Dice Score Research Articles

Semi-Self-Supervised Domain Adaptation: Developing Deep Learning Models with Limited Annotated Data for Wheat Head Segmentation

Precision agriculture involves the application of advanced technologies to improve agricultural productivity, efficiency, and profitability while minimizing waste and environmental impacts. Deep learning approaches enable automated decision-making for many visual tasks. However, in the agricultural domain, variability in growth stages and environmental conditions, such as weather and lighting, presents significant challenges to developing deep-learning-based techniques that generalize across different conditions. The resource-intensive nature of creating extensive annotated datasets that capture these variabilities further hinders the widespread adoption of these approaches. To tackle these issues, we introduce a semi-self-supervised domain adaptation technique based on deep convolutional neural networks with a probabilistic diffusion process, requiring minimal manual data annotation. Using only three manually annotated images and a selection of video clips from wheat fields, we generated a large-scale computationally annotated dataset of image–mask pairs and a large dataset of unannotated images extracted from video frames. We developed a two-branch convolutional encoder–decoder model architecture that uses both synthesized image–mask pairs and unannotated images, enabling effective adaptation to real images. The proposed model achieved a Dice score of 80.7% on an internal test dataset and a Dice score of 64.8% on an external test set composed of images from five countries and spanning 18 domains, indicating its potential to develop generalizable solutions that could encourage the wider adoption of advanced technologies in agriculture.

VESCL: an open source 2D vessel contouring library.

VESCL (pronounced 'vessel') is a novel vessel contouring library for computer-assisted 2D vessel contouring and segmentation. VESCL facilitates manual vessel segmentation in 2D medical images to generate gold-standard datasets for training, testing, and validating automatic vessel segmentation. VESCL is an open-source C++ library designed for easy integration into medