrospun polymer graft could be removed. The mandrel was a 5-part piece consisting of the main PA (MPA), distal right PA (RPA), proximal RPA, distal left PA (LPA), and proximal LPA pinned together that could be taken apart for the graft removal step. An additional pin is added to the mandrel to clamp it into the electrospinning setup.
The biodegradable nanofiber material composed of a 1:1 ratio of polycaprolactone (PCL) and poly-l-lactide-co-ε-caprolactone (PLCL) were electrospun to coat the 3D-printed mandrel (Nanofiber Solutions, Hilliard, Ohio) (Figure E1). By applying high voltage to the polymer solution, polymer fibers are deposited on the mandrel to create the graft. After deposition of the fibers, the mandrel was disassembled, and the electrospun graft was removed. The graft was placed in a standard Tyvek pouch and subjected to low-temperature sterilization with vaporized hydrogen peroxide and ozone (STERIZONE VP4; Getinge, Getinge, Sweden).
Graft Implantation in Vivo
The Animal Care and Use Committee at Johns Hopkins Hospital approved the care, use, and monitoring of animals for these experiments. The animals received humane care in compliance with the National Institutes of Health's Guide for the Care and Use of Laboratory Animals. The graft was implanted as central PA reconstruction in the porcine model (mean body weight, 35 kg). The animal size increased by roughly one-third over the 1-month study period (mean body weight, 45 kg). The pigs were anesthetized with 1.5% isoflurane during surgery. The surgery was performed at the left thoracotomy position, the LPA and RPA were exposed and isolated, and heparin (100 IU/kg) was administered intravenously. During the procedure, cardiopulmonary bypass (CPB) was performed. The graft was implanted to replace main and branch PAs as an interposition graft using 6-0 Prolene suture. Antibiotic treatment (cefazolin 22 mg/kg intravenously) was administered intraoperatively and at 7 days postoperatively. All pigs were maintained on a daily oral dose of aspirin (81 mg/day) until the 1-month endpoint. The animals were imaged with an identical cardiac MRI protocol at 1 month postoperatively, prior to sacrifice.
Histology and Immunohistochemistry
Explanted TEVG samples were fixed in 10% formalin for 24 hours at 4°C and then embedded in paraffin. Standard histology was done using staining of tissue sections with hematoxylin and eosin, Masson trichrome, Verhoeff-Van Gieson, and von Kossa stains. Immunohistochemistry was performed using the following primary antibodies: von Willebrand factor (1:2000; Dako, catalog A0082), α-smooth muscle actin (SMA) (1:500; Abcam, catalog ab5694), and CD68 (1:200; Abcam, ab31630). Detection of antibody binding was done using biotinylated secondary antibodies (Vector Laboratories, Burlingame, Calif), followed by incubation with streptavidinated horseradish peroxidase (Vector Laboratories). A chromogenic reaction with 3,3-diaminobenzidine (Vector Laboratories) was performed for the development of immunohistochemistry. Counterstaining of the nuclei was done with Gill's hematoxylin (Vector Laboratories).
Histological and Quantitative Analyses
The remaining scaffold area was measured by hematoxylin and eosin staining and plain (polarized light) histology. The thickness of the smooth muscle layer from the graft was measured by SMA staining using ImageJ software (National Institutes of Health, Bethesda, Md). The thickness of the muscle layer was quantified by analyzing the muscular thickness–to–full-thickness ratio of the graft to that of the native tissue from a representative section of each sample and averaging the value for all samples (n = 5). The remaining scaffold area was quantified by analyzing the light intensity under polarized light microscopy the same exposure time and the pixel calculation from a representative section of each sample and averaging the value for all samples (n = 5).
The Sircol colorimetric assay (Biocolor, Carrickfergus, United Kingdom) was used to evaluate the collagen content. In accordance with the assay's protocol, we measured 100 mg dry weight of each sample and transferred it to low-protein-binding 1.5-mL conical microcentrifuge tubes with 1.0 mL of pepsin (Sigma-Aldrich, St Louis, Mo), with a concentration of 0.1 mg/mL of 0.5 M acetic acid to solubilize the collagen by overnight incubation. The collagen content in each sample was evaluated by assessing the absorbance intensity at 555 nm after dye binding according to the manufacturer's protocol.
