Demons Deformable Registration Research Articles

Dosimetric evaluation of 4D‐CBCT reconstructed by Simultaneous Motion Estimation and Image Reconstruction (SMEIR) for carbon ion therapy of lung cancer

Motion management is critical for the efficacy of carbon ion therapy for moving targets such as lung tumors. We evaluated the feasibility of using four-dimensional cone beam computed tomography (4D-CBCT) reconstructed by Simultaneous Motion Estimation and Image Reconstruction (SMEIR) for dose calculation and accumulation in carbon ion treatment of lung cancer. Motion-compensated 4D-CBCT images were reconstructed with the SMEIR algorithm to capture the most updated anatomy and motion with an updated interphase motion model on the treatment day. Projections of all CBCT phases were simulated from the planning 4D-CT by the ray tracing technique. Treatment planning and dose calculation were performed with a GPU-based Monte Carlo dose calculation software for carbon ion therapy. The treatment plan was optimized on the average computed tomography (CT) to obtain optimal intensity of the carbon ions. From the optimized plan, dose distributions on individual phases of 4D-CT and 4D-CBCT were calculated by the Monte Carlo-based dose engine. Dose accumulation was performed on 4D-CBCT images using deformable vector fields (DVF) generated by SMEIR. The accumulated planning target volume (PTV) dose based on 4D-CBCT was then compared to the accumulated dose calculated on 4D-CT, where the DVFs between different phases were obtained by the demons deformable registration algorithm. Dose value histograms (DVH) as well as absolute deviations of the maximum dose ( ), mean dose ( ), and dose coverage metrics ( and ) for PTV were quantitatively evaluated for the two sets of plans. Good agreement was found between the 4D-CT and 4D-CBCT-based PTV-DVH curves. The average values of , , and calculated between the 4D-CT and SMEIR-4D-CBCT-based plans were , , 2.12%, and , respectively, for the PTVs of ten patient case studies. Based on these results, SMEIR-reconstructed 4D-CBCTs can potentially be used for motion estimation, dose evaluation, and adaptive treatment planning in lung cancer carbon ion therapy.

Novel Super-Resolution Approach to Time-Resolved Volumetric 4-Dimensional Magnetic Resonance Imaging With High Spatiotemporal Resolution for Multi-Breathing Cycle Motion Assessment

To develop and evaluate a super-resolution approach to reconstruct time-resolved 4-dimensional magnetic resonance imaging (TR-4DMRI) with a high spatiotemporal resolution for multi-breathing cycle motion assessment. A super-resolution approach was developed to combine fast 3-dimensional (3D) cine MRI with low resolution during free breathing (FB) and high-resolution 3D static MRI during breath hold (BH) using deformable image registration. A T1-weighted, turbo field echo sequence, coronal 3D cine acquisition, partial Fourier approximation, and SENSitivity Encoding parallel acceleration were used. The same MRI pulse sequence, field of view, and acceleration techniques were applied in both FB and BH acquisitions; the intensity-based Demons deformable image registration method was used. Under an institutional review board-approved protocol, 7 volunteers were studied with 3D cine FB scan (voxel size: 5×5×5mm3) at 2Hz for 40seconds and a 3D static BH scan (2×2×2mm3). To examine the image fidelity of 3D cine and super-resolution TR-4DMRI, a mobile gel phantom with multi-internal targets was scanned at 3 speeds and compared with the 3D static image. Image similarity among 3D cine, 4DMRI, and 3D static was evaluated visually using difference image and quantitatively using voxel intensity correlation and Dice index (phantom only). Multi-breathing-cycle waveforms were extracted and compared in both phantom and volunteer images using the 3D cine as the references. Mild imaging artifacts were found in the 3D cine and TR-4DMRI of the mobile gel phantom with a Dice index of >0.95. Among 7 volunteers, the super-resolution TR-4DMRI yielded high voxel-intensity correlation (0.92±0.05) and low voxel-intensity difference (<0.05). The detected motion differences between TR-4DMRI and 3D cine were -0.2±0.5mm (phantom) and -0.2±1.9mm (diaphragms). Super-resolution TR-4DMRI has been reconstructed with adequate temporal (2Hz) and spatial (2×2×2mm3) resolutions. Further TR-4DMRI characterization and improvement are necessary before clinical applications. Multi-breathing cycles can be examined, providing patient-specific breathing irregularities and motion statistics for future 4D radiation therapy.

