3D U-Net Research Articles

Enhancing brain image quality with 3D U-net for stripe removal in light sheet fluorescence microscopy

Light Sheet Fluorescence Microscopy (LSFM) is increasingly popular in neuroimaging for its ability to capture high-resolution 3D neural data. However, the presence of stripe noise significantly degrades image quality, particularly in complex 3D stripes with varying widths and brightness, posing challenges in neuroscience research. Existing stripe removal algorithms excel in suppressing noise and preserving details in 2D images with simple stripes but struggle with the complexity of 3D stripes. To address this, we propose a novel 3D U-net model for Stripe Removal in Light sheet fluorescence microscopy (USRL). This approach directly learns and removes stripes in 3D space across different scales, employing a dual-resolution strategy to effectively handle stripes of varying complexities. Additionally, we integrate a nonlinear mapping technique to normalize high dynamic range and unevenly distributed data before applying the stripe removal algorithm. We validate our method on diverse datasets, demonstrating substantial improvements in peak signal-to-noise ratio (PSNR) compared to existing algorithms. Moreover, our algorithm exhibits robust performance when applied to real LSFM data. Through extensive validation experiments, both on test sets and real-world data, our approach outperforms traditional methods, affirming its effectiveness in enhancing image quality. Furthermore, the adaptability of our algorithm extends beyond LSFM applications to encompass other imaging modalities. This versatility underscores its potential to enhance image usability across various research disciplines.

Cross-site Validation of AI Segmentation and Harmonization in Breast MRI.

This work aims to perform a cross-site validation of automated segmentation for breast cancers in MRI and to compare the performance to radiologists. A three-dimensional (3D) U-Net was trained to segment cancers in dynamic contrast-enhanced axial MRIs using a large dataset from Site 1 (n = 15,266; 449 malignant and 14,817 benign). Performance was validated on site-specific test data from this and two additional sites, and common publicly available testing data. Four radiologists from each of the three clinical sites provided two-dimensional (2D) segmentations as ground truth. Segmentation performance did not differ between the network and radiologists on the test data from Sites 1 and 2 or the common public data (median Dice score Site 1, network 0.86 vs. radiologist 0.85, n = 114; Site 2, 0.91 vs. 0.91, n = 50; common: 0.93 vs. 0.90). For Site 3, an affine input layer was fine-tuned using segmentation labels, resulting in comparable performance between the network and radiologist (0.88 vs. 0.89, n = 42). Radiologist performance differed on the common test data, and the network numerically outperformed 11 of the 12 radiologists (median Dice: 0.85-0.94, n = 20). In conclusion, a deep network with a novel supervised harmonization technique matches radiologists' performance in MRI tumor segmentation across clinical sites. We make code and weights publicly available to promote reproducible AI in radiology.

Large-language-model empowered 3D dose prediction for intensity-modulated radiotherapy.

Treatment planning is currently a patient specific, time-consuming, and resource demanding task in radiotherapy. Dose-volume histogram (DVH) prediction plays a critical role in automating this process. The geometric relationship between DVHs in radiotherapy plans and organs-at-risk (OAR) and planning target volume (PTV) has been well established. This study explores the potential of deep learning models for predicting DVHs using images and subsequent human intervention facilitated by a large-language model (LLM) to enhance the planning quality. We propose a pipeline to convert unstructured images to a structured graph consisting of image-patch nodes and dose nodes. A novel Dose Graph Neural Network (DoseGNN) model is developed for predicting DVHs from the structured graph. The proposed DoseGNN is enhanced with the LLM to encode massive knowledge from prescriptions and interactive instructions from clinicians. In this study, we introduced an online human-AI collaboration (OHAC) system as a practical implementation of the concept proposed for the automation of intensity-modulated radiotherapy (IMRT) planning. The proposed DoseGNN model was compared to widely employed DL models used in radiotherapy, including Swin Transformer, 3D U-Net CNN, and vanilla MLP. For PTV, DoseGNN achieved the mean absolute error (MAE) of , , , and between true plans and predicted plans that were 64%, 53%, 64%, 61% of the best baseline model. For the worst case among OARs (left lung, right lung, chest wall, heart, spinal cord), DoseGNN achieved the mean absolute error of , , that were 85%, 91%, 80% of the best baseline model. Moreover, the LLM-empowered DoseGNN model facilitates seamless adjustment to treatment plans through interaction with clinicians using natural language. We developed DoseGNN, a novel deep learning model for predicting delivered radiation doses from medical images, enhanced by LLM to allow adjustment through seamless interaction with clinicians. The preliminary results confirm DoseGNN's superior accuracy in DVH prediction relative to typical DL methods, highlighting its potential to facilitate an online clinician-AI collaboration system for streamlined treatment planning automation.

