Semi-automated brain tumor segmentation on multi-parametric MRI using regularized non-negative matrix factorization

BMC Medical Imaging

Research Article

Frederik Maes¹  Nicolas Sauwen²  Diana M. Sima²  Sabine Van Huffel²  Uwe Himmelreich³  Jelle Veraart⁴  Eric Achten⁵  Marjan Acou⁵ 

[1] Department of Electrical Engineering (ESAT), PSI Centre for Processing Speech and Images, KULeuven, Kasteelpark Arenberg 10, 3001, Leuven, Belgium;Department of Electrical Engineering (ESAT), STADIUS Centre for Dynamical Systems, Signal Processing and Data Analytics, KULeuven, Kasteelpark Arenberg, Leuven, Belgium;imec, Kapeldreef 75, 3001, Leuven, Belgium;Department of Imaging and Pathology, Biomedical MRI/MoSAIC, KULeuven, Herestraat 49, 3000, Leuven, Belgium;Department of Physics, iMinds Vision Lab, University of Antwerp, Edegemsesteenweg 200–240, 2610, Antwerp, Belgium;Department of Radiology, Ghent University Hospital, De Pintelaan 185, 9000, Ghent, Belgium;

关键词: MRI;    Segmentation;    Brain tumors;    Non-negative matrix factorization;    Unsupervised classification;   

DOI  :  10.1186/s12880-017-0198-4

 received in 2016-09-21, accepted in 2017-04-11,  发布年份 2017

来源: Springer

【 摘 要 】

BackgroundSegmentation of gliomas in multi-parametric (MP-)MR images is challenging due to their heterogeneous nature in terms of size, appearance and location. Manual tumor segmentation is a time-consuming task and clinical practice would benefit from (semi-) automated segmentation of the different tumor compartments.MethodsWe present a semi-automated framework for brain tumor segmentation based on non-negative matrix factorization (NMF) that does not require prior training of the method. L1-regularization is incorporated into the NMF objective function to promote spatial consistency and sparseness of the tissue abundance maps. The pathological sources are initialized through user-defined voxel selection. Knowledge about the spatial location of the selected voxels is combined with tissue adjacency constraints in a post-processing step to enhance segmentation quality. The method is applied to an MP-MRI dataset of 21 high-grade glioma patients, including conventional, perfusion-weighted and diffusion-weighted MRI. To assess the effect of using MP-MRI data and the L1-regularization term, analyses are also run using only conventional MRI and without L1-regularization. Robustness against user input variability is verified by considering the statistical distribution of the segmentation results when repeatedly analyzing each patient’s dataset with a different set of random seeding points.ResultsUsing L1-regularized semi-automated NMF segmentation, mean Dice-scores of 65%, 74 and 80% are found for active tumor, the tumor core and the whole tumor region. Mean Hausdorff distances of 6.1 mm, 7.4 mm and 8.2 mm are found for active tumor, the tumor core and the whole tumor region. Lower Dice-scores and higher Hausdorff distances are found without L1-regularization and when only considering conventional MRI data.ConclusionsBased on the mean Dice-scores and Hausdorff distances, segmentation results are competitive with state-of-the-art in literature. Robust results were found for most patients, although careful voxel selection is mandatory to avoid sub-optimal segmentation.

【 授权许可】

CC BY   
© The Author(s) 2017

【 图 表 】

Fig. 3

Fig. 2

Fig. 7

12951_2017_255_Article_IEq54.gif

Fig. 5

12951_2017_255_Article_IEq53.gif

Fig. 2

12888_2023_5292_Article_IEq1.gif

Fig. 1

Fig. 1

Fig. 2

12951_2016_246_Article_IEq16.gif

12951_2016_246_Article_IEq15.gif

Fig. 6

12951_2016_246_Article_IEq13.gif

12951_2016_246_Article_IEq12.gif

Fig. 1

Fig. 1

Fig. 9

Fig. 4

12951_2017_255_Article_IEq52.gif

12951_2017_255_Article_IEq51.gif

Fig. 1

Fig. 4

Fig. 10