Search

PURPOSES: Preimplant diagnostic magnetic resonance imaging is the gold standard for image-guided tandem-and-ovoids (T&O) brachytherapy for cervical cancer. However, high dose rate brachytherapy planning is typically done on postimplant CT-based high-risk clinical target volume (HR-CTV(CT)) because the transfer of preimplant Magnetic resonance (MR)-based HR-CTV (HR-CTV(MR)) to the postimplant planning CT is difficult due to anatomical changes caused by applicator insertion, vaginal packing, and the filling status of the bladder and rectum. This study aims to train a dual-path convolutional neural network (CNN) for automatic segmentation of HR-CTV(CT) on postimplant planning CT with guidance from preimplant diagnostic MR. METHODS: Preimplant T2-weighted MR and postimplant CT images for 65 (48 for training, eight for validation, and nine for testing) patients were retrospectively solicited from our institutional database. MR was aligned to the corresponding CT using rigid registration. HR-CTV(CT) and HR-CTV(MR) were manually contoured on CT and MR by an experienced radiation oncologist. All images were then resampled to a spatial resolution of 0.5 × 0.5 × 1.25 mm. A dual-path 3D asymmetric CNN architecture with two encoding paths was built to extract CT and MR image features. The MR was masked by HR-CTV(MR) contour while the entire CT volume was included. The network put an asymmetric weighting of 18:6 for CT: MR. Voxel-based dice similarity coefficient (DSC(V)), sensitivity, precision, and 95% Hausdorff distance (95-HD) were used to evaluate model performance. Cross-validation was performed to assess model stability. The study cohort was divided into a small tumor group (40 cc) based on the HR-CTV(CT) for model evaluation. Single-path CNN models were trained with the same parameters as those in dual-path models. RESULTS: For this patient cohort, the dual-path CNN model improved each of our objective findings, including DSC(V), sensitivity, and precision, with an average improvement of 8%, 7%, and 12%, respectively. The 95-HD was improved by an average of 1.65 mm compared to the single-path model with only CT images as input. In addition, the area under the curve for different networks was 0.86 (dual-path with CT and MR) and 0.80 (single-path with CT), respectively. The dual-path CNN model with asymmetric weighting achieved the best performance with DSC(V) of 0.65 ± 0.03 (0.61–0.70), 0.79 ± 0.02 (0.74–0.85), and 0.75 ± 0.04 (0.68–0.79) for small, medium, and large group. 95-HD were 7.34 (5.35–10.45) mm, 5.48 (3.21–8.43) mm, and 6.21 (5.34–9.32) mm for the three size groups, respectively. CONCLUSIONS: An asymmetric CNN model with two encoding paths from preimplant MR (masked by HR-CTV(MR)) and postimplant CT images was successfully developed for automatic segmentation of HR-CTV(CT) for T&O brachytherapy patients.

Purpose In this investigation, we sought to characterize X-ray beam qualities and quantify percent depth dose (PDD) curves for fluoroscopic X-ray beams incorporating added copper (Cu) filtration, such as those commonly used in fluoroscopically guided interventions (FGI). The intended application of this research is for dosimetry in soft tissue from FGI procedures using these data. Methods All measurements in this study were acquired on a Siemens (Erlangen, Germany) Artis zeego fluoroscope. X-ray beam characteristics of first half-value layer (HVL), second HVL, homogeneity coefficients (HCs), backscatter factors (BSFs) and kVp accuracy and precision were determined to characterize the X-ray beams used for the PDD measurements. A scanning water tank was used to measure PDD curves for 60, 80, 100, and 120 kVp X-ray beams with Cu filtration thicknesses of 0.0, 0.1, 0.3, 0.6, and 0.9 mm at 11 cm, 22 cm, and 42 cm nominal fields of view, in water depths of 0 to 150 mm. Results X-ray beam characteristics of first HVLs and HCs differed from previous published research of fluoroscopic X-ray beam qualities without Cu filtration. PDDs for 60, 80, 100, and 120 kVp with 0 mm of Cu filtration were comparable to previous published research, accounting for differences in fluoroscopes, geometric orientation, type of ionization chamber, X-ray beam quality, and the water tank used for data collection. PDDs and X-ray beam characteristics for beam qualities with Cu filtration are presented, which have not been previously reported. Conclusions The data sets of X-ray beam characteristics and PDDs presented in this study can be used to estimate organ or soft tissue doses at depth involving similar beam qualities or to compare with mathematical models. This article is protected by copyright. All rights reserved.

