The goal of this study is to demonstrate the feasibility of a novel fully-convolutional volumetric dose prediction neural network (DoseNet) and test its performance on a cohort of prostate ...stereotactic body radiotherapy (SBRT) patients. DoseNet is suggested as a superior alternative to U-Net and fully connected distance map-based neural networks for non-coplanar SBRT prostate dose prediction. DoseNet utilizes 3D convolutional downsampling with corresponding 3D deconvolutional upsampling to preserve memory while simultaneously increasing the receptive field of the network. DoseNet was implemented on 2 Nvidia 1080 Ti graphics processing units and utilizes a 3 phase learning protocol to help achieve convergence and improve generalization. DoseNet was trained, validated, and tested with 151 patients following Kaggle completion rules. The dosimetric quality of DoseNet was evaluated by comparing the predicted dose distribution with the clinically approved delivered dose distribution in terms of conformity index, heterogeneity index, and various clinically relevant dosimetric parameters. The results indicate that the DoseNet algorithm is a superior alternative to U-Net and fully connected methods for prostate SBRT patients. DoseNet required ~50.1 h to train, and ~0.83 s to make a prediction on a 128 × 128 × 64 voxel image. In conclusion, DoseNet is capable of making accurate volumetric dose predictions for non-coplanar SBRT prostate patients, while simultaneously preserving computational efficiency.
An adaptive proton therapy workflow using cone beam computed tomography (CBCT) is proposed. It consists of an online evaluation of a fast range-corrected dose distribution based on a virtual CT (vCT) ...scan. This can be followed by more accurate offline dose recalculation on the vCT scan, which can trigger a rescan CT (rCT) for replanning.
The workflow was tested retrospectively for 20 consecutive lung cancer patients. A diffeomorphic Morphon algorithm was used to generate the lung vCT by deforming the average planning CT onto the CBCT scan. An additional correction step was applied to account for anatomic modifications that cannot be modeled by deformation alone. A set of clinical indicators for replanning were generated according to the water equivalent thickness (WET) and dose statistics and compared with those obtained on the rCT scan. The fast dose approximation consisted of warping the initial planned dose onto the vCT scan according to the changes in WET. The potential under- and over-ranges were assessed as a variation in WET at the target's distal surface.
The range-corrected dose from the vCT scan reproduced clinical indicators similar to those of the rCT scan. The workflow performed well under different clinical scenarios, including atelectasis, lung reinflation, and different types of tumor response. Between the vCT and rCT scans, we found a difference in the measured 95% percentile of the over-range distribution of 3.4 ± 2.7 mm. The limitations of the technique consisted of inherent uncertainties in deformable registration and the drawbacks of CBCT imaging. The correction step was adequate when gross errors occurred but could not recover subtle anatomic or density changes in tumors with complex topology.
A proton therapy workflow based on CBCT provided clinical indicators similar to those using rCT for patients with lung cancer with considerable anatomic changes.
The purpose of the work is to develop a deep unsupervised learning strategy for cone-beam CT (CBCT) to CT deformable image registration (DIR). This technique uses a deep convolutional inverse ...graphics network (DCIGN) based DIR algorithm implemented on 2 Nvidia 1080 Ti graphics processing units. The model is comprised of an encoding and decoding stage. The fully-convolutional encoding stage learns hierarchical features and simultaneously forms an information bottleneck, while the decoding stage restores the original dimensionality of the input image. Activations from the encoding stage are used as the input channels to a sparse DIR algorithm. DCIGN was trained using a distributive learning-based convolutional neural network architecture and used 285 head and neck patients to train, validate, and test the algorithm. The accuracy of the DCIGN algorithm was evaluated on 100 synthetic cases and 12 hold out test patient cases. The results indicate that DCIGN performed better than rigid registration, intensity corrected Demons, and landmark-guided deformable image registration for all evaluation metrics. DCIGN required ~14 h to train, and ~3.5 s to make a prediction on a 512 × 512 × 120 voxel image. In conclusion, DCIGN is able to maintain high accuracy in the presence of CBCT noise contamination, while simultaneously preserving high computational efficiency.
Machine learning (ML) has the potential to revolutionize the field of radiation oncology, but there is much work to be done. In this article, we approach the radiotherapy process from a workflow ...perspective, identifying specific areas where a data-centric approach using ML could improve the quality and efficiency of patient care. We highlight areas where ML has already been used, and identify areas where we should invest additional resources. We believe that this article can serve as a guide for both clinicians and researchers to start discussing issues that must be addressed in a timely manner.