Performance evaluation of MIND demons deformable registration of MR and CT images in spinal interventions

Accurate intraoperative localization of target anatomy and adjacent nervous and vascular tissue is essential to safe, effective surgery, and multimodality deformable registration can be used to identify such anatomy by fusing preoperative CT or MR images with intraoperative images. A deformable image registration method has been developed to estimate viscoelastic diffeomorphisms between preoperative MR and intraoperative CT using modality-independent neighborhood descriptors (MIND) and a Huber metric for robust registration. The method, called MIND Demons, optimizes a constrained symmetric energy functional incorporating priors on smoothness, geodesics, and invertibility by alternating between Gauss–Newton optimization and Tikhonov regularization in a multiresolution scheme. Registration performance was evaluated for the MIND Demons method with a symmetric energy formulation in comparison to an asymmetric form, and sensitivity to anisotropic MR voxel-size was analyzed in phantom experiments emulating image-guided spine-surgery in comparison to a free-form deformation (FFD) method using local mutual information (LMI). Performance was validated in a clinical study involving 15 patients undergoing intervention of the cervical, thoracic, and lumbar spine. The target registration error (TRE) for the symmetric MIND Demons formulation (1.3 ± 0.8 mm (median ± interquartile)) outperformed the asymmetric form (3.6 ± 4.4 mm). The method demonstrated fairly minor sensitivity to anisotropic MR voxel size, with median TRE ranging 1.3–2.9 mm for MR slice thickness ranging 0.9–9.9 mm, compared to TRE = 3.2–4.1 mm for LMI FFD over the same range. Evaluation in clinical data demonstrated sub-voxel TRE (<2 mm) in all fifteen cases with realistic deformations that preserved topology with sub-voxel invertibility (0.001 mm) and positive-determinant spatial Jacobians. The approach therefore appears robust against realistic anisotropic resolution characteristics in MR and yields registration accuracy suitable to application in image-guided spine-surgery.

Temporal subtraction contrast-enhanced dedicated breast CT

The development of a framework of deformable image registration and segmentation for the purpose of temporal subtraction contrast-enhanced breast CT is described.An iterative histogram-based two-means clustering method was used for the segmentation. Dedicated breast CT images were segmented into background (air), adipose, fibroglandular and skin components. Fibroglandular tissue was classified as either normal or contrast-enhanced then divided into tiers for the purpose of categorizing degrees of contrast enhancement. A variant of the Demons deformable registration algorithm, intensity difference adaptive Demons (IDAD), was developed to correct for the large deformation forces that stemmed from contrast enhancement. In this application, the accuracy of the proposed method was evaluated in both mathematically-simulated and physically-acquired phantom images. Clinical usage and accuracy of the temporal subtraction framework was demonstrated using contrast-enhanced breast CT datasets from five patients. Registration performance was quantified using normalized cross correlation (NCC), symmetric uncertainty coefficient, normalized mutual information (NMI), mean square error (MSE) and target registration error (TRE).The proposed method outperformed conventional affine and other Demons variations in contrast enhanced breast CT image registration. In simulation studies, IDAD exhibited improvement in MSE (0–16%), NCC (0–6%), NMI (0–13%) and TRE (0–34%) compared to the conventional Demons approaches, depending on the size and intensity of the enhancing lesion. As lesion size and contrast enhancement levels increased, so did the improvement. The drop in the correlation between the pre- and post-contrast images for the largest enhancement levels in phantom studies is less than 1.2% (150 Hounsfield units). Registration error, measured by TRE, shows only submillimeter mismatches between the concordant anatomical target points in all patient studies. The algorithm was implemented using a parallel processing architecture resulting in rapid execution time for the iterative segmentation and intensity-adaptive registration techniques.Characterization of contrast-enhanced lesions is improved using temporal subtraction contrast-enhanced dedicated breast CT. Adaptation of Demons registration forces as a function of contrast-enhancement levels provided a means to accurately align breast tissue in pre- and post-contrast image acquisitions, improving subtraction results. Spatial subtraction of the aligned images yields useful diagnostic information with respect to enhanced lesion morphology and uptake.

Mooney-Rivlin biomechanical modeling of lung with Inhomogeneous material.

