In the sphere of Cone Beam Computed Tomography (CBCT), acquiring X-ray projections from sufficient angles is indispensable for traditional image reconstruction methods to accurately reconstruct 3D anatomical intricacies. However, this acquisition procedure for the linear accelerator-mounted CBCT systems in radiotherapy takes approximately one minute, impeding its use for ultra-fast intra-fractional motion monitoring during treatment delivery. To address this challenge, we introduce the Patient-specific Cycle-domain Geometric-integrated Denoising Diffusion Probabilistic Model (CG-DDPM). This model aims to leverage patient-specific priors from patient’s CT/4DCT images, which are acquired for treatment planning purposes, to reconstruct 3D CBCT from a single-view 2D CBCT projection of any arbitrary angle during treatment, namely single-view reconstructed CBCT (svCBCT). The CG-DDPM framework encompasses a dual DDPM structure: the Projection-DDPM for synthesizing comprehensive full-view projections and the CBCT-DDPM for creating CBCT images. A key innovation is our Cycle-Domain Geometry-Integrated (CDGI) method, incorporating a Cone Beam X-ray Geometric Transformation Module (GTM) to ensure precise, synergistic operation between the dual DDPMs, thereby enhancing reconstruction accuracy and reducing artifacts. Evaluated in a study involving 37 lung cancer patients, the method demonstrated its ability to reconstruct CBCT not only from simulated X-ray projections but also from real-world data. The CG-DDPM significantly outperforms existing V-shape convolutional neural networks (V-nets), Generative Adversarial Networks (GANs), and DDPM methods in terms of reconstruction fidelity and artifact minimization. This was confirmed through extensive voxel-level, structural, visual, and clinical assessments. The capability of CG-DDPM to generate high-quality reconstructed CBCT from a single-view projection at any arbitrary angle using a single model opens the door for ultra-fast, in-treatment volumetric imaging. This is especially beneficial for radiotherapy at motion-associated cancer sites and image-guided interventional procedures.
Objective. High-dose-rate (HDR) brachytherapy is integral to the standard-of-care for locally advanced cervical cancer (LACC). Currently, selection of brachytherapy applicators relies on physician’s clinical experience, which can lead to variability in treatment quality and outcomes. This study presents a deep learning-based decision-support tool for selecting between interstitial Syed applicators and intracavitary tandem & ovoids applicators. Approach. The network architecture consists of six 3D convolutional-pooling-rectified linear unit blocks, followed by a fully connected block. The input to the network includes three channels: a 3D contour mask of clinical target volume (CTV), organs at risk (OAR), and central tandem, and two 3D distance maps of CTV and OAR voxels relative to the tandem’s central axis. The network outputs a probability score, indicating the suitability of Syed applicators. Binary cross-entropy loss combined with L1 regularization was used for network training. Main results. A retrospective study was performed on 184 LACC patients with 422 instances of applicator insertion. The data was divided into three sets: Dataset-1 of 163 patients with 372 insertions for training and hyperparameter tuning, Dataset-2 of 17 patients with 36 insertions and Dataset-3 of four complex cases with 14 insertions for testing. Five-fold cross-validation was performed on Dataset-1, during which hyperparameters were heuristically tuned to optimize classification accuracy across the folds. The highest average accuracy was 92.1 ± 3.8%. Using the hyperparameters that resulted in this highest accuracy, the final model was then trained on the full Dataset-1, and evaluated on the other two independent datasets, achieving 96.0% accuracy, 90.9% sensitivity, and 97.4% specificity. Significance. These results demonstrate the potential of our model as a quality assurance tool in LACC HDR brachytherapy, providing feedback on physicians’ applicator choice and supporting continuous improvement in decision-making. Future work will focus on collecting more data for further validation and extending its application for prospective applicator selection.
