Better understanding of the dose-toxicity relationship is critical for safe dose escalation to improve local control in late-stage cervical cancer radiotherapy. In this study, we introduced a convolutional neural network (CNN) model to analyze rectum dose distribution and predict rectum toxicity. Forty-two cervical cancer patients treated with combined external beam radiotherapy (EBRT) and brachytherapy (BT) were retrospectively collected, including twelve toxicity patients and thirty non-toxicity patients. We adopted a transfer learning strategy to overcome the limited patient data issue. A 16-layers CNN developed by the visual geometry group (VGG-16) of the University of Oxford was pre-trained on a large-scale natural image database, ImageNet, and fine-tuned with patient rectum surface dose maps (RSDMs), which were accumulated EBRT + BT doses on the unfolded rectum surface. We used the adaptive synthetic sampling approach and the data augmentation method to address the two challenges, data imbalance and data scarcity. The gradient-weighted class activation maps (Grad-CAM) were also generated to highlight the discriminative regions on the RSDM along with the prediction model. We compare different CNN coefficients fine-tuning strategies, and compare the predictive performance using the traditional dose volume parameters, e.g. D 0.1/1/2cc, and the texture features extracted from the RSDM. Satisfactory prediction performance was achieved with the proposed scheme, and we found that the mean Grad-CAM over the toxicity patient group has geometric consistence of distribution with the statistical analysis result, which indicates possible rectum toxicity location. The evaluation results have demonstrated the feasibility of building a CNN-based rectum dose-toxicity prediction model with transfer learning for cervical cancer radiotherapy.
more »
« less
Post-Radiotherapy PET Image Outcome Prediction by Deep Learning Under Biological Model Guidance: A Feasibility Study of Oropharyngeal Cancer Application
Purpose To develop a method of biologically guided deep learning for post-radiation 18 FDG-PET image outcome prediction based on pre-radiation images and radiotherapy dose information. Methods Based on the classic reaction–diffusion mechanism, a novel biological model was proposed using a partial differential equation that incorporates spatial radiation dose distribution as a patient-specific treatment information variable. A 7-layer encoder–decoder-based convolutional neural network (CNN) was designed and trained to learn the proposed biological model. As such, the model could generate post-radiation 18 FDG-PET image outcome predictions with breakdown biological components for enhanced explainability. The proposed method was developed using 64 oropharyngeal patients with paired 18 FDG-PET studies before and after 20-Gy delivery (2 Gy/day fraction) by intensity-modulated radiotherapy (IMRT). In a two-branch deep learning execution, the proposed CNN learns specific terms in the biological model from paired 18 FDG-PET images and spatial dose distribution in one branch, and the biological model generates post-20-Gy 18 FDG-PET image prediction in the other branch. As in 2D execution, 718/233/230 axial slices from 38/13/13 patients were used for training/validation/independent test. The prediction image results in test cases were compared with the ground-truth results quantitatively. Results The proposed method successfully generated post-20-Gy 18 FDG-PET image outcome prediction with breakdown illustrations of biological model components. Standardized uptake value (SUV) mean values in 18 FDG high-uptake regions of predicted images (2.45 ± 0.25) were similar to ground-truth results (2.51 ± 0.33). In 2D-based Gamma analysis, the median/mean Gamma Index (<1) passing rate of test images was 96.5%/92.8% using the 5%/5 mm criterion; such result was improved to 99.9%/99.6% when 10%/10 mm was adopted. Conclusion The developed biologically guided deep learning method achieved post-20-Gy 18 FDG-PET image outcome predictions in good agreement with ground-truth results. With the breakdown biological modeling components, the outcome image predictions could be used in adaptive radiotherapy decision-making to optimize personalized plans for the best outcome in the future.
more »
« less
- Award ID(s):
- 1952339
- NSF-PAR ID:
- 10332064
- Date Published:
- Journal Name:
- Frontiers in Oncology
- Volume:
- 12
- ISSN:
- 2234-943X
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
more » « less
Abstract Purpose Synthetic digital mammogram (SDM) is a 2D image generated from digital breast tomosynthesis (DBT) and used as a substitute for a full‐field digital mammogram (FFDM) to reduce the radiation dose for breast cancer screening. The previous deep learning‐based method used FFDM images as the ground truth, and trained a single neural network to directly generate SDM images with similar appearances (e.g., intensity distribution, textures) to the FFDM images. However, the FFDM image has a different texture pattern from DBT. The difference in texture pattern might make the training of the neural network unstable and result in high‐intensity distortion, which makes it hard to decrease intensity distortion and increase perceptual similarity (e.g., generate similar textures) at the same time. Clinically, radiologists want to have a 2D synthesized image that feels like an FFDM image in vision and preserves local structures such as both mass and microcalcifications (MCs) in DBT because radiologists have been trained on reading FFDM images for a long time, while local structures are important for diagnosis. In this study, we proposed to use a deep convolutional neural network to learn the transformation to generate SDM from DBT.
Method To decrease intensity distortion and increase perceptual similarity, a multi‐scale cascaded network (MSCN) is proposed to generate low‐frequency structures (e.g., intensity distribution) and high‐frequency structures (e.g., textures) separately. The MSCN consist of two cascaded sub‐networks: the first sub‐network is used to predict the low‐frequency part of the FFDM image; the second sub‐network is used to generate a full SDM image with textures similar to the FFDM image based on the prediction of the first sub‐network. The mean‐squared error (MSE) objective function is used to train the first sub‐network, termed low‐frequency network, to generate a low‐frequency SDM image. The gradient‐guided generative adversarial network's objective function is to train the second sub‐network, termed high‐frequency network, to generate a full SDM image with textures similar to the FFDM image.
