Automatic localization of cochlear implant electrodes using cone beam computed tomography images.
https://doi.org/10.1007/s11192-009-2063-x
Jasmin Thormählen Department of Otolaryngology, Hannover Medical School, Karl-Wiechert-Allee 3, 30625, Hannover, Germany.
Benjamin Krüger Department of Otolaryngology, Hannover Medical School, Karl-Wiechert-Allee 3, 30625, Hannover, Germany.
Cluster of Excellence Hearing4all, Hannover, Germany.
Waldo Nogueira Department of Otolaryngology, Hannover Medical School, Karl-Wiechert-Allee 3, 30625, Hannover, Germany. nogueiravazquez.waldo@mh-hannover.de.
Cluster of Excellence Hearing4all, Hannover, Germany. nogueiravazquez.waldo@mh-hannover.de.
Cochlear implants (CI) are implantable medical devices that enable the perception of sounds and the understanding of speech by electrically stimulating the auditory nerve in case of inner ear damage. The stimulation takes place via an array of electrodes surgically inserted in the cochlea. After CI implantation, cone beam computed tomography (CBCT) is used to evaluate the position of the electrodes. Moreover, CBCT is used in research studies to investigate the relationship between the position of the electrodes and the hearing outcome of CI user. In clinical routine, the estimation of the position of the CI electrodes is done manually, which is very time-consuming.
The aim of this study was to optimize procedures of automatic electrode localization from CBCT data following CI implantation. For this, we analyzed the performance of automatic electrode localization for 150 CBCT data sets of 10 different types of electrode arrays. Our own implementation of the method by Noble and Dawant (Lecture notes in computer science (Including subseries lecture notes in artificial intelligence and lecture notes in bioinformatics), Springer, pp 152-159, 2015. https://doi.org/10.1007/978-3-319-24571-3_19 ) for automated electrode localization served as a benchmark for evaluation. Differences in the detection rate and the localization accuracy across types of electrode arrays were evaluated and errors were classified. Based on this analysis, we developed a strategy to optimize procedures of automatic electrode localization. It was shown that particularly distantly spaced electrodes in combination with a deep insertion can lead to apical-basal confusions in the localization procedure. This confusion prevents electrodes from being detected or assigned correctly, leading to a deterioration in localization accuracy.
We propose an extended cost function for automatic electrode localization methods that prevents double detection of electrodes to avoid apical-basal confusions. This significantly increased the detection rate by 11.15 percent points and improved the overall localization accuracy by 0.53 mm (1.75 voxels). In comparison to other methods, our proposed cost function does not require any prior knowledge about the individual cochlea anatomy.
Google Scholar (2007), Retrieved December 20, 2007, from http://scholar.google.com
Ma, F. C., Song, E. M. (2006), An author co-citation analysis of information science in China, Journal of the China Society for Scientific and Technical Information, 25(3): 259–268.
10.1002/(SICI)1097-4571(199009)41:6<433::AID-ASI11>3.0.CO;2-Q
Qiu, J. P., Informatrics. Wuhan: Wuhan University Press, 2007.
Qiu, J. P., Ma, R. M., Li, Y. J. (2008). New thoughts about co-citation analysis, Journal of the China Society for Scientific and Technical Information (in Chinese), 27(1): 69–74.
Su, X. L. (2006), Report on academic influence in library, information and documentation science (2000–2004), Journal of the China Society for Scientific and Technical Information, 25(2): 131–153.
White, H. D., Mccain, K. W. (1998), Visualizing a discipline: An author cocitation analysis of information science, 1972–1995, Journal of the American Society for Information Science, 49(4): 327–355.