Performance of 12 DIR algorithms in low-contrast regions for mass and density conserving deformation

Abstract

The accuracy of DIR algorithms was quantitatively evaluated using a tissue equivalent, mass, and density conserving DEFGEL phantom. For the model studied, optical flow algorithms performed better than demons algorithms, with the original Horn and Schunck performing best. The degree of error is influenced more by the magnitude of displacement than the geometric complexity of the deformation. As might be expected, deformation is estimated less accurately for low-contrast regions than for high-contrast features, and the method presented here allows quantitative analysis of the differences. The evaluation of registration accuracy through observation of the same high contrast features that drive the DIR calculation is shown to be circular and hence misleading.

Recent Advances and Challenges in Adaptive Radiotherapy for Patients with Locally Advanced NSCLC

“…The accuracy of dose accumulation depends on the DIR and dose mapping methods used. Currently most registration algorithms could be accurate within 2 mm to 3 mm on average [40][41][42], which is comparable to the resolution of dose grids often used in clinic [43,44]. The spatial uncertainties may result in dose mapping errors up to 3 Gy/mm [45], but in clinical scenarios the impact of these errors could be limited [46].…”

“…It has been illustrated that intensity-based DIR algorithms are prone to have errors in regions with low image contrasts [40,42,83], and consequently, errors in dose reconstruction and response assessment may exist in these regions. Also due to the lack of knowledge on the pattern of tumor regression, how to deform the anatomical structures near by the tumor is still unclear [74,80].…”

Evaluation of dosimetric misrepresentations from 3D conventional planning of liver SBRT using 4D deformable dose integration

“…The latter were morphed to approximate the target image using the optical flow method of deformable image registration (DIR). Selection of the algorithm employed for this study was justified based on our previous study, 25 , 26 as well as those of other groups 15 , 17 who have similarly used it in the context of lung treatments. We have used the DIRART implementations (27) of these algorithms as these are freely available in the public domain and, thus, the present approach may be easily reproduced by interested readers.…”

Evaluation of various deformable image registrations for point and volume variations

“…Figure 5 shows the axial, sagital and coronal images of V id depending on the DIR algorithms and the optimized parameters of each DIR algorithms in DIRART were decided by the pre-study by Yeo et al [25].…”

“…On the other hand, this study generated the Yeo et al designed three types of deformation in order to evaluate the accuracy of DIR algorithms [1,25]. In this study, the reference volume was deformed into two types of deformation; relaxation and contraction.…”