Patrik Kunz

Automatic segmentation of thoracic and pelvic CT images for radiotherapy planning using implicit anatomic knowledge and organ-specific segmentation strategies

Automatic segmentation of anatomical structures in medical images is a valuable tool for efficient computer-aided radiotherapy and surgery planning and an enabling technology for dynamic adaptive radiotherapy. This paper presents the design, algorithms and validation of new software for the automatic segmentation of CT images used for radiotherapy treatment planning. A coarse to fine approach is followed that consists of presegmentation, anatomic orientation and structure segmentation. No user input or a priori information about the image content is required. In presegmentation, the body outline, the bones and lung equivalent tissue are detected. Anatomic orientation recognizes the patients position, orientation and gender and creates an elastic mapping of the slice positions to a reference scale. Structure segmentation is divided into localization, outlining and refinement, performed by procedures with implicit anatomic knowledge using standard image processing operations. The presented version of algorithms automatically segments the body outline and bones in any gender and patient position, the prostate, bladder and femoral heads for male pelvis in supine position, and the spinal canal, lungs, heart and trachea in supine position. The software was developed and tested on a collection of over 600 clinical radiotherapy planning CT stacks. In a qualitative validation on this test collection, anatomic orientation correctly detected gender, patient position and body region in 98% of the cases, a correct mapping was produced for 89% of thorax and 94% of pelvis cases. The average processing time for the entire segmentation of a CT stack was less than 1 min on a standard personal computer. Two independent retrospective studies were carried out for clinical validation. Study I was performed on 66 cases (30 pelvis, 36 thorax) with dosimetrists, study II on 52 cases (39 pelvis, 13 thorax) with radio-oncologists as experts. The experts rated the automatically produced structures on the scale 1-excellent (no corrections necessary, maximum time saving), 2-good (corrections necessary for up to 1/3 of slices), 3-acceptable (major corrections necessary, but still time saving), 4-not acceptable (manual redrawing more efficient, no time saving). A rating<or=3 indicates a time saving in the treatment planning process and was given for pelvis segmentation in 70% (I) and 68% (II) of the cases, with average ratings 2.9 (I) and 2.6 (II). For the thorax, a rating<or=3 was given in 94% and 91% of the cases, with average ratings 2.1 and 1.9, respectively. For quantitative validation, automatically generated structures were compared geometrically in 2D and 3D to manually drawn structures created by experts on seven randomly selected cases. The quantitative validation was limited to pelvis structures. The results indicate that the accuracy of the algorithms is within the bandwidth of manual segmentation by experts, except for specific erroneous situations. Even though manual review and corrections of automatically segmented structures are still mandatory, it can be expected that due to the speed of the presented software and the quality of its results, its introduction in the radiotherapy treatment planning process will lead to a considerable amount of time being saved.

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

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 than3°∕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.

Self‐adapting cyclic registration for motion‐compensated cone‐beam CT in image‐guided radiation therapy

PURPOSE In image-guided radiation therapy an additional kV imaging system next to the linear particle accelerator provides information for an accurate patient positioning. However, the acquisition time of the system is much longer than the patients breathing cycle due to the low gantry rotation speed. Our purpose is a cyclic registration in the context of motion-compensated image reconstruction that provides high quality respiratory-correlated 4D volumes for on-board flat panel detector cone-beam CT scans. METHODS Based on the small motion assumption, widely used within registration algorithms, a strategy is developed for motion estimation. In this strategy temporal restrictions are incorporated, for example, the cyclic motion patterns of respiration. The resultant cyclic registration method is to show less sensitivity on image artifacts, in particular on artifacts due to projection data sparsification. Using a new cyclic registration method a motion estimation is performed on respiratory-correlated reconstructions, and the obtained motion vector fields are used for motion compensation. RESULTS The proposed cyclic registration is evaluated in the context of motion-compensated image reconstruction using simulated data and patient data. Motion artifacts of 3D standard reconstructions can be significantly reduced by the resulting cyclic motion compensation. The method outperforms the respiratory-correlated reconstructions regarding sparse-view artifacts and maintains the high temporal resolution at the same time. Image artifacts show only minor to almost no effect on the motion estimation using the cyclic registration. CONCLUSIONS The cyclic motion compensation approach provides respiratory-correlated volumes with high image quality. The cyclic motion estimation is of such low sensitivity to sparse-view artifacts, that it is capable to determine high quality motion vector fields based on reconstructions of low sampled data.