Deep learning–based statistical robustness evaluation of intensity-modulated proton therapy for head and neck cancer

A dense, dilated 3D U-net was trained to predict the 5th and 95th percentile distributions of planning dose plan using the nominal dose and planning CT images. The data set comprised proton therapy plans for 582 H&N cancer patients. Ground truth percentile values were estimated for each patient through 600 dose recalculations, representing randomly sampled uncertainty scenarios. The comprehensive comparisons of different models were conducted for H&N cancer patients, considering those with and without a beam mask (BM) and diverse beam configurations, including varying beam angles, couch angles, and beam numbers. The performance of our model trained based on a mixture of patients with H&N and prostate cancer was also assessed in contrast with models trained based on data specific for patients with cancer at either site. The DL-based model's predictions of percentile dose distributions exhibited excellent agreement with the ground truth dose distributions. The average gamma index with 2 mm/2%, consistently exceeded 97% for both 5th and 95th percentile dose volumes. Mean dose-volume histogram error analysis revealed that predictions from the combined training set yielded mean errors and standard deviations that were generally similar to those in the specific patient training data sets. Our proposed DL-based method for evaluation of the robustness of proton therapy plans provides precise, rapid predictions of percentile dose for a given confidence level regardless of the beam arrangement and cancer site. This versatility positions our model as a valuable tool for evaluating the robustness of proton therapy across various cancer sites.

Intelligent diagnostic method for developmental hip dislocation

BackgroundDevelopmental dislocation of the hip joint (DDH) is a condition that severely threatens children’s healthy growth. Without timely and correct treatment, it will lead to osteoarthritis and hip dysfunction in the evolution of children.ObjectiveIt is essential to develop an intelligent model for diagnosing hip dislocation and performing accurate quantitative analysis.MethodsIn this paper, 46 cases of computed tomography (CT) images were retrospectively collected, including 19 cases of hip dislocation and 27 cases of healthy people. The experiment first uses ITK-SNAP to sketch the ilium and femoral head in the original image. Then, it uses 3D U-Net to send the label of the background, ilium, and femoral head into three channels, respectively, to realize the three-dimensional segmentation of the ilium and femoral head. Next, the extraction of the surface of the acetabulum and femoral head is performed. Subsequently, the erroneous points are eliminated, and the spherical surfaces of the acetabulum and femoral head are fitted using the least squares method. Ultimately, the spherical center distance is calculated quantitatively to predict whether the hip joint is dislocated.ResultsUnder the independent test set, the segmentation average dice coefficients of the ilium and femoral head are 89% and 93%, respectively. The spherical center distance between the acetabulum and femoral head is calculated quantitatively. If the value exceeds 10 mm, it is considered a hip dislocation. Compared with the doctor’s diagnosis, the accuracy result is 94.4%.ConclusionThis paper successfully implements a precise and automated intelligent diagnostic system for the identification of hip dislocation. Commencing with the development of a 3D segmentation algorithm for the ilium and femoral head, we further introduce a novel method that computes the spherical distance for the prediction of hip dislocation. This approach provides robust quantitative analysis, thereby facilitating more informed clinical decision-making.

Exploring the impact of network depth on 3D U-Net-based dose prediction for cervical cancer radiotherapy.

