Baozhou Sun
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Baozhou Sun has a total of 102 co-authors. Here are the top co-authors by number of shared publications:
Sasa Mutic
9 shared publicationsTianyu Zhao
Department of Radiation Oncology, Physics Division, Washington University in St. Louis School of Medicine, St. Louis, MO 63108, United States of America
7 shared publications- Sridhar Yaddanapudi
Department of Radiation Oncology, Mayo Clinic, Jacksonville, FL, USA
5 shared publications Dao Lam
Radiation Oncology, Washington University in Saint Louis School of Medicine, Saint Louis, Missouri, UNITED STATES
4 shared publications- Sasa Mutic
Washington University School of Medicine, St. Louis, MO
4 shared publications - Bin Cai
Department of Radiation Oncology, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA
3 shared publications
Latest Publications
- 1 June 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (3) , 035084
Dosimetric comparison of high dose rate interstitial brachytherapy and stereotactic body radiotherapy as alternative treatment methods for advanced-stage cervical cancer
Purpose. High-dose-rate interstitial brachytherapy (HDR-IBT) is an essential component of curative treatment for locally advanced cervical cancer, but it is invasive and resource-intensive. Alternative boost techniques are needed when brachytherapy is not feasible. This study compares the dosimetric performance of stereotactic body radiotherapy (SBRT) and Syed-based HDR-IBT to evaluate the potential role of SBRT as a non-invasive boost modality. Materials and Methods. Eighteen patients with stage III–IV cervical cancer previously treated with external beam radiotherapy (EBRT) (45 Gy/25 fx ± nodal boost) followed by Syed HDR-IBT (28 Gy/4 fx) in our clinic were retrospectively analyzed. For each patient, SBRT plans prescribing 28 Gy in 4 fractions were generated on EBRT computed tomography images using fused Syed contours. Dosimetric parameters for targets and organs-at-risk (OARs), such as D 98 , D 90 , mean dose, maximum dose, D 2cc , D 1cc , were compared using the Wilcoxon signed-rank test. Results. For the high-risk clinical target volume (CTV_HR), HDR-IBT provided significantly superior coverage: mean 27 Gy D 98 and 32 Gy D 90 (Syed) vs. mean 25 Gy D 98 and 29 Gy D 90 (SBRT) ( p = 0.04 and p < 0.001). For the intermediate-risk CTV, SBRT plans achieved significantly better coverage: mean 18 Gy D 98 and 22 Gy D 90 (SBRT) vs. mean 14 Gy D 98 and 18 Gy D 90 (Syed) ( p < 0.001 for both). OAR doses were largely comparable between the two modalities, except that HDR-IBT resulted in significantly lower mean bladder dose and improved bowel sparing, whereas SBRT yielded significantly lower bladder maximum dose and reduced rectum D 2cc . Overall, SBRT exhibited greater low-dose spread to surrounding tissues, while HDR-IBT produced higher intratumoral dose hot spots. Conclusion. HDR-IBT remains superior for CTV_HR dose escalation, but SBRT can achieve acceptable target coverage and OAR sparing when brachytherapy is not feasible. SBRT may serve as a practical alternative for selected patients who are medically inoperable, anatomically unsuitable, or lack access to brachytherapy services. Prospective studies are needed to optimize SBRT planning strategies and define appropriate patient selection criteria.
- 1 June 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (3) , 035084
Dosimetric comparison of high dose rate interstitial brachytherapy and stereotactic body radiotherapy as alternative treatment methods for advanced-stage cervical cancer
Purpose. High-dose-rate interstitial brachytherapy (HDR-IBT) is an essential component of curative treatment for locally advanced cervical cancer, but it is invasive and resource-intensive. Alternative boost techniques are needed when brachytherapy is not feasible. This study compares the dosimetric performance of stereotactic body radiotherapy (SBRT) and Syed-based HDR-IBT to evaluate the potential role of SBRT as a non-invasive boost modality. Materials and Methods. Eighteen patients with stage III–IV cervical cancer previously treated with external beam radiotherapy (EBRT) (45 Gy/25 fx ± nodal boost) followed by Syed HDR-IBT (28 Gy/4 fx) in our clinic were retrospectively analyzed. For each patient, SBRT plans prescribing 28 Gy in 4 fractions were generated on EBRT computed tomography images using fused Syed contours. Dosimetric parameters for targets and organs-at-risk (OARs), such as D 98 , D 90 , mean dose, maximum dose, D 2cc , D 1cc , were compared using the Wilcoxon signed-rank test. Results. For the high-risk clinical target volume (CTV_HR), HDR-IBT provided significantly superior coverage: mean 27 Gy D 98 and 32 Gy D 90 (Syed) vs. mean 25 Gy D 98 and 29 Gy D 90 (SBRT) ( p = 0.04 and p < 0.001). For the intermediate-risk CTV, SBRT plans achieved significantly better coverage: mean 18 Gy D 98 and 22 Gy D 90 (SBRT) vs. mean 14 Gy D 98 and 18 Gy D 90 (Syed) ( p < 0.001 for both). OAR doses were largely comparable between the two modalities, except that HDR-IBT resulted in significantly lower mean bladder dose and improved bowel sparing, whereas SBRT yielded significantly lower bladder maximum dose and reduced rectum D 2cc . Overall, SBRT exhibited greater low-dose spread to surrounding tissues, while HDR-IBT produced higher intratumoral dose hot spots. Conclusion. HDR-IBT remains superior for CTV_HR dose escalation, but SBRT can achieve acceptable target coverage and OAR sparing when brachytherapy is not feasible. SBRT may serve as a practical alternative for selected patients who are medically inoperable, anatomically unsuitable, or lack access to brachytherapy services. Prospective studies are needed to optimize SBRT planning strategies and define appropriate patient selection criteria.
