Search

Deformable image registration (DIR) of fractional high-dose-rate (HDR) CT images is challenging due to the presence of applicators in the brachytherapy image. Point-to-point correspondence fails because of the undesired deformation vector fields (DVF) propagated from the applicator region (AR) to the surrounding tissues, which can potentially introduce significant DIR errors in dose mapping. This paper proposes a novel segmentation and point-matching enhanced efficient DIR (named SPEED) scheme to facilitate dose accumulation among HDR treatment fractions. In SPEED, a semi-automatic seed point generation approach is developed to obtain the incremented fore/background point sets to feed the random walks algorithm, which is used to segment and remove the AR, leaving empty AR cavities in the HDR CT images. A feature-based 'thin-plate-spline robust point matching' algorithm is then employed for AR cavity surface points matching. With the resulting mapping, a DVF defining on each voxel is estimated by B-spline approximation, which serves as the initial DVF for the subsequent Demons-based DIR between the AR-free HDR CT images. The calculated DVF via Demons combined with the initial one serve as the final DVF to map doses between HDR fractions. The segmentation and registration accuracy are quantitatively assessed by nine clinical HDR cases from three gynecological cancer patients. The quantitative analysis and visual inspection of the DIR results indicate that SPEED can suppress the impact of applicator on DIR, and accurately register HDR CT images as well as deform and add interfractional HDR doses.

Deep learning (DL) models for medical image segmentation are highly influenced by intensity variations of input images and lack generalization due to primarily utilizing pixels' intensity information for inference. Acquiring sufficient training data is another challenge limiting models' applications. Here, we proposed to leverage the consistency of organs' anatomical position and shape information in medical images. We introduced a framework leveraging recurring anatomical patterns through global binary masks for organ segmentation. Two scenarios were studied: (1) global binary masks were the only input for the U-Net based model, forcing exclusively encoding organs' position and shape information for rough segmentation or localization. (2) Global binary masks were incorporated as an additional channel providing position/shape clues to mitigate training data scarcity. Two datasets of the brain and heart computed tomography (CT) images with their ground-truth were split into (26:10:10) and (12:3:5) for training, validation, and test respectively. The two scenarios were evaluated using full training split as well as reduced subsets of training data. In scenario (1), training exclusively on global binary masks led to Dice scores of 0.77 ± 0.06 and 0.85 ± 0.04 for the brain and heart structures respectively. Average Euclidian distance of 3.12 ± 1.43 mm and 2.5 ± 0.93 mm were obtained relative to the center of mass of the ground truth for the brain and heart structures respectively. The outcomes indicated encoding a surprising degree of position and shape information through global binary masks. In scenario (2), incorporating global binary masks led to significantly higher accuracy relative to the model trained on only CT images in small subsets of training data; the performance improved by 4.3%-125.3% and 1.3%-48.1% for 1-8 training cases of the brain and heart datasets respectively. The findings imply the advantages of utilizing global binary masks for building models that are robust to image intensity variations as well as an effective approach to boost performance when access to labeled training data is highly limited., (© 2023 Institute of Physics and Engineering in Medicine.)

Accurate and efficient delineation of the clinical target volume (CTV) is of utmost significance in post-operative breast cancer radiotherapy. However, CTV delineation is challenging as the exact extent of microscopic disease encompassed by CTV is not visualizable in radiological images and remains uncertain. We proposed to mimic physicians' contouring practice for CTV segmentation in stereotactic partial breast irradiation (S-PBI) where CTV is derived from tumor bed volume (TBV) via a margin expansion followed by correcting the extensions for anatomical barriers of tumor invasion (e.g. skin, chest wall). We proposed a deep-learning model, where CT images and the corresponding TBV masks formed a multi-channel input for a 3D U-Net based architecture. The design guided the model to encode the location-related image features and directed the network to focus on TBV to initiate CTV segmentation. Gradient weighted class activation map (Grad-CAM) visualizations of the model predictions revealed that the extension rules and geometric/anatomical boundaries were learnt during model training to assist the network to limit the expansion to a certain distance from the chest wall and the skin. We retrospectively collected 175 prone CT images from 35 post-operative breast cancer patients who received 5-fraction partial breast irradiation regimen on GammaPod. The 35 patients were randomly split into training (25), validation (5) and test (5) sets. Our model achieved mean (standard deviation) of 0.94 (±0.02), 2.46 (±0.5) mm, and 0.53 (±0.14) mm for Dice similarity coefficient, 95th percentile Hausdorff distance, and average symmetric surface distance respectively on the test set. The results are promising for improving the efficiency and accuracy of CTV delineation during on-line treatment planning procedure., (© 2023 Institute of Physics and Engineering in Medicine.)

Objective : Gliomas are the most common primary brain tumors. Approximately 70% of the glioma patients diagnosed with glioblastoma have an averaged overall survival (OS) of only ∼16 months. Early survival prediction is essential for treatment decision-making in glioma patients. Here we proposed an ensemble learning approach to predict the post-operative OS of glioma patients using only pre-operative MRIs. Approach : Our dataset was from the Medical Image Computing and Computer Assisted Intervention Brain Tumor Segmentation challenge 2020, which consists of multimodal pre-operative MRI scans of 235 glioma patients with survival days recorded. The backbone of our approach was a Siamese network consisting of twinned ResNet-based feature extractors followed by a 3-layer classifier. During training, the feature extractors explored traits of intra and inter-class by minimizing contrastive loss of randomly paired 2D pre-operative MRIs, and the classifier utilized the extracted features to generate labels with cost defined by cross-entropy loss. During testing, the extracted features were also utilized to define distance between the test sample and the reference composed of training data, to generate an additional predictor via K-NN classification. The final label was the ensemble classification from both the Siamese model and the K-NN model. Main results : Our approach classifies the glioma patients into 3 OS classes: long-survivors (>15 months), mid-survivors (between 10 and 15 months) and short-survivors (<10 months). The performance is assessed by the accuracy (ACC) and the area under the curve (AUC) of 3-class classification. The final result achieved an ACC of 65.22% and AUC of 0.81. Significance : Our Siamese network based ensemble learning approach demonstrated promising ability in mining discriminative features with minimal manual processing and generalization requirement. This prediction strategy can be potentially applied to assist timely clinical decision-making., (© 2022 Institute of Physics and Engineering in Medicine.)

Stereotactic radiosurgery (SRS) is now the standard of care for brain metastases (BMs) patients. The SRS treatment planning process requires precise target delineation, which in clinical workflow for patients with multiple (>4) BMs (mBMs) could become a pronounced time bottleneck. Our group has developed an automated BMs segmentation platform to assist in this process. The accuracy of the auto-segmentation, however, is influenced by the presence of false-positive segmentations, mainly caused by the injected contrast during MRI acquisition. To address this problem and further improve the segmentation performance, a deep-learning and radiomics ensemble classifier was developed to reduce the false-positive rate in segmentations. The proposed model consists of a Siamese network and a radiomic-based support vector machine (SVM) classifier. The 2D-based Siamese network contains a pair of parallel feature extractors with shared weights followed by a single classifier. This architecture is designed to identify the inter-class difference. On the other hand, the SVM model takes the radiomic features extracted from 3D segmentation volumes as the input for twofold classification, either a false-positive segmentation or a true BM. Lastly, the outputs from both models create an ensemble to generate the final label. The performance of the proposed model in the segmented mBMs testing dataset reached the accuracy (ACC), sensitivity (SEN), specificity (SPE) and area under the curve of 0.91, 0.96, 0.90 and 0.93, respectively. After integrating the proposed model into the original segmentation platform, the average segmentation false negative rate (FNR) and the false positive over the union (FPoU) were 0.13 and 0.09, respectively, which preserved the initial FNR (0.07) and significantly improved the FPoU (0.55). The proposed method effectively reduced the false-positive rate in the BMs raw segmentations indicating that the integration of the proposed ensemble classifier into the BMs segmentation platform provides a beneficial tool for mBMs SRS management., (© 2022 Institute of Physics and Engineering in Medicine.)