In this study, the Mooney-Rivlin material with hyperelastic strain energy was proposed for biomechanical modeling of the lung. We modeled the lung as an inhomogeneous Mooney-Rivlin material with uncoupled deviatoric and volumetric behavior. The proposed method was evaluated on the lungs of eight lung cancer patients. For each patient, the lung was segmented from the 4D-CT images and tetrahedral volume mesh of the lung in phase 50% was created by using the adaptive mesh generation toolkit. The demons deformable registration algorithm was used to extract the displacement vector fields (DVFs). The Jacobian of the deformation gradient was calculated from DVFs, and the lung strain energy function was optimized to improve the tumor center of mass (TCM) motion simulation accuracy between respiratory phase 50% and 0%. The average TCM motion simulation error for the proposed strategy is 1.95 mm for eight patients. We observed 13% improvement in the TCM position prediction compared with the homogeneous Mooney-Rivlin modeling.

TU‐F‐CAMPUS‐J‐02: Mooney‐Rivlin Biomechanical Modeling of Lung with Inhomogeneous Material Property

Purpose:The Mooney‐Rivlin material with hyperelastic strain energy has been proposed for realistic biomechanical modeling of lung. In this study, the lung is modeled as an inhomogeneous Mooney‐Rivlin material with the incompressibility factors being optimized to improve the tumor center of mass (TCM) motion simulation accuracy during respiration.Method:ITK‐SNAP was used to segment lungs of eight lung cancer patients from the 4D‐CT images and tetrahedral volume meshes of the lungs in phase 50% were created by using adaptive mesh generation toolkit. The interphase deformation vector fields (DVFs) are calculated by demons deformable registration algorithm and the barycentric coordinate system of tetrahedral elements is obtained from the resulted DVFs. Mooney‐Rivlin hyperelastic material is used to model the lung volume. Each element is considered unique where the incompressibility factor (k‐factor) for each element is assumed to be proportional to the magnitude of normalized DVF. The incompressibility factor for each element was optimized by minimizing the tumor center of mass motion simulation error.Results:If lung is considered as a homogenous material in Mooney‐Rivlin modeling, the average TCM motion simulation error is 2.26 mm. By considering inhomogeneous properties of lung in the proposed strategy, the average TCM motion simulation error is reduced to 2.04 mm.Conclusions:We proposed a method for assigning the inhomogeneous biomechanical material in the Mooney‐Rivlin model of lung based on the lung regional deformation vector fields. Inhomogeneous material property of lung improves the simulation accuracy.

Effect of deformable registration on the dose calculated in radiation therapy planning CT scans of lung cancer patients.

To characterize the effects of deformable image registration of serial computed tomography (CT) scans on the radiation dose calculated from a treatment planning scan. Eighteen patients who received curative doses (≥ 60 Gy, 2 Gy/fraction) of photon radiation therapy for lung cancer treatment were retrospectively identified. For each patient, a diagnostic-quality pretherapy (4-75 days) CT scan and a treatment planning scan with an associated dose map were collected. To establish correspondence between scan pairs, a researcher manually identified anatomically corresponding landmark point pairs between the two scans. Pretherapy scans then were coregistered with planning scans (and associated dose maps) using the demons deformable registration algorithm and two variants of the Fraunhofer MEVIS algorithm ("Fast" and "EMPIRE10"). Landmark points in each pretherapy scan were automatically mapped to the planning scan using the displacement vector field output from each of the three algorithms. The Euclidean distance between manually and automatically mapped landmark points (dE) and the absolute difference in planned dose (|ΔD|) were calculated. Using regression modeling, |ΔD| was modeled as a function of dE, dose (D), dose standard deviation (SD(dose)) in an eight-pixel neighborhood, and the registration algorithm used. Over 1400 landmark point pairs were identified, with 58-93 (median: 84) points identified per patient. Average |ΔD| across patients was 3.5 Gy (range: 0.9-10.6 Gy). Registration accuracy was highest using the Fraunhofer MEVIS EMPIRE10 algorithm, with an average dE across patients of 5.2 mm (compared with >7 mm for the other two algorithms). Consequently, average |ΔD| was also lowest using the Fraunhofer MEVIS EMPIRE10 algorithm. |ΔD| increased significantly as a function of dE (0.42 Gy/mm), D (0.05 Gy/Gy), SD(dose) (1.4 Gy/Gy), and the algorithm used (≤ 1 Gy). An average error of <4 Gy in radiation dose was introduced when points were mapped between CT scan pairs using deformable registration, with the majority of points yielding dose-mapping error <2 Gy (approximately 3% of the total prescribed dose). Registration accuracy was highest using the Fraunhofer MEVIS EMPIRE10 algorithm, resulting in the smallest errors in mapped dose. Dose differences following registration increased significantly with increasing spatial registration errors, dose, and dose gradient (i.e., SDdose). This model provides a measurement of the uncertainty in the radiation dose when points are mapped between serial CT scans through deformable registration.

