Coronary artery properties in atherosclerosis: A deep learning predictive model

Abstract

In this work an Artificial Neural Network (ANN) was developed to help in the diagnosis of plaque vulnerability by predicting the Young modulus of the core (Ecore) and the plaque (Eplaque) of atherosclerotic coronary arteries. A representative in silico database was constructed to train the ANN using Finite Element simulations covering the ranges of mechanical properties present in the bibliography. A statistical analysis to pre-process the data and determine the most influential variables was performed to select the inputs of the ANN. The ANN was based on Multilayer Perceptron architecture and trained using the developed database, resulting in a Mean Squared Error (MSE) in the loss function under 10–7, enabling accurate predictions on the test dataset for Ecore and Eplaque. Finally, the ANN was applied to estimate the mechanical properties of 10,000 realistic plaques, resulting in relative errors lower than 3%.

Characterizing atherosclerotic tissues: in silico analysis of mechanical properties using intravascular ultrasound and inverse finite element methods

“…On the other hand, in the material analysis, the elasticity modules were 516, 708.6, and 1,404.5 kPa for hypocellular, cellular, and calcified tissues respectively. The resulting elasticity for fibrotic tissues exceeds the range previously proposed (Caballero et al, 2023). Therefore, the method failed to accurately estimate the stiffness of the lipid tissue, resulting in softer values than the actual ones.…”

“…In these cases, it was not possible to directly compare the estimated Young’s modulus with the GOH material parameters. However, stiffness values were found to be over the limits ( Caballero et al, 2023 ). Due to the high Young’s modulus estimation of fibrotic tissues, lipids appeared to be softer than their actual stiffness values.…”

“…The robustness of this approach was tested with five geometries and different material combinations of lipid and fibrotic tissues to ensure that the results were consistent regardless of the geometry or material properties. In previous work, Caballero et al (2023) conducted a study on the elastic modulus ranges for lipid and fibrotic tissues through the literature. They collected a range of [1–100 kPa] for lipid elastic modulus and [390–1,200 kPa] for fibrotic elastic modulus.…”

“…So, the ranges for fibrotic GOH parameters were , , obtained from curve fitting of literature data ( Loree et al, 1994 ; Versluis et al, 2006 ), and giving more relevance to k 1 which proved to be more important in the mechanical response at physiological pressures. Both the lipid and the calcification ranges were adjusted considering the proposed range of Caballero et al (2023) , with the lipid parameters and and the elastic modulus of the calcification with a range of . • Third Step.…”