Professional societies, which advocated for increased opioid use, received funding from the pharmaceutical industry. 2 In 2012 alone, over 250 million prescriptions were written for opioids. 3,4 From ...1999 to 2011, the use of prescription opioids, oxycontin, oxycodone, and hydrocodone, skyrocketed, as did deaths due to opioid-related overdoses. 5 Policy changes beginning at individual State levels began to reduce the availability of prescription opioids, resulting in a dramatic increase of the price on the illegal market, and a switch to lower-cost alternatives of heroin and more recently, fentanyl. The goal of the Sunshine Act was to increase financial transparency and to reveal potential conflicts of interest. 10 Data collection began on 1 August 1 2013; the process is now supported the Open Payments system run by CMS (https://openpaymentsdata.cms.gov/) and includes a tool for searching the database. In 2019, using data from CMS Open, ProPublica reported that more than 2500 physicians had received at least half a million dollars apiece from the pharmaceutical and medical device industries in the prior 5 years. 11 In the medical device industry, Rachal and Lim reported that in the 5 years since the Sunshine Act came into effect, contributions to doctors from the 20 top-spending medical technology companies more than tripled. 12,13 In our own profession, Marshall et al. have reported that payments to radiation oncologists have actually increased since the Open Payments system was implemented and suggested that increased emphasis on financial conflicts of interest is needed. 14 Lexchin and Fugh-Berman performed an exhaustive review of the literature and concluded that “… there is no evidence that physician behavior regarding conflicts of interest has changed,” and that disclosure is “…not sufficient to address the damage that industry relationships causes to medical knowledge and public health.” Additionally, the perverse incentives created can lead to wasteful health spending, with studies suggesting 25%−30% of overall health care spending could be considered waste. 16 And as we witnessed from the opioid crisis, perverse incentives can have profound public health consequences.
Deep learning algorithms have recently been developed that utilize patient anatomy and raw imaging information to predict radiation dose, as a means to increase treatment planning efficiency and ...improve radiotherapy plan quality. Current state-of-the-art techniques rely on convolutional neural networks (CNNs) that use pixel-to-pixel loss to update network parameters. However, stereotactic body radiotherapy (SBRT) dose is often heterogeneous, making it difficult to model using pixel-level loss. Generative adversarial networks (GANs) utilize adversarial learning that incorporates image-level loss and is better suited to learn from heterogeneous labels. However, GANs are difficult to train and rely on compromised architectures to facilitate convergence. This study suggests an attention-gated generative adversarial network (DoseGAN) to improve learning, increase model complexity, and reduce network redundancy by focusing on relevant anatomy. DoseGAN was compared to alternative state-of-the-art dose prediction algorithms using heterogeneity index, conformity index, and various dosimetric parameters. All algorithms were trained, validated, and tested using 141 prostate SBRT patients. DoseGAN was able to predict more realistic volumetric dosimetry compared to all other algorithms and achieved statistically significant improvement compared to all alternative algorithms for the V
and V
of the PTV, V
of the rectum, and heterogeneity index.
To report the first clinical results and value assessment of prompt gamma imaging for in vivo proton range verification in pencil beam scanning mode.
A stand-alone, trolley-mounted, prototype prompt ...gamma camera utilizing a knife-edge slit collimator design was used to record the prompt gamma signal emitted along the proton tracks during delivery of proton therapy for a brain cancer patient. The recorded prompt gamma depth detection profiles of individual pencil beam spots were compared with the expected profiles simulated from the treatment plan.
In 6 treatment fractions recorded over 3 weeks, the mean (± standard deviation) range shifts aggregated over all spots in 9 energy layers were -0.8 ± 1.3 mm for the lateral field, 1.7 ± 0.7 mm for the right-superior-oblique field, and -0.4 ± 0.9 mm for the vertex field.
This study demonstrates the feasibility and illustrates the distinctive benefits of prompt gamma imaging in pencil beam scanning treatment mode. Accuracy in range verification was found in this first clinical case to be better than the range uncertainty margin applied in the treatment plan. These first results lay the foundation for additional work toward tighter integration of the system for in vivo proton range verification and quantification of range uncertainties.
A deeply supervised attention-enabled boosted convolutional neural network (DAB-CNN) is presented as a superior alternative to current state-of-the-art convolutional neural networks (CNNs) for ...semantic CT segmentation. Spatial attention gates (AGs) were incorporated into a novel 3D cascaded CNN framework to prioritize relevant anatomy and suppress redundancies within the network. Due to the complexity and size of the network, incremental channel boosting was used to decrease memory usage and facilitate model convergence. Deep supervision was used to encourage semantically meaningful deep features and mitigate local minima traps during training. The accuracy of DAB-CNN is compared to seven architectures: a variation of U-Net (UNet), attention-enabled U-Net (A-UNet), boosted U-Net (B-UNet), deeply-supervised U-Net (D-UNet), U-Net with ResNeXt blocks (ResNeXt), life-long learning segmentation CNN (LL-CNN), and deeply supervised attention-enabled U-Net (DA-UNet). The accuracy of each method was assessed based on Dice score compared to manually delineated contours as the gold standard. One hundred and twenty patients who had definitive prostate radiotherapy were used in this study. Training, validation, and testing followed Kaggle competition rules, with 80 patients used for training, 20 patients used for internal validation, and 20 test patients used to report final accuracies. Comparator p -values indicate that DAB-CNN achieved significantly superior Dice scores than all alternative algorithms for the prostate, rectum, and penile bulb. This study demonstrated that attention-enabled boosted convolutional neural networks (CNNs) using deep supervision are capable of achieving superior prediction accuracy compared to current state-of-the-art automatic segmentation methods.
PDF