The 3D U-Net deep neural network structure is widely employed for dose prediction in radiotherapy. However, the attention to the network depth and its impact on the accuracy and robustness of dose prediction remains inadequate. 92 cervical cancer patients who underwent Volumetric Modulated Arc Therapy (VMAT) are geometrically augmented to investigate the effects of network depth on dose prediction by training and testing three different 3D U-Net structures with depths of 3, 4, and 5. For planning target volume (PTV), the differences between predicted and true values of D98, D99, and Homogeneity were statistically 1.00 ± 0.23, 0.32 ± 0.72, and -0.02 ± 0.02 for the model with a depth of 5, respectively. Compared to the other two models, these parameters were also better. For most of the organs at risk, the mean and maximum differences between the predicted values and the true values for the model with a depth of 5 were better than for the other two models. The results reveal that the network model with a depth of 5 exhibits superior performance, albeit at the expense of the longest training time and maximum computational memory in the three models. A small server with two NVIDIA GeForce RTX 3090 GPUs with 24 G of memory was employed for this training. For the 3D U-Net model with a depth of more than 5 cannot be supported due to insufficient training memory, the 3D U-Net neural network with a depth of 5 is the commonly used and optimal choice for small servers.

An hetero-modal deep learning framework for medical image synthesis applied to contrast and non-contrast MRI

Some pathologies such as cancer and dementia require multiple imaging modalities to fully diagnose and assess the extent of the disease. Magnetic resonance imaging offers this kind of polyvalence, but examinations take time and can require contrast agent injection. The flexible synthesis of these imaging sequences based on the available ones for a given patient could help reduce scan times or circumvent the need for contrast agent injection. In this work, we propose a deep learning architecture that can perform the synthesis of all missing imaging sequences from any subset of available images. The network is trained adversarially, with the generator consisting of parallel 3D U-Net encoders and decoders that optimally combines their multi-resolution representations with a fusion operation learned by an attention network trained conjointly with the generator network. We compare our synthesis performance with 3D networks using other types of fusion and a comparable number of trainable parameters, such as the mean/variance fusion. In all synthesis scenarios except one, the synthesis performance of the network using attention-guided fusion was better than the other fusion schemes. We also inspect the encoded representations and the attention network outputs to gain insights into the synthesis process, and uncover desirable behaviors such as prioritization of specific modalities, flexible construction of the representation when important modalities are missing, and modalities being selected in regions where they carry sequence-specific information. This work suggests that a better construction of the latent representation space in hetero-modal networks can be achieved by using an attention network.

Using 3D U-Net for Brain Tumour Segmentation from Magnetic Resonance Images

Brain tumours within the skull can lead to serious health issues. The rapid and accurate detection and segmentation of tumour regions allow patients to receive appropriate treatment at an early stage, increasing their chances of recovery and survival. Various medical imaging methods, such as Magnetic Resonance Imaging (MRI), Positron and digital pathology, Emission Tomography (PET), Computed Tomography (CT) are used for the detection of brain tumours. Nowadays, with advancing technology and hardware, concepts like artificial intelligence and deep learning (DL) are becoming increasingly popular. Many artificial intelligence methods are also being utilized in studies on brain tumour segmentation. This paper proposes a 3D U-Net DL model for brain tumour segmentation. The training and testing processes are carried out on the Brain Tumour Segmentation (BraTS) 2020 dataset, which is widely used in the literature. As a result, an Intersection over Union (IoU) score of 0.81, a dice score of 0.87 and a pixel accuracy of 0.99 are achieved. The proposed model has the potential to assist experts in diagnosing the disease and developing appropriate treatment plans, thanks to its ability to segment brain tumours quickly and with high accuracy.

Evaluation of an automated laminar cartilage T2 relaxation time analysis method in an early osteoarthritis model.

