Development of a quasi-3D dosimeter using radiochromic plastic for patient-specific quality assurance.

BACKGROUND: Patient-specific QA verification ensures patient safety and treatment by verifying radiation delivery and dose calculations in treatment plans for errors. However, a two-dimensional (2D) dose distribution is insufficient for detecting information on the three-dimensional (3D) dose delivered to the patient. In addition, 3D radiochromic plastic dosimeters (RPDs) such as PRESAGE® represent the volume effect in which the dosimeters have different sensitivities according to the size of the dosimeters. Therefore, to solve the volume effect, a Quasi-3D dosimetry system was proposed to perform patient-specific QA using predetermined-sized and multiple RPDs. PURPOSE: For patient-specific quality assurance (QA) in radiation treatment, this study aims to assess a quasi-3D dosimetry system using an RPD. METHODS: Gamma analysis was performed to verify the agreement between the measured and estimated dose distributions of intensity-modulated radiotherapy (IMRT) and volumetric modulated arc therapy (VMAT). We fabricated cylindrical RPDs and a quasi-3D dosimetry phantom. A practicability test for a pancreatic patient utilized a quasi-3D dosimetry device, an in-house RPD, and a quasi-3D phantom. The dose distribution of the VMAT design dictated the placement of nine RPDs. Moreover, a 2D diode array detector was used for 2D gamma analysis (MapCHECK2). The patient-specific QA was performed for IMRT, VMAT, and stereotactic ablative radiotherapy (SABR) in 20 prostate and head-and-neck patients. For each patient, six RPDs were positioned according to the dose distribution. VMAT SABR and IMRT/VMAT plans employed a 2%/2 mm gamma criterion, whereas IMRT/VMAT plans used a 3%/2 mm gamma criterion, a 10% threshold value, and a 90% passing rate tolerance. 3D gamma analysis was conducted using the 3D Slicer software. RESULTS: The average gamma passing rates with 2%/2 mm and 3%/3 mm criteria for relative dose distribution were 91.6% ± 1.4% and 99.4% ± 0.7% for the 3D gamma analysis using the quasi-3D dosimetry system, respectively, and 97.5% and 99.3% for 2D gamma analysis using MapCHECK2, respectively. The 3D gamma analysis for patient-specific QA of 20 patients showed passing rates of over 90% with 2%/2 mm, 3%/2 mm, and 3%/3 mm criteria. CONCLUSIONS: The quasi-3D dosimetry system was evaluated by performing patient-specific QAs with RPDs and quasi-3D phantom. The gamma indices for all RPDs showed more than 90% for 2%/2 mm, 3%/2 mm, and 3%/3 mm criteria. We verified the feasibility of a quasi-3D dosimetry system by performing the conventional patient-specific QA with the quasi-3D dosimeters.

Evaluation of initial patient setup methods for breast cancer between surface-guided radiation therapy and laser alignment based on skin marking in the Halcyon system.

BACKGROUND: This study was conducted to evaluate the efficiency and accuracy of the daily patient setup for breast cancer patients by applying surface-guided radiation therapy (SGRT) using the Halcyon system instead of conventional laser alignment based on the skin marking method. METHODS AND MATERIALS: We retrospectively investigated 228 treatment fractions using two different initial patient setup methods. The accuracy of the residual rotational error of the SGRT system was evaluated by using an in-house breast phantom. The residual translational error was analyzed using the couch position difference in the vertical, longitudinal, and lateral directions between the reference computed tomography and daily kilo-voltage cone beam computed tomography acquired from the record and verification system. The residual rotational error (pitch, yaw, and roll) was also calculated using an auto rigid registration between the two images based on Velocity. The total setup time, which combined the initial setup time and imaging time, was analyzed to evaluate the efficiency of the daily patient setup for SGRT. RESULTS: The average residual rotational errors using the in-house fabricated breast phantom for pitch, roll, and yaw were 0.14°, 0.13°, and 0.29°, respectively. The average differences in the couch positions for laser alignment based on the skin marking method were 2.7 ± 1.6 mm, 2.0 ± 1.2 mm, and 2.1 ± 1.0 mm for the vertical, longitudinal, and lateral directions, respectively. For SGRT, the average differences in the couch positions were 1.9 ± 1.2 mm, 2.9 ± 2.1 mm, and 1.9 ± 0.7 mm for the vertical, longitudinal, and lateral directions, respectively. The rotational errors for pitch, yaw, and roll without the surface-guided radiation therapy approach were 0.32 ± 0.30°, 0.51 ± 0.24°, and 0.29 ± 0.22°, respectively. For SGRT, the rotational errors were 0.30 ± 0.22°, 0.51 ± 0.26°, and 0.19 ± 0.13°, respectively. The average total setup times considering both the initial setup time and imaging time were 314 s and 331 s, respectively, with and without SGRT. CONCLUSION: We demonstrated that using SGRT improves the accuracy and efficiency of initial patient setups in breast cancer patients using the Halcyon system, which has limitations in correcting the rotational offset.

