Room: Exhibit Hall | Forum 2
Purpose: When patients require adaptive radiotherapy or a replan-CT, clinical target volumes(CTV) and organs at-risk(OARs) must be re-delineated. This work evaluates and validates the auto-segmentation based on self-registration and atlas in cervical cancer using MIM-Maestro software(version 6.6.5).
Methods: Sixty cervical patients were enrolled to establish the atlas in MIM-Maestroï¼Œanother 15 patients with planning CT (pCT) and replan CT (rCT) were selected randomly.One experienced radiation oncologist contoured CTV and 6 OARs for the pCT and rCT according to published guidelines and the contours of the rCT were defined as reference volumes. The rCT of 15 patients were auto-contoured using atlas-based auto-segmentation (atlas group), and mapping contours from the pCT to the rCT by rigid and deformable image registration (rigid group and deformable group).The time for segmentation was also recorded. The similarity of the auto-contours and reference contours was assessed using dice similarity coefficient(DSC), overlap index(OI), mean hausdorff distance(HDmean) and deviation of centroid, and the results of three groups were evaluated statistically using one-way Analysis of Variance.
Results: The mean time was 89.2s, 22.4s and 42.6s for atlas group, rigid group and deformable group. The most significant difference was found between rigid group, deformable group and atlas group in DSC, OI and HDmean(p<0.001) for CTV and rectum,while only in OI(p=0.02,0.014) for bowel. The mean DSC for CTV is 0.89(rigid group and deformable group), 0.76(atlas group). The deformable group agreed more closely with the reference volumes for bladder, pelvis and femoral heads.
Conclusion: The three methods of auto-segmentation have the excellent ability to contour CTV and OARs,and the deformable group is better than rigid group and atlas group. The automatic contouring workflow can decrease the delineating time significantly and it can be used in adaptive and replan radiotherapy.
Segmentation, Radiation Therapy