Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics
The argument on the effectiveness of the Generalized Transfer Function (GTF) technique is currently ongoing. To resolve the dispute in experiment based studies, this thesis aims to use a well-validated numerical model as an alternative to experimental studies to test the validity of the GTF method....
| Main Author: | |
|---|---|
| Format: | Thesis (University of Nottingham only) |
| Language: | English |
| Published: |
2019
|
| Subjects: | |
| Online Access: | https://eprints.nottingham.ac.uk/56109/ |
| _version_ | 1848799273586524160 |
|---|---|
| author | Butt, Ahmed Tamkin |
| author_facet | Butt, Ahmed Tamkin |
| author_sort | Butt, Ahmed Tamkin |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | The argument on the effectiveness of the Generalized Transfer Function (GTF) technique is currently ongoing. To resolve the dispute in experiment based studies, this thesis aims to use a well-validated numerical model as an alternative to experimental studies to test the validity of the GTF method. This thesis divides the work into four main inter-disciplinary areas of research.
Development of an existing non-linear one-dimensional (1-D) mathematical model that can comprehensively compute the propagation of blood in the human arterial network. The model developed was divided into large arteries and small arteries. The large arteries are based on physiological data while the small arteries were based on statistical relations. Instead of the more commonly used Windkessel model, the structured tree outflow boundary condition was used as the computation of pressure and flow in the small arteries provides a more dynamic and physiological boundary condition to the large arteries.
A multi-level validation of the developed model was undertaken in order to demonstrate the robustness and the applicability of the developed 1-D model to real life situations. The model was used to simulate pulse wave propagation along a single vessel (aorta) and the results compared against in-vivo data. The in-vivo systolic and diastolic pressures were 16.8 ± 0.4 kPa and 9.5 ± 0.4 kPa while the model estimated were 16.89 kPa and 10.94 kPa, respectively, showing excellent agreement. Simulation of pulse wave propagation in the entire arterial tree was then undertaken with two different geometries, from a 3-D model and physiological data. Comparison against the 3-D model showed a maximum percentage error of 2.5% while the excellent waveform amplitude and shape comparison with in-vivo data, confirmed the validity of the 1-D model.
The multi-level validation confirmed the robustness of the 1-D model to accurately simulate pulse propagation under varying geometric, elastic and fluid properties. This allowed the use of the 1-D model to create a database that recorded several different cardiovascular responses due to several physiological and pathological conditions. The physiological conditions simulated were the variation in cardiac output and the variation in arterial stiffness while the pathological conditions simulated were abdominal aortic aneurysm and the coarctation of aorta. All physiological and pathological conditions agreed well with literature and were extremely well captured by the 1-D model.
Half of the pressure response database was used to estimate the GTFs between the ascending aorta and four different peripheral anatomical locations namely, the carotid artery, brachial artery, radial artery and the femoral artery. The estimated GTFs were multiplied with pulse pressures (PP) from the respective locations of the remaining half of the database and the yielding GTF-estimated Central Aortic Pressure (CAP) were statistically compared with the known, model-generated CAP to evaluate the validity of the GTF technique. The Pearson’s r values for the carotid, brachial, radial and femoral artery generated CAP of 0.991, 0.981, 0.978 and 0.873 (p < 0.001), 0.996, 0.996, 0.993 and 0.971 (p < 0.001) and 0.999, 1.000, 1.000 and 0.934 (p < 0.001) for the systolic, diastolic and mean pressures, respectively, showed that the GTF technique is capable of estimating the CAP with extremely high accuracy. These results were further cemented by carrying out a Bland-Altman analysis as well as a linear regression which demonstrate that the GTF estimated CAP are highly correlated with model-generated CAP with the carotid artery being the most preferable and femoral artery being the least preferable site of PP measurement.
