Autoadaptive phase-correlated (AAPC) reconstruction for 4D CBCT

Abstract

Purpose: Kilovoltage cone-beam computed tomography (CBCT) is widely used in image-guided radiation therapy for exact patient positioning prior to the treatment. However, producing time series of volumetric images (4D CBCT) of moving anatomical structures remains challenging. The presented work introduces a novel method, combining high temporal resolution inside anatomical regions with strong motion and image quality improvement in regions with little motion.

Methods: In the proposed method, the projections are divided into regions that are subject to motion and regions at rest. The latter ones will be shared among phase bins, leading thus to an overall reduction in artifacts and noise. An algorithm based on the concept of optical flow was developed to analyze motion-induced changes between projections. The technique was optimized to distinguish patient motion and motion deriving from gantry rotation. The effectiveness of the method is shown in numerical simulations and patient data.

Results: The images reconstructed from the presented method yield an almost the same temporal resolution in the moving volume segments as a conventional phase-correlated reconstruction, while reducing the noise in the motionless regions down to the level of a standard reconstruction without phase correlation. The proposed simple motion segmentation scheme is yet limited to rotation speeds of less than 3 degrees/s.

Conclusions: The method reduces the noise in the reconstruction and increases the image quality. More data are introduced for each phase-correlated reconstruction, and therefore the applied dose is used more efficiently.

Figure 1

Standard reconstruction: Motionless phantom (upper left) and patient image (lower left), phantom with motion of ellipsoids and inserts (upper right) and patient image with respiratory motion artifacts (lower right). The arrows point to moving and therefore blurred tissue. Also, streak artifacts are visible ($C=-200\mathrm{HU}$, $W=1000\mathrm{HU}$).

Figure 2

Observation window for the segmentation of motionless areas. If a possible displacement within the cone can be found, then the pixel at the center is segmented as motionless.

Figure 3

Overview over the 4D thorax phantom.

Figure 4

Evaluation of the segmentation performance for a varying number of motion cycles. The dashed line shows the percentage of pixel in the segmentation that was additionally marked as motion affected although being motionless. The dotted line shows the percentage of pixel that was falsely segmented as motionless, although being in the motion affected region. The black line shows the corresponding rotation speed for 360°, assuming a fixed motion frequency of $0.25\mathrm{Hz}$. In the dashed region, the segmentation quality is degraded.

Figure 5

Projections for simulated dataset (left). The corresponding segmentations (right) for estimated motion affected regions are shown. A bright color indicates that the region will be subject to the phase-correlated weighting, while dark regions are used in each reconstruction phase.

Figure 6

Reconstruction of various slices for the standard (Std), the conventional phase-correlated (PC), and the AAPC reconstruction for the slow 360° scan. Difference images of AAPC and conventional phase-correlated reconstructions are shown to distinguish the different levels of noise and the different temporal resolutions. ROIs are placed in various positions in motion affected and motionless regions in order to compare the image noise. The noise was calculated in difference images of two identical reconstructions with different noise realizations (images: $C=0\mathrm{HU}$, $W=1000\mathrm{HU}$), (difference image: $C=0\mathrm{HU}$, $W=200\mathrm{HU}$).

Figure 7

Slice with number of backprojected values (left, arbitrary units, and windowing). Reconstructed images using a simple integral approximation and AAPC (middle) and AAPC with the trapezoidal integral approximation (right) ($C=0\mathrm{HU}$, $W=1000\mathrm{HU}$).

Figure 8

Projections for simulated dataset (upper left) and a human thorax scan (lower left). On the right side the corresponding segmentations for estimated motion affected regions are shown. A bright color indicates that the region will be subject to the phase-correlated weighting, while dark regions are used in each reconstruction phase.

Figure 9

Reconstruction of slices in different orientations: Nonphase-correlated (Std), phase correlated with weighting of whole projection (PC), phase-correlated with motion-detection weighting (AAPC), difference image of both phase-correlated reconstructions (images: $C=0\mathrm{HU}$, $W=1000\mathrm{HU}$) (difference image: $C=0\mathrm{HU}$, $W=500\mathrm{HU}$).

Figure 10

Reconstruction of the coronal slice through the lung mass in different motion phases: Phase correlated with weighting of the whole projection (PC), phase correlated with motion-detection weighting (AAPC), and difference image of both phase-correlated reconstructions. The white arrows indicate the motion direction in the reconstructed phases (images: $C=0\mathrm{HU}$, $W=1000\mathrm{HU}$) (difference image: $C=0\mathrm{HU}$, $W=500\mathrm{HU}$).

Figure 11

Geometry for the CBCT reconstruction.