Development of a deep learning‐based patient‐specific target contour prediction model for markerless tumor positioning

Objective. Digitally reconstructed radiography (DRR) plays an important role in the registration of intraoperative x-ray and preoperative CT images. However, existing DRR algorithms often neglect the critical isocentric fixed angle irradiation (IFAI) principle in C-arm imaging, resulting in inaccurate simulation of x-ray images. This limitation degrades registration algorithms relying on DRR image libraries or employing DRR images (DRRs) to train neural network models. To address this issue, we propose a novel IFAI-based DRR method that accurately captures the true projection transformation during x-ray imaging of the human body. Approach. By strictly adhering to the IFAI principle and utilizing known parameters from intraoperative x-ray images paired with CT scans, our method successfully simulates the real projection transformation and generates DRRs that closely resemble actual x-ray images. Main result. Experimental results validate the effectiveness of our IFAI-based DRR method by successfully registering intraoperative x-ray images with preoperative CT images from multiple patients who underwent thoracic endovascular aortic procedures. Significance. The proposed IFAI-based DRR method enhances the quality of DRR images, significantly accelerates the construction of DRR image libraries, and thereby improves the performance of x-ray and CT image registration. Additionally, the method has the generality of registering CT and x-ray images generated by large C-arm devices.

BackgroundBy integrating Digitally Reconstructed Radiograph (DRR) images of pulmonary tumors with Electronic Portal Imaging Device (EPID) images to assist in target segmentation, and subsequently comparing morphological changes in segmented targets across different radiotherapy stages, this approach enables precise quantification of dynamic variations in target volume and shape. This methodological integration provides objective evidence for treatment response evaluation and dynamic optimization of treatment plans, thereby significantly enhancing the precision of radiotherapy delivery.MethodsThe proposed multimodal segmentation framework, named EPIDSeg-Net, comprises an encoder, a multi-scale feature layer, and a decoder. The encoder utilizes a dual-branch architecture: a CNN branch for extracting local texture features and a Swin-Transformer branch for capturing global semantic features. The model first calibrates multimodal input features through a Dual Attention Mechanism (DAM) to adaptively adjust modality-specific weights, thereby enhancing tolerance to missing image information in multi-sequence segmentation. Subsequently, two key modules are implemented within the multi-scale feature layer: a Large-Kernel Grouped Attention Gating (LKG-Gate) module to strengthen local contextual awareness, and a Multi-Path Feature Extraction (MPFE) module to improve feature robustness via a parallel structure. These designs enable the model to effectively focus on lung tumor target regions, optimize segmentation accuracy, and achieve high-performance reconstruction.ResultsThe framework effectively integrates multimodal features, enabling high-precision localization and sharp boundary delineation while preserving anatomical details. Quantitative evaluations demonstrate superior performance: DICE = 93.2 (92.4∼93.9), CE = 0.352, HD95 = 9.42 (6.03∼12.8), IOU = 86.0 (84.1∼87.9), and SENCE = 0.828. Overall, the model excels at preserving gradient information, regional integrity, and fine details; effectively suppresses feature loss; and reduces missed segmentation rates, leading to improvements in both subjective and objective performance metrics.ConclusionThe proposed segmentation method effectively integrates information from EPID and DRR images, enabling more precise localization and segmentation of lesion regions within EPID images while enhancing segmentation accuracy.

Respiratory gating based on external surrogates is performed in many clinics. We have developed a new technique for treatment verification using an electronic portal imaging device (EPID) in cine mode for gated 3D conformal therapy. Implanted radiopaque fiducial markers inside or near the target are required for this technique. The markers are contoured on the planning CT set, enabling us to create digitally reconstructed radiographs (DRRs) for each treatment beam. During the treatment, a sequence of EPID images can be acquired without disrupting the treatment. Implanted markers are visualized in the images and their positions in the beam's eye view are calculated off-line and compared to the reference position by matching the field apertures in corresponding EPID and DRR images. The precision of the patient set-up, the placement of the beam-gating window, as well as the residual tumour motion can be assessed for each treatment fraction. This technique has been demonstrated with a case study patient, who had three markers implanted in his liver. For this patient, the intra-fractional variation of all marker positions in the gating window had a 95% range of 4.8 mm in the SI direction (the primary axis of motion). This was about the same (5 mm) as the residual motion considered in the planning process. The inter-fractional variation of the daily mean positions of the markers, which indicates the uncertainty in the set-up procedure, was within +8.3 mm/−4.5 mm (95% range) in the SI direction for this case.