In this experiment, data from CFD simulation and histological analysis are represented graphically as bar or line charts with error bars representing the mean with standard error of the mean. The unpaired 2-tailed t test was used for the analysis of collagen. The Pearson correlation coefficient was used to determine the significance of any correlation of SMA/high-power field. A P value <.05 was considered statistically significant. Statistical analysis was performed using Prism version 8 (GraphPad Software, La Jolla, Calif).
Optimization of the Custom-Made Graft With CFD Simulation
Optimization of the graft was done using an iterative strategy as described previously.
First, with all the designed CAD models, the model of the surrounding heart and vasculature anatomy was overlaid with the new graft designs, to minimize any overlap between the optimized graft and other thoracic structures (Figure 3, A). Grafts with different feasible anatomic designs were chosen. The CFD simulation optimized the graft based on 3 parameters: (1) reduction of power loss across the conduit inlet and outlet, (2) adjustment of the even fluid flow between the right and left branches, and (3) wall shear stress distribution optimization. The design with the best performance was determined for the next iteration of the simulation (Figure 3, B).
The CFD analysis was performed with the hemodynamic parameters considered in this study, including power loss, flow distribution (LPA:RPA), and wall shear stress. The power loss was calculated based on the changes in pressure and flow rates on inlets and outlet of each model as described previously.
The wall shear stress distribution of the porcine model is shown in Figure 3, C and D. In the porcine models 1 and 2, compared with the preoperative native model, the simulation model has a similar power loss between the inlet and the outlet of the graft (pig 1: native vs designed, 1.5 × 10−2 W vs 1.4 × 10−2 W; pig 2: native vs designed, 2.35 × 10−3 W vs 1.65 × 10−2 W). The flow distribution was maintained with a similar flow in the both the native and designed conduits (pig 1: LPA:RPA, 47:53; pig 2: LPA:RPA, 42:58). The wall shear stress was comparable in the native tissue and graft (pig 1: native vs designed, 2 Pa vs 2.16 Pa; pig 2: native vs designed, 1.14 Pa vs 1.02 Pa) (Figure 3, C and D). Overall, the designed models demonstrated comparable hemodynamic and wall shear stress to the native models.
In Vivo Implantation of the Customized Conduit
The preoperative vasculature images were prepared with MRI for CAD image creation (Figure 4, A). Preoperative LPA:RPA blood flow balance was recorded (LPA:RPA, 45:55) (Figure 4, F). The customized branched TEVG was fabricated via 3D electrospinning technology (Figure 4, B). Graft implantation surgery was performed as described above (Figure 4, C). At 1 month after graft implantation, postoperative MRI was performed (Figure 4, D), and LPA:RPA blood flow balance in the optimized graft design was measured (LPA:RPA, 52:48) (Figure 4, G). At the end of the study, the graft was explanted for macroscopic inspection (Figure 4, E), and direct blood pressure measurements in the LPA and RPA were obtained (Figure 4, H). The magnitudes of LPA and RPA pressure were roughly 11 to 12 mm Hg. The absence of clinically significant stenosis was demonstrated by an almost equal flow distribution and similar pressure within the PA branches.
Biocompatibility of the Graft
The biocompatibility of the graft, including patency and tissue degradation, was evaluated after 1-month. No graft-related stenosis, dilation, or rupture occurred by the 1-month endpoint of the study, and there was no clot formation macroscopically or microscopically (Figure 5, A-D). The presence of collagen in the graft was verified by collagen assay, suggesting ongoing collagen tissue formation (Figure 5, E). The graft degradation was detected by the presence of remaining scaffold area from the polarized light microscopy (Figure 5, F). The remaining scaffold area in the individual parts of the graft was measured; the averages of the remaining scaffold area in the models were 11.01 ± 2.64% for of MPA, 9.76 ± 3.18% for RPA, and 8.93 ± 2.14% for LPA (Figure 5, G).
