Limits...
Segmentation of corpus callosum using diffusion tensor imaging: validation in patients with glioblastoma.

Nazem-Zadeh MR, Saksena S, Babajani-Fermi A, Jiang Q, Soltanian-Zadeh H, Rosenblum M, Mikkelsen T, Jain R - BMC Med Imaging (2012)

Bottom Line: In this algorithm, diffusion pattern of corpus callosum was used as prior information.Dice coefficients, estimated to compare automatic and manual segmentation results for Witelson subdivisions, ranged from 94% to 98% for control subjects and from 81% to 95% for tumor patients, illustrating closeness of automatic and manual segmentations.Studying the effect of corpus callosum rotation by different Euler angles showed that although segmentation results were more sensitive to azimuth and elevation than skew, rotations caused by brain tumors do not have major effects on the segmentation results.

View Article: PubMed Central - HTML - PubMed

Affiliation: Control and Intelligent Processing Center of Excellence, School of Electrical and Computer Engineering, University of Tehran, Tehran 14399, Iran.

ABSTRACT

Background: This paper presents a three-dimensional (3D) method for segmenting corpus callosum in normal subjects and brain cancer patients with glioblastoma.

Methods: Nineteen patients with histologically confirmed treatment naïve glioblastoma and eleven normal control subjects underwent DTI on a 3T scanner. Based on the information inherent in diffusion tensors, a similarity measure was proposed and used in the proposed algorithm. In this algorithm, diffusion pattern of corpus callosum was used as prior information. Subsequently, corpus callosum was automatically divided into Witelson subdivisions. We simulated the potential rotation of corpus callosum under tumor pressure and studied the reproducibility of the proposed segmentation method in such cases.

Results: Dice coefficients, estimated to compare automatic and manual segmentation results for Witelson subdivisions, ranged from 94% to 98% for control subjects and from 81% to 95% for tumor patients, illustrating closeness of automatic and manual segmentations. Studying the effect of corpus callosum rotation by different Euler angles showed that although segmentation results were more sensitive to azimuth and elevation than skew, rotations caused by brain tumors do not have major effects on the segmentation results.

Conclusions: The proposed method and similarity measure segment corpus callosum by propagating a hyper-surface inside the structure (resulting in high sensitivity), without penetrating into neighboring fiber bundles (resulting in high specificity).

Show MeSH

Related in: MedlinePlus

Comparing of corpus callosum segmentations by the proposed method and the Jonasson's method [25]. In (a) for the Jonasson's method, the speed threshold was chosen low enough to capture the corpus callosum structure. However, the segmentation front propagates inside irrelevant fiber structures such as superior longitudinal fasciculus (cyan rounded rectangles), cingulum (green rounded rectangles), minor forceps (red rounded rectangles), and tracts of corona-radiata (yellow rounded rectangles). In (b), our method segments the corpus callosum structure with high sensitivity (segmenting the whole structure) and high specificity (without major leakage into the neighboring structures).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3368740&req=5

Figure 4: Comparing of corpus callosum segmentations by the proposed method and the Jonasson's method [25]. In (a) for the Jonasson's method, the speed threshold was chosen low enough to capture the corpus callosum structure. However, the segmentation front propagates inside irrelevant fiber structures such as superior longitudinal fasciculus (cyan rounded rectangles), cingulum (green rounded rectangles), minor forceps (red rounded rectangles), and tracts of corona-radiata (yellow rounded rectangles). In (b), our method segments the corpus callosum structure with high sensitivity (segmenting the whole structure) and high specificity (without major leakage into the neighboring structures).

Mentions: The tensor-based method proposed by Jonasson et al. [32] with NTPP similarity measure is quite sensitive to the speed threshold. If the speed threshold is selected high enough to prevent the front from propagating into the neighboring structures, corpus callosum is not segmented entirely (low sensitivity). However, if the speed threshold is chosen low enough to segment the entire corpus callosum, the front propagates into irrelevant fiber structures such as superior longitudinal fasciculus, cingulum, minor forceps, and tracts of corona-radiata (low specificity). Figure 4a shows the results of our implementation of their method. In this figure, we tuned the speed threshold so that the whole structure is segmented. Figure 4b demonstrates high sensitivity (segmenting almost the whole structure) and high specificity (without major penetration into the neighbouring structures) of our proposed method compared to the Jonasson's method.


