P152. Quantitative Characterization of Transmural Pore Networks Predicts Endothelialization of Cardiovascular Prostheses
Presented During: Adult Cardiac Poster Session II
Groote Schuur Hospital, Observatory, Cape Town
Cape Town
South Africa - Contact Me
Dr Andrea Tonelli is a MD-PhD student at the University of Cape Town, South Africa, supervised by Emeritus Professor Peter Zilla. His research focuses on the surface endothelialisation of polymeric heart valves and vascular graft prostheses.
Monday, May 4, 2026: 9:00 AM - 4:00 PM
McCormick Place Lakeside Center
Room: Exhibit Hall, Poster Area
Description
Objective: Endothelialization of cardiovascular conduits may be crucial for overcoming thrombogenic, haemodynamic, and degradative complications. Current paradigms emphasize transmural vessel ingrowth via pores as big as 50-80µm, yet biomaterial characterization often lacks the resolution to elucidate and quantify continuous, full thickness 3D pore architecture, leading to variable in vivo outcomes. This study establishes a novel, quantitative approach to non-destructively reconstruct conduit pore space, assess its 3D spatial porosity relevant to blood vessel permissivity, and validate findings in vivo.
Methods: Electrospun cardiovascular conduits were non-destructively scanned using micro-computed tomography (µCT) and processed with deep-learning super-resolution and segmentation to isolate the 3D pore phase. Pore network modelling (PNM) using a custom Python script reconstructed the pore space as an interconnected network, quantifying pore and throat dimensions. Scaffolds were implanted for 1 and 3 weeks in subcutaneous rodent model; explants were assessed for cell ingrowth, extracellular matrix maturation, and blood vessel density.
Results: Deep-learning-enhanced micro-computed tomography (µCT) enabled high-resolution (0.54 µm) imaging across large fields of view (2.8 mm²), achieving reconstructions exceeding 1 x 10⁸ µm³ in ≈70 minutes – a ≈57-fold increase in speed over a brute force approach, and ≈100-fold increase in volume compared to conventional methods like focused-ion beam scanning electron microscopy (FIB-SEM). Validation demonstrated improved accuracy in quantifying material metrics relative to confocal laser scanning microscopy (CLSM) and SEM. Pore network modelling (PNM) revealed median pore inscribed diameters of 5.51 µm (IQR: 5.15) for 16% polymer weight, 5.40 µm (IQR: 6.23) for 18%, and 5.40 µm (IQR: 4.22) for 20%. Thresholding revealed the highest density of interconnected, 3D ingrowth-permissive porosity within the 18% group, which translated at 3 weeks post-implantation, exhibiting collagen density comparable to controls and significantly increased vessel area versus 16% (p < 0.001) and 20% (p < 0.05).
Conclusions: Traditional characterization techniques may inadequately predict transmural blood vessel regeneration in cardiovascular conduits. This study demonstrates that µCT, deep learning, and PNM provide a reliable, efficient, and validated means to predict endothelial regeneration potential of implanted prostheses.
Authors
Andrea Tonelli (1), Timothy Pennel (1), Jaco Theron (1), Francesco Iacoviello (2), Peter Zilla (1)
Institutions
(1) Chris Barnard Division of Cardiothoracic Surgery, University of Cape Town, Cape Town, South Africa, (2) Correlative X-Ray Laboratory, University College of London, London, United Kingdom
Methods: Electrospun cardiovascular conduits were non-destructively scanned using micro-computed tomography (µCT) and processed with deep-learning super-resolution and segmentation to isolate the 3D pore phase. Pore network modelling (PNM) using a custom Python script reconstructed the pore space as an interconnected network, quantifying pore and throat dimensions. Scaffolds were implanted for 1 and 3 weeks in subcutaneous rodent model; explants were assessed for cell ingrowth, extracellular matrix maturation, and blood vessel density.
Results: Deep-learning-enhanced micro-computed tomography (µCT) enabled high-resolution (0.54 µm) imaging across large fields of view (2.8 mm²), achieving reconstructions exceeding 1 x 10⁸ µm³ in ≈70 minutes – a ≈57-fold increase in speed over a brute force approach, and ≈100-fold increase in volume compared to conventional methods like focused-ion beam scanning electron microscopy (FIB-SEM). Validation demonstrated improved accuracy in quantifying material metrics relative to confocal laser scanning microscopy (CLSM) and SEM. Pore network modelling (PNM) revealed median pore inscribed diameters of 5.51 µm (IQR: 5.15) for 16% polymer weight, 5.40 µm (IQR: 6.23) for 18%, and 5.40 µm (IQR: 4.22) for 20%. Thresholding revealed the highest density of interconnected, 3D ingrowth-permissive porosity within the 18% group, which translated at 3 weeks post-implantation, exhibiting collagen density comparable to controls and significantly increased vessel area versus 16% (p < 0.001) and 20% (p < 0.05).
Conclusions: Traditional characterization techniques may inadequately predict transmural blood vessel regeneration in cardiovascular conduits. This study demonstrates that µCT, deep learning, and PNM provide a reliable, efficient, and validated means to predict endothelial regeneration potential of implanted prostheses.
Authors
Andrea Tonelli (1), Timothy Pennel (1), Jaco Theron (1), Francesco Iacoviello (2), Peter Zilla (1)
Institutions
(1) Chris Barnard Division of Cardiothoracic Surgery, University of Cape Town, Cape Town, South Africa, (2) Correlative X-Ray Laboratory, University College of London, London, United Kingdom
Presentation Duration
There is no formal presentation for posters. Your poster will be on display on your assigned day from 9:00AM - 4:00PM
View Submission
ADULT CARDIAC
Basic and Translational Research