Automatic measuring of quality criteria for heart valvesAlexandru Paul Condurache a , Tobias Hahn a , Ulrich G. Hofmann a , Michael Scharfschwerdt b ,Martin Misfeld b and Til Aach ca Institute for Signal Processing, University of Luebeck, Ratzeburger Allee 160, 23538 Luebeck,Germany;b Clinic for Heart Surgery, University Clinic Schleswig-Holstein, Ratzeburger Allee 160, 23538Luebeck, Germany;cInstitute of Imaging and Computer Vision, RWTH-Aachen University, Templergraben 55,52056 Aachen, GermanyABSTRACTPatients suffering from a heart valve deficiency are often treated by replacing the valve with an artificial orbiological implant. In case of biological implants, the use of porcine heart valves is common. Quality assessmentand inspection methods are mandatory to supply the patients (and also medical research) with only the bestsuch xenograft implants thus reducing the number of follow-up surgeries to replace worn-up valves. We describean approach for automatic in-vitro evaluation of prosthetic heart valves in an artificial circulation system. Weshow how to measure the orifice area during a heart cycle to obtain an orifice curve. Different quality parametersare then estimated on such curves.Keywords: prosthetic heart valves, quality control, valve-orifice area, thresholding, snake1. INTRODUCTIONReplacing a deficient heart valve with a xenograft implant is a common procedure in cardiac surgery. 1 Porcineheart valves are preferred for this procedure because of their similarity to the human heart valve and the easyaccess to these valves. Typically a large number of porcine valves are available from specially bread animals.Therefore, quality assessment becomes a vital task to provide patients, surgeons and researchers with the mostsuitable material. Major quality criteria are the maximum value of the orifice area and the evolution of thisarea over a valve cycle – i.e. the time interval between the moment the valve opens until it closes. Althoughthere are already approaches to determine the valve-orifice area automatically, 2 this is still done manually inmost cases. In either case a sequence of images showing a valve cycle is used. The images are acquired withthe help of a test setup. A human operator has to select a set of points at the leaflets tips and connectionsto define six triangles. Summed areas give an approximation to the orifice area. This approximation neglectsthe rather curved boundary of the leaflets. Also the area is measured in only a few images of the sequence andfrom there, orifice curves – i.e. curves of area over frame index, showing the behavior of the valve over a cycle– are interpolated. The methods described in this contribution are intended to provide the researcher with amore precise, reproducible and robust technique to calculate the orifice area automatically. We first segment theorifice and then compute its area by counting all object pixels. To obtain an orifice curve, segmentation of allimages in an analyzed sequence is done by a combination of threshold- and snake-based methods.1.1. Data acquisitionThe area of the valve orifice is measured in each of a sequence of images acquired in vitro in a special test setup, 3which is shown in Figure 1 (a) and (b). From a reservoir (1) a transparent fluid – usually sterile water – istransported through a disc valve (4) by a piston pump (2), which is driven by a waveform adapted cam plate(3). After passing an input compliance (5) the fluid is pressed through the inspected heart valve (11) into avisualization chamber (7) located in another fluid reservoir (6). Pressure sensors (10) are installed below andA.P.C. and T.H. contributed equally. Send correspondence to A.P.C.: E-mail: condura@isip.uni-luebeck.de, Tele-phone: +49 (0)451 500 5801