- Open Access
- 6 January 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (1) , 015039
Optimizing and evaluating robustness of AI for brain metastasis detection and segmentation via loss functions and multi-dataset training
Purpose . Accurate detection and segmentation of brain metastases (BM) from MRI are critical for the appropriate management of cancer patients. This study investigates strategies to enhance the robustness of artificial intelligence (AI)-based BM detection and segmentation models. Method . A DeepMedic-based network with a loss function, tunable with a sensitivity/specificity tradeoff weighting factor α - was trained on T1 post-contrast MRI datasets from two institutions (514 patients, 4520 lesions). Robustness was evaluated on an external dataset from a third institution dataset (91 patients, 397 lesions), featuring ground truth annotations from two physicians. We investigated the impact of loss function weighting factor, α and training dataset combinations. Detection performance (sensitivity, precision, F1 score) and segmentation accuracy (Dice similarity, and 95% Hausdorff distance (HD95)) were evaluated using one physician’s contours as the reference standard. The optimal AI model was then directly compared to the performance of the second physician. Results . Varying α demonstrated a trade-off between sensitivity (higher α ) and precision (lower α ), with α = 0.5 yielding the best F1 score (0.80 versus 0.78 for α = 0.95 and 0.72 for α = 0.99) on the external dataset. The optimally trained model achieved detection performance comparable to the physician (F1: AI = 0.83, Physician = 0.83), but slightly underperformed in segmentation (Dice: 0.81 versus AI = 0.69; HD95: 3.0 mm versus AI = 4.94 mm, p < 0.05). Conclusion . The derived optimal model achieves detection and segmentation performance comparable to an expert physician in a parallel comparison.
- 6 January 2026
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 12 (1) , 015039
Open AccessOptimizing and evaluating robustness of AI for brain metastasis detection and segmentation via loss functions and multi-dataset training
Purpose . Accurate detection and segmentation of brain metastases (BM) from MRI are critical for the appropriate management of cancer patients. This study investigates strategies to enhance the robustness of artificial intelligence (AI)-based BM detection and segmentation models. Method . A DeepMedic-based network with a loss function, tunable with a sensitivity/specificity tradeoff weighting factor α - was trained on T1 post-contrast MRI datasets from two institutions (514 patients, 4520 lesions). Robustness was evaluated on an external dataset from a third institution dataset (91 patients, 397 lesions), featuring ground truth annotations from two physicians. We investigated the impact of loss function weighting factor, α and training dataset combinations. Detection performance (sensitivity, precision, F1 score) and segmentation accuracy (Dice similarity, and 95% Hausdorff distance (HD95)) were evaluated using one physician’s contours as the reference standard. The optimal AI model was then directly compared to the performance of the second physician. Results . Varying α demonstrated a trade-off between sensitivity (higher α ) and precision (lower α ), with α = 0.5 yielding the best F1 score (0.80 versus 0.78 for α = 0.95 and 0.72 for α = 0.99) on the external dataset. The optimally trained model achieved detection performance comparable to the physician (F1: AI = 0.83, Physician = 0.83), but slightly underperformed in segmentation (Dice: 0.81 versus AI = 0.69; HD95: 3.0 mm versus AI = 4.94 mm, p < 0.05). Conclusion . The derived optimal model achieves detection and segmentation performance comparable to an expert physician in a parallel comparison.