Automated landmark-guided deformable image registration

The purpose of this work is to develop an automated landmark-guided deformable image registration (LDIR) algorithm between the planning CT and daily cone-beam CT (CBCT) with low image quality. This method uses an automated landmark generation algorithm in conjunction with a local small volume gradient matching search engine to map corresponding landmarks between the CBCT and the planning CT. The landmarks act as stabilizing control points in the following Demons deformable image registration. LDIR is implemented on graphics processing units (GPUs) for parallel computation to achieve ultra fast calculation. The accuracy of the LDIR algorithm has been evaluated on a synthetic case in the presence of different noise levels and data of six head and neck cancer patients. The results indicate that LDIR performed better than rigid registration, Demons, and intensity corrected Demons for all similarity metrics used. In conclusion, LDIR achieves high accuracy in the presence of multimodality intensity mismatch and CBCT noise contamination, while simultaneously preserving high computational efficiency.

Super-resolution reconstruction for 4D computed tomography of the lung via the projections onto convex sets approach.

The use of 4D computed tomography (4D-CT) of the lung is important in lung cancer radiotherapy for tumor localization and treatment planning. Sometimes, dense sampling is not acquired along the superior-inferior direction. This disadvantage results in an interslice thickness that is much greater than in-plane voxel resolutions. Isotropic resolution is necessary for multiplanar display, but the commonly used interpolation operation blurs images. This paper presents a super-resolution (SR) reconstruction method to enhance 4D-CT resolution. The authors assume that the low-resolution images of different phases at the same position can be regarded as input "frames" to reconstruct high-resolution images. The SR technique is used to recover high-resolution images. Specifically, the Demons deformable registration algorithm is used to estimate the motion field between different "frames." Then, the projection onto convex sets approach is implemented to reconstruct high-resolution lung images. The performance of the SR algorithm is evaluated using both simulated and real datasets. Their method can generate clearer lung images and enhance image structure compared with cubic spline interpolation and back projection (BP) method. Quantitative analysis shows that the proposed algorithm decreases the root mean square error by 40.8% relative to cubic spline interpolation and 10.2% versus BP. A new algorithm has been developed to improve the resolution of 4D-CT. The algorithm outperforms the cubic spline interpolation and BP approaches by producing images with markedly improved structural clarity and greatly reduced artifacts.

TU‐A‐12A‐04: Quantitative Texture Features Calculated in Lung Tissue From CT Scans Demonstrate Consistency Between Two Databases From Different Institutions

Purpose:To evaluate the consistency of computed tomography (CT) scan texture features, previously identified as stable in a healthy patient cohort, in esophageal cancer patient CT scans.Methods:116 patients receiving radiation therapy (median dose: 50.4Gy) for esophageal cancer were retrospectively identified. For each patient, diagnostic‐quality pre‐therapy (0‐183 days) and post‐therapy (5‐120 days) scans (mean voxel size: 0.8mm×0.8mm×2.5mm) and a treatment planning scan and associated dose map were collected. An average of 501 32×32‐pixel ROIs were placed randomly in the lungs of each pre‐therapy scan. ROI centers were mapped to corresponding locations in post‐therapy and planning scans using the displacement vector field output by demons deformable registration. Only ROIs with mean dose &lt;5Gy were analyzed, as these were expected to contain minimal post‐treatment damage. 140 texture features were calculated in pre‐therapy and post‐therapy scan ROIs and compared using Bland‐Altman analysis. For each feature, the mean feature value change and the distance spanned by the 95% limits of agreement were normalized to the mean feature value, yielding normalized range of agreement (nRoA) and normalized bias (nBias). Using Wilcoxon signed rank tests, nRoA and nBias were compared with values computed previously in 27 healthy patient scans (mean voxel size: 0.67mm×0.67mm×1mm) acquired at a different institution.Results:nRoA was significantly (p&lt;0.001) larger in cancer patients than healthy patients. Differences in nBias were not significant (p=0.23). The 20 features identified previously as having nRoA&lt;20% for healthy patients had the lowest nRoA values in the current database, with an average increase of 5.6%.Conclusion:Despite differences in CT scanner type, scan resolution, and patient health status, the same 20 features remained stable (i.e., low variability and bias) in the absence of disease changes for databases from two institutions. Identification of these features is the first step towards quantifying radiation‐induced changes between preand post‐therapy scans.Supported, in part, by NIH Grant Nos. S10 RR021039, and P30 CA14599, the Virginia and D. K. Ludwig Fund for Cancer Research, Imaging Research Institute, Biological Sciences Division, The University of Chicago, and The Institute for Translational Medicine Pilot Award, The University of Chicago.