A fully automated laminar cartilage composition (MRI-based T2) analysis method was technically and clinically validated by comparing radiographically normal knees with (CL-JSN) and without contra-lateral joint space narrowing or other signs of radiographic osteoarthritis (OA, CL-noROA). 2D U-Nets were trained from manually segmented femorotibial cartilages (n = 72) from all 7 echoes (AllE), or from the 1st echo only (1stE) of multi-echo-spin-echo (MESE) MRIs acquired by the Osteoarthritis Initiative (OAI). Because of its greater accuracy, only the AllE U-Net was then applied to knees from the OAI healthy reference cohort (n = 10), CL-JSN (n = 39), and (1:1) matched CL-noROA knees (n = 39) that all had manual expert segmentation, and to 982 non-matched CL-noROA knees without expert segmentation. The agreement (Dice similarity coefficient) between automated vs. manual expert cartilage segmentation was between 0.82 ± 0.05/0.79 ± 0.06 (AllE/1stE) and 0.88 ± 0.03/0.88 ± 0.03 (AllE/1stE) across femorotibial cartilage plates. The deviation between automated vs. manually derived laminar T2 reached up to - 2.2 ± 2.6ms/ + 4.1 ± 10.2ms (AllE/1stE). The AllE U-Net showed a similar sensitivity to cross-sectional laminar T2 differences between CL-JSN and CL-noROA knees in the matched (Cohen's D ≤ 0.54) and the non-matched (D ≤ 0.54) comparison as the matched manual analyses (D ≤ 0.48). Longitudinally, the AllE U-Net also showed a similar sensitivity to CL-JSN vs. CS-noROA differences in the matched (D ≤ 0.51) and the non-matched (D ≤ 0.43) comparison as matched manual analyses (D ≤ 0.41). The fully automated T2 analysis showed a high agreement, acceptable accuracy, and similar sensitivity to cross-sectional and longitudinal laminar T2 differences in an early OA model, compared with manual expert analysis. Clinicaltrials.gov identification: NCT00080171.

A 3D U-NET BASED ON EARLY FUSION MODEL: IMPROVEMENT, COMPARATIVE ANALYSIS WITH STATE-OF-THE-ART MODELS AND FINE-TUNING

Multi-organ segmentation is the process of identifying and separating multiple organs in medical images. This segmentation allows for the detection of structural abnormalities by examining the morphological structure of organs. Carrying out the process quickly and precisely has become an important issue in today's conditions. In recent years, researchers have used various technologies for the automatic segmentation of multiple organs. In this study, improvements were made to increase the multi-organ segmentation performance of the 3D U-Net based fusion model combining HSV and grayscale color spaces and compared with state-of-the-art models. Training and testing were performed on the MICCAI 2015 dataset published at Vanderbilt University, which contains 3D abdominal CT images in NIfTI format. The model's performance was evaluated using the Dice similarity coefficient. In the tests, the liver organ showed the highest Dice score. Considering the average Dice score of all organs, and comparing it with other models, it has been observed that the fusion approach model yields promising results.

Edit Temporal-Consistent Videos with Image Diffusion Model

Large-scale text-to-image (T2I) diffusion models have been extended for text-guided video editing, yielding impressive zero-shot video editing performance. Nonetheless, the generated videos usually show spatial irregularities and temporal inconsistencies as the temporal characteristics of videos have not been faithfully modeled. In this paper, we propose an elegant yet effective Temporal-Consistent Video Editing (TCVE) method to mitigate the temporal inconsistency challenge for robust text-guided video editing. In addition to the utilization of a pretrained T2I 2D Unet for spatial content manipulation, we establish a dedicated temporal Unet architecture to faithfully capture the temporal coherence of the input video sequences. Furthermore, to establish coherence and interrelation between the spatial-focused and temporal-focused components, a cohesive spatial-temporal modeling unit is formulated. This unit effectively interconnects the temporal Unet with the pretrained 2D Unet, thereby enhancing the temporal consistency of the generated videos while preserving the capacity for video content manipulation. Quantitative experimental results and visualization results demonstrate that TCVE achieves state-of-the-art performance in both video temporal consistency and video editing capability, surpassing existing benchmarks in the field. Codes are released at https://github.com/mdswyz/TCVE .

Enhancing brain tumor segmentation in MRI images: A hybrid approach using UNet, attention mechanisms, and transformers

Accurate brain tumor segmentation in MRI images is crucial for effective treatment planning and monitoring. Traditional methods often encounter challenges due to the complexity and variability of tumor shapes and textures. Consequently, there is a growing need for automated solutions to assist healthcare professionals in segmentation tasks, improving efficiency and reducing workload. This study introduces an innovative method for accurately segmenting brain tumors in MRI images by employing a refined 3D UNet model integrated with a Transformer. The goal is to leverage self-attention mechanisms to enhance segmentation capabilities. The proposed model combines Contextual Transformer (CoT) and Double Attention (DA) architectures. CoT is extended to a 3D format and integrated with the baseline model to exploit intricate contextual details in MRI images. DA blocks in skip connections aggregate and distribute long-range features, emphasizing inter-dependencies within an expanded spatial scope. Experimental results demonstrate superior segmentation performance compared to current state-of-the-art methods. With its ability to accurately segment and delineate tumors in 3D, our segmentation model promises to be a powerful tool for medical image processing and performance optimization, saving time for healthcare professionals and healthcare systems.

