Kai Ritschel1, Ioannis Pechlivanis, Susanne Winter. 1. Institut für Neuroinformatik, Ruhr-Universität Bochum, Universitätsstr. 150, 44780 , Bochum, Germany, kai.ritschel@ini.rub.de.
Abstract
PURPOSE: Contrast-enhanced ultrasound (CEUS) imaging of tissue perfusion is based on microbubble echo detection. CEUS can visualize tumors based on local perfusion variations. The acquired video data are qualitatively interpreted by subjective visualization in clinical practice. An automated CEUS classifier was developed for intraoperative identification of tumor tissue and especially tumor borders. METHODS: Support vector machines (SVM) were trained using CEU data sets to differentiate tumor and non-tumor tissue in glioblastoma patients. The classification was based on features derived from model functions approximated to time courses for each pixel in the video data. Classification performance was evaluated with single and cross- patient training data sets. RESULTS: The minimum mean classification error (14.6 %) with single patient data set training was achieved by SVM training using a sigmoid combination of model function parameter sets. A comparison of different model functions showed that the minimum average classification error (17.4 %) in a cross-validation study with 13 patients was achieved with the sigmoid model using an automatic relevance detection kernel. CONCLUSION: CEUS-based classification map images derived from approximated model functions can be generated with moderate accuracy and have significant potential to support intraoperative decisions concerning glioblastoma tumor borders.
PURPOSE: Contrast-enhanced ultrasound (CEUS) imaging of tissue perfusion is based on microbubble echo detection. CEUS can visualize tumors based on local perfusion variations. The acquired video data are qualitatively interpreted by subjective visualization in clinical practice. An automated CEUS classifier was developed for intraoperative identification of tumor tissue and especially tumor borders. METHODS: Support vector machines (SVM) were trained using CEU data sets to differentiate tumor and non-tumor tissue in glioblastomapatients. The classification was based on features derived from model functions approximated to time courses for each pixel in the video data. Classification performance was evaluated with single and cross- patient training data sets. RESULTS: The minimum mean classification error (14.6 %) with single patient data set training was achieved by SVM training using a sigmoid combination of model function parameter sets. A comparison of different model functions showed that the minimum average classification error (17.4 %) in a cross-validation study with 13 patients was achieved with the sigmoid model using an automatic relevance detection kernel. CONCLUSION: CEUS-based classification map images derived from approximated model functions can be generated with moderate accuracy and have significant potential to support intraoperative decisions concerning glioblastoma tumor borders.
Authors: Ursula Nemec; Stefan F Nemec; Clemens Novotny; Michael Weber; Christian Czerny; Christian R Krestan Journal: Eur Radiol Date: 2012-02-10 Impact factor: 5.315
Authors: M Claudon; D Cosgrove; T Albrecht; L Bolondi; M Bosio; F Calliada; J-M Correas; K Darge; C Dietrich; M D'Onofrio; D H Evans; C Filice; L Greiner; K Jäger; N de Jong; E Leen; R Lencioni; D Lindsell; A Martegani; S Meairs; C Nolsøe; F Piscaglia; P Ricci; G Seidel; B Skjoldbye; L Solbiati; L Thorelius; F Tranquart; H P Weskott; T Whittingham Journal: Ultraschall Med Date: 2008-02 Impact factor: 6.548
Authors: H Hashizume; P Baluk; S Morikawa; J W McLean; G Thurston; S Roberge; R K Jain; D M McDonald Journal: Am J Pathol Date: 2000-04 Impact factor: 4.307
Authors: Edoardo Vicenzini; Roberto Delfini; Francesca Magri; Francesco Puccinelli; Marta Altieri; Antonio Santoro; Maria Fabrizia Giannoni; Luigi Bozzao; Vittorio Di Piero; Gian Luigi Lenzi Journal: J Ultrasound Med Date: 2008-05 Impact factor: 2.153
Authors: Justin S Smith; Edward F Chang; Kathleen R Lamborn; Susan M Chang; Michael D Prados; Soonmee Cha; Tarik Tihan; Scott Vandenberg; Michael W McDermott; Mitchel S Berger Journal: J Clin Oncol Date: 2008-03-10 Impact factor: 44.544
Authors: Zoe Z Zhang; Lisa B E Shields; David A Sun; Yi Ping Zhang; Matthew A Hunt; Christopher B Shields Journal: Front Oncol Date: 2015-07-30 Impact factor: 6.244
Authors: Ling-Gang Cheng; Wen He; Hong-Xia Zhang; Qian Song; Bin Ning; Hui-Zhan Li; Yan He; Song Lin Journal: Biomed Res Int Date: 2016-03-16 Impact factor: 3.411