Pelvic Bone Tumor and Cortex Visualization from 3D CT Data

A single-scale sheet filter was applied to pelvic CT images for bone cortex enhancement. The purpose was to visualize the distribution of bone tumors and localize them in relation to the pelvic structure for biopsy planning as well as diagnosis [32]. Healthy bone cortex tissues and bone tumors have similar original CT values. However, bone cortices are sheet-like in structure, while tumors are not. Thus, enhanced bone cortices using sheet enhancement filtering are expected to be discriminated from tumors which are not enhanced.

The CT dataset consisted of 40 slices with a 512 x 512 matrix (Fig. 10.6(a)); the pixel dimensions were 0.82 mm2. The slice thickness and reconstruction pitch were 5 mm. The matrix was reduced to half in the xy-plane, and thus the pixel interval was 1.64 mm. Sheet filtering was applied to the sinc-interpolated images using af = 1.0 pixel, y23 = y13 = 0.5, and a = 0.25.

Figure 10.6(b) shows the color volume renderings of bone tumors (pink) and cortices (white). In the left frame, the opacity and color functions were adjusted using only CT values of the original images. In the right frame, both the original and sheet filtered images were used, where voxels having high intensities

Figure 10.6: Visualization of pelvic bone tumors from CT data. (a) Original CT slice image. (b) Volume rendered images of bone tumors and cortices. Left: Using only original images. Right: Using original and sheet-filtered images. (c) Manually traced tumor regions. (( 2004 IEEE). A color version of this figure will appear on the CD that accompanies the volume.

Figure 10.6: Visualization of pelvic bone tumors from CT data. (a) Original CT slice image. (b) Volume rendered images of bone tumors and cortices. Left: Using only original images. Right: Using original and sheet-filtered images. (c) Manually traced tumor regions. (( 2004 IEEE). A color version of this figure will appear on the CD that accompanies the volume.

both in sheet-filtered and original images were assigned as cortices (white), and those having high in original but low in sheet-filtered images were assigned as tumors (pink). Figure 10.6(c) shows the rendered color image generated from the tumor regions manually traced by a radiology specialist, which is regarded as an ideal visualization. The color rendering of the left frame of Fig. 10.6(b) was well correlated with Fig. 10.6(c) (the "ideal" image), and the bone tumors were visualized considerably better than only using original CT images. However, nontumor regions around articular spaces were also detected mainly due to the partial volume effect (by a large slice thickness).

0 0

Post a comment