Results 1646 cases with FFDM and DBT were retrospectively collected from the Hologic Selenia system for training and validation dataset, and 145 cases with masses or MC clusters were independently collected from the Hologic Selenia system for testing dataset. For comparison, the baseline network has the same architecture as the high‐frequency network and directly generates a full SDM image. Compared to the baseline method, the proposed MSCN improves the peak‐to‐noise ratio from 25.3 to 27.9 dB and improves the structural similarity from 0.703 to 0.724, and significantly increases the perceptual similarity.
Conclusions The proposed method can stabilize the training and generate SDM images with lower intensity distortion and higher perceptual similarity.
-
more » « less
Abstract Purpose To determine the optimal dose‐volume constraint for laryngeal sparing using three commonly employed intensity modulated radiation therapy (IMRT) approaches in patients with oropharyngeal cancer treated to the bilateral neck.
Materials and methods Thirty patients with stage II‐IVA oropharynx cancers received definitive radiotherapy with split‐field IMRT (SF‐IMRT) to the bilateral neck between 2008 and 2013. Each case was re‐planned using whole‐field IMRT (WF‐IMRT) and volumetric modulated arc therapy (VMAT) and plan quality metrics and dose to laryngeal structures was evaluated. Two larynx volumes were defined and compared on the current study: the Radiation Therapy Oncology Group (RTOG) larynx as defined per the RTOG 1016 protocol and the MDACC larynx defined as the components of the larynx bounded by the superior and inferior extent of the thyroid cartilage.
Results Target coverage, conformity, and heterogeneity indices were similar in all techniques. The RTOG larynx mean dose was lower with WF‐IMRT than SF‐IMRT (22.1 vs 25.8 Gy;
P < 0.01). The MDACC larynx mean dose was 17.5 Gy ± 5.4 Gy with no differences between the 3 techniques. WF‐IMRT and VMAT plans were associated with lower mean doses to the supraglottic larynx (42.1 vs 41.2 vs 54.8 Gy;P < 0.01) and esophagus (18.1 vs 18.2 vs 36 Gy;P < 0.01).Conclusions Modern whole field techniques can provide effective laryngeal sparing in patients receiving radiotherapy to the bilateral neck for advanced oropharyngeal cancers.
Summary We evaluated laryngeal dose in patients with locally advanced oropharyngeal cancer treated to the bilateral neck using split‐field IMRT (SF‐IMRT), whole‐field IMRT (WF‐IMRT) and volumetric arc therapy (VMAT). All three techniques provided good sparing of laryngeal structures and were able to achieve a mean larynx dose < 33 Gy. There were no significant differences in dose to target structures or non‐laryngeal organs at risk among techniques.
-
more » « less
Abstract Tumors exhibit high molecular, phenotypic, and physiological heterogeneity. In this effort, we employ quantitative magnetic resonance imaging (MRI) data to capture this heterogeneity through imaging-based subregions or “habitats” in a murine model of glioma. We then demonstrate the ability to model and predict the growth of the habitats using coupled ordinary differential equations (ODEs) in the presence and absence of radiotherapy. Female Wistar rats (N = 21) were inoculated intracranially with 106C6 glioma cells, a subset of which received 20 Gy (N = 5) or 40 Gy (N = 8) of radiation. All rats underwent diffusion-weighted and dynamic contrast-enhanced MRI at up to seven time points. All MRI data at each visit were subsequently clustered using
k -means to identify physiological tumor habitats. A family of four models consisting of three coupled ODEs were developed and calibrated to the habitat time series of control and treated rats and evaluated for predictive capability. The Akaike Information Criterion was used for model selection, and the normalized sum-of-square-error (SSE) was used to evaluate goodness-of-fit in model calibration and prediction. Three tumor habitats with significantly different imaging data characteristics (p < 0.05) were identified: high-vascularity high-cellularity, low-vascularity high-cellularity, and low-vascularity low-cellularity. Model selection resulted in a five-parameter model whose predictions of habitat dynamics yielded SSEs that were similar to the SSEs from the calibrated model. It is thus feasible to mathematically describe habitat dynamics in a preclinical model of glioma using biology-based ODEs, showing promise for forecasting heterogeneous tumor behavior. -
more » « less
Purpose The combination of nonhuman primates (NHPs) with the state‐of‐the‐art molecular imaging technologies allows for within‐subject longitudinal research aiming at gaining new insights into human normal and disease conditions and provides an ideal foundation for future translational studies of new diagnostic tools, medical interventions, and therapies. However, radiation dose estimations for nonhuman primates from molecular imaging probes are lacking and are difficult to perform experimentally. The aim of this work is to construct age‐dependent NHP computational model series to estimate the absorbed dose to NHP specimens in common molecular imaging procedures.
Materials and methods A series of NHP models from baby to adult were constructed based on nonuniform rational B‐spline surface (NURBS) representations. Particle transport was simulated using Monte Carlo calculations to estimate S‐values from nine positron‐emitting radionuclides and absorbed doses from PET radiotracers.
Results Realistic age‐dependent NHP computational model series were developed. For most source‐target pairs in computational NHP models, differences between C‐11 S‐values were between −13.4% and −8.8%/kg difference in body weight while differences between F‐18 S‐values were between −12.9% and −8.0%/kg difference in body weight. The absorbed doses of11C‐labeled brain receptor substances,18F‐labeled brain receptor substances, and18F‐FDG in the brain ranged within 0.047–0.32 mGy/MBq, 0.25–1.63 mGy/MBq, and 0.32–2.12 mGy/MBq, respectively.
Conclusion The absorbed doses to organs are significantly higher in the baby NHP model than in the adult model. These results can be used in translational longitudinal studies to estimate the cumulated absorbed organ doses in NHPs at various ages.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.3389/fonc.2022.895544