- 23 August 2022
- Published by IOP Publishing in Physics in Medicine & Biology
- Vol. 67 (17) , 175006
A multi-layer strip ionization chamber (MLSIC) device for proton pencil beam scan quality assurance
Objective. Proton pencil beam scanning (PBS) treatment fields needs to be verified before treatment deliveries to ensure patient safety. In current practice, treatment beam quality assurance (QA) is measured at a few selected depths using film or a 2D detector array, which is insensitive and time-consuming. A QA device that can measure all key dosimetric characteristics of treatment beams spot-by-spot within a single beam delivery is highly desired. Approach. We developed a multi-layer strip ionization chamber (MLSIC) prototype device that comprises of two layers of strip ionization chambers (IC) plates for spot position measurement and 64 layers of plate IC for beam energy measurement. The 768-channel strip ion chamber signals are integrated and sampled at a speed of 3.125 kHz. It has a 25.6 cm × 25.6 cm maximum measurement field size and 2 mm spatial resolution for spot position measurement. The depth resolution and maximum depth were 2.91 mm and 18.6 cm for 1.6 mm thick IC plate, respectively. The relative weight of each spot was determined from total charge by all IC detector channels. Main results. The MLSIC is able to measure ionization currents spot-by-spot. The depth dose measurement has a good agreement with the ground truth measured using a water tank and commercial one-dimensional (1D) multi-layer plate chamber. It can verify the spot position, energy, and relative weight of clinical PBS beams and compared with the treatment plans. Significance. The MLSIC is a highly efficient QA device that can measure the key dosimetric characteristics of proton treatment beams spot-by-spot with a single beam delivery. It may improve the quality and efficiency of clinical proton treatments.
- 23 August 2022
- Published by IOP Publishing in Physics in Medicine & Biology
- Vol. 67 (17) , 175006
A multi-layer strip ionization chamber (MLSIC) device for proton pencil beam scan quality assurance
Objective. Proton pencil beam scanning (PBS) treatment fields needs to be verified before treatment deliveries to ensure patient safety. In current practice, treatment beam quality assurance (QA) is measured at a few selected depths using film or a 2D detector array, which is insensitive and time-consuming. A QA device that can measure all key dosimetric characteristics of treatment beams spot-by-spot within a single beam delivery is highly desired. Approach. We developed a multi-layer strip ionization chamber (MLSIC) prototype device that comprises of two layers of strip ionization chambers (IC) plates for spot position measurement and 64 layers of plate IC for beam energy measurement. The 768-channel strip ion chamber signals are integrated and sampled at a speed of 3.125 kHz. It has a 25.6 cm × 25.6 cm maximum measurement field size and 2 mm spatial resolution for spot position measurement. The depth resolution and maximum depth were 2.91 mm and 18.6 cm for 1.6 mm thick IC plate, respectively. The relative weight of each spot was determined from total charge by all IC detector channels. Main results. The MLSIC is able to measure ionization currents spot-by-spot. The depth dose measurement has a good agreement with the ground truth measured using a water tank and commercial one-dimensional (1D) multi-layer plate chamber. It can verify the spot position, energy, and relative weight of clinical PBS beams and compared with the treatment plans. Significance. The MLSIC is a highly efficient QA device that can measure the key dosimetric characteristics of proton treatment beams spot-by-spot with a single beam delivery. It may improve the quality and efficiency of clinical proton treatments.
- Open Access
- 15 October 2021
- Published by Wiley in Medical Physics
- Vol. 48 (11) , 7052-7062
Weaving attention U‐net: A novel hybrid CNN and attention‐based method for organs‐at‐risk segmentation in head and neck CT images
Purpose: In radiotherapy planning, manual contouring is labor‐intensive and time‐consuming. Accurate and robust automated segmentation models improve the efficiency and treatment outcome. We aim to develop a novel hybrid deep learning approach, combining convolutional neural networks (CNNs) and the self‐attention mechanism, for rapid and accurate multi‐organ segmentation on head and neck computed tomography (CT) images.Methods: Head and neck CT images with manual contours of 115 patients were retrospectively collected and used. We set the training/validation/testing ratio to 81/9/25 and used the 10‐fold cross‐validation strategy to select the best model parameters. The proposed hybrid model segmented 10 organs‐at‐risk (OARs) altogether for each case. The performance of the model was evaluated by three metrics, that is, the Dice Similarity Coefficient (DSC), Hausdorff distance 95% (HD95), and mean surface distance (MSD). We also tested the performance of the model on the head and neck 2015 challenge dataset and compared it against several state‐of‐the‐art automated segmentation algorithms.Results: The proposed method generated contours that closely resemble the ground truth for 10 OARs. On the head and neck 2015 challenge dataset, the DSC scores of these OARs were 0.910.02, 0.730.10, 0.950.03, 0.760.08, 0.790.05, 0.870.05, 0.860.08, 0.870.03, and 0.870.07 for brain stem, chiasm, mandible, left/right optic nerve, left/right submandibular, and left/right parotid, respectively. Our results of the new weaving attention U‐net (WAU‐net) demonstrate superior or similar performance on the segmentation of head and neck CT images.Conclusions: We developed a deep learning approach that integrates the merits of CNNs and the self‐attention mechanism. The proposed WAU‐net can efficiently capture local and global dependencies and achieves state‐of‐the‐art performance on the head and neck multi‐organ segmentation task.
