Evaluation of accuracy of B-spline transformation-based deformable image registration with different parameter settings for thoracic images

Abstract

Deformable image registration (DIR) is fundamental technique for adaptive radiotherapy and image-guided radiotherapy. However, further improvement of DIR is still needed. We evaluated the accuracy of B-spline transformation-based DIR implemented in elastix. This registration package is largely based on the Insight Segmentation and Registration Toolkit (ITK), and several new functions were implemented to achieve high DIR accuracy. The purpose of this study was to clarify whether new functions implemented in elastix are useful for improving DIR accuracy. Thoracic 4D computed tomography images of ten patients with esophageal or lung cancer were studied. Datasets for these patients were provided by DIR-lab (dir-lab.com) and included a coordinate list of anatomical landmarks that had been manually identified. DIR between peak-inhale and peak-exhale images was performed with four types of parameter settings. The first one represents original ITK (Parameter 1). The second employs the new function of elastix (Parameter 2), and the third was created to verify whether new functions improve DIR accuracy while keeping computational time (Parameter 3). The last one partially employs a new function (Parameter 4). Registration errors for these parameter settings were calculated using the manually determined landmark pairs. 3D registration errors with standard deviation over all cases were 1.78 (1.57), 1.28 (1.10), 1.44 (1.09) and 1.36 (1.35) mm for Parameter 1, 2, 3 and 4, respectively, indicating that the new functions are useful for improving DIR accuracy, even while maintaining the computational time, and this B-spline-based DIR could be used clinically to achieve high-accuracy adaptive radiotherapy.

Evaluation of a deformable registration algorithm for subsequent lung computed tomography imaging during radiochemotherapy

“…28,45 Matching manually placed anatomical landmark pairs in corresponding CT data sets has often been used to determine DIR accuracy in the lung region. 13,14,23,28,33,46 In addition to landmarks at bronchial bifurcations and smaller tubes, landmarks from adjacent bone structures 23 and naturally occurring fiducial markers in nearby organs 46 were included in other studies. Because the investigated DIR algorithm is optimized for registration inside the lungs and manual definition of landmarks is a time-consuming process, our landmark-based analysis was limited to 25 landmarks per patient, but summed to a total of 1100 landmark pairs from the different CT combinations.…”

“…It is more difficult to identify useful landmarks in CT images with relatively low resolution (5 mm slice distance and 1.37 mm pixel spacing) acquired at different time points. In previous studies, 13,14,23,28,46 registration was primarily performed between different breathing phases of 4D CTs with slice distances ≤2.5 mm and pixel spacing <1 mm. Our investigated registration algorithm was previously tested against the same 4D CT benchmark datasets including several hundred predefined landmarks (available at www.dir-lab.…”

“…7 The accuracy of ROI propagation via DIR in the lung region has been primarily investigated for automatic contour transfer between different phases of the same 4D-CT for which anatomical differences are only correlated to respiration. [8][9][10][11][12][13][14] DIR for ROI propagation between planning and intratherapy CT has often been reported for ART in head and neck cancer, [15][16][17][18][19][20][21] but rarely for ART in lung lesions. 22,23 Several standards for validating a DIR algorithm via physical or virtual phantoms [24][25][26][27] and imaging data of exemplary patients have been reported.…”

“…In the latter, clinically relevant case, an expert clinician's contours and landmarks are commonly considered as the ground truth to be matched by the algorithm. Thus, averages of contour similarity metrics [9][10][11][12][16][17][18]21,22 and landmark distances 13,14,20,23,28 are typically reported to describe the general DIR quality. Unfortunately, neither landmarks nor contours assure sufficient registration quality evaluation within homogeneous tissue, because landmarks may not be set with reliable accuracy due to low contrast.…”

Quantifying local tumor morphological changes with Jacobian map for prediction of pathologic tumor response to chemo-radiotherapy in locally advanced esophageal cancer

“…On the other hand, diverging vectors create a source point that represents a morphological expansion. Usually 0 ≤ α ≤ 1.0 (Kanai et al , 2014; Shusharina and Sharp, 2012) and in this study α was experimentally set to 0.1.…”