This thesis, in addition to comprehensively validating the 1-D model with structured tree outflow condition and demonstrating disease modelling, uses an alternative to experimental studies, which is free from human and calibration errors, to exhibit the accuracy of the GTF technique. The pressure response database created using the validated 1-D model for 194 physiological and pathological conditions introduces variations in PP as well as CAP. The GTFs estimated using half of these responses agrees well with GTFs found in literature and when put to test for CAP estimation using the remaining half of the responses, performs extremely well. |
| first_indexed | 2025-11-14T20:33:03Z |
| format | Thesis (University of Nottingham only) |
| id | nottingham-56109 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-14T20:33:03Z |
| publishDate | 2019 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-561092025-02-28T14:24:26Z https://eprints.nottingham.ac.uk/56109/ Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics Butt, Ahmed Tamkin The argument on the effectiveness of the Generalized Transfer Function (GTF) technique is currently ongoing. To resolve the dispute in experiment based studies, this thesis aims to use a well-validated numerical model as an alternative to experimental studies to test the validity of the GTF method. This thesis divides the work into four main inter-disciplinary areas of research. Development of an existing non-linear one-dimensional (1-D) mathematical model that can comprehensively compute the propagation of blood in the human arterial network. The model developed was divided into large arteries and small arteries. The large arteries are based on physiological data while the small arteries were based on statistical relations. Instead of the more commonly used Windkessel model, the structured tree outflow boundary condition was used as the computation of pressure and flow in the small arteries provides a more dynamic and physiological boundary condition to the large arteries. A multi-level validation of the developed model was undertaken in order to demonstrate the robustness and the applicability of the developed 1-D model to real life situations. The model was used to simulate pulse wave propagation along a single vessel (aorta) and the results compared against in-vivo data. The in-vivo systolic and diastolic pressures were 16.8 ± 0.4 kPa and 9.5 ± 0.4 kPa while the model estimated were 16.89 kPa and 10.94 kPa, respectively, showing excellent agreement. Simulation of pulse wave propagation in the entire arterial tree was then undertaken with two different geometries, from a 3-D model and physiological data. Comparison against the 3-D model showed a maximum percentage error of 2.5% while the excellent waveform amplitude and shape comparison with in-vivo data, confirmed the validity of the 1-D model. The multi-level validation confirmed the robustness of the 1-D model to accurately simulate pulse propagation under varying geometric, elastic and fluid properties. This allowed the use of the 1-D model to create a database that recorded several different cardiovascular responses due to several physiological and pathological conditions. The physiological conditions simulated were the variation in cardiac output and the variation in arterial stiffness while the pathological conditions simulated were abdominal aortic aneurysm and the coarctation of aorta. All physiological and pathological conditions agreed well with literature and were extremely well captured by the 1-D model. Half of the pressure response database was used to estimate the GTFs between the ascending aorta and four different peripheral anatomical locations namely, the carotid artery, brachial artery, radial artery and the femoral artery. The estimated GTFs were multiplied with pulse pressures (PP) from the respective locations of the remaining half of the database and the yielding GTF-estimated Central Aortic Pressure (CAP) were statistically compared with the known, model-generated CAP to evaluate the validity of the GTF technique. The Pearson’s r values for the carotid, brachial, radial and femoral artery generated CAP of 0.991, 0.981, 0.978 and 0.873 (p < 0.001), 0.996, 0.996, 0.993 and 0.971 (p < 0.001) and 0.999, 1.000, 1.000 and 0.934 (p < 0.001) for the systolic, diastolic and mean pressures, respectively, showed that the GTF technique is capable of estimating the CAP with extremely high accuracy. These results were further cemented by carrying out a Bland-Altman analysis as well as a linear regression which demonstrate that the GTF estimated CAP are highly correlated with model-generated CAP with the carotid artery being the most preferable and femoral artery being the least preferable site of PP measurement. This thesis, in addition to comprehensively validating the 1-D model with structured tree outflow condition and demonstrating disease modelling, uses an alternative to experimental studies, which is free from human and calibration errors, to exhibit the accuracy of the GTF technique. The pressure response database created using the validated 1-D model for 194 physiological and pathological conditions introduces variations in PP as well as CAP. The GTFs estimated using half of these responses agrees well with GTFs found in literature and when put to test for CAP estimation using the remaining half of the responses, performs extremely well. 2019-07-28 Thesis (University of Nottingham only) NonPeerReviewed application/pdf en arr https://eprints.nottingham.ac.uk/56109/1/AHMED%20TAMKIN%20BUTT%20%2818016700%29-%20Corrected.pdf Butt, Ahmed Tamkin (2019) Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics. PhD thesis, University of Nottingham. cardiovascular diagnostics; Generalized Transfer Function (GTF); physiological data; human arterial network |
| spellingShingle | cardiovascular diagnostics; Generalized Transfer Function (GTF); physiological data; human arterial network Butt, Ahmed Tamkin Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics |
| title | Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics |
| title_full | Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics |
| title_fullStr | Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics |
| title_full_unstemmed | Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics |
| title_short | Numerical modelling of the human arterial network in conjunction with the GTF technique to improve cardiovascular diagnostics |
| title_sort | numerical modelling of the human arterial network in conjunction with the gtf technique to improve cardiovascular diagnostics |
| topic | cardiovascular diagnostics; Generalized Transfer Function (GTF); physiological data; human arterial network |
| url | https://eprints.nottingham.ac.uk/56109/ |