Purpose: The purpose of this investigation is to characterize and quantify X-ray beam profiles for fluoroscopic x-ray beam spectra incorporating spectral (copper) filtration. Methods: A PTW (Freiburg, Germany) type 60016 silicon diode detector and PTW MP3 water tank were used to measure X-ray beam profiles for 60, 80, 100 and 120 kVp x-ray beams at five different copper filtration thicknesses ranging from 0–0.9 mm at 22 and 42 cm fields of view and depths of 1, 5, and 10 cm in both the anode-cathode axis (inplane) and cross-plane directions. All measurements were acquired on a Siemens (Erlangen, Germany) Artis ZeeGo fluoroscope inverted from the typical orientation providing an x-ray beam originating from above the water surface with the water level set at 60 cm from the focal spot. Results: X-ray beam profiles for beam spectra without copper filtration compared well to previously published data by Fetterly et al. [Med Phys, 28, 205 (2001)]. Our data collection benefited from the geometric orientation of the fluoroscope, providing a beam perpendicular to the tank water surface, rather than through a thin side wall as did the previously mentioned study. Profiles for beams with copper filtration were obtained which have not been previously investigated and published. Beam profiles in the anode-cathode axis near the surface and at lower x-ray energy exhibited substantial heel effect, which became less pronounced at greater depth. At higher energy with copper filtration in the beam, the dose falloff out-of-field became less pronounced, as would be anticipated given higher scatter photon energy. Conclusion: The x-ray beam profile data for the fluoroscopic x-ray beams incorporating copper filtration are intended for use as reference data for estimating doses to organs or soft tissue, including fetal dose, involving similar beam qualities or for comparison with mathematical models.

Purpose: The purpose of this investigation is to quantify various first half-value-layers (HVLs), second HVLs and homogeneity coefficients (HCs) for a state-of-the-art fluoroscope utilizing spectral (copper) filtration. Methods: A Radcal (Monrovia, Ca) AccuPro dosimeter with a 10×6-6 calibrated ionization chamber was used to measure air kerma for radiographic x-ray exposures made on a Siemens (Erlangen, Germany) Artis ZeeGo fluoroscope operated in the service mode. The ionization chamber was centered in the x-ray beam at 72 cm from the focal spot with a source-to-image-distance of 120 cm. The collimators were introduced to limit the x-ray field to approximately 5 cm × 5 cm at the ionization chamber plane. Type-1100 aluminum filters, in 0.5 mm increments, were used to determine the HVL. Two HVL calculation methods were used, log-linear interpolation and Lambert-W interpolation as described by Mathieu [Med Phys, 38(8), 4546 (2011)]. Multiple measurements were made at 60, 80, 100, 120 kVp at spectral filtration thicknesses of 0, 0.1, 0.3, 0.6 and 0.9 mm. Results: First HVL, second HVL, and HCs are presented for the fluoroscopic x-ray beam spectra indicated above, with nearly identical results from the two interpolation methods. Accuracy of the set kVp was also determined and deviated less than 2%. First HVLs for fluoroscopic x-ray beam spectra without spectral filtration determined in our study were 7%–16% greater than previously published data by Fetterly et al. [Med Phys, 28, 205 (2001)]. However, the FDA minimum HVL requirements changed since that publication, requiring larger HVLs as of 2006. Additionally, x-ray tube and generator architecture have substantially changed over the last 15 years providing different beam spectra. Conclusion: X-ray beam quality characteristics for state-of-the-art fluoroscopes with spectral filtration have not been published. This study provides reference data which will be useful for defining beam qualities encountered on fluoroscopes using spectral filtration.

Purpose: ABC is a breath-hold technique for SBRT treatments. At our institution, three CTs with ABC are acquired at simulation to check the breath-hold consistency because tumor position may deviate among three scans. The purpose of this study was to determine whether RPM can predict tumor position deviation in SBRT breath-hold treatments with ABC. Methods: Five SBRT lung and five SBRT liver patients were selected. An RPM block was placed on the patients between the xiphoid and the umbilicus. RPM recorded patients’ anterior-posterior skin surface positions during breath holds with ABC. For each patient, three CTs (ABC1, ABC2, ABC3) were acquired at the same breath-hold level. ABC2 and ABC3 CTs were rigidly registered to ABC1 CT. Tumor position deviations were measured on the fused CTs (ABC2 with ABC1, ABC3 with ABC1) and compared to the corresponding RPM position shifts. Linear regression, k-means clustering, and decision tree classifier were used to determine the relationship between RPM position shifts and tumor position deviations. Results: Tumor position deviations were 0.4–8.7 mm for lung patients and 1.6–9.1 mm for liver patients. There were no significant linear correlations between RPM position shifts and tumor position deviations for both lung (R2=0.20) and liver (R2=0.45) patients. In contrast, decision tree classification showed that the RPM position shift of 3.45 mm or more indicated tumor position deviation >5 mm for lower lobe lung tumors and liver tumors that are close to the diaphragm. Conclusion: RPM position shifts are not linearly correlated with tumor position deviations. However, RPM can be used to predict large tumor position deviations (>5 mm) among three ABC CTs in SBRT for lower lobe lung tumors and liver tumors and to validate the breath-hold consistency with the ABC technique. Therefore, RPM has the potential to replace the additional verification CT scans.

