Predictive modeling in glioma grading from mr perfusion images using support vector machines.
October 6th, 2008 by 
admin | Related Articles |
Predictive moulding in glioma rating from MR perfusion images using hold agent machines.
Magn Reson Med. 2008 Sep 24;60(4):945-952
Authors: Emblem KE, Zoellner FG, Tennoe B, Nedregaard B, town T, Due-Tonnessen P, Hald JK, Scheie D, Bjornerud A
The advantages of prophetic moulding in glioma rating from MR perfusion images hit not still been explored. The intend of the underway think was to compel a prophetic help supported on hold agent machines (SVM) for glioma rating using growth murder intensity histogram signatures derivative from MR perfusion images and to set the characteristic quality of the help and the sense to distribution size. A amount of 86 patients with histologically-confirmed gliomas were imaged using impulsive status oppositeness (DSC) tomography at 1.5T. Histogram signatures from 53 of the 86 patients were analyzed independently by quaternary neuroradiologists and utilised as a foundation for the prophetic SVM model. The resulting SVM help was proven on the remaining 33 patients and analyzed by a ordinal neuroradiologist. At best SVM parameters, the genuine constructive evaluate (TPR) and genuine perverse evaluate (TNR) of the SVM help on the 33 patients was 0.76 and 0.82, respectively. The interobserver commendation and the TPR accumulated significantly when the SVM help was supported on an crescendo distribution filler (P < 0.001). This termination suggests that a prophetic SVM help crapper assistance in the identification of glioma evaluate from MR perfusion images and that the help improves with crescendo distribution size. Magn Reson Med 60:945-952, 2008. (c) 2008 Wiley-Liss, Inc.
PMID: 18816815 [PubMed - as supplied by publisher]
(Source: Magnetic Resonance in Medicine)
Posted in Cancer | 