Performance evaluation of a visual guidance patient-controlled respiratory gating system for respiratory-gated magnetic-resonance image-guided radiation therapy.

The performance of a visual guidance patient-controlled (VG-PC) respiratory gating system for magnetic-resonance (MR) image-guided radiation therapy (MR-IGRT) was evaluated through a clinical trial of patients with either lung or liver cancer. Patients can voluntarily control their respiration utilizing the VG-PC respiratory gating system. The system enables patients to view near-real-time cine planar MR images projected inside the bore of MR-IGRT systems or an external screen. Twenty patients who had received stereotactic ablative radiotherapy (SABR) for lung or liver cancer were prospectively selected for this study. Before the first treatment, comprehensive instruction on the VG-PC respiratory gating system was provided to the patients. Respiratory-gated MR-IGRT was performed for each patient with it in the first fraction and then without it in the second fraction. For both the fractions, the total treatment time, beam-off time owing to the respiratory gating, and number of beam-off events were analyzed. The average total treatment time, beam-off time, and number of beam-off events with the system were 1507.3 s, 679.5 s, and 185, respectively, and those without the system were 2023.7 s (p < 0.001), 1195.0 s (p < 0.001), and 380 times (p < 0.001), respectively. The VG-PC respiratory gating system improved treatment efficiency through a reduction in the beam-off time, the number of beam-off events, and consequently the total treatment time when performing respiratory-gated MR-IGRT for lung and liver SABR.

Development of an anthropomorphic multimodality pelvic phantom for quantitative evaluation of a deep-learning-based synthetic computed tomography generation technique.

PURPOSE: The objective of this study was to fabricate an anthropomorphic multimodality pelvic phantom to evaluate a deep-learning-based synthetic computed tomography (CT) algorithm for magnetic resonance (MR)-only radiotherapy. METHODS: Polyurethane-based and silicone-based materials with various silicone oil concentrations were scanned using 0.35 T MR and CT scanner to determine the tissue surrogate. Five tissue surrogates were determined by comparing the organ intensity with patient CT and MR images. Patient-specific organ modeling for three-dimensional printing was performed by manually delineating the structures of interest. The phantom was finally fabricated by casting materials for each structure. For the quantitative evaluation, the mean and standard deviations were measured within the regions of interest on the MR, simulation CT (CTsim ), and synthetic CT (CTsyn ) images. Intensity-modulated radiation therapy plans were generated to assess the impact of different electron density assignments on plan quality using CTsim and CTsyn . The dose calculation accuracy was investigated in terms of gamma analysis and dose-volume histogram parameters. RESULTS: For the prostate site, the mean MR intensities for the patient and phantom were 78.1 ± 13.8 and 86.5 ± 19.3, respectively. The mean intensity of the synthetic image was 30.9 Hounsfield unit (HU), which was comparable to that of the real CT phantom image. The original and synthetic CT intensities of the fat tissue in the phantom were -105.8 ± 4.9 HU and -107.8 ± 7.8 HU, respectively. For the target volume, the difference in D95% was 0.32 Gy using CTsyn with respect to CTsim values. The V65Gy values for the bladder in the plans using CTsim and CTsyn were 0.31% and 0.15%, respectively. CONCLUSION: This work demonstrated that the anthropomorphic phantom was physiologically and geometrically similar to the patient organs and was employed to quantitatively evaluate the deep-learning-based synthetic CT algorithm.

Synthetic CT generation from weakly paired MR images using cycle-consistent GAN for MR-guided radiotherapy.

Although MR-guided radiotherapy (MRgRT) is advancing rapidly, generating accurate synthetic CT (sCT) from MRI is still challenging. Previous approaches using deep neural networks require large dataset of precisely co-registered CT and MRI pairs that are difficult to obtain due to respiration and peristalsis. Here, we propose a method to generate sCT based on deep learning training with weakly paired CT and MR images acquired from an MRgRT system using a cycle-consistent GAN (CycleGAN) framework that allows the unpaired image-to-image translation in abdomen and thorax. Data from 90 cancer patients who underwent MRgRT were retrospectively used. CT images of the patients were aligned to the corresponding MR images using deformable registration, and the deformed CT (dCT) and MRI pairs were used for network training and testing. The 2.5D CycleGAN was constructed to generate sCT from the MRI input. To improve the sCT generation performance, a perceptual loss that explores the discrepancy between high-dimensional representations of images extracted from a well-trained classifier was incorporated into the CycleGAN. The CycleGAN with perceptual loss outperformed the U-net in terms of errors and similarities between sCT and dCT, and dose estimation for treatment planning of thorax, and abdomen. The sCT generated using CycleGAN produced virtually identical dose distribution maps and dose-volume histograms compared to dCT. CycleGAN with perceptual loss outperformed U-net in sCT generation when trained with weakly paired dCT-MRI for MRgRT. The proposed method will be useful to increase the treatment accuracy of MR-only or MR-guided adaptive radiotherapy. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s13534-021-00195-8.