Segmentation of corpus callosum using diffusion tensor imaging: validation in patients with glioblastoma.

Nazem-Zadeh MR, Saksena S, Babajani-Fermi A, Jiang Q, Soltanian-Zadeh H, Rosenblum M, Mikkelsen T, Jain R - BMC Med Imaging (2012)

Comparing of corpus callosum segmentations by the proposed method and the Jonasson's method [25]. In (a) for the Jonasson's method, the speed threshold was chosen low enough to capture the corpus callosum structure. However, the segmentation front propagates inside irrelevant fiber structures such as superior longitudinal fasciculus (cyan rounded rectangles), cingulum (green rounded rectangles), minor forceps (red rounded rectangles), and tracts of corona-radiata (yellow rounded rectangles). In (b), our method segments the corpus callosum structure with high sensitivity (segmenting the whole structure) and high specificity (without major leakage into the neighboring structures).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3368740&req=5

Figure 4: Comparing of corpus callosum segmentations by the proposed method and the Jonasson's method [25]. In (a) for the Jonasson's method, the speed threshold was chosen low enough to capture the corpus callosum structure. However, the segmentation front propagates inside irrelevant fiber structures such as superior longitudinal fasciculus (cyan rounded rectangles), cingulum (green rounded rectangles), minor forceps (red rounded rectangles), and tracts of corona-radiata (yellow rounded rectangles). In (b), our method segments the corpus callosum structure with high sensitivity (segmenting the whole structure) and high specificity (without major leakage into the neighboring structures).
Mentions: The tensor-based method proposed by Jonasson et al. [32] with NTPP similarity measure is quite sensitive to the speed threshold. If the speed threshold is selected high enough to prevent the front from propagating into the neighboring structures, corpus callosum is not segmented entirely (low sensitivity). However, if the speed threshold is chosen low enough to segment the entire corpus callosum, the front propagates into irrelevant fiber structures such as superior longitudinal fasciculus, cingulum, minor forceps, and tracts of corona-radiata (low specificity). Figure 4a shows the results of our implementation of their method. In this figure, we tuned the speed threshold so that the whole structure is segmented. Figure 4b demonstrates high sensitivity (segmenting almost the whole structure) and high specificity (without major penetration into the neighbouring structures) of our proposed method compared to the Jonasson's method.

Bottom Line: In this algorithm, diffusion pattern of corpus callosum was used as prior information.Dice coefficients, estimated to compare automatic and manual segmentation results for Witelson subdivisions, ranged from 94% to 98% for control subjects and from 81% to 95% for tumor patients, illustrating closeness of automatic and manual segmentations.Studying the effect of corpus callosum rotation by different Euler angles showed that although segmentation results were more sensitive to azimuth and elevation than skew, rotations caused by brain tumors do not have major effects on the segmentation results.

View Article: PubMed Central - HTML - PubMed

Affiliation: Control and Intelligent Processing Center of Excellence, School of Electrical and Computer Engineering, University of Tehran, Tehran 14399, Iran.

ABSTRACT

Background: This paper presents a three-dimensional (3D) method for segmenting corpus callosum in normal subjects and brain cancer patients with glioblastoma.

Methods: Nineteen patients with histologically confirmed treatment naïve glioblastoma and eleven normal control subjects underwent DTI on a 3T scanner. Based on the information inherent in diffusion tensors, a similarity measure was proposed and used in the proposed algorithm. In this algorithm, diffusion pattern of corpus callosum was used as prior information. Subsequently, corpus callosum was automatically divided into Witelson subdivisions. We simulated the potential rotation of corpus callosum under tumor pressure and studied the reproducibility of the proposed segmentation method in such cases.

Results: Dice coefficients, estimated to compare automatic and manual segmentation results for Witelson subdivisions, ranged from 94% to 98% for control subjects and from 81% to 95% for tumor patients, illustrating closeness of automatic and manual segmentations. Studying the effect of corpus callosum rotation by different Euler angles showed that although segmentation results were more sensitive to azimuth and elevation than skew, rotations caused by brain tumors do not have major effects on the segmentation results.

Conclusions: The proposed method and similarity measure segment corpus callosum by propagating a hyper-surface inside the structure (resulting in high sensitivity), without penetrating into neighboring fiber bundles (resulting in high specificity).

Show MeSH
Related in: MedlinePlus