- 15 October 2021
- Published by Wiley in Medical Physics
- Vol. 48 (11) , 7052-7062
Open AccessWeaving attention U‐net: A novel hybrid CNN and attention‐based method for organs‐at‐risk segmentation in head and neck CT images
Purpose: In radiotherapy planning, manual contouring is labor‐intensive and time‐consuming. Accurate and robust automated segmentation models improve the efficiency and treatment outcome. We aim to develop a novel hybrid deep learning approach, combining convolutional neural networks (CNNs) and the self‐attention mechanism, for rapid and accurate multi‐organ segmentation on head and neck computed tomography (CT) images.Methods: Head and neck CT images with manual contours of 115 patients were retrospectively collected and used. We set the training/validation/testing ratio to 81/9/25 and used the 10‐fold cross‐validation strategy to select the best model parameters. The proposed hybrid model segmented 10 organs‐at‐risk (OARs) altogether for each case. The performance of the model was evaluated by three metrics, that is, the Dice Similarity Coefficient (DSC), Hausdorff distance 95% (HD95), and mean surface distance (MSD). We also tested the performance of the model on the head and neck 2015 challenge dataset and compared it against several state‐of‐the‐art automated segmentation algorithms.Results: The proposed method generated contours that closely resemble the ground truth for 10 OARs. On the head and neck 2015 challenge dataset, the DSC scores of these OARs were 0.910.02, 0.730.10, 0.950.03, 0.760.08, 0.790.05, 0.870.05, 0.860.08, 0.870.03, and 0.870.07 for brain stem, chiasm, mandible, left/right optic nerve, left/right submandibular, and left/right parotid, respectively. Our results of the new weaving attention U‐net (WAU‐net) demonstrate superior or similar performance on the segmentation of head and neck CT images.Conclusions: We developed a deep learning approach that integrates the merits of CNNs and the self‐attention mechanism. The proposed WAU‐net can efficiently capture local and global dependencies and achieves state‐of‐the‐art performance on the head and neck multi‐organ segmentation task.
- 5 October 2021
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 7 (6) , 065023
Semi-supervised semantic segmentation of prostate and organs-at-risk on 3D pelvic CT images
The recent development of deep learning approaches has revoluted medical data processing, including semantic segmentation, by dramatically improving performance. Automated segmentation can assist radiotherapy treatment planning by saving manual contouring efforts and reducing intra-observer and inter-observer variations. However, training effective deep learning models usually Requires a large amount of high-quality labeled data, often costly to collect. We developed a novel semi-supervised adversarial deep learning approach for 3D pelvic CT image semantic segmentation. Unlike supervised deep learning methods, the new approach can utilize both annotated and un-annotated data for training. It generates un-annotated synthetic data by a data augmentation scheme using generative adversarial networks (GANs). We applied the new approach to segmenting multiple organs in male pelvic CT images. CT images without annotations and GAN-synthesized un-annotated images were used in semi-supervised learning. Experimental results, evaluated by three metrics (Dice similarity coefficient, average Hausdorff distance, and average surface Hausdorff distance), showed that the new method achieved comparable performance with substantially fewer annotated images or better performance with the same amount of annotated data, outperforming the existing state-of-the-art methods.
- 5 October 2021
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 7 (6) , 065023
Semi-supervised semantic segmentation of prostate and organs-at-risk on 3D pelvic CT images
The recent development of deep learning approaches has revoluted medical data processing, including semantic segmentation, by dramatically improving performance. Automated segmentation can assist radiotherapy treatment planning by saving manual contouring efforts and reducing intra-observer and inter-observer variations. However, training effective deep learning models usually Requires a large amount of high-quality labeled data, often costly to collect. We developed a novel semi-supervised adversarial deep learning approach for 3D pelvic CT image semantic segmentation. Unlike supervised deep learning methods, the new approach can utilize both annotated and un-annotated data for training. It generates un-annotated synthetic data by a data augmentation scheme using generative adversarial networks (GANs). We applied the new approach to segmenting multiple organs in male pelvic CT images. CT images without annotations and GAN-synthesized un-annotated images were used in semi-supervised learning. Experimental results, evaluated by three metrics (Dice similarity coefficient, average Hausdorff distance, and average surface Hausdorff distance), showed that the new method achieved comparable performance with substantially fewer annotated images or better performance with the same amount of annotated data, outperforming the existing state-of-the-art methods.