Efficient, reliable and reproducible target volume delineation is a key step in the effective planning of breast radiotherapy. However, post-operative breast target delineation is challenging as the contrast between the tumor bed volume (TBV) and normal breast tissue is relatively low in CT images. In this study, we propose to mimic the marker-guidance procedure in manual target delineation. We developed a saliency-based deep learning segmentation (SDL-Seg) algorithm for accurate TBV segmentation in post-operative breast irradiation. The SDL-Seg algorithm incorporates saliency information in the form of markers' location cues into a U-Net model. The design forces the model to encode the location-related features, which underscores regions with high saliency levels and suppresses low saliency regions. The saliency maps were generated by identifying markers on CT images. Markers' location were then converted to probability maps using a distance transformation coupled with a Gaussian filter. Subsequently, the CT images and the corresponding saliency maps formed a multi-channel input for the SDL-Seg network. Our in-house dataset was comprised of 145 prone CT images from 29 post-operative breast cancer patients, who received 5-fraction partial breast irradiation (PBI) regimen on GammaPod. The 29 patients were randomly split into training (19), validation (5) and test (5) sets. The performance of the proposed method was compared against basic U-Net. Our model achieved mean (standard deviation) of 76.4(±2.7) %, 6.76(±1.83) mm, and 1.9(±0.66) mm for Dice similarity coefficient, 95 percentile Hausdorff distance, and average symmetric surface distance respectively on the test set with computation time of below 11 seconds per one CT volume. SDL-Seg showed superior performance relative to basic U-Net for all the evaluation metrics while preserving low computation cost. The findings demonstrate that SDL-Seg is a promising approach for improving the efficiency and accuracy of the on-line treatment planning procedure of PBI, such as GammaPod based PBI., (© 2021 Institute of Physics and Engineering in Medicine.)

We developed a fluence map prediction method that directly generates fluence maps for a given desired dose distribution without optimization for volumetric modulated arc therapy (VMAT) planning. The prediction consists of two steps. First, projections of the desired dose are calculated and then inversely mapped to fluence maps in the phantom geometry by a deep neural network. Second, a plan scaling technique is applied to scale fluence maps from phantom to patient geometry. We evaluated the performance of the proposed fluence map prediction method for 102 head and neck (H&N) and 14 prostate cancer VMAT plans by comparing the patient doses calculated from the predicted fluence maps with the given desired dose distributions. The mean dose differences were 1.42% ± 0.37%, 1.53% ± 0.44% and 1.25% ± 0.44% for the planning target volume (PTV), the region from the PTV boundary to the 50% isodose line, and the region from the 50% to the 20% isodose line, respectively. The gamma passing rate was 98.06% ± 2.64% with the 3 mm/3% criterion. The prediction time for a single VMAT plan was less than one second. In conclusion, we developed an inverse mapping-based method that predicts fluence maps for desired dose distributions with high accuracy. Our method is effectively an optimization-free inverse planning approach, which was orders of magnitude faster than fluence map optimization. Combining the proposed method with leaf sequencing has the potential to dramatically speed up VMAT treatment planning., (© 2020 Institute of Physics and Engineering in Medicine.)

Deep learning (DL)-based auto-segmentation has the potential for accurate organ delineation in radiotherapy applications but requires large amounts of clean labeled data to train a robust model. However, annotating medical images is extremely time-consuming and requires clinical expertise, especially for segmentation that demands voxel-wise labels. On the other hand, medical images without annotations are abundant and highly accessible. To alleviate the influence of the limited number of clean labels, we propose a weakly supervised DL training approach using deformable image registration (DIR)-based annotations, leveraging the abundance of unlabeled data. We generate pseudo-contours by utilizing DIR to propagate atlas contours onto abundant unlabeled images and train a robust DL-based segmentation model. With 10 labeled TCIA dataset and 50 unlabeled CT scans from our institution, our model achieved Dice similarity coefficient of 87.9%, 73.4%, 73.4%, 63.2% and 61.0% on mandible, left & right parotid glands and left & right submandibular glands of TCIA test set and competitive performance on our institutional clinical dataset and a third party (PDDCA) dataset. Experimental results demonstrated the proposed method outperformed traditional multi-atlas DIR methods and fully supervised limited data training and is promising for DL-based medical image segmentation application with limited annotated data., (© 2020 Institute of Physics and Engineering in Medicine.)