Rapid identification of time-frequency domain gravitational wave signals from binary black holes using deep learning

Abstract Recent developments in deep learning techniques have offered an alternative and complementary approach to traditional matched filtering methods for the identification of gravitational wave (GW) signals. The rapid and accurate identification of GW signals is crucial for the progress of GW physics and multi-messenger astronomy, particularly in light of the upcoming fourth and fifth observing runs of LIGO-Virgo-KAGRA. In this work, we use the 2D U-Net algorithm to identify the time-frequency domain GW signals from stellar-mass binary black hole (BBH) mergers. We simulate BBH mergers with component masses from 7 to 50 $M_{\odot}$ and account for the LIGO detector noise. We find that the GW events in the first and second observation runs could all be clearly and rapidly identified. For the third observing run, about 80\% GW events could be identified. Compared to the traditional convolutional neural network, the U-Net algorithm can output time-frequency domain signal images corresponding to probabilities, providing a more intuitive analysis. In conclusion, the U-Net algorithm can rapidly identify the time-frequency domain GW signals from BBH mergers.

Performance enhancement of deep learning based solutions for pharyngeal airway space segmentation on MRI scans

The automatic segmentation of the pharyngeal airway space has many potential medical use, one of which is to help facilitate the creation of the Tubingen Palatal Plate. Therefore, it is of great importance to understand which methods are suitable for this task. Here, neural network based solutions available in the literature are compared to find the best methods. Neural network models were chosen to encompass a diverse landscape. Some models were taken from the general semantic segmentation literature, while others were taken from the medical or pharyngeal airway space segmentation literature. Some models are convolutional neural networks, while others are transformer-based model or a mix of both convolutional and transformer-based model. These models include 2d/3d U-Net, Deeplabv3, YOLOv8, Swinv2 UNETR, SegFormer, and 3D MRU-Net. Furthermore, additional strategies to enhance performance were also considered. These strategies consisted of training two separate networks in multiple stages as well leveraging unlabeled data to pretrain the neural networks before fine-tuning them on the labeled data. It was found that out of all the models considered here, the 2d U-Net performed the best achieving an average dice score of 0.9180 ± 0.0111. Out of all the strategies to enhance performance, only two strategies improve the actual results but only by a small margin. Therefore, these strategies can be consider if a small increase in performance is desired from the 2d U-Net at the expense of computational resource.

A hierarchical deep learning approach for diagnosing impacted canine-induced root resorption via cone-beam computed tomography

ObjectivesCanine-induced root resorption (CIRR) is caused by impacted canines and CBCT images have shown to be more accurate in diagnosing CIRR than panoramic and periapical radiographs with the reported AUCs being 0.95, 0.49, and 0.57, respectively. The aim of this study was to use deep learning to automatically evaluate the diagnosis of CIRR in maxillary incisors using CBCT images.MethodsA total of 50 cone beam computed tomography (CBCT) images and 176 incisors were selected for the present study. The maxillary incisors were manually segmented and labeled from the CBCT images by two independent radiologists as either healthy or affected by root resorption induced by the impacted canines. We used five different strategies for training the model: (A) classification using 3D ResNet50 (Baseline), (B) classification of the segmented masks using the outcome of a 3D U-Net pretrained on the 3D MNIST, (C) training a 3D U-Net for the segmentation task and use its outputs for classification, (D) pretraining a 3D U-Net for the segmentation and transfer of the model, and (E) pretraining a 3D U-Net for the segmentation and fine-tuning the model with only the model encoder. The segmentation models were evaluated using the mean intersection over union (mIoU) and Dice coefficient (DSC). The classification models were evaluated in terms of classification accuracy, precision, recall, and F1 score.ResultsThe segmentation model achieved a mean intersection over union (mIoU) of 0.641 and a DSC of 0.901, indicating good performance in segmenting the tooth structures from the CBCT images. For the main classification task of detecting CIRR, Model C (classification of the segmented masks using 3D ResNet) and Model E (pretraining on segmentation followed by fine-tuning for classification) performed the best, both achieving 82% classification accuracy and 0.62 F1-scores on the test set. These results demonstrate the effectiveness of the proposed hierarchical, data-efficient deep learning approaches in improving the accuracy of automated CIRR diagnosis from limited CBCT data compared to the 3D ResNet baseline model.ConclusionThe proposed approaches are effective at improving the accuracy of classification tasks and are helpful when the diagnosis is based on the volume and boundaries of an object. While the study demonstrated promising results, future studies with larger sample size are required to validate the effectiveness of the proposed method in enhancing the medical image classification tasks.