Purpose: The purpose of this investigation was to quantify percent depth dose (PDD) curves for fluoroscopic x-ray beam qualities incorporating added copper filtration. Methods: A PTW (Freiburg, Germany) MP3 water tank was used with a Standard Imaging (Middleton, WI) Exradin Model 11 Spokas Chamber to measure PDD curves for 60, 80, 100 and 120 kVp x-ray beams with copper filtration ranging from 0.0–0.9 mm at 22cm and 42cm fields of view from 0 to 150 mm of water. A free-in-air monitor chamber was used to normalize the water tank data to fluctuations in output from the fluoroscope. The measurements were acquired on a Siemens (Erlangen, Germany) Artis ZeeGo fluoroscope. The fluoroscope was inverted from the typical orientation providing an x-ray beam originating from above the water tank. The water tank was positioned so that the water level was located at 60cm from the focal spot; which also represents the focal spot to interventional reference plane distance for that fluoroscope. Results: PDDs for 60, 80, 100, and 120 kVp with 0 mm of copper filtration compared well to previously published data by Fetterly et al. [Med Phys, 28, 205 (2001)] for those beam qualities given differences in fluoroscopes, geometric orientation, type of ionization chamber, and the water tank used for data collection. PDDs for 60, 80, 100, and 120 kVp with copper filtration were obtained and are presented, which have not been previously investigated and published. Conclusion: The equipment and processes used to acquire the reported data were sound and compared well with previously published data for PDDs without copper filtration. PDD data for the fluoroscopic x-ray beams incorporating copper filtration can be used as reference data for estimating organ or soft tissue dose at depth involving similar beam qualities or for comparison with mathematical models.

Purpose: To create digital phantoms using paired planning and replanning CTs from patients with H'N cancer and to comprehensively evaluate three commonly-used deformable image registration (DIR) algorithms with known ROIs and volumetric doses from the established phantoms. Methods: We have identified five H'N patients with adaptive-IMRT. For each patient, manually-adjusted meshes were created for selected physicianapproved ROIs. These meshes were applied to drive the thin-plate spline algorithm to deform the planning CT to match the replanning CT. The digital phantom consisted of the planning CT and simulated replanning CT, which were linked by a known deformation vector field (DVF). Three DIR algorithms (Fast-Symmetric Demons, B-Spline, and intensity-based) were applied to these five digital phantoms. The images, ROIs, and volumetric doses were mapped from the planning CT to the simulated replanning CT using the DVFs computed by these three DIRs and compared to those using the known DVF. Results: The average normalized cross-correlation (NCC) value between planning and replanning CTs with rigid registration was 0.924, indicating large anatomical changes. The NCC values were improved to 0.998, 0.974, and 0.996, with Demons, B-Spline, and intensity-based DIRs, respectively. The mean Dice coefficients for selected ROIs (e.g. PTV, brainstem, spinal cord, mandible, parotids) were in the range of 0.87-0.95, 0.84–0.94, and 0.85–0.95 for Demons, B-Spline, and intensity-based DIRs, respectively. The mean slice-wise Hausdorff distances for selected ROIs were in the range of 1.8–4.9 mm, 2.2–5.4 mm, and 2.4–6.6 mm for Demons, B-Spline, and intensity-based DIRs, respectively. The average percentage of voxels in the external contours that have dose errors

Purpose: To evaluate the impact of a commercial orthopedic metal artifact reduction (O-MAR) algorithm on CT image quality and dose calculation for patients with spinal prostheses near spinal tumors. Methods: A CT electron density phantom was scanned twice: with tissue-simulating inserts only, and with a titanium insert replacing solid water. A patient plan was mapped to the phantom images in two ways: with the titanium inside or outside of the spinal tumor. Pinnacle and Eclipse were used to evaluate the dosimetric effects of O-MAR on 12-bit and 16-bit CT data, respectively. CT images from five patients with spinal prostheses were reconstructed with and without O-MAR. Two observers assessed the image quality improvement from O-MAR. Both pencil beam and Monte Carlo dose calculation in iPlan were used for the patient study. The percentage differences between non-OMAR and O-MAR datasets were calculated for PTV-min, PTV-max, PTV-mean, PTV-V100, PTV-D90, OAR-V10Gy, OAR-max, and OAR-D0.1cc. Results: O-MAR improved image quality but did not significantly affect the dose distributions and DVHs for both 12-bit and 16- bit CT phantom data. All five patient cases demonstrated some degree of image quality improvement from O-MAR, ranging from small to large metal artifact reduction. For pencil beam, the largestmore » discrepancy was observed for OARV-10Gy at 5.4%, while the other seven parameters were ≤0.6%. For Monte Carlo, the differences between non-O-MAR and O-MAR datasets were ≤3.0%. Conclusion: Both phantom and patient studies indicated that O-MAR can substantially reduce metal artifacts on CT images, allowing better visualization of the anatomical structures and metal objects. The dosimetric impact of O-MAR was insignificant regardless of the metal location, image bit-depth, and dose calculation algorithm. O-MAR corrected images are recommended for radiation treatment planning on patients with spinal prostheses because of the improved image quality and no need to modify current dose constraints. This work was supported by a research grant from Philips Healthcare. Paul Klahr is an employee of Philips Healthcare.« less