- 5 November 2020
- Published by Wiley in Medical Physics
- Vol. 48 (1) , 227-237
ARPM‐net: A novel CNN‐based adversarial method with Markov random field enhancement for prostate and organs at risk segmentation in pelvic CT images
Purpose: The research is to develop a novel CNN‐based adversarial deep learning method to improve and expedite the multi‐organ semantic segmentation of CT images and to generate accurate contours on pelvic CT images.Methods: Planning CT and structure datasets for 120 patients with intact prostate cancer were retrospectively selected and divided for tenfold cross‐validation. The proposed adversarial multi‐residual multi‐scale pooling Markov random field (MRF) enhanced network (ARPM‐net) implements an adversarial training scheme. A segmentation network and a discriminator network were trained jointly, and only the segmentation network was used for prediction. The segmentation network integrates a newly designed MRF block into a variation of multi‐residual U‐net. The discriminator takes the product of the original CT and the prediction/ground‐truth as input and classifies the input into fake/real. The segmentation network and discriminator network can be trained jointly as a whole, or the discriminator can be used for fine‐tuning after the segmentation network is coarsely trained. Multi‐scale pooling layers were introduced to preserve spatial resolution during pooling using less memory compared to atrous convolution layers. An adaptive loss function was proposed to enhance the training on small or low contrast organs. The accuracy of modeled contours was measured with the dice similarity coefficient (DSC), average Hausdorff distance (AHD), average surface Hausdorff distance (ASHD), and relative volume difference (VD) using clinical contours as references to the ground‐truth. The proposed ARPM‐net method was compared to several state‐of‐the‐art deep learning methods.Results: ARPM‐net outperformed several existing deep learning approaches and MRF methods and achieved state‐of‐the‐art performance on a testing dataset. On the test set with 20 cases, the average DSC on the prostate, bladder, rectum, left femur, and right femur were 0.88 (0.11), 0.97 (0.07), 0.86 (0.12), 0.97 (0.01), and 0.97 (0.01), respectively. The average HD (mm) on these organs were 1.58 ( 1.77), 1.91 (1.29), 3.14 (2.39), 1.76 (1.57), and 1.92 (1.01). The average surface HD (mm) on these organs are 2.11 (2.03), 2.36 (2.43), 3.05 (2.11), 1.99 (1.66), and 2.00 (2.07).Conclusion: ARPM‐net was designed for the automatic segmentation of pelvic CT images. With adversarial fine‐tuning, ARPM‐net produces state‐of‐the‐art accurate contouring of multiple organs on CT images and has the potential to facilitate routine pelvic cancer radiation therapy planning process.
- 5 November 2020
- Published by Wiley in Medical Physics
- Vol. 48 (1) , 227-237
ARPM‐net: A novel CNN‐based adversarial method with Markov random field enhancement for prostate and organs at risk segmentation in pelvic CT images
Purpose: The research is to develop a novel CNN‐based adversarial deep learning method to improve and expedite the multi‐organ semantic segmentation of CT images and to generate accurate contours on pelvic CT images.Methods: Planning CT and structure datasets for 120 patients with intact prostate cancer were retrospectively selected and divided for tenfold cross‐validation. The proposed adversarial multi‐residual multi‐scale pooling Markov random field (MRF) enhanced network (ARPM‐net) implements an adversarial training scheme. A segmentation network and a discriminator network were trained jointly, and only the segmentation network was used for prediction. The segmentation network integrates a newly designed MRF block into a variation of multi‐residual U‐net. The discriminator takes the product of the original CT and the prediction/ground‐truth as input and classifies the input into fake/real. The segmentation network and discriminator network can be trained jointly as a whole, or the discriminator can be used for fine‐tuning after the segmentation network is coarsely trained. Multi‐scale pooling layers were introduced to preserve spatial resolution during pooling using less memory compared to atrous convolution layers. An adaptive loss function was proposed to enhance the training on small or low contrast organs. The accuracy of modeled contours was measured with the dice similarity coefficient (DSC), average Hausdorff distance (AHD), average surface Hausdorff distance (ASHD), and relative volume difference (VD) using clinical contours as references to the ground‐truth. The proposed ARPM‐net method was compared to several state‐of‐the‐art deep learning methods.Results: ARPM‐net outperformed several existing deep learning approaches and MRF methods and achieved state‐of‐the‐art performance on a testing dataset. On the test set with 20 cases, the average DSC on the prostate, bladder, rectum, left femur, and right femur were 0.88 (0.11), 0.97 (0.07), 0.86 (0.12), 0.97 (0.01), and 0.97 (0.01), respectively. The average HD (mm) on these organs were 1.58 ( 1.77), 1.91 (1.29), 3.14 (2.39), 1.76 (1.57), and 1.92 (1.01). The average surface HD (mm) on these organs are 2.11 (2.03), 2.36 (2.43), 3.05 (2.11), 1.99 (1.66), and 2.00 (2.07).Conclusion: ARPM‐net was designed for the automatic segmentation of pelvic CT images. With adversarial fine‐tuning, ARPM‐net produces state‐of‐the‐art accurate contouring of multiple organs on CT images and has the potential to facilitate routine pelvic cancer radiation therapy planning process.