SU‐E‐J‐08: A Hybrid Three Dimensional Registration Framework for Image‐Guided Accurate Radiotherapy System ARTS‐IGRT

Purpose:The purpose of this work was to develop a registration framework and method based on the software platform of ARTS‐IGRT and implement in C++ based on ITK libraries to register CT images and CBCT images. ARTS‐IGRT was a part of our self‐developed accurate radiation planning system ARTS.Methods:Mutual information (MI) registration treated each voxel equally. Actually, different voxels even having same intensity should be treated differently in the registration procedure. According to their importance values calculated from self‐information, a similarity measure was proposed which combined the spatial importance of a voxel with MI (S‐MI). For lung registration, Firstly, a global alignment method was adopted to minimize the margin error and achieve the alignment of these two images on the whole. The result obtained at the low resolution level was then interpolated to become the initial conditions for the higher resolution computation. Secondly, a new similarity measurement S‐MI was established to quantify how close the two input image volumes were to each other. Finally, Demons model was applied to compute the deformable map.Results:Registration tools were tested for head‐neck and lung images and the average region was 128*128*49. The rigid registration took approximately 2 min and converged 10% faster than traditional MI algorithm, the accuracy reached 1mm for head‐neck images. For lung images, the improved symmetric Demons registration process was completed in an average of 5 min using a 2.4GHz dual core CPU.Conclusion:A registration framework was developed to correct patient's setup according to register the planning CT volume data and the daily reconstructed 3D CBCT data. The experiments showed that the spatial MI algorithm can be adopted for head‐neck images. The improved Demons deformable registration was more suitable to lung images, and rigid alignment should be applied before deformable registration to get more accurate result.Supported by National Natural Science Foundation of China (NO.81101132) and Natural Science Foundation of Anhui Province (NO.11040606Q55)

Application of three-dimensional medical image registration algorithm in image-guided radiotherapy

To acquire an accurate patient positioning in image-guided radiotherapy, an improved Demons deformable registration method was developed. The FDK algorithm was adopted to reconstruct Cone Beam CT( CBCT) and the reconstruction result was visualized by a volume rendering method with Visualization ToolKit( VTK). Based on the Insight segmentation and registration ToolKit( ITK), the Demons algorithm was completed incorporating the gradient information of fixed image and floating image by the concept of symmetric gradient, and a new formula of Demons force was demonstrated.Registrion experiments were carried out using medical images both from single modality and multi-modality. The results show that the improved Demons algorithm achieves a faster convergence speed and a higher precision compared with the original demons algorithm, which indicates that the Demons algorithm based on symmetric gradient is more suitable for the registration of CBCT reconstruction image and CT plan image in image-guided radiotherapy.

SU‐E‐J‐71: Dosimetric Evaluation of Image Based and Anatomical/physiological Model Based Deformable Registration Methods

Purpose: To study the dosimetric differences of five image and anatomical/physiological model based registration methods. Methods: A lung patient with a tumor in the right middle lobe was studied. GTV and normal structures were contoured individually on ten breathing phases of 4DCT. Volume of GTV ranged from 23.3 cc on end‐inhale phase p0 to 28.0 cc on end‐exhale phase p50. The range of GTV motion was 0.37 cm in L‐R, 0.28 cm in A‐P and 0.54 cm in S‐I directions. VMAT plans were generated on ten breathing phases and 4D composite doses were calculated by registering doses to p0. Five different registration methods were compared: (1) Rigid registration aligning the center of mass of the tumor; (2) B‐spline deformable image registration; (3) Demons deformable image registration; (4) A physiological lung model based on 4DCT, measured pressure‐volume data and finite element analysis and (5) a whole body non‐uniform rational B‐spline model based on 4DCT and contours of structures. The parameters of image registration and anatomical/physiological modeling were optimized so that the volume of the GTV and the lungs and the boundary between the GTV and the lungs were preserved. Results: All image and model based deformable registration methods gave very similar dose coverage to GTV and PTV and similar doses to normal structures including lungs, spinal cord, heart and esophagus. Rigid registration has higher minimum dose to the targets due to the fact that it does not account for tumor deformation. Conclusion: All deformable registration methods with parameters chosen to preserve the volume of target and normal structures and the boundary between them will yield similar dosimetric results. For tumors whose volume changes among phases, deformable registration will give different results compared to rigid registration.

WE-C-103-09: Investigation of Demons Deformable Registration-Based Methods to Measure Lung CT Texture Change Over Time

Purpose: To compare three demons registration-based methods to identify spatially matched regions in serial computed tomography (CT) scans for use in texture analysis. Methods: Two normal thoracic CT scans were collected from 27 patients. Over 1,000 regions of interest (ROIs) were randomly placed in the lungs of each baseline scan. Anatomically matched ROIs in the corresponding follow-up scan were placed by mapping the baseline scan ROI center pixel to (1) the original follow-up scan, (2) the follow-up scan resampled to match the baseline scan voxel size, and (3) the follow-up scan aligned to the baseline scan through affine registration. Mappings used the vector field obtained through demons deformable registration of each follow-up scan variant to the baseline scan. 140 texture features were calculated in all ROIs. Feature value differences between paired ROIs were evaluated using Bland-Altman analysis. For each feature, (1) the mean feature value change and (2) the distance spanned by the 95% limits of agreement were normalized to the mean feature value to obtain, respectively, normalized bias (nBias) and normalized range of agreement (nRoA). Paired Student's t-tests were used to compare nBias across the three methods and nRoA across the three methods. Results: For 20 features with low variability (nRoA<20%), significant differences in nBias existed among the three methods. For every feature, the lowest nBias was achieved when feature values were calculated on original follow-up scans. nRoA was not significantly increased, indicating low variability in feature value change. Conclusion: Three methods to facilitate texture analysis of serial CT scans were evaluated. The bias in feature value change between matched ROIs was minimized with original follow-up scans, using demons deformation to identify corresponding ROIs between scans. This approach could facilitate future measurement of pathologic change between CT scans without necessitating calculation of feature values on deformed scans. This work was supported, in part, by The Coleman Endowment through The University of Chicago Comprehensive Cancer Center, National Science Foundation Research Experience for Undergraduates (NSF REU) Award No. 1062909, and National Institutes of Health (NIH) Grant Nos. S10 RR021039, P30 CA14599, and T32 EB002103-23.

Lung texture in serial thoracic CT scans: Registration‐based methods to compare anatomically matched regions

The aim of this study was to compare three demons registration-based methods to identify spatially matched regions in serial computed tomography (CT) scans for use in texture analysis. Two thoracic CT scans containing no lung abnormalities and acquired during serial examinations separated by at least one week were retrospectively collected from 27 patients. Over 1000 regions of interest (ROIs) were randomly placed in the lungs of each baseline scan. Anatomically matched ROIs in the corresponding follow-up scan were placed by mapping the baseline scan ROI center pixel to (1) the original follow-up scan, (2) the follow-up scan resampled to match the baseline scan voxel size, and (3) the follow-up scan aligned to the baseline scan through affine registration. Mappings used the vector field obtained through demons deformable registration of each follow-up scan variant to the baseline scan. 140 texture features distributed among five feature classes were calculated in all ROIs. Feature value differences between paired ROIs were evaluated using Bland-Altman 95% limits of agreement. For each feature, (1) the mean feature value change and (2) the difference between the upper and lower limits of agreement were normalized to the mean feature value to obtain, respectively, the normalized bias and normalized range of agreement (nRoA). Nonparametric tests were used to evaluate differences in normalized bias and nRoA across the three methods. Because patient CT scans contained no pathology, minimal changes in feature values were expected (i.e., low nRoA and normalized bias). Seventy-five features with very large feature value variability (nRoA ≥ 100%) were excluded from further analysis. Across the remaining 65 features, significant differences in normalized bias were observed among the three methods. The lowest normalized bias (median: 0.06%) was achieved when feature values were calculated on original follow-up scans. The affine registration method achieved the lowest nRoA, though nRoA was not significantly increased using original follow-up scans. Features with low nRoA values also had low normalized bias, though the converse was not necessarily true. Using nRoA as a metric, a set of 20 features having both low nRoA and normalized bias were identified. Three methods to facilitate texture analysis of serial CT scans using demons registration for ROI placement were evaluated. The bias in feature value change between matched ROIs was minimized when feature values were calculated on original baseline and follow-up scans. A set of features that had both low bias and variability (nRoA) in feature value change using this method were identified. This texture analysis approach could facilitate future measurement of pathologic changes between CT scans without necessitating calculation of feature values on deformed scans.

Stopping Criteria for Log-Domain Diffeomorphic Demons Registration

A crucial issue in deformable image registration is achieving a robust registration algorithm at a reasonable computational cost. Given the iterative nature of the optimization procedure an algorithm must automatically detect convergence, and stop the iterative process when most appropriate. This paper ranks the performances of three stopping criteria and six stopping value computation strategies for a Log-Domain Demons Deformable registration method simulating both a coarse and a fine registration. The analyzed stopping criteria are: (a) velocity field update magnitude, (b) mean squared error, and (c) harmonic energy. Each stoping condition is formulated so that the user defines a threshold ∊, which quantifies the residual error that is acceptable for the particular problem and calculation strategy. In this work, we did not aim at assigning a value to e, but to give insights in how to evaluate and to set the threshold on a given exit strategy in a very popular registration scheme. Experiments on phantom and patient data demonstrate that comparing the optimization metric minimum over the most recent three iterations with the minimum over the fourth to sixth most recent iterations can be an appropriate algorithm stopping strategy. The harmonic energy was found to provide best trade-off between robustness and speed of convergence for the analyzed registration method at coarse registration, but was outperformed by mean squared error when all the original pixel information is used. This suggests the need of developing mathematically sound new convergence criteria in which both image and vector field information could be used to detect the actual convergence, which could be especially useful when considering multi-resolution registrations. Further work should be also dedicated to study same strategies performances in other deformable registration methods and body districts.

High-quality four-dimensional cone-beam CT by deforming prior images

Due to a limited number of projections at each phase, severe view aliasing artifacts are present in four-dimensional cone beam computed tomography (4D-CBCT) when reconstruction is performed using conventional algorithms. In this work, we aim to obtain high-quality 4D-CBCT of lung cancer patients in radiation therapy by deforming the planning CT. The deformation vector fields (DVF) to deform the planning CT are estimated through matching the forward projection of the deformed prior image and measured on-treatment CBCT projection. The estimation of the DVF is formulated as an unconstrained optimization problem, where the objective function to be minimized is the sum of the squared difference between the forward projection of the deformed planning CT and the measured 4D-CBCT projection. A nonlinear conjugate gradient method is used to solve the DVF. As the number of the variables in the DVF is much greater than the number of measurements, the solution to such a highly ill-posed problem is very sensitive to the initials during the optimization process. To improve the estimation accuracy of DVF, we proposed a new strategy to obtain better initials for the optimization. In this strategy, 4D-CBCT is first reconstructed by total variation minimization. Demons deformable registration is performed to register the planning CT and the 4D-CBCT reconstructed by total variation minimization. The resulted DVF from demons registration is then used as the initial parameters in the optimization process. A 4D nonuniform rotational B-spline-based cardiac-torso (NCAT) phantom and a patient 4D-CBCT are used to evaluate the algorithm. Image quality of 4D-CBCT is substantially improved by using the proposed strategy in both NCAT phantom and patient studies. The proposed method has the potential to improve the temporal resolution of 4D-CBCT. Improved 4D-CBCT can better characterize the motion of lung tumors and will be a valuable tool for image-guided adaptive radiation therapy.

Extra‐dimensional Demons: A method for incorporating missing tissue in deformable image registration

