• Nobuhiko Hata, Yoshihiro Muragaki, Takashi Inomata, Takashi Maruyama, Hiroshi Iseki, Tomokatsu Hori, and Takeyoshi Dohi.
• Graduate School of Information Science and Technology, The University of Tokyo, Tokyo, Japan. noby@atre.t.u-tokyo.ac.jp
• Acad Radiol. 2005 Jan 1; 12 (1): 116-22.

Rationale And ObjectivesGross-total surgery under intraoperative magnetic resonance imaging (MRI) is a promising method of glioma removal. The purpose of this article is intraoperative measurement of resected tumor volume in MRI-guided glioma surgery using semiautomatic image segmentation to unbiased resection rate control.Materials And MethodsA newly developed software program based on a fuzzy connectedness (FC) segmentation algorithm was used to achieve fast and semiautomatic tumor segmentation and tumor volume measurement. The program was validated by retrospective study of eight glioma cases and then applied to seven glioma cases. All clinical cases underwent actual MRI-guided surgery using 0.3-T open magnets.ResultsThe volume of the tumor before resection ranged from 10.1 to 206.7 mL. A comparison of the results of manual segmentation with those of the semiautomatic FC-based segmentation gave an average dice similarity coefficient of 0.80 and an average match of 76%. Volume measurement combined with a developed software program enabled quantitative monitoring of tumor removal, which was critical in the near-total resection of glioma in MRI-guided surgery.ConclusionThe FC-based tumor segmentation method can be used for intraoperative tumor segmentation and volume measurement in MRI-guided glioma surgery using 0.3-T open magnets. This method is useful for objective resection rate monitoring, which may ultimately minimize the amount of residual tumor in glioma surgery.

31 keywords...

hide…