- 22 September 2020
- Published by IOP Publishing in Journal of Radiological Protection
- Vol. 40 (4) , 980-996
A Monte Carlo based analytic model of the in-room neutron ambient dose equivalent for a Mevion gantry-mounted passively scattered proton system
The goal of this study was to develop a Monte Carlo (MC)-based analytical model that can predict the in-room ambient dose equivalent from a Mevion gantry-mounted passively scattered proton system. The Mevion S250 and treatment vault were simulated using the MCNPX MC code. The results of the in-room neutron dose measurements, using an FHT 762 WENDI-II detector, were employed to benchmark the MC-derived values. After tuning the MCNPX MC code, for the same beam delivery parameters, the code was used to calculate the neutron spectra and ambient dose equivalent in the vault and at varying angles from the isocenter. Then, based on the calculations, an analytical model was reconstructed and data were fitted to derive the model parameters at 95% confidence intervals (CI). The MCNPX codes were tuned to within about 19% of the measured values for most of the measurements in the vault. For the maze, up to 0.08 mSv Gy −1 discrepancies were found between the experimental measurements and MCNPX calculated results. The analytical model showed up to 18% discrepancy for distances between 100 and 600 cm from the isocenter compared to the MC calculations. The model may underestimate the neutron ambient dose equivalent up to 21% for distances less than 100 cm from the isocenter. The proposed analytical model can be used to estimate the contribution of the secondary neutron dose from the Mevion S250 for the design of local shielding inside the proton therapy treatment vault.
- 22 September 2020
- Published by IOP Publishing in Journal of Radiological Protection
- Vol. 40 (4) , 980-996
A Monte Carlo based analytic model of the in-room neutron ambient dose equivalent for a Mevion gantry-mounted passively scattered proton system
The goal of this study was to develop a Monte Carlo (MC)-based analytical model that can predict the in-room ambient dose equivalent from a Mevion gantry-mounted passively scattered proton system. The Mevion S250 and treatment vault were simulated using the MCNPX MC code. The results of the in-room neutron dose measurements, using an FHT 762 WENDI-II detector, were employed to benchmark the MC-derived values. After tuning the MCNPX MC code, for the same beam delivery parameters, the code was used to calculate the neutron spectra and ambient dose equivalent in the vault and at varying angles from the isocenter. Then, based on the calculations, an analytical model was reconstructed and data were fitted to derive the model parameters at 95% confidence intervals (CI). The MCNPX codes were tuned to within about 19% of the measured values for most of the measurements in the vault. For the maze, up to 0.08 mSv Gy −1 discrepancies were found between the experimental measurements and MCNPX calculated results. The analytical model showed up to 18% discrepancy for distances between 100 and 600 cm from the isocenter compared to the MC calculations. The model may underestimate the neutron ambient dose equivalent up to 21% for distances less than 100 cm from the isocenter. The proposed analytical model can be used to estimate the contribution of the secondary neutron dose from the Mevion S250 for the design of local shielding inside the proton therapy treatment vault.
- Open Access
ARPM-net: A novel CNN-based adversarial method with Markov Random Field enhancement for prostate and organs at risk segmentation in pelvic CT images
Purpose: The research is to develop a novel CNN-based adversarial deep learning method to improve and expedite the multi-organ semantic segmentation of CT images, and to generate accurate contours on pelvic CT images. Methods: Planning CT and structure datasets for 120 patients with intact prostate cancer were retrospectively selected and divided for 10-fold cross-validation. The proposed adversarial multi-residual multi-scale pooling Markov Random Field (MRF) enhanced network (ARPM-net) implements an adversarial training scheme. A segmentation network and a discriminator network were trained jointly, and only the segmentation network was used for prediction. The segmentation network integrates a newly designed MRF block into a variation of multi-residual U-net. The discriminator takes the product of the original CT and the prediction/ground-truth as input and classifies the input into fake/real. The segmentation network and discriminator network can be trained jointly as a whole, or the discriminator can be used for fine-tuning after the segmentation network is coarsely trained. Multi-scale pooling layers were introduced to preserve spatial resolution during pooling using less memory compared to atrous convolution layers. An adaptive loss function was proposed to enhance the training on small or low contrast organs. The accuracy of modeled contours was measured with the Dice similarity coefficient (DSC), Average Hausdorff Distance (AHD), Average Surface Hausdorff Distance (ASHD), and relative Volume Difference (VD) using clinical contours as references to the ground-truth. The proposed ARPM-net method was compared to several stateof-the-art deep learning methods.