Gadolinium-based contrast agents (GBCAs) are commonly employed with T1-weighted (T1w) MRI to enhance lesion visualization but are restricted in patients at risk of nephrogenic systemic fibrosis. In addition, variations in GBCA administration can introduce imaging inconsistencies. This study develops an efficient 3D deep-learning framework to generate T1-contrast enhanced images (T1C) from pre-contrast multiparametric MRI. We propose the 3D latent rectified flow (T1C-RFlow) model for generating high-quality T1C images. First, T1w and T2-FLAIR images are input into a pretrained autoencoder to acquire an efficient latent space representation. A rectified flow diffusion model is then trained in this latent space representation. The T1C-RFlow model was trained on a curated dataset comprised of the Brain Tumor Segmentation (BraTS) 2024 glioma (GLI; 1480 patients), meningioma (MEN; 1141 patients), and metastases (MET; 1475 patients) datasets. Selected patients were split into training (N = 2860), test (N = 614), and validation (N = 612) sets. Model performance was evaluated with the normalized mean squared error (NMSE) and structural similarity index measure (SSIM). Both qualitative and quantitative results demonstrate that the T1C-RFlow model outperforms benchmark 3D models (pix2pix, denoising diffusion probability models (DDPM), Diffusion Transformers (DiT-3D) trained in the same latent space. T1C-RFlow achieved the following metrics - GLI: NMSE 0.044 ± 0.047, SSIM 0.935 ± 0.025; MEN: NMSE 0.046 ± 0.029, SSIM 0.937 ± 0.021; MET: NMSE 0.098 ± 0.088, SSIM 0.905 ± 0.082. In a blinded reader study of 15 patients (5 GLI, 5 MEN, 5 MET), T1C-RFlow received the highest diagnostic-quality scores across all tumor types (3.80 ± 0.45, 3.20 ± 0.45, and 2.60 ± 0.89 on a 5-point Likert scale), significantly outperforming all baseline methods (p < .05). Further studies showed T1C-RFlow to have the best tumor reconstruction performance and significantly faster denoising times (6.9 s volume −1 , 200 steps) than conventional DDPM models in both latent space (37.7 s, 1000 steps) and patch-based in image space (4.3 h volume −1 ). Our proposed method generates synthetic T1C images that closely resemble radiological features of ground truth T1C in much less time than previous diffusion models. Further development may permit a practical method for contrast-agent-free MRI for brain tumors. Code is made available at https://github.com/zacheidex/An-Efficient-3D-Latent-Diffusion-Model-for-T1-contrast-Enhanced-MRI-Generation.
High-dose-rate (HDR) brachytherapy plays a critical role in the treatment of locally advanced cervical cancer but remains highly dependent on manual treatment planning expertise. The objective of this study is to develop a fully automated HDR brachytherapy planning framework that integrates reinforcement learning (RL) and dose-based optimization to generate clinically acceptable treatment plans with improved consistency and efficiency. We propose a hierarchical two-stage autoplanning framework. In the first stage, a deep Q-network (DQN)-based RL agent iteratively selects treatment planning parameters (TPPs), which control the trade-offs between target coverage and organ-at-risk (OAR) sparing. The agent's state representation includes both dose-volume histogram (DVH) metrics and current TPP values, while its reward function incorporates clinical dose objectives and safety constraints, including D90, V150, V200 for targets, and D2cc for all relevant OARs (bladder, rectum, sigmoid, small bowel, and large bowel). In the second stage, a customized Adam-based optimizer computes the corresponding dwell time distribution for the selected TPPs using a clinically informed loss function. The framework was evaluated on a cohort of patients with complex applicator geometries. The proposed framework successfully learned clinically meaningful TPP adjustments across diverse patient anatomies. For the unseen test patients, the RL-based automated planning method achieved an average score of 93.89%, outperforming the clinical plans which averaged 91.86%. These findings are notable given that score improvements were achieved while maintaining full target coverage and reducing CTV hot spots in most cases.
Background/Objectives: GK plan quality is strongly affected by lesion size and shape, and the same evaluation metrics may not be directly comparable across patients with different anatomies. This study proposes a deep learning-based method to predict achievable, clinically acceptable plan quality from patient-specific geometry. Methods: A hierarchically densely connected U-Net (HD-U-Net) was trained at the lesion level to predict 3D dose distributions for the estimation of plan quality metrics, including coverage, selectivity, gradient index (GI), and conformity index at a 50% prescription dose (CI50). To improve the prediction accuracy of plan quality metrics, Dice similarity coefficient losses for the 100% and 50% isodose lines were incorporated with conventional mean squared error (MSE) loss. Results: Ten-fold cross-validation on 463 brain metastases (BMs) from 175 patients showed that our method achieved smaller mean absolute errors across all four metrics than the HD-U-Net baseline trained with MSE loss. Improvements were pronounced in all metrics for small metastases, and were observed primarily in GI and CI50 for medium and large lesions. Paired Wilcoxon signed-rank tests confirmed the statistical significance of these improvements (p < 0.05). Conclusions: The proposed method outperformed the baseline model in capturing overall trends, improving per-lesion accuracy, and enhancing robustness to dataset variability. It can serve as a pre-planning tool to guide planners in constraint setting and priority tuning, a post-planning quality control tool to identify subpar plans that could be substantially improved, and as a foundation for developing deep reinforcement learning-based automated planning of GK treatments for brain metastases.