Partly due to the use of exhaustive-annotated data, deep networks have achieved impressive performance on medical image segmentation. Medical imaging data paired with noisy annotation are, however, ubiquitous, but little is known about the effect of noisy annotation on deep learning based medical image segmentation. We studied the effect of noisy annotation in the context of mandible segmentation from CT images. First, 202 images of head and neck cancer patients were collected from our clinical database, where the organs-at-risk were annotated by one of twelve planning dosimetrists. The mandibles were roughly annotated as the planning avoiding structure. Then, mandible labels were checked and corrected by a head and neck specialist to get the reference standard. At last, by varying the ratios of noisy labels in the training set, deep networks were trained and tested for mandible segmentation. The trained models were further tested on other two public datasets. Experimental results indicated that the network trained with noisy labels had worse segmentation than that trained with reference standard, and in general, fewer noisy labels led to better performance. When using 20% or less noisy cases for training, no significant difference was found on the segmentation results between the models trained by noisy or reference annotation. Cross-dataset validation results verified that the models trained with noisy data achieved competitive performance to that trained with reference standard. This study suggests that the involved network is robust to noisy annotation to some extent in mandible segmentation from CT images. It also highlights the importance of labeling quality in deep learning. In the future work, extra attention should be paid to how to utilize a small number of reference standard samples to improve the performance of deep learning with noisy annotation.

We propose a novel BIRADS-SSDL network that integrates clinically-approved breast lesion characteristics (BIRADS features) into task-oriented semi-supervised deep learning (SSDL) for accurate diagnosis of ultrasound (US) images with a small training dataset. Breast US images are converted to BIRADS-oriented feature maps (BFMs) using a distance-transformation coupled with a Gaussian filter. Then, the converted BFMs are used as the input of an SSDL network, which performs unsupervised stacked convolutional auto-encoder (SCAE) image reconstruction guided by lesion classification. This integrated multi-task learning allows SCAE to extract image features with the constraints from the lesion classification task, while the lesion classification is achieved by utilizing the SCAE encoder features with a convolutional network. We trained the BIRADS-SSDL network with an alternative learning strategy by balancing the reconstruction error and classification label prediction error. To demonstrate the effectiveness of our approach, we evaluated it using two breast US image datasets. We compared the performance of the BIRADS-SSDL network with conventional SCAE and SSDL methods that use the original images as inputs, as well as with an SCAE that use BFMs as inputs. The experimental results on two breast US datasets show that BIRADS-SSDL ranked the best among the four networks, with a classification accuracy of around 94.23 ± 3.33% and 84.38 ± 3.11% on two datasets. In the case of experiments across two datasets collected from two different institutions/and US devices, the developed BIRADS-SSDL is generalizable across the different US devices and institutions without overfitting to a single dataset and achieved satisfactory results. Furthermore, we investigate the performance of the proposed method by varying the model training strategies, lesion boundary accuracy, and Gaussian filter parameters. The experimental results showed that a pre-training strategy can help to speed up model convergence during training but with no improvement of the classification accuracy on the testing dataset. The classification accuracy decreases as the segmentation accuracy decreases. The proposed BIRADS-SSDL achieves the best results among the compared methods in each case and has the capacity to deal with multiple different datasets under one model. Compared with state-of-the-art methods, BIRADS-SSDL could be promising for effective breast US computer-aided diagnosis using small datasets.

Throughout the course of delivering a radiation therapy treatment, which may take several weeks, a patient's anatomy may change drastically, and adaptive radiation therapy (ART) may be needed. Cone-beam computed tomography (CBCT), which is often available during the treatment process, can be used for both patient positioning and ART re-planning. However, due to the prominent amount of noise, artifacts, and inaccurate Hounsfield unit (HU) values, the dose calculation based on CBCT images could be inaccurate for treatment planning. One way to solve this problem is to convert CBCT images to more accurate synthesized CT (sCT) images. In this work, we have developed a cycle-consistent generative adversarial network framework (CycleGAN) to synthesize CT images from CBCT images. This model is capable of image-to-image translation using unpaired CT and CBCT images in an unsupervised learning setting. The sCT images generated from CBCT through this CycleGAN model are visually and quantitatively similar to real CT images with decreased mean absolute error (MAE) from 69.29 HU to 29.85 HU for head-and-neck (H&N) cancer patients. The dose distributions calculated on the sCT by CycleGAN demonstrated a higher accuracy than those on CBCT in a 3D gamma index analysis with increased gamma index pass rate from 86.92% to 96.26% under 1 mm/1% criteria, when using the deformed planning CT image (dpCT) as the reference. We also compared the CycleGAN model with other unsupervised learning methods, including deep convolutional generative adversarial networks (DCGAN) and progressive growing of GANs (PGGAN), and demonstrated that CycleGAN outperformed the other two models. A phantom study has been conducted to compare sCT with dpCT, and the increase of structural similarity index from 0.91 to 0.93 shows that CycleGAN performed better than DIR in terms of preserving anatomical accuracy.