Purpose: 4D cone beam computed tomography (4DCBCT) suffers from poor image quality, which limits its use as a reference image for image‐guided procedures that use radiography and fluoroscopy. This abstract demonstrates improved image quality in Digitally Reconstructed Radiographs (DRR) derived from thoracic 4DCBCT. Methods: In our method, projection images are used to create both a phase‐rebinned 4D volume and streak‐free 3D volume. These volumes are combined into a new 4DCBCT that includes stationary voxels from 3D volume and moving voxels from 4D volume. Stationary voxels include the chest wall, which has only small motion during breathing, and moving voxels include the soft tissues inside chest wall, such as lung, heart, and abdomen. The chest wall was segmented from 3D volume, and the segmented Result was used to replace corresponding voxels in 4D volume. To generate the DRR, only high intensity voxels, including chest wall, heart, abdomen, main vessels and bronchi, and tumor, were used. This method reduced streak artifacts of 4DCBCT volume, but preserved the motion information present within the lungs and diaphragm. Results: DRR images computed from 4DCBCT are visually poor because of the streak artifacts, while DRR images computed from 3DCBCT suffer from motion blur. The DRR computed from the combined 3D and 4D volumes are visually more similar to the original projection images at corresponding projection angles. Conclusion: DRR image quality declined because of the streak artifacts in 4DCBCT. By combining 4D volume and streak‐free 3D volume and extracting the high intensity voxels to generate DRR, the streak artifacts can be effectively reduced, which results in improvement of the DRR image quality. This work was supported by the National Basic Research Program of China (973) under Grant 2011CB707701 and research foundation of Tsinghua University.

BackgroundNasopharyngeal carcinoma (NPC), an endemic malignancy in Southeast Asia and southern China, necessitates precise delineation of the gross tumor volume (GTV) for radiotherapy to optimize patient prognosis. Positron emission tomography (PET) imaging offers valuable metabolic insights to guide radiotherapy planning but is hampered by manual GTV segmentation, which is labor-intensive and prone to inter- and intra-observer variability. Conventional segmentation methods are inherently limited, and comparative evaluations of deep learning (DL) frameworks for PET-based primary NPC lesion segmentation remain scarce. Thus, this study aimed to optimize the delineation of primary NPC radiotherapy target volumes by leveraging PET-image-based DL segmentation technology, with the goal of enhancing both the precision and efficiency of lesion delineation.MethodsResearchers retrospectively collected PET imaging data from 212 NPC patients at the Affiliated Hospital of Guangdong Medical University. The patients were randomly divided into training (170 patients) and testing (42 patients) datasets. A radiation oncologist and a nuclear medicine physician collaboratively delineated the cancer lesion boundaries through consensus. After data preprocessing, three DL models (Res-Unet, Nn-Unet, and Nn-Former) were used to automatically segment the lesions. Training was based on the training dataset. During evaluation, the models’ segmentation performance was comprehensively assessed using Dice similarity coefficient (DSC) and Hausdorff distance (HD). A visual analysis of the results was also conducted to intuitively understand the models’ segmentation capabilities.ResultsIn the training set, Nn-Unet achieved the highest DSC of 0.869, whereas in the testing set, its DSC was 0.833. The DSC for Nn-Former hovered around 0.8 in both the training and testing sets. In contrast, Res-Unet demonstrated the lowest DSC values among the three models, particularly in the testing set (DSC =0.794). Statistical analysis revealed that the DSC of Nn-Unet was significantly higher than that of both Res-Unet and Nn-Former in the training set (P<0.01 for both comparisons). Regarding the 95% HD and average surface distance (ASD) values, Nn-Unet outperformed the other two models in both the training and testing sets. However, Res-Unet exhibited a much higher HD value in the testing set compared to the training set. The Loss curves for all models gradually decreased as training progressed, indicating that the models were learning relevant features of NPC lesions. The final Loss value for Res-Unet was approximately 0.7, whereas Nn-Unet and Nn-Former had final Loss values below 0.8.ConclusionsThe PET-image-based automatic segmentation model for the primary tumor of NPC established in this study demonstrates the clinical potential of Nn-Unet in primary lesion of NPC segmentation tasks.