A deformable registration method capable of accounting for missing tissue (e.g., excision) is reported for application in cone-beam CT (CBCT)-guided surgical procedures. Excisions are identified by a segmentation step performed simultaneous to the registration process. Tissue excision is explicitly modeled by increasing the dimensionality of the deformation field to allow motion beyond the dimensionality of the image. The accuracy of the model is tested in phantom, simulations, and cadaver models. A variant of the Demons deformable registration algorithm is modified to include excision segmentation and modeling. Segmentation is performed iteratively during the registration process, with initial implementation using a threshold-based approach to identify voxels corresponding to "tissue" in the moving image and "air" in the fixed image. With each iteration of the Demons process, every voxel is assigned a probability of excision. Excisions are modeled explicitly during registration by increasing the dimensionality of the deformation field so that both deformations and excisions can be accounted for by in- and out-of-volume deformations, respectively. The out-of-volume (i.e., fourth) component of the deformation field at each voxel carries a magnitude proportional to the excision probability computed in the excision segmentation step. The registration accuracy of the proposed "extra-dimensional" Demons (XDD) and conventional Demons methods was tested in the presence of missing tissue in phantom models, simulations investigating the effect of excision size on registration accuracy, and cadaver studies emulating realistic deformations and tissue excisions imparted in CBCT-guided endoscopic skull base surgery. Phantom experiments showed the normalized mutual information (NMI) in regions local to the excision to improve from 1.10 for the conventional Demons approach to 1.16 for XDD, and qualitative examination of the resulting images revealed major differences: the conventional Demons approach imparted unrealistic distortions in areas around tissue excision, whereas XDD provided accurate "ejection" of voxels within the excision site and maintained the registration accuracy throughout the rest of the image. Registration accuracy in areas far from the excision site (e.g., > ∼5 mm) was identical for the two approaches. Quantitation of the effect was consistent in analysis of NMI, normalized cross-correlation (NCC), target registration error (TRE), and accuracy of voxels ejected from the volume (true-positive and false-positive analysis). The registration accuracy for conventional Demons was found to degrade steeply as a function of excision size, whereas XDD was robust in this regard. Cadaver studies involving realistic excision of the clivus, vidian canal, and ethmoid sinuses demonstrated similar results, with unrealistic distortion of anatomy imparted by conventional Demons and accurate ejection and deformation for XDD. Adaptation of the Demons deformable registration process to include segmentation (i.e., identification of excised tissue) and an extra dimension in the deformation field provided a means to accurately accommodate missing tissue between image acquisitions. The extra-dimensional approach yielded accurate "ejection" of voxels local to the excision site while preserving the registration accuracy (typically subvoxel) of the conventional Demons approach throughout the rest of the image. The ability to accommodate missing tissue volumes is important to application of CBCT for surgical guidance (e.g., skull base drillout) and may have application in other areas of CBCT guidance.

A Framework of Whole Heart Extracellular Volume Fraction Estimation for Low-Dose Cardiac CT Images

Cardiac CT (CCT) is widely available and has been validated for the detection of focal myocardial scar using a delayed enhancement technique in this paper. CCT, however, has not been previously evaluated for quantification of diffuse myocardial fibrosis. In our investigation, we sought to evaluate the potential of low-dose CCT for the measurement of myocardial whole heart extracellular volume (ECV) fraction. ECV is altered under conditions of increased myocardial fibrosis. A framework consisting of three main steps was proposed for CCT whole heart ECV estimation. First, a shape-constrained graph cut (GC) method was proposed for myocardium and blood pool segmentation on postcontrast image. Second, the symmetric demons deformable registration method was applied to register precontrast to postcontrast images. So the correspondences between the voxels from precontrast to postcontrast images were established. Finally, the whole heart ECV value was computed. The proposed method was tested on 20 clinical low-dose CCT datasets with precontrast and postcontrast images. The preliminary results demonstrated the feasibility and efficiency of the proposed method.

WE-E-213CD-02: Gaussian Weighted Multi-Atlas Based Segmentation for Head and Neck Radiotherapy Planning.

To develop a multi-atlas segmentation strategy for IMRT head and neck therapy planning. The method was tested on thirty-one head and neck simulation CTs, without demographic or pathology pre-clustering. We compare Fixed Number (FN) and Thresholding (TH) selection (based on normalized mutual information ranking) of the atlases to be included for current patient segmentation. Next step is a pairwise demons Deformable Registration (DR) onto current patient CT. DR was extended to automatically compensate for patient different field of view. Propagated labels are combined according to a Gaussian Weighted (GW) fusion rule, adapted to poor soft tissues contrast. Agreement with manual segmentation was quantified in terms of Dice Similarity Coefficient (DSC). Selection methods, number of atlases used, as well as GW, average and majority voting fusion were discriminated by means of Friedman Test (a=5%). Experimental tuning of the algorithm parameters was performed on five patients, deriving an optimal configuration for each structure. DSC reduction was not significant when ten or more atlases are selected, whereas DSC for single most similar atlas selection is 10% lower in median. DSC of FN selection rule were significantly higher for most structures. Tubular structures may benefit from computing average contour rather than looking at the singular voxel contribution, whereas the best performing strategy for all other structures was GW. When half database is selected, final median DSC were 0.86, 0.80, 0.51, 0.81, 0.69 and 0.79 for mandible, spine, optical nerves, eyes, parotids and brainstem respectively. We developed an efficient algorithm for multiatlas based segmentation of planning CT volumes, based on DR and GW. FN selection of database atlases is foreseen to increase computational efficiency. The absence of clinical pre-clustering and specific imaging protocol on database subjects makes the results closer to real clinical application. "Progetto Roberto Rocca" funded by the Fondazione Fratelli Agostino and Enrico Rocca, Italy.