A machine learning approach for planning valve-sparing aortic root reconstruction

Jannis Hagenah, Alexander Schlaefer, Christoph Metzner, Michael Scharfschwerdt


Choosing the optimal prosthesis size and shapeis a difficult task during surgical valve-sparing aortic rootreconstruction. Hence, there is a need for surgery plan-ning tools. Common surgery planning approaches try tomodel the mechanical behaviour of the aortic valve andits leaflets. However, these approaches suffer from inac-curacies due to unknown biomechanical properties andfrom a high computational complexity. In this paper, wepresent a new approach based on machine learning thatavoids these problems. The valve geometry is described bygeometrical features obtained from ultrasound images. Weinterpret the surgery planning as a learning problem, inwhich the features of the healthy valve are predicted fromthese of the dilated valve using support vector regression(SVR). Our first results indicate that a machine learningbased surgery planning can be possible.
Original languageEnglish
JournalCurrent Directions in Biomedical Engineering
Pages (from-to)361-365
Number of pages5
Publication statusPublished - 2015


Dive into the research topics of 'A machine learning approach for planning valve-sparing aortic root reconstruction'. Together they form a unique fingerprint.

Cite this