Dosimetric Implications of Margin-Reduced MRI-Guided Stereotactic Body Radiotherapy to the Prostate Bed Following Radical Prostatectomy: Post-Hoc Analysis of a Phase II Study

IntroductionThis study aimed to evaluate the effect of image-guided radiation therapy (IGRT) on the margin between the clinical target volume (CTV) and planning target volume (PTV) in lung cancer.MethodsThe CTV and PTV margin were determined in three dimensions by four radiation oncologists using a standard method in 10 lung cancer patients, and compared to consensus values. Transfer error was measured using a rigid phantom containing gold markers. Systematic error () and random error () set up errors were calculated in three dimensions from pre-treatment and post-treatment cone beam CT scans. Finally, the margin between the CTV and PTV was corrected for set up error and calculated.ResultsThe margins between the CTV and PTV with IGRT (and without IGRT) were 0.88 cm (0.96 cm), 0.99 cm (1.08 cm) and 1.28 cm (1.82 cm) in the anterior and posterior (AP), left and right (LR) and superior and inferior (SI) directions, respectively. Images from two other patients verified the validity of the corrected margin. The target delineation errors of the radiation oncologists are considered to be the largest compared with the set up errors. The application of IGRT reduced the set up errors and the margins between CTV and PTV.ConclusionsThe delineation errors of radiation oncologists are the most important factor to consider for the margin between CTV and PTV for lung cancer. IGRT can reduce the margins by reducing the set up errors, especially in the SI direction. Further research is required to assess whether the reduction in the margin is solely based on set up errors.

Radiomics-guided Automatic Delineation for Clinical Target Volume of Endometrial Cancer: Limited-sample Multi-center Study.

A major objective of this project was to apply GE’s commercial machine learning and data analytics toolsets to large-scale, real-world, anonymized Phasor Measurement Unit (PMU) datasets in order to extract signatures, correlated and/or causal factors, and precursor patterns associated with significant power system phenomena. The project had a particular emphasis on extraction of insights relevant to asset health monitoring, real-time load modeling and cybersecurity monitoring. Additionally, the team was directed to undertake a comprehensive data quality analysis for the provided datasets and encouraged to estimate the ‘machine-learning readiness’ of the datasets by documenting any major obstacles to the application of commercial machine learning algorithms. To accomplish the aforementioned objectives, the project team’s work centered around the identification of key event signatures and application of the identified event signatures for event detection and event classification. The industry-validated, semi-supervised machine learning strategy employed for event signature identification involved several major tasks, including data-preprocessing, generation of an overabundance of features, normal data identification, normality modeling, and event signature identification through a methodical, quantitative ranking of features in order of relevance to each studied event type. Throughout the project, data quality issues and mitigation techniques were investigated. In this report, insights are provided regarding the readiness of the provided synchrophasor datasets for application of machine learning and data analytics. The methodologies employed for this technical strategy are summarized in this report. With regards to data preprocessing and feature generation, the provided Training and Test Datasets were ingested into GE’s big data environment. Subsequently, the team applied bad data cleansing and data imputation scripts, event detection scripts, and application programming interfaces (APIs) to the datasets for convenient data access. The project team completed development and validation of dozens of physics-based, statistics-based and transformation-based feature functions used for the extraction of over 60 synchrophasor features. Using a new parallel feature generation technology developed on this project, over 60 features have been rapidly generated for the full two years’ worth of Training and Test Dataset data associated with both the Eastern and Western interconnects. Even accommodating for temporal down-sampling inherent to the feature extraction procedure, this parallel feature generation activity resulted in a massive feature set with a storage requirement approximately equal to that of the raw training dataset itself. With regards to normal data identification and normality modeling, a normality model was built using the feature data extracted from the Training Dataset and iteratively refined subsequent to incremental adjustments and expansions of the Training Dataset feature data. With respect to event characterization and signature identification, an event signature identification pipeline was developed and used in conjunction with the normality model to identify over 15 event signatures for key event categories within the Training Dataset. The identified event signatures were used to characterize hundreds of key events in terms of relative severity, duration, and location of the event. An investigation was undertaken to identify correlated and causal factors involved in transformer events. A separate investigation into temporal trends in ring-down analysis results was undertaken to determine possible associations between system dynamics and various other factors such as loading, season or year. To validate the identified event signatures, additional work was undertaken to develop signature-based anomaly detection and classification tools suitable for convenient application to the synchrophasor datasets. The anomaly detection and classification tools, suitable for online application, were then applied to the entirety of the Eastern Interconnect Training and Test Datasets. Performance of the event detection and classification tools was evaluated upon receipt of the Test Dataset event logs (i.e., the labels for events contained in the Test Dataset), and promising results were obtained despite several challenges (documented herein) associated with application of supervised or semi-supervised machine learning methods to large-scale, anonymized datasets. Finally, the detection and classification tools were used to detect, classify, and characterize thousands of new events not included in the original event logs provided by the DOE within both the Training and Test Datasets.