Evaluation of a new six degrees of freedom couch for radiation therapy

PURPOSE The aim of this work is to evaluate the geometric accuracy of a prerelease version of a new six degrees of freedom (6DoF) couch. Additionally, a quality assurance method for 6DoF couches is proposed. METHODS The main principle of the performance tests was to request a known shift for the 6DoF couch and to compare this requested shift with the actually applied shift by independently measuring the applied shift using different methods (graph paper, laser, inclinometer, and imaging system). The performance of each of the six axes was tested separately as well as in combination with the other axes. Functional cases as well as realistic clinical cases were analyzed. The tests were performed without a couch load and with a couch load of up to 200 kg and shifts in the range between -4 and +4 cm for the translational axes and between -3° and +3° for the rotational axes were applied. The quality assurance method of the new 6DoF couch was performed using a simple cube phantom and the imaging system. RESULTS The deviations (mean ± one standard deviation) accumulated over all performance tests between the requested shifts and the measurements of the applied shifts were -0.01 ± 0.02, 0.01 ± 0.02, and 0.01 ± 0.02 cm for the longitudinal, lateral, and vertical axes, respectively. The corresponding values for the three rotational axes couch rotation, pitch, and roll were 0.03° ± 0.06°, -0.04° ± 0.12°, and -0.01° ± 0.08°, respectively. There was no difference found between the tests with and without a couch load of up to 200 kg. CONCLUSIONS The new 6DoF couch is able to apply requested shifts with high accuracy. It has the potential to be used for treatment techniques with the highest demands in patient setup accuracy such as those needed in stereotactic treatments. Shifts can be applied efficiently and automatically. Daily quality assurance of the 6DoF couch can be performed in an easy and efficient way. Long-term stability has to be evaluated in further tests.

Motion-compensated 4D cone-beam computed tomography

In image-guided radiation therapy (IGRT) beside the linear particle accelerator an additional kV system provides information for an accurate patient positioning. The precise execution of the treatment plan is assured that way. An extra goal is the refinement of further planned treatment fractions based on the on-board imaging system information. Additional inter-fractional planning CTs become thus obsolete. However, the acquisition time of the system is much longer than the patients breathing cycle due to technical limitations on the gantry rotation speed. Severe artifacts like blurring or streaks are the consequence. They affect image quality and thus also the refinement of the treatment plan.

Cone-beam CT sequence scan reconstruction with improved dose usage and scan coverage

For circular cone-beam CT often one scan covers not the complete z-range of interest. If this is the case two or more circle scans are made. These sequence scans are typically reconstructed by separately reconstructing each circle scan followed by combining the resulting partial volumes. This image-based concatenation method uses only those data that are needed for each partial volume, the contribution of rays to neighboring volumes are ignored, redundancies are not used, and dose is wasted. Therefore an algorithm that uses all rays that run through a voxel by appropriately weighting the rays followed by filtered backprojection was developed and evaluated. This leads to improved dose usage and increases the overlap region of neighboring volumes, potentially leading to reduced artifacts in this region. Alternatively, our approach allows to increase the table increment between adjacent circle acquisitions, and thereby the scan coverage, without impairing image quality or increasing dose. To evaluate our method we use the geometry of the Varian OBI flat panel detector CT scanner. Simulated and measured data are processed at varying table increment values and an evaluation of image noise, spatial resolution and artifacts has been performed. The method shows good image quality on simulated phantom data as well as on clinical patient data. In this paper the algorithm demonstrates its ability to extend the z-range of sequence scans and to improve the image quality in the overlap region and turns out to be usable on devices in the clinical practice.

A comparison of 4D cone-beam CT algorithms for slowly rotating scanners

The study compares several algorithms for the 4D reconstruction of cone-beam computed tomography (CBCT) data that were recently proposed and which can be used from slowly rotating devices. In our case the imaging units are mounted to linear particle accelerators (LINAC). The algorithms are the conventional phase-correlated reconstruction (PC), the McKinnon/Bates-Algorithm, the prior image constrained compressed sensing (PICCS) algorithm, the total-variation minimization (TV) algorithm, and our auto-adaptive phase-correlation (AAPC) algorithm. For each algorithm the same motion-affected rawdata are used and the reconstruction results compared to each other regarding their noise and artifact levels, as well as temporal resolution, and computational complexity and convergence. These criteria result in a discussion of the advantages and disadvantages of each algorithm. The temporal resolution is best in the algorithms which exclusively use data from a single motion phase only. The iterative algorithms show lower noise and artifact levels but are computationally complex and therefore may have a limited usage in the clinical application. Algorithms which include image enhancements beside a faster reconstruction represent a suitable trade-off for the clinical workflow.