Open AccessARPM-net: A novel CNN-based adversarial method with Markov Random Field enhancement for prostate and organs at risk segmentation in pelvic CT images
Purpose: The research is to develop a novel CNN-based adversarial deep learning method to improve and expedite the multi-organ semantic segmentation of CT images, and to generate accurate contours on pelvic CT images. Methods: Planning CT and structure datasets for 120 patients with intact prostate cancer were retrospectively selected and divided for 10-fold cross-validation. The proposed adversarial multi-residual multi-scale pooling Markov Random Field (MRF) enhanced network (ARPM-net) implements an adversarial training scheme. A segmentation network and a discriminator network were trained jointly, and only the segmentation network was used for prediction. The segmentation network integrates a newly designed MRF block into a variation of multi-residual U-net. The discriminator takes the product of the original CT and the prediction/ground-truth as input and classifies the input into fake/real. The segmentation network and discriminator network can be trained jointly as a whole, or the discriminator can be used for fine-tuning after the segmentation network is coarsely trained. Multi-scale pooling layers were introduced to preserve spatial resolution during pooling using less memory compared to atrous convolution layers. An adaptive loss function was proposed to enhance the training on small or low contrast organs. The accuracy of modeled contours was measured with the Dice similarity coefficient (DSC), Average Hausdorff Distance (AHD), Average Surface Hausdorff Distance (ASHD), and relative Volume Difference (VD) using clinical contours as references to the ground-truth. The proposed ARPM-net method was compared to several stateof-the-art deep learning methods.
- 1 January 2020
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 6 (1) , 015025
Long-term Inter-protocol kV CBCT image quality assessment for a ring-gantry linac via automated QA approach
We develop a fully automated QA process to compare the image quality of all kV CBCT protocols on a Halcyon linac with ring gantry design, and evaluate image quality stability over a 10-month period. A total of 19 imaging scan and reconstruction protocols were characterized with measurement on a newly released QUART phantom. A set of image analysis algorithms were developed and integrated into an automated analysis suite to derive key image quality metrics, including HU value accuracy on density inserts, HU uniformity using the background plate, high contrast resolution with the modulation transfer function (MTF) from the edge profiles, low contrast resolution using the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR), slice thickness with the air gap modules, and geometric accuracy with the diameter of the phantom. Image quality data over 10 months was tracked and analyzed to evaluate the stability of the Halcyon kV imaging system. The HU accuracy over all 19 protocols is within tolerance (±50HU). The maximum uniformity deviation is 12.2 HU. The SNR and CNR, depending on the protocol selected, range from 18.5–911.9 and 1.9–102.8, respectively. A much-improved SNR and CNR were observed for iterative reconstruction (iCBCT) modes and protocols designed for large subjects over low dose and fast scanning modes. The Head and Image Gently protocols have the greatest high contrast resolution with MTF 10% over 1 lp/mm and MTF 50% over 0.6 lp/mm. The iCBCT mode slightly improved the MTF 10% and MTF 50% compared to the Feldkamp-Davis-Kress approach. The slice thickness (maximum error of 0.31 mm) and geometry metrics (maximum error of 0.7 mm) are all within tolerance (±0.5 mm for slice thickness and ±1 mm for geometry metrics). The long-term study over 10-month showed no significant drift for all key image quality metrics, which indicated the kV CBCT image quality is stable over time.