Purpose:To assess changes in the volumes and spatial locations of tumors and surrounding organs by cone beam computed tomography during treatment for cervical cancer.Materials and Methods:Sixteen patients with cervical cancer had intensity-modulated radiotherapy and off-line cone beam computed tomography during chemotherapy and/or radiation therapy. The gross tumor volume (GTV-T) and clinical target volumes (CTVs) were contoured on the planning computed tomography and weekly cone beam computed tomography image, and changes in volumes and spatial locations were evaluated using the volume difference method and Dice similarity coefficients.Results:The GTV-T was 79.62 cm3 at prior treatment (0f) and then 20.86 cm3 at the end of external-beam chemoradiation. The clinical target volume changed slightly from 672.59 cm3 to 608.26 cm3, and the uterine volume (CTV-T) changed slightly from 83.72 cm3 to 80.23 cm3. There were significant differences in GTV-T and CTV-T among the different groups (P < .001), but the clinical target volume was not significantly different in volume (P > .05). The mean percent volume changes ranged from 23.05% to 70.85% for GTV-T, 4.71% to 6.78% for CTV-T, and 5.84% to 9.59% for clinical target volume, and the groups were significantly different (P < .05). The Dice similarity coefficient of GTV-T decreased during the course of radiation therapy (P < .001). In addition, there were significant differences in GTV-T among different groups (P < .001), and changes in GTV-T correlated with the radiotherapy (P < .001). There was a negative correlation between volume change rate (DV) and Dice similarity coefficient in the GTV-T and organs at risk (r < 0; P < .05).Conclusion:The volume, volume change rate, and Dice similarity coefficient of GTV-T were all correlated with increase in radiation treatment. Significant variations in tumor regression and spatial location occurred during radiotherapy for cervical cancer. Adaptive radiotherapy approaches are needed to improve the treatment accuracy for cervical cancer.

Objective To investigate the impact of anatomical landmarks on registration in image-guided radiotherapy (IGRT) for central and peripheral lung cancer. Methods Twenty-five patients with central or peripheral lung cancer for IGRT were enrolled in this study. Kilo-voltage cone-beam CT (kV-CBCT) scanning was acquired before irradiotion. Tumor coverage on CBCT was assessed using gross tumor volume (GTV), clinical target volume (CTV), and planning target volume (PTV) contours according to tumor alignment, carina registration, and spine registration, respectively. The grading analysiswas based on visual tumor assessment as follows: grade 0, tumor within GTV; grade 1, tumor outside GTV but inside CTV; grade 2, tumor outside CTV but inside PTV; and grade 3, tumor outside PTV. Results Totally 177 sets of kV-CBCT of 25 patients was collected. According to the registration landmarks of the tumor, carina and spine for central lung cancer, the percentages were 57.55%, 53.77% and 16.04% in grade 0, 39.62%, 45.28% and 58.49% in grade 1, and 1.89%, 0.94% and 25.47% in grade 2, respectively. For peripheral lung cancer, the percentages were 47.89%, 14.08% and 2.82% in grade 0, 43.66%, 29.58% and 45.07% in grade 1, and 8.45%, 40.85% and 35.21% in grade 2, respectively. Conclusions For central lung cancer, the tumor was recommended as the best registration landmark, and the carina was recommended as well, while the spine was not recommended. For peripheral lung cancer, the tumor was recommended as the best registration landmark, while the spine and the carina were not recommended. Key words: Lung neoplasms, central; Lung neoplasms, peripheral; Image-guided intensity-modulated radiotherapy; Registration landmarks; Grade

Many surrogate-based motion models (SMMs), proposed to guide motion management in radiotherapy, are constructed by correlating motion of an external surrogate and internal anatomy during CT-simulation. Changes in this correlation define model break down. We validate a methodology that incorporates fluoroscopic (FL) images acquired during treatment for SMM construction and update. Under a prospective IRB, 4DCT scans, VisionRT (VRT) surfaces, and orthogonal FLs were collected from five lung cancer patients. VRT surfaces and two FL time-series were acquired pre- and post-treatment. A simulated annealing optimization scheme was used to estimate optimal lung deformations by maximizing the mutual information (MI) between digitally reconstructed radiographs (DRRs) of the SMM-estimated 3D images and FLs. Our SMM used partial-least-regression and was trained using the optimal deformations and VRT surfaces from the first breathing-cycle. SMM performance was evaluated using the MI score between reference FLs and the corresponding SMM or phase-assigned 4DCT DRRs. The Hausdorff distance for contoured landmarks was used to evaluate target position estimation error. For four out of five patients, two principal components approximated lung surface deformations with submillimeter accuracy. Analysis of the MI score between more than 4000 pairs of FL and DRR demonstrated that our model led to more similarity between the FL and DRR images compared to 4DCT and DRR images from a model based on an a priori correlation model. Our SMM consistently displayed lower mean and 95th percentile Hausdorff distances. For one patient, 95th percentile Hausdorff distance was reduced by 11 mm. Patient-averaged reductions in mean and 95th percentile Hausdorff distances were 3.6 mm and 7 mm for right-lung, and 3.1 mm and 4 mm for left-lung targets. FL data were used to evaluate model performance and investigate the feasibility of model update. Despite variability in breathing, use of post-treatment FL preserved model fidelity and consistently outperformed 4DCT for position estimation.