The aim of this work is to develop a novel recursive ensemble OARs segmentation (REOS) framework for accurate organs-at-risk (OARs) automatic segmentation. The REOS recursively segment individual OARs by ensembling images features extracted from an organ localization module and a contour detection module. Both modules are based on a 3D U-Net architecture. The organ localization module is trained for rough segmentation to localize a region of interest (ROI) that encompasses the to-be-delineated OAR, while the contour detection module is trained to segment the OAR within the identified ROI. In this study, the developed REOS framework is applied for brain radiotherapy on segmenting six OARs including the eyes, the brainstem (BS), the optical nerves and the chiasm. Eighty T1-weighted magnetic resonance images (MRI) from 80 brain cancer patients' cases with OARs' gold standard contours were collected for training and testing REOS. On 20 testing cases, the REOS achieve a high segmentation accuracy with Dice similarity coefficient (DSC) mean and standard deviation of 93.9%  ±  1.4%, 94.5%  ±  2.0%, 90.6%  ±  2.7%, on the left and right eyes and the BS, respectively. On small and segmentation-challenging organs, the left and right optical nerves and the chiasm, the REOS achieves DSC of 78.0%  ±  10.5%, 82.2%  ±  5.9% and 71.1%  ±  9.1%. The satisfactory performances demonstrated the effectiveness of the REOS in OARs segmentation.

Deep inspiration breath hold (DIBH) with surface supervising is a common technique for cardiac dose reduction in left breast cancer radiotherapy. Surface supervision accuracy relies on the characteristics of surface region. In this study, a convolutional neural network (CNN) based automatic region-of-interest (ROI) selection method was proposed to select an optimal surface ROI for DIBH surface monitoring. The curvature entropy and the normal of each vertex on the breast cancer patient surface were calculated and formed as representative maps for ROI selection learning. 900 ROIs were randomly extracted from each patient's surface representative map, and the corresponding rigid ROI registration errors (REs) were calculated. The VGG-16 (a 16-layer network structure developed by Visual Geometry Group(VGG) from University of Oxford) pre-trained on a large natural image database ImageNet were fine-tuned using 27 thousand extracted ROIs and the corresponding REs from thirty patients. The RE prediction accuracy of the trained model was validated on additional ten patients. Satisfactory RE predictive accuracies were achieved with the root mean square error (RMSE)/mean absolute error (MAE) smaller than 1 mm/0.7 mm in translations and 0.45°/0.35° in rotations, respectively. The REs of the model selected ROIs on ten testing cases is close to the minimal predicted RE with mean RE differences  <1 mm and  <0.5° for translation and rotation, respectively. The proposed RE predictive model can be utilized for selecting a quasi-optimal ROI in left breast cancer DIBH radiotherapy (DIBH-RT).

Accurate dose delivery in stereotactic partial breast irradiation (S-PBI) is challenging because of the target position uncertainty caused by breast deformation, the target volume changes caused by lumpectomy cavity shrinkage, and the target delineation uncertainty on simulation computed tomography (CT) images caused by poor soft tissue contrast. We have developed a volumetric ultrasound tomography (UST) image guidance system for prone position S-PBI. The system is composed of a novel 3D printed rotation water tank, a patient-specific resin breast immobilization cup, and a 1D array ultrasound transducer. Coronal 2D US images were acquired in 5° increments over a 360° range, and planes were acquired every 2 mm in elevation. A super-compounding technique was used to reconstruct the image volume. The image quality of UST was evaluated with a BB-1 breast phantom and BioZorb surgical marker, and the results revealed that UST offered better soft tissue contrast than CT and similar image quality to MR. In the evaluated plane, the size and location of five embedded objects were measured and compared to MR, which is considered as the ground truth. Objects' diameters and the distances between objects in UST differ by approximately 1 to 2 mm from those in MR, which showed that UST offers the image quality required for S-PBI. In future work we will develop a robotic system that will be ultimately implemented in the clinic.

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.