Dose calculation of 3D printing lead shield covered by biocompatible silicone for electron beam therapy.

This study aims to calculate the dose delivered to the upstream surface of a biocompatible flexible absorber covering lead for electron beam treatment of skin and subcutaneous tumour lesions for head and neck. Silicone (Ecoflex™ 00-30, Smooth-On, Easton, PA, USA) was used to cover the lead to absorb backscattered electrons from lead. A 3D printer (Zortrax M300, Zortrax, Olsztyn, Poland) was used to fabricate the lead shield. Analytic calculation, simplified Monte Carlo (MC) simulation, and detailed MC simulation which includes a modeling of metal-oxide-semiconductor field-effect transistor (MOSFET) detector were performed to determine the electron backscatter factor (EBF) for 6 MeV and 9 MeV electron beams of a Varian iX Silhouette. MCNP6.2 was used to calculate the EBF and corresponding measurements were carried out by using MOSFET detectors. The EBF was experimentally measured by the ratio of dose at the upstream surface of the silicone to the same point without the presence of the lead shield. The results derived by all four methods agreed within 2.8% for 6 MeV and 3.4% for 9 MeV beams. In detailed MC simulations, for 6 MeV, dose to the surface of 7-mm-thick absorber was 103.7 [Formula: see text] 1.9% compared to dose maximum (Dmax) without lead. For 9 MeV, the dose to the surface of the 10-mm-thick absorber was 104.1 [Formula: see text] 2.1% compared to Dmax without lead. The simplified MC simulation was recommended for practical treatment planning due to its acceptable calculation accuracy and efficiency. The simplified MC simulation was completed within 20 min using parallel processing with 80 CPUs, while the detailed MC simulation required 40 h to be done. In this study, we outline the procedures to use the lead shield covered by silicone in clinical practice from fabrication to dose calculation.

Inter-institutional Variation in Intensity-modulated Radiotherapy for Breast Cancer in Korea (KROG 19-01).

BACKGROUND/AIM: To present the variations in the target delineation and the planning results of intensity-modulated radiation therapy (IMRT) for breast cancers. PATIENTS AND METHODS: We requested the target volumes and organs at risk delineation for two cases of left breast cancers, and evaluated the IMRT plans including the supraclavicular and internal mammary node irradiation. RESULTS: Twenty-one institutions participated in this study. Differences in the planning target volume among institutions reached up to three-times for breast-conserving surgery (BCS) case and five-times for mastectomy case. Mean heart doses ranged from 3.3 to 24.1 Gy for BCS case and from 5.0 to 26.5 Gy for mastectomy case. Ipsilateral lung volumes receiving more than 20 Gy ranged from 4.7 to 57.4% for BCS case and from 16.4 to 55.5% for mastectomy case. CONCLUSION: There were large variations in the target delineation and planning results of IMRT for breast cancers among institutions. Considering the increased use of breast IMRT, more standardized protocols are needed.

Deep learning-enabled accurate normalization of reconstruction kernel effects on emphysema quantification in low-dose CT.

Lung densitometry is being frequently adopted in CT-based emphysema quantification, yet known to be affected by the choice of reconstruction kernel. This study presents a two-step deep learning architecture that enables accurate normalization of reconstruction kernel effects on emphysema quantification in low-dose CT. Deep learning is used to convert a CT image of a sharp kernel to that of a standard kernel with restoration of truncation artifacts and smoothing-free pixel size normalization. We selected 353 scans reconstructed by both standard and sharp kernels from four different CT scanners from the United States National Lung Screening Trial program database. A truncation artifact correction model was constructed with a combination of histogram extrapolation and a deep learning model trained with truncated and non-truncated image sets. Then, we performed frequency domain zero-padding to normalize reconstruction field of view effects while preventing image smoothing effects. The kernel normalization model has a U-Net based architecture trained for each CT scanner dataset. Three lung density measurements including relative lung area under 950 HU (RA950), lower 15th percentile threshold (perc15), and mean lung density were obtained in the datasets from standard, sharp, and normalized kernels. The effect of kernel normalization was evaluated with pair-wise differences in lung density metrics. The mean of pair-wise differences in RA950 between standard and sharp kernel reconstructions was reduced from 10.75% to -0.07% using kernel normalization. The difference for perc15 decreased from -31.03 HU to -0.30 HU after kernel normalization. Our study demonstrated the feasibility of applying deep learning techniques for normalizing CT kernel effects, thereby reducing the kernel-induced variability in lung density measurements. The deep learning model could increase the accuracy of emphysema quantification, thereby allowing reliable surveillance of emphysema in lung cancer screening even when follow-up CT scans are acquired with different reconstruction kernels.