Delineating the clinical target volume (CTV) and organs at risk (OARs) is crucial in rectal cancer radiotherapy. However, the accuracy of manual delineation (MD) is variable and the process is time consuming. Automatic delineation (AD) may be a solution to produce quicker and more accurate contours. In the present study, a convolutional neural network (CNN)-based AD tool was clinically evaluated to analyze its accuracy and efficiency in rectal cancer. CT images were collected from 148 supine patients in whom tumor stage and type of surgery were not differentiated. The rectal cancer contours consisted of CTV and OARs, where the OARs included the bladder, left and right femoral head, left and right kidney, spinal cord and bowel bag. The MD contours reviewed and modified together by a senior radiation oncologist committee were set as the reference values. The Dice similarity coefficient (DSC), Jaccard coefficient (JAC) and Hausdorff distance (HD) were used to evaluate the AD accuracy. The correlation between CT slice number and AD accuracy was analyzed, and the AD accuracy for different contour numbers was compared. The time recorded in the present study included the MD time, AD time for different CT slice and contour numbers and the editing time for AD contours. The Pearson correlation coefficient, paired-sample t-test and unpaired-sample t-test were used for statistical analyses. The results of the present study indicated that the DSC, JAC and HD for CTV using AD were 0.80±0.06, 0.67±0.08 and 6.96±2.45 mm, respectively. Among the OARs, the highest DSC and JAC using AD were found for the right and left kidney, with 0.91±0.06 and 0.93±0.04, and 0.84±0.09 and 0.88±0.07, respectively, and HD was lowest for the spinal cord with 2.26±0.82 mm. The lowest accuracy was found for the bowel bag. The more CT slice numbers, the higher the accuracy of the spinal cord analysis. However, the contour number had no effect on AD accuracy. To obtain qualified contours, the AD time plus editing time was 662.97±195.57 sec, while the MD time was 3294.29±824.70 sec. In conclusion, the results of the present study indicate that AD can significantly improve efficiency and a higher number of CT slices and contours can reduce AD efficiency. The AD tool provides acceptable CTV and OARs for rectal cancer and improves efficiency for delineation.

Image Guided Radiotherapy (IGRT) for Cervical Cancer: The Impact of Tumor Regression and Internal Organ Motion

Purpose: Measuring the setup errors and organ movements of patients with esophagus carcinoma during the radiotherapy, finding a reasonable margin from the Clinic Target Volume (CTV) to the Planning Target Volume (PTV). Methods and methods: 1. Set‐up verification: Forty‐two cases of untreated esophageal cancer were enrolled into this study. The Physicist firstly made the planning according to the doctor requests and ensure the best distribution at the target. Thereafter, the 0o and 90o Digital Reconstructed Radiograph(DRR)were transmitted to the iView GT workshop. Meanwhile, two copies of cross‐cut Electronic Portal Image (EPI), were required before radiotherapy. Two doctors confirm the variance of the osteal mark from the EPI and DRR, and output a 3 D direction (left to right, superior to inferior, anterior to posterior) of the setup errors through the iView GT software. 2. Breathing motion: Ten cases of untreated esophageal cancer were enrolled into this study. Three distinct breathing levels were defined: FB, EBH and IBH. We give the treatment planning in FB, then by moving the isocenter to EBH and IBH, we can recalculate the dose distribution without changing the field angle, shape and weighing(Mus). Displacements were analyzed at four points (anterior, posterior, right lateral and left lateral) and five levels of target (upper, quarter, isocenter, three‐quarter and lower). Result: (1) The systematic setup errors were −2.31mm, −0.16mm and −0.55mm, the random errors were 4.42mm, 4.48mm and 4.35mm at the direction of left to right(LR), superior to inferior (SI), anterior to posterior(AP) respectively. (2) The organ movements were 2.2mm, 4.0mm and 2.0mm at the LR, SI, AP respectively. Conclusion: As an alternative, the root‐sum‐of‐squares of set‐up error and organ motion is suggested, e.g. _tot=_(_ITV 2+_SM 2). The CTV to PTV margin were 8mm left to right, 5.3mm superior to inferior, 4.6mm anterior to posterior.