TU‐A‐204B‐01: The CBCT Performance of a New Treatment Platform

Purpose: To characterize the CBCT performance of a new radiation therapy platform (Trilogy MX). Methods: An entirely new CBCT system has been developed for Trilogy MX. The new CBCT system differs from that of the On‐Board Imager® (OBI) in the use of a beam hardening filter, which reduces patient dose, and an improved reconstructor, which uses scatter correction algorithms to account for the x‐ray scatter caused by the cone‐beam geometry. Scans of Catphan® and electron density (Model 062A, CIRS) phantoms have been compared with OBI scans. Hounsfield unit (HU) accuracy was checked by changing the z scan length and the phantom diameters for pelvis (125kVp, 680mAs, 45cm dia.) acquisitions. The Catphan was imaged using doses ranging between 5–20mGy (CTDIw). Projections acquired using clinical OBI units were also reconstructed for comparison using the new reconstructor.Results: The new CBCT system has higher dose efficiency and higher HU accuracy. When the volume length is reduced from 160mm to 90mm or when the phantom diameter is reduced from 330mm to 180mm, the HU values measured for the same inserts differ by −190 to +80HU for OBI scans and by –60 to +50HU for Trilogy MX scans (for electron densities between 0.2 and 1.2). The contrast detectability of the 1% contrast objects in the Catphan phantom improves from 9mm to 4mm diameter when using the same CTDIwdose as OBI. Clinical images exhibit much better uniformity, elimination of streaks and better definition of the skin surface. Conclusions: The new reconstruction algorithm makes substantial improvements in CBCTimage quality, reduces patient dose and increases HU accuracy. The new system produces CBCTimages, which are much better suited to image guidance and which may be suitable for other tasks such as adaptive RT planning.

Voxel-based reconstruction combined with motion detection for slow rotating 4D CBCT

Flat panel detector (FPD) cone-beam computed tomography (CBCT) systems, such as C-arm CT scanners or onboard imaging systems, rotate far slower than a typical motion cycle of the heart or lung. Therefore 4D CBCT is more complicated with flat panel detectors than with clinical CT. Recent approaches for 4D imaging from FPD CBCT either use multiple scans over the same angular range or a single slow rotation, and they perform a motion phase-dependent weighting of complete projections (e.g. using a respiratory monitor). This leads to a relatively high temporal resolution but also to high image noise and in the second approach to artifacts due to sparse angular sampling. Our proposed method also uses the data from several scans without causing streak artifacts. But instead of weighting the whole projections the weighting is applied to the motion affected areas in the projection only. These regions are automatically estimated using motion detection techniques between consecutive projections. Thus, projection data are used in our algorithm that are completely ignored in conventional approaches. To evaluate our method simulated data of a breathing thorax phantom and measurements acquired with the OBI scanner (Varian Medical Systems, Palo Alto, CA) were reconstructed. Image quality was compared with the projection-weighting approach for whole projections and standard reconstructions without phase-correlation.

Iterative motion-compensated reconstruction for image-guided radiation therapy

In image-guided radiation therapy (IGRT) an additional kV imaging system next to the linear particle accelerator provides information for an accurate patient positioning. However, due to the limited gantry rotation speed during treatment the typical acquisition time is much longer than the patients breathing cycle resulting in low image quality. In particular, respiratory motion causes severe artifacts such as blurring and streaks in tomographic images. Compensating for motion is an interesting option and capable of providing high quality respiratory-correlated 4D volumes. The particular challenge is to do this without knowledge from prior scans and without specific requirements on the acquisition. State-of-the-art methods for estimation of the motion vector fields suffer from image artifacts that appear in intermediated image frames. In applications like ours conventional registration algorithms tend to register artifacts rather than anatomy. Our proposed registration method addresses the image artifact problem by additionally using temporal constraints within the registration process like the cyclic motion patterns of respiration. This cyclic regularization avoids the algorithm being sensitive to the above-mentioned streak artifacts. Hence, this registration algorithm consists of two components, the spatial and the temporal part, and one of them might be dominating. For this very reason the cyclic registration method is supplemented by an iterative refinement to avoid a potential underestimation of motion and simultaneously to comply with the temporal constraints. The iterative and cyclic registration method is verified using simulated rawdata and patient data. It shows low sensitivity on image artifacts and deals with potential underestimation of motion at the same time. In this way, the motion is accurately estimated and a motion compensation corrects for it. Our iterative and cyclic registration method is capable of high quality motion estimation only on basis of respiratory-correlated reconstructions. Thus, motion-compensated image reconstruction without knowledge from prior scans and without specific requirements on the acquisition becomes feasible in image-guided radiation therapy.