Received: 29 October 2017 Revised: 7 February 2018 Accepted: 23 February 2018
MACHINE LEARNING AND DEEP LEARNING APPROACHES
Machine learning–aided exploration of relationship
between strength and elastic properties in ascending
Department of Mechanical and Industrial
Engineering, The University of Iowa, Iowa
City, IA 52242, USA
Department of Computer Sciences,
University of Wisconsin-Madison, 1210 W.
Dayton St, Madison, WI 53706-1613, USA
Jia Lu, Department of Mechanical and
Industrial Engineering, The University of
Iowa, Iowa City, IA 52242, USA.
Machine learning was applied to classify tension-strain curves harvested from
inflation tests on ascending thoracic aneurysm samples. The curves were clas-
sified into rupture and nonrupture groups using prerupture response features.
Two groups of features were used as the basis for classification. The first was
the constitutive parameters fitted from the tension-strain data, and the second
was geometric parameters extracted from the tension-strain curve. Based on
the importance scores provided by the machine learning, implications of some
features were interrogated. It was found that (1) the value of a constitutive
parameter is nearly the same for all members in the rupture group and (2) the
strength correlates strongly with a tension in the early phase of response as well
as with the end stiffness. The study suggests that the strength, which is not avail-
able without rupturing the tissue, may be indirectly inferred from prerupture
ATAA, machine learning, random forest, rupture, strength
Ascending thoracic aortic aneurysm (ATAA) is an abnormal dilation in the ascending aorta. It is ranked as the 15th
leading cause of death in the United States. A typical ATAA grows silently over many years until sudden rupture.
indicate that only 41% of the patients with the ruptured ATAA were alive on arrival at an emergency hospital.
size of the aneurysm has been a primary factor in monitoring and intervention decision, while it has been known that
aneurysms smaller than the selected threshold rupture, and some aneurysms larger than the threshold do not.
biomechanical analysis has been used to evaluate the stress state in the aneurysm wall. Mechanics-based indicators such
as the peak wall stress and peak wall rupture risk index have been suggested and shown to improve the risk assessment.
One of the major challenges in stress-based risk evaluation is lack of tissue properties on a patient-specific basis. There is
a large body of literature pertaining to the in vitro mechanical properties of ATAA (see, eg, previous works
these contributions shed significant light on understanding the behavior of this diseased tissue. It was commonly reported
that ATAAs exhibit a significant level of intersubject and intrasubject heterogeneities in their mechanical properties. The
properties can vary regionally,
They can be affected by disease conditions such as Marfan
and aortic valve phenotype.
Age, gender, family history and aortic diseases and other clinical
Int J Numer Meth Biomed Engng. 2018;34:e2977. wileyonlinelibrary.com/journal/cnm Copyright © 2018 John Wiley & Sons, Ltd. 1of12