Formation of Neotissue and Vasculature
Extracellular matrix formation was evaluated with Mason Trichrome staining. Compared with the native tissue (Figure 6, A, F, and K), the vascular graft showed ongoing extracellular matrix formation. (Figure 6, B, G, and L). Von Kossa staining demonstrated no ectopic calcification (Figure 6, C, H, and M). SMC layers contribute to vascular function and were evaluated by SMA immunohistochemistry staining. The presence of multilayered SMA-positive cells in the graft compared with the native pulmonary tissue suggests ongoing active vascular muscle remodeling (Figure 6, D, I, and N).
Endothelial cell formation in the graft was shown with the von Willebrand factor–positive cells in the graft (Figure 6, E, J, and O). The presence of CD68+ macrophages with graft degradation is indicative of the inflammatory process of tissue remodeling in the graft (Figure 7, A, center of the graft; B, edge of the graft). The percent thickness of the smooth muscle layer (percentage = thickness of smooth muscle/total thickness of the vessel × 100%) demonstrated some neotissue formation (Figure 7, C).
Mechanical Properties of the Customized TEVG Conduit
The mechanical properties of the graft were measured at different time points, including before implantation and 1 month after implantation. The circumferential tensile strength, an analog of the ultimate tensile strength in the tubular structure, of the native tissue and graft after implantation have comparable values: native vs graft (1 month), 0.65 N/mm vs 0.45 N/mm (Figure 7, D). The compliance of the graft after implantation was slightly higher than that of the native tissue: compliance, native vs graft (1 month), 9.89% mm Hg vs 21.10% mm Hg (Figure 7, E). At the endpoint, the inner diameter of the graft was similar to the inner diameter of the native PA: native vs graft (1 month), 10 mm vs 10 mm (Figure 7, F). Due to the short follow-up, the mechanical property measurements of the graft indicate that the customized graft degradation process was in progress with a higher compliance from the remaining scaffold though a similar tensile strength was noted.
The results of this study validate the feasibility of the fabrication of the complex-shaped 3D-printed cell-free nanofiber TEVG in an in vivo experiment. We previously demonstrated the capability of creating customized TEVGs using 3D printing and electrospinning technology for straight IVC conduits.
This study demonstrates that a complex curved and branched “Y” shape conduit can be fabricated by the electrospinning process. In combination with CFD technology, the development of more physiologically compatible patient specific hemodynamics can be achieved.
The materials used in this study showed adequate physical properties in a low-pressure venous system in this in vivo experiment with a 1-month time frame. By optimizing the ratios of the combination of different polymers, the materials can exhibit the advantageous mechanical properties of each individual polymer. The biocompatibility profile of some of the polymers, such as polyglycolic acid (PGA), has been investigated in vivo.
We tested the graft with a similar diameter and demonstrated adequate mechanical properties and 1-month survival without graft-related complications. To extend the potential clinical use of this combination of polymers, the in vivo degradation profile of the TEVG made from the combined polymers will be studied in a long-term animal survival study.
One of the innovations of this study is the bifurcated shape of the TEVG. Sugiura and colleagues
reported promising results with no graft-related fatal complications in a clinical trial of TEVG use in children with congenital heart disease. That study followed 25 patients who underwent implantation of a linear TEVG as an extracardiac total cavopulmonary conduit for 11 years. The potential benefit would be extended to a larger population of patients who may need curved or even complex-shaped TEVGs. Best and colleagues
but unfortunately, the excessively complex anatomy made the patient ineligible for linear TEVG clinical use.
This study establishes that the production of a complex-shaped, customized 3D printed and electrospun TEVG is a feasible technology to manufacture a customized vascular graft with predictable flow dynamics. It would provide the possibility of future development of complex-shaped grafts to fit the anatomy of individual patients.
The goal of TEVG use is to support autologous tissue growth and to remain patent without affecting the blood flow inside the conduit and causing diameter mismatch.
In this study, we used a cell-free electrospun nanofiber vascular graft, which demonstrated satisfactory patency and tissue remodeling by 6 months in our previous study of straight TEVG in a sheep model.
In this study, we demonstrate complex graft in vivo implantation without stenosis formation over a 1-month period. Our promising data support the need for a study with a longer follow-up period.
