The performance evaluation of a semi-supervised fuzzy c-means (SFCM) c
lustering method for monitoring brain tumor volume changes during the
course of routine clinical radiation-therapeutic and chemo-therapeutic
regimens is presented, The tumor volume determined using the SFCM met
hod was compared with the volume estimates obtained using three other
methods: (a) a k nearest neighbor (kNN) classifier, b) a grey level th
resholding and seed growing (ISG-SG) method and c) a manual pixel labe
ling (GT) method for ground truth estimation, The SFCM and kNN methods
are applied to the multispectral, contrast enhanced T-1, proton densi
ty, and T-2 weighted, magnetic resonance images (MRI) whereas the ISG-
SG and GT methods are applied only to the contrast enhanced T-1 weight
ed image, Estimations of tumor volume were made on eight patient cases
with follow-up MRI scans performed over a 32 week interval during tre
atment, The tumor cases studied include one meningioma, two brain meta
stases and five gliomas, Comparisons with manually labeled ground trut
h estimations showed that there is a limited agreement between the seg
mentation methods for absolute tumor volume measurements when using im
ages of patients after treatment, The average intraobserver reproducib
ility for the SFCM, kNN and ISG-SG methods was found to be 5.8%, 6.6%
and 8.9%, respectively, The average of the interobserver reproducibili
ty of these methods was found to be 5.5%, 6.5% and 11.4%, respectively
, For the measurement of relative change of tumor volume as required f
or the response assessment, the multi-spectral methods kNN and SFCM ar
e therefore preferred over the seed growing method. (C) 1997 Elsevier
Science Inc.