A machine learning approach to investigate the relationship between shape features and numerically predicted risk of ascending aortic aneurysm

Abstract

Geometric features of the aorta are linked to patient risk of rupture in the clinical decision to electively repair an ascending aortic aneurysm (AsAA). Previous approaches have focused on relationship between intuitive geometric features (e.g. diameter and curvature) and wall stress. This work investigates the feasibility of a machine learning approach to establish the linkages between shape features and FEA predicted AsAA rupture risk, and it may serve as a faster surrogate for FEA associated with long simulation time and numerical convergence issues. This method consists of four main steps: (1) constructing a statistical shape model (SSM) from clinical 3D CT images of AsAA patients; (2) generating a dataset of representative aneurysm shapes and obtaining FEA predicted risk scores defined as systolic pressure divided by rupture pressure (rupture is determined by a threshold criterion); (3) establishing relationship between shape features and risk by using classifiers and regressors; and (4) evaluating such relationship in cross validation. The results show that SSM parameters can be used as strong shape features to make predictions of risk scores consistent with FEA, which lead to an average risk classification accuracy of 95.58% by using support vector machine and an average regression error of 0.0332 by using support vector regression, while intuitive geometric features have relatively weak performance. Compared to FEA, this machine learning approach is magnitudes faster. In our future studies, material properties and inhomogeneous thickness will be incorporated into the models and learning algorithms, which may lead to a practical system for clinical applications.

A machine learning approach as a surrogate of finite element analysis–based inverse method to estimate the zero‐pressure geometry of human thoracic aorta

“…The proposed ML model is not ready for clinical applications yet and needs to be further enhanced to handle a large range of shape variations. Since the SSM only covers some major shape variations, a larger amount of patient data is need to enhance the ML model. As suggested by a recent study, DNN performance can be significantly improved by using more training data and fine tuning the structure, which will be applied in our future work.…”

“…In the literature, several methods are available to estimate zero‐pressure organ geometries. These methods usually take as input a loaded geometry reconstructed from in vivo images and assume the material properties are known, then either (1) estimate the zero‐pressure geometry by adjusting a candidate geometry and running forward finite element (FE) simulations, or (2) estimate the prestress/prestrain field on the loaded configuration, which can be used to back out the zero‐pressure geometry (eg, by depressurizing the FE model), or (3) estimate the zero‐pressure geometry and the prestress field using an inverse FE formulation . All of these methods rely on FEA and most of them require many iterations of FE simulations, which makes these methods very time‐consuming and inefficient.…”

“…8,9 Such 3D-printed models, if built with accurate zero-pressure geometries, would enable more realistic in vitro simulation platforms for preoperative planning, such as in vitro simulation of transcatheter aortic valve replacement. 8 In the literature, several methods [10][11][12][13][14][15][16][17][18][19][20] are available to estimate zero-pressure organ geometries. These methods usually take as input a loaded geometry reconstructed from in vivo images and assume the material properties are known, then either (1) estimate the zero-pressure geometry 10,14,[16][17][18]20 by adjusting a candidate geometry and running forward finite element (FE) simulations, or (2) estimate the prestress/prestrain field on the loaded configuration, 12,13,15,19,21 which can be used to back out the zero-pressure geometry (eg, by depressurizing the FE model), or (3) estimate the zero-pressure geometry and the prestress field using an inverse FE formulation.…”

A deep learning approach to estimate chemically-treated collagenous tissue nonlinear anisotropic stress-strain responses from microscopy images

“…For each tissue sample, the resampled strain values from the two curves were assembled as a vector of 126 numbers, Y . By using principle component analysis (PCA) [31–33], the vector Y of a tissue sample can be decomposed as $$Y\cong Y_{\mathit{PCA}}=\overline{Y}+{\alpha }_1{V}_1+{\alpha }_2{V}_2+{\alpha }_3{V}_3$$where $\overline{Y}$ is the population mean, { V i } are the modes of variation, and { α i } are the coefficients. Here, { α i } can vary, while $\overline{Y}$ and { V i } are the same for all tissue samples.…”

“…Here, { α i } can vary, while $\overline{Y}$ and { V i } are the same for all tissue samples. From the PCA calculation [33], the first three modes of variation { V 1 , V 2 , V 3 } with { α 1 , α 2 , α 3 } can describe 99% of the total variation of the stress-strain curves, which means each stress-strain curve can be almost perfectly reconstructed by using Eq. (1) as shown in Figure 3b.…”