- 1 January 2020
- Published by IOP Publishing in Biomedical Physics & Engineering Express
- Vol. 6 (1) , 015025
Long-term Inter-protocol kV CBCT image quality assessment for a ring-gantry linac via automated QA approach
We develop a fully automated QA process to compare the image quality of all kV CBCT protocols on a Halcyon linac with ring gantry design, and evaluate image quality stability over a 10-month period. A total of 19 imaging scan and reconstruction protocols were characterized with measurement on a newly released QUART phantom. A set of image analysis algorithms were developed and integrated into an automated analysis suite to derive key image quality metrics, including HU value accuracy on density inserts, HU uniformity using the background plate, high contrast resolution with the modulation transfer function (MTF) from the edge profiles, low contrast resolution using the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR), slice thickness with the air gap modules, and geometric accuracy with the diameter of the phantom. Image quality data over 10 months was tracked and analyzed to evaluate the stability of the Halcyon kV imaging system. The HU accuracy over all 19 protocols is within tolerance (±50HU). The maximum uniformity deviation is 12.2 HU. The SNR and CNR, depending on the protocol selected, range from 18.5–911.9 and 1.9–102.8, respectively. A much-improved SNR and CNR were observed for iterative reconstruction (iCBCT) modes and protocols designed for large subjects over low dose and fast scanning modes. The Head and Image Gently protocols have the greatest high contrast resolution with MTF 10% over 1 lp/mm and MTF 50% over 0.6 lp/mm. The iCBCT mode slightly improved the MTF 10% and MTF 50% compared to the Feldkamp-Davis-Kress approach. The slice thickness (maximum error of 0.31 mm) and geometry metrics (maximum error of 0.7 mm) are all within tolerance (±0.5 mm for slice thickness and ±1 mm for geometry metrics). The long-term study over 10-month showed no significant drift for all key image quality metrics, which indicated the kV CBCT image quality is stable over time.
- 6 August 2019
- Published by Wiley in Medical Physics
- Vol. 46 (10) , 4666-4675
Predicting gamma passing rates for portal dosimetry‐based IMRT QA using machine learning
Purpose: Intensity‐modulated radiation therapy (IMRT) quality assurance (QA) measurements are routinely performed prior to treatment delivery to verify dose calculation and delivery accuracy. In this work, we applied a machine learning‐based approach to predict portal dosimetry based IMRT QA gamma passing rates.Methods: 182 IMRT plans for various treatment sites were planned and delivered with portal dosimetry on two TrueBeam and two Trilogy LINACs. A total of 1497 beams were collected and analyzed using gamma criteria of 2%/2 mm with a 5% threshold. The datasets for building the machine learning models consisted of 1269 beams. Ten‐fold cross‐validation was utilized to tune the model and prevent “overfitting.” A separate test set with the remaining 228 beams was used to evaluate model performance. Each beam was characterized by a set of 31 features including both plan complexity metrics and machine characteristics. Three tree‐based machine learning algorithms (AdaBoost, Random Forest, and XGBoost) were used to train the models and predict gamma passing rates.Results: Both AdaBoost and Random Forest had 98% of predictions within 3% of the measured 2%/2 mm gamma passing rates with a maximum error less than 4% and a mean absolute error < 1%. XGBoost showed a slightly worse prediction accuracy with 95% of the predictions within 3% of the measured gamma passing rates and a maximum error of 4.5%. The three models identified the same nine features in the top 10 most important ones that are related to plan complexity and maximum aperture displacement from the central axis or the maximum jaw size in a beam.Conclusion: We have demonstrated that portal dosimetry IMRT QA gamma passing rates can be accurately predicted using tree‐based ensemble learning models. The machine learning based approach allows physicists to better identify the failures of IMRT QA measurements and to develop proactive QA approaches.
- 6 August 2019
- Published by Wiley in Medical Physics
- Vol. 46 (10) , 4666-4675
Predicting gamma passing rates for portal dosimetry‐based IMRT QA using machine learning
Purpose: Intensity‐modulated radiation therapy (IMRT) quality assurance (QA) measurements are routinely performed prior to treatment delivery to verify dose calculation and delivery accuracy. In this work, we applied a machine learning‐based approach to predict portal dosimetry based IMRT QA gamma passing rates.Methods: 182 IMRT plans for various treatment sites were planned and delivered with portal dosimetry on two TrueBeam and two Trilogy LINACs. A total of 1497 beams were collected and analyzed using gamma criteria of 2%/2 mm with a 5% threshold. The datasets for building the machine learning models consisted of 1269 beams. Ten‐fold cross‐validation was utilized to tune the model and prevent “overfitting.” A separate test set with the remaining 228 beams was used to evaluate model performance. Each beam was characterized by a set of 31 features including both plan complexity metrics and machine characteristics. Three tree‐based machine learning algorithms (AdaBoost, Random Forest, and XGBoost) were used to train the models and predict gamma passing rates.Results: Both AdaBoost and Random Forest had 98% of predictions within 3% of the measured 2%/2 mm gamma passing rates with a maximum error less than 4% and a mean absolute error < 1%. XGBoost showed a slightly worse prediction accuracy with 95% of the predictions within 3% of the measured gamma passing rates and a maximum error of 4.5%. The three models identified the same nine features in the top 10 most important ones that are related to plan complexity and maximum aperture displacement from the central axis or the maximum jaw size in a beam.Conclusion: We have demonstrated that portal dosimetry IMRT QA gamma passing rates can be accurately predicted using tree‐based ensemble learning models. The machine learning based approach allows physicists to better identify the failures of IMRT QA measurements and to develop proactive QA approaches.