The tissue remodeling potential in the graft was demonstrated by the presence of a single layer of endothelial cells, an organized SMC layer, and collagen deposition. As a pilot study, we aimed to demonstrate the feasibility of neotissue formation in a custom-made graft using a porcine model with short term survival of 1 month. Specifically, we found no graft-related mortality and no cases of aneurysm formation, graft rupture, or ectopic calcification using routine imaging modalities for TEVG observation and histological assessment.
In this study, we have proven the feasibility of the in vivo implantation of a complex-shaped, patient-specific TEVG. We have developed several key components, including specific parameters for optimal graft design, unique electrospinning technology, and the technique for surgical experiments using CPB. The novelty of this experiment may provide a potential clinical application in congenital heart surgery, which demands complex-shaped grafts with optimal hemodynamic profiles. Our future work will include a study with a larger number of animals and a longer follow-up period. Our CFD optimization strategy will be refined to evaluate the low-flow regions and areas of low wall shear stress for the susceptibility of thrombosis formation. We will also include different simulation strategies, such as pulsatile flow circulation and hyperdynamic conditions. In the long term, we aim to create an animal model like Fontan physiology that will be amenable to the real clinical encounters, such as vessel stenosis or hypoplasia, to provide a possible application to clinical needs.
In conclusion, we have demonstrated the manufacturability and biocompatibility of the virtual surgical optimization strategy of the 3D-printed patient-specific TEVG conduit in a large animal model. Although this study was limited by its short survival duration and small sample number, the results are promising and definitely warrant further investigation.
J. Opfermann, J. Johnson, A. Krieger, and N. Hibino are inventors listed on International Patent WO/2017/035500Al (Patient-Specific Tissue Engineered Vascular Graft Utilizing Electrospinning). The patent filing has been disclosed for grant applications and to institutions. J. Johnson and N. Hibino are equity holders in Nanofiber Solutions. All other authors have nothing to disclose with regard to commercial support.
We thank Dr Henry Halperin, Dr Cecillia Lui, Dr Sara Abdollahi, Tom Loke, Melissa Jones, and Sean Kearney for their surgical support and Dr Ehud Schmidt, Michael Guttman, Rick Tunin, and Sarah Fink for perioperative surgical support and veterinary imaging technical expertise. We also acknowledge the University of Maryland supercomputing resources (http://hpcc.umd.edu/) that were made available for conducting the research reported in this article.
Appendix E1. Calculation Method for Remaining Scaffold
The remaining scaffold percentage was calculated as follows. The remaining scaffold was measured as light intensity under standardized condition including same light exposure and the exposure time. The intensity measured was converted into pixel count by ImageJ, with the exclusion of background noise in each individual slide owing to the intrinsic impurity of the slides. The maximal pixel count determined by ImageJ was set to be the original undegraded graft pixel. The percentage was calculated as pixel of graft/pixel of the undegraded graft × 100% .
Heart disease and stroke statistics—2017 update: a report from the American Heart Association.
Vascular grafts and, more commonly, patches are commonly used in the repair of congenital heart defects. Non-autologous, circumferential vascular grafts have key shortcomings, such as the lack of growth potential and need for anticoagulation, that prohibit their universal adoption.1 Another shortcoming of grafts and patches is that tailoring of their shape and size before implantation is according to the surgeon's eye or crude measurements in 1 or 2 dimensions. Although this tailoring may be satisfactory for simple geometries, achieving the most efficient hemodynamic result for complex pathways and baffles in growing children is sometimes difficult, and the development of late obstruction may necessitate reoperation.
I congratulate Yeung and colleagues1 for their development of an elegant, sophisticated, and innovative technology and also commend their collaboration across multiple disciplines and institutions. The basic concept they promote is intuitively very appealing—use patient-specific anatomy, guided by computational fluid dynamics, to develop 3-dimensional (3D) vascular grafts tailored to a given patient's specific needs—in this particular case, the creation of bifurcated pulmonary artery grafts. The reader should look beyond the fact that both animals originally had normal anatomy, as I presume the computer-aided design phase would permit creation of any desired anatomy (for missing parts), which could then be modified to fit the patient's true anatomy.