The purpose of this work is to develop an automated landmark-guided deformable image registration (LDIR) algorithm between the planning CT and daily cone-beam CT (CBCT) with low image quality. This method uses an automated landmark generation algorithm in conjunction with a local small volume gradient matching search engine to map corresponding landmarks between the CBCT and the planning CT. The landmarks act as stabilizing control points in the following Demons deformable image registration. LDIR is implemented on graphics processing units (GPUs) for parallel computation to achieve ultra fast calculation. The accuracy of the LDIR algorithm has been evaluated on a synthetic case in the presence of different noise levels and data of six head and neck cancer patients. The results indicate that LDIR performed better than rigid registration, Demons, and intensity corrected Demons for all similarity metrics used. In conclusion, LDIR achieves high accuracy in the presence of multimodality intensity mismatch and CBCT noise contamination, while simultaneously preserving high computational efficiency.

Due to a limited number of projections at each phase, severe view aliasing artifacts are present in four-dimensional cone beam computed tomography (4D-CBCT) when reconstruction is performed using conventional algorithms. In this work, we aim to obtain high-quality 4D-CBCT of lung cancer patients in radiation therapy by deforming the planning CT. The deformation vector fields (DVF) to deform the planning CT are estimated through matching the forward projection of the deformed prior image and measured on-treatment CBCT projection. The estimation of the DVF is formulated as an unconstrained optimization problem, where the objective function to be minimized is the sum of the squared difference between the forward projection of the deformed planning CT and the measured 4D-CBCT projection. A nonlinear conjugate gradient method is used to solve the DVF. As the number of the variables in the DVF is much greater than the number of measurements, the solution to such a highly ill-posed problem is very sensitive to the initials during the optimization process. To improve the estimation accuracy of DVF, we proposed a new strategy to obtain better initials for the optimization. In this strategy, 4D-CBCT is first reconstructed by total variation minimization. Demons deformable registration is performed to register the planning CT and the 4D-CBCT reconstructed by total variation minimization. The resulted DVF from demons registration is then used as the initial parameters in the optimization process. A 4D nonuniform rotational B-spline-based cardiac-torso (NCAT) phantom and a patient 4D-CBCT are used to evaluate the algorithm. Image quality of 4D-CBCT is substantially improved by using the proposed strategy in both NCAT phantom and patient studies. The proposed method has the potential to improve the temporal resolution of 4D-CBCT. Improved 4D-CBCT can better characterize the motion of lung tumors and will be a valuable tool for image-guided adaptive radiation therapy.

Online adaptive radiation therapy (ART) promises the ability to deliver an optimal treatment in response to daily patient anatomic variation. A major technical barrier for the clinical implementation of online ART is the requirement of rapid image segmentation. Deformable image registration (DIR) has been used as an automated segmentation method to transfer tumor/organ contours from the planning image to daily images. However, the current computational time of DIR is insufficient for online ART. In this work, this issue is addressed by using computer graphics processing units (GPUs). A gray-scale-based DIR algorithm called demons and five of its variants were implemented on GPUs using the compute unified device architecture (CUDA) programming environment. The spatial accuracy of these algorithms was evaluated over five sets of pulmonary 4D CT images with an average size of 256 x 256 x 100 and more than 1100 expert-determined landmark point pairs each. For all the testing scenarios presented in this paper, the GPU-based DIR computation required around 7 to 11 s to yield an average 3D error ranging from 1.5 to 1.8 mm. It is interesting to find out that the original passive force demons algorithms outperform subsequently proposed variants based on the combination of accuracy, efficiency and ease of implementation.

Online adaptive radiation therapy (ART) is an attractive concept that promises the ability to deliver an optimal treatment in response to the inter-fraction variability in patient anatomy. However, it has yet to be realized due to technical limitations. Fast dose deposit coefficient calculation is a critical component of the online planning process that is required for plan optimization of intensity-modulated radiation therapy (IMRT). Computer graphics processing units (GPUs) are well suited to provide the requisite fast performance for the data-parallel nature of dose calculation. In this work, we develop a dose calculation engine based on a finite-size pencil beam (FSPB) algorithm and a GPU parallel computing framework. The developed framework can accommodate any FSPB model. We test our implementation in the case of a water phantom and the case of a prostate cancer patient with varying beamlet and voxel sizes. All testing scenarios achieved speedup ranging from 200 to 400 times when using a NVIDIA Tesla C1060 card in comparison with a 2.27 GHz Intel Xeon CPU. The computational time for calculating dose deposition coefficients for a nine-field prostate IMRT plan with this new framework is less than 1 s. This indicates that the GPU-based FSPB algorithm is well suited for online re-planning for adaptive radiotherapy.