Multi-Stage Feature Fusion of Multispectral and SAR Satellite Images for Seasonal Crop-Type Mapping at Regional Scale Using an Adapted 3D U-Net Model

Multi-Stage Feature Fusion of Multispectral and SAR Satellite Images for Seasonal Crop-Type Mapping at Regional Scale Using an Adapted 3D U-Net Model

An end-to-end deep learning pipeline to derive blood input with partial volume corrections for automated parametric brain PET mapping

Dynamic 2-[18F] fluoro-2-deoxy-D-glucose positron emission tomography (dFDG-PET) for human brain imaging has considerable clinical potential, yet its utilization remains limited. A key challenge in the quantitative analysis of dFDG-PET is characterizing a patient-specific blood input function, traditionally reliant on invasive arterial blood sampling. This research introduces a novel approach employing non-invasive deep learning model-based computations from the internal carotid arteries (ICA) with partial volume (PV) corrections, thereby eliminating the need for invasive arterial sampling. We present an end-to-end pipeline incorporating a 3D U-Net based ICA-net for ICA segmentation, alongside a Recurrent Neural Network (RNN) based MCIF-net for the derivation of a model-corrected blood input function (MCIF) with PV corrections. The developed 3D U-Net and RNN was trained and validated using a 5-fold cross-validation approach on 50 human brain FDG PET scans. The ICA-net achieved an average Dice score of 82.18% and an Intersection over Union of 68.54% across all tested scans. Furthermore, the MCIF-net exhibited a minimal root mean squared error of 0.0052. The application of this pipeline to ground truth data for dFDG-PET brain scans resulted in the precise localization of seizure onset regions, which contributed to a successful clinical outcome, with the patient achieving a seizure-free state after treatment. These results underscore the efficacy of the ICA-net and MCIF-net deep learning pipeline in learning the ICA structure’s distribution and automating MCIF computation with PV corrections. This advancement marks a significant leap in non-invasive neuroimaging.

Development of deep learning-based novel auto-segmentation for the prostatic urethra on planning CT images for prostate cancer radiotherapy.

Urinary toxicities are one of the serious complications of radiotherapy for prostate cancer, and dose-volume histogram of prostatic urethra has been associated with such toxicities in previous reports. Previous research has focused on estimating the prostatic urethra, which is difficult to delineate in CT images; however, these studies, which are limited in number, mainly focused on cases undergoing brachytherapy uses low-dose-rate sources and do not involve external beam radiation therapy (EBRT). In this study, we aimed to develop a deep learning-based method of determining the position of the prostatic urethra in patients eligible for EBRT. We used contour data from 430 patients with localized prostate cancer. In all cases, a urethral catheter was placed when planning CT to identify the prostatic urethra. We used 2D and 3D U-Net segmentation models. The input images included the bladder and prostate, while the output images focused on the prostatic urethra. The 2D model determined the prostate's position based on results from both coronal and sagittal directions. Evaluation metrics included the average distance between centerlines. The average centerline distances for the 2D and 3D models were 2.07 ± 0.87mm and 2.05 ± 0.92mm, respectively. Increasing the number of cases while maintaining equivalent accuracy as we did in this study suggests the potential for high generalization performance and the feasibility of using deep learning technology for estimating the position of the prostatic urethra.

First implementation of artificial intelligence empowered all-in-one radiotherapy workflow for nasopharyngeal carcinoma.

134 Background: To establish an artificial intelligence (AI)-empowered All-In-One (AIO) workflow for providing one-stop radiotherapy to patients with nasopharyngeal carcinoma (NPC), and evaluate its feasibility and performance. Methods: The NPC-specific AIO workflow operates on a CT-integrated linac and integrates AI contouring and AI planning modules. A 3D U-net was employed for contours generation and 3D dose distribution. Knowledge-based planning approach and a protocol-based approach were used for plan optimization. MRI datasets with contours of primary gross tumors, CT datasets with contours of targets and organs at risk, and radiotherapy treatment planning datasets were collected from 877, 139, and 243 patients, respectively, for the training, validation, and testing of AI contouring and planning models. The NPC-specific AIO workflow underwent real-world clinical testing with an end-to-end evaluation involving 120 patients with NPC. The technical characteristics, implementation procedures and performance of AIO workflow were described and evaluated. Results: From March 2022 to October 2023, 120 patients with newly-diagnosed, non-metastatic NPC underwent AIO workflow, and 117 patients completed it in a median time of 23.2 min (range: 16.3 to 45.8 min). Median translation motions were 0.2 mm (CT scan to planning approval) and 0.1 mm (during beam delivery). AI-generated contours required little revision for clinical target volume (CTV1) and organs at risk (OARs), minor revision for cervical lymph nodes (GTVn) and clinical target volume (CTV2), with median Dice similarity coefficients (DSCs) of 0.98 and 0.94, respectively; more revision for GTVp, with a median DSC of 0.84. Of the 117 AI-generated plannings, 108 (92.3%) passed after initial auto-optimization. Volumes receiving ≥ 100% of the prescribed dose for all planning target volumes was ≥ 97.8%. Mean dosimetric parameters met dose constraints for all OARs except the thyroid and submandibulars. One patient died from acute renal failure after 10 fractions of radiotherapy. At week 12 after radiotherapy completion, 115 patients achieved complete response;1 had stable disease. Grade 3-4 acute effects were reported by 54 of 115 patients (36%). Conclusions: We successfully established an AI-empowered AIO workflow for initial radiotherapy of NPC, confirming its feasibility and safety for clinical application.

Anatomy-constrained synthesis for spleen segmentation improvement in unpaired mouse micro-CT scans with 3D CycleGAN

Objective. Auto-segmentation in mouse micro-CT enhances the efficiency and consistency of preclinical experiments but often struggles with low-native-contrast and morphologically complex organs, such as the spleen, resulting in poor segmentation performance. While CT contrast agents can improve organ conspicuity, their use complicates experimental protocols and reduces feasibility. We developed a 3D Cycle Generative Adversarial Network (CycleGAN) incorporating anatomy-constrained U-Net models to leverage contrast-enhanced CT (CECT) insights to improve unenhanced native CT (NACT) segmentation. Approach. We employed a standard CycleGAN with an anatomical loss function to synthesize virtual CECT images from unpaired NACT scans at two different resolutions. Prior to training, two U-Nets were trained to automatically segment six major organs in NACT and CECT datasets, respectively. These pretrained 3D U-Nets were integrated during the CycleGAN training, segmenting synthetic images, and comparing them against ground truth annotations. The compound loss within the CycleGAN maintained anatomical fidelity. Full image processing was achieved for low-resolution datasets, while high-resolution datasets employed a patch-based method due to GPU memory constraints. Automated segmentation was applied to original NACT and synthetic CECT scans to evaluate CycleGAN performance using the Dice Similarity Coefficient (DSC) and the 95th percentile Hausdorff Distance (HD95p). Main results. High-resolution scans showed improved auto-segmentation, with an average DSC increase from 0.728 to 0.773 and a reduced HD95p from 1.19 mm to 0.94 mm. Low-resolution scans benefited more from synthetic contrast, showing a DSC increase from 0.586 to 0.682 and an HD95p reduction from 3.46 mm to 1.24 mm. Significance. Implementing CycleGAN to synthesize CECT scans substantially improved the visibility of the mouse spleen, leading to more precise auto-segmentation. This approach shows the potential in preclinical imaging studies where contrast agent use is impractical.