Dataset Information

Dose Super-Resolution in Prostate Volumetric Modulated Arc Therapy Using Cascaded Deep Learning Networks.

ABSTRACT: Purpose: This study proposes a cascaded network model for generating high-resolution doses (i.e., a 1 mm grid) from low-resolution doses (i.e., ?3 mm grids) with reduced computation time.

Methods: Using the anisotropic analytical algorithm with three grid sizes (1, 3, and 5 mm) and the Acuros XB algorithm with two grid sizes (1 and 3 mm), dose distributions were calculated for volumetric modulated arc therapy plans for 73 prostate cancer patients. Our cascaded network model consisted of a hierarchically densely connected U-net (HD U-net) and a residual dense network (RDN), which were trained separately following a two-dimensional slice-by-slice procedure. The first network (HD U-net) predicted the downsampled high-resolution dose (generated through bicubic downsampling of the baseline high-resolution dose) using the low-resolution dose; subsequently, the second network (RDN) predicted the high-resolution dose from the output of the first network. Further, the predicted high-resolution dose was converted to its absolute value. We quantified the network performance using the spatial/dosimetric parameters (dice similarity coefficient, mean dose, maximum dose, minimum dose, homogeneity index, conformity index, and V_95%, V_70%, V_50%, and V_30%) for the low-resolution and predicted high-resolution doses relative to the baseline high-resolution dose. Gamma analysis (between the baseline dose and the low-resolution dose/predicted high-resolution dose) was performed with a 2%/2 mm criterion and 10% threshold.

Results: The average computation time to predict a high-resolution axial dose plane was <0.02 s. The dice similarity coefficient values for the predicted doses were closer to 1 when compared to those for the low-resolution doses. Most of the dosimetric parameters for the predicted doses agreed more closely with those for the baseline than for the low-resolution doses. In most of the parameters, no significant differences (p-value of >0.05) between the baseline and predicted doses were observed. The gamma passing rates for the predicted high-resolution does were higher than those for the low-resolution doses.

Conclusion: The proposed model accurately predicted high-resolution doses for the same dose calculation algorithm. Our model uses only dose data as the input without additional data, which provides advantages of convenience to user over other dose super-resolution methods.

SUBMITTER: Shin DS

PROVIDER: S-EPMC7701297 | biostudies-literature | 2020

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Dose Super-Resolution in Prostate Volumetric Modulated Arc Therapy Using Cascaded Deep Learning Networks.

Shin Dong-Seok DS Kim Kyeong-Hyeon KH Kang Sang-Won SW Kang Seong-Hee SH Kim Jae-Sung JS Kim Tae-Ho TH Kim Dong-Su DS Cho Woong W Suh Tae Suk TS Chung Jin-Beom JB

Frontiers in oncology 20201116

<h4>Purpose</h4>This study proposes a cascaded network model for generating high-resolution doses (i.e., a 1 mm grid) from low-resolution doses (i.e., ≥3 mm grids) with reduced computation time.<h4>Methods</h4>Using the anisotropic analytical algorithm with three grid sizes (1, 3, and 5 mm) and the Acuros XB algorithm with two grid sizes (1 and 3 mm), dose distributions were calculated for volumetric modulated arc therapy plans for 73 prostate cancer patients. Our cascaded network model consiste ...[more]

PMID: 33304852

Similar Datasets

Project description:The purpose of this work is to evaluate the volumetric-modulated arc therapy (VMAT) multicriteria optimization (MCO) algorithm clinically available in the RayStation treatment planning system (TPS) and its ability to reduce treatment planning time while providing high dosimetric plan quality. Nine patients with localized prostate cancer who were previously treated with 78 Gy in 39 fractions using VMAT plans and rayArc system based on the direct machine parameter optimization (DMPO) algorithm were selected and replanned using the VMAT-MCO system. First, the dosimetric quality of the plans was evaluated using multiple conformity metrics that account for target coverage and sparing of healthy tissue, used in our departmental clinical protocols. The conformity and homogeneity index, number of monitor units, and treatment planning time for both modalities were assessed. Next, the effects of the technical plan parameters, such as constraint leaf motion CLM (cm/°) and maximum arc delivery time T (s), on the accuracy of delivered dose were evaluated using quality assurance passing rates (QAs) measured using the Delta4 phantom from ScandiDos. For the dosimetric plan's quality analysis, the results show that the VMAT-MCO system provides plans comparable to the rayArc system with no statistical difference for V95% (p < 0.01), D1% (p < 0.01), CI (p < 0.01), and HI (p < 0.01) of the PTV, bladder (p < 0.01), and rectum (p < 0.01) constraints, except for the femoral heads and healthy tissues, for which a dose reduction was observed using MCO compared with rayArc (p < 0.01). The technical parameter study showed that a combination of CLM equal to 0.5 cm/degree and a maximum delivery time of 72 s allowed the accurate delivery of the VMAT-MCO plan on the Elekta Versa HD linear accelerator. Planning evaluation and dosimetric measurements showed that VMAT-MCO can be used clinically with the advantage of enhanced planning process efficiency by reducing the treatment planning time without impairing dosimetric quality.

Project description:To investigate boosting dominant intraprostatic lesions (DILs) in the context of stereotactic ablative radiation therapy (SABR) and to examine the impact on tumor control probability (TCP) and normal tissue complication probability (NTCP).Ten prostate datasets were selected. DILs were defined using T2-weighted, dynamic contrast-enhanced and diffusion-weighted magnetic resonance imaging. Four plans were produced for each dataset: (1) no boost to DILs; (2) boost to DILs, no seminal vesicles in prescription; (3) boost to DILs, proximal seminal vesicles (proxSV) prescribed intermediate dose; and (4) boost to DILs, proxSV prescribed higher dose. The prostate planning target volume (PTV) prescription was 42.7 Gy in 7 fractions. DILs were initially prescribed 115% of the PTV(Prostate) prescription, and PTV(DIL) prescriptions were increased in 5% increments until organ-at-risk constraints were reached. TCP and NTCP calculations used the LQ-Poisson Marsden, and Lyman-Kutcher-Burman models respectively.When treating the prostate alone, the median PTV(DIL) prescription was 125% (range: 110%-140%) of the PTV(Prostate) prescription. Median PTV(DIL) D50% was 55.1 Gy (range: 49.6-62.6 Gy). The same PTV(DIL) prescriptions and similar PTV(DIL) median doses were possible when including the proxSV within the prescription. TCP depended on prostate ?/? ratio and was highest with an ?/? ratio = 1.5 Gy, where the additional TCP benefit of DIL boosting was least. Rectal NTCP increased with DIL boosting and was considered unacceptably high in 5 cases, which, when replanned with an emphasis on reducing maximum dose to 0.5 cm(3) of rectum (Dmax(0.5cc)), as well as meeting existing constraints, resulted in considerable rectal NTCP reductions.Boosting DILs in the context of SABR is technically feasible but should be approached with caution. If this therapy is adopted, strict rectal constraints are required including Dmax(0.5cc). If the ?/? ratio of prostate cancer is 1.5 Gy or less, then high TCP and low NTCP can be achieved by prescribing SABR to the whole prostate, without the need for DIL boosting.

Project description:Hypofractionated treatments generally increase the complexity of a treatment plan due to the more stringent constraints of normal tissues and target coverage. As a result, treatment plans contain more modulated MLC motions that may require extra efforts for accurate dose calculation. This study explores methods to minimize the differences between in-house dose calculation and actual delivery of hypofractionated volumetric-modulated arc therapy (VMAT), by focusing on arc approximation and tongue-and-groove (TG) modeling. For dose calculation, the continuous delivery arc is typically approximated by a series of static beams with an angular spacing of 2°. This causes significant error when there is large MLC movement from one beam to the next. While increasing the number of beams will minimize the dose error, calculation time will increase significantly. We propose a solution by inserting two additional apertures at each of the beam angle for dose calculation. These additional apertures were interpolated at two-thirds' degree before and after each beam. Effectively, there were a total of three MLC apertures at each beam angle, and the weighted average fluence from the three apertures was used for calculation. Because the number of beams was kept the same, calculation time was only increased by about 6%-8%. For a lung plan, areas of high local dose differences (> 4%) between film measurement and calculation with one aperture were significantly reduced in calculation with three apertures. Ion chamber measurement also showed similar results, where improvements were seen with calculations using additional apertures. Dose calculation accuracy was further improved for TG modeling by developing a sampling method for beam fluence matrix. Single element point sampling for fluence transmitted through MLC was used for our fluence matrix with 1 mm resolution. For Varian HDMLC, grid alignment can cause fluence sampling error. To correct this, transmission volume averaging was applied. For three paraspinal HDMLC cases, the average dose difference was greatly reduced in film and calculation comparisons with our new approach. The gamma (3%, 3 mm) pass rates have improved significantly from 74.1%, 90.0%, and 90.4% to 99.2%, 97.9%, and 97.3% for three cases, for calculation without volume averaging and calculation with volume averaging, respectively. Our results indicate that more accurate MLC leaf position and transmission sampling can improve accuracy and agreement between calculation and measurement, and are particularly important for hypofractionated VMAT that consists of large MLC movement.

Project description:IntroductionProstate bed radiotherapy (RT) is a major affecter of patients' long-term quality of life (QoL). To ensure the best possible outcome of these patients, dose constraints are key for optimal RT planning and delivery. However, establishing refined dose constraints requires access to patient-level data. Therefore, we aimed to provide such data on the relationship between OAR and gastrointestinal (GI) as well as genitourinary (GU) QoL outcomes of a homogenous patient cohort who received dose-intensified post-operative RT to the prostate bed. Furthermore, we aimed to conduct an exploratory analysis of the resulting data.MethodsPatients who were treated with prostate bed RT between 2010 and 2020 were inquired about their QoL based on the Expanded Prostate Cancer Index Composite (EPIC). Those (n = 99) who received volumetric arc therapy (VMAT) of at least 70 Gy to the prostate bed were included. Dose-volume histogram (DVH) parameters were gathered and correlated with the EPIC scores.ResultsThe median age at the time of prostate bed RT was 68.9 years, and patients were inquired about their QoL in the median 2.3 years after RT. The median pre-RT prostate-specific antigen (PSA) serum level was 0.35 ng/mL. The median duration between surgery and RT was 1.5 years. The median prescribed dose to the prostate bed was 72 Gy. A total of 61.6% received prostate bed RT only. For the bladder, the highest level of statistical correlation (p < 0.01) was seen for V10-20Gy, Dmean and Dmedian with urinary QoL. For bladder wall, the highest level of statistically significant correlation (p < 0.01) was seen for V5-25Gy, Dmean and Dmedian with urinary QoL. Penile bulb V70Gy was statistically significantly correlated with sexual QoL (p < 0.05). A larger rectal volume was significantly correlated with improved bowel QoL (p < 0.05). Sigmoid and urethral DVH parameters as well as the surgical approach were not statistically significantly correlated with QoL.ConclusionSpecific dose constraints for bladder volumes receiving low doses seem desirable for the further optimization of prostate bed RT. This may be particularly relevant in the context of the aspiration of establishing focal RT of prostate cancer and its local recurrences. Our comprehensive dataset may aid future researchers in achieving these goals.

Project description:PurposeTo develop a 4D dose reconstruction method and to evaluate the delivered dose in respiratory-gated volumetric modulated arc therapy (VMAT).Materials and methodsA total 112 treatment sessions of gated VMAT for 30 stereotactic body radiotherapy (SBRT) patients (10 lung, 10 liver, and 10 pancreas) were evaluated. For respiratory-gated SBRT, 4DCT was acquired, and the CT data at the end-exhale phase was used for a VMAT plan. The delivered dose was reconstructed using a patient's respiratory motion and machine motion acquired during the beam delivery. The machine motion was obtained from the treatment log file, while the target position was estimated from an external respiratory marker position. The target position was divided into 1-mm position bins, and sub-beams with beam isocenters corresponding to each position bin were created in a motion mimicking plan, reflecting motion data including MLC leaf positions and gantry angle and target position data during beam treatment. The reconstructed 4D dose was compared with the dose of the original plan using these dosimetric parameters; the maximum dose (Dmax) and mean dose (Dmean) of gross target volume (GTV) or organs at risk (spinal cord, esophagus, heart, duodenum, kidney, spinal cord, and stomach). The minimum dose (Dmin) to GTV was also calculated to verify cold spots in tumors.ResultsThere was no significant difference of dose parameters regard to the GTV in all tumors. For the liver cases, there were significant differences in the Dmax of duodenum (-4.2 ± 1.4%), stomach (-3.5 ± 4.2%), left kidney (-4.1 ± 2.8%), and right kidney (-3.2 ± 1.3%), and in the Dmean of duodenum (-3.8 ± 1.4%), stomach (-3.9 ± 2.2%), left kidney (-3.1 ± 2.8%), and right kidney (-4.1 ± 2.6%). For the pancreas cases, there were significant differences in the Dmax of stomach (2.1 ± 3.0%), and in the Dmean of liver (1.5 ± 0.6%), duodenum (-1.0 ± 1.4%), stomach (2.1 ± 1.6%), and right kidney (-1.3 ± 0.9%). The average gamma pass rates were 97.6 ± 4.8% for lung cases, 99.6 ± 0.5% for liver cases, and 99.5 ± 0.5% for pancreas cases. Most cases showed insignificant dose variation, with gamma pass rates higher than 98%, except for two lung cases with gamma pass rates of 86.9% and 90.6%. The low gamma pass rates showed larger global motion ranges resulting from the baseline shift during beam delivery.ConclusionThe actual delivered dose in thoracic and abdominal VMAT under breathing motion was verified by 4D dose reconstruction using typical treatment equipment and software. The proposed method provides a verification method for the actual delivered dose and could be a dosimetric verification QA tool for radiation treatment under various respiratory management techniques.

Dataset Information

Dose Super-Resolution in Prostate Volumetric Modulated Arc Therapy Using Cascaded Deep Learning Networks.

Publications

Dose Super-Resolution in Prostate Volumetric Modulated Arc Therapy Using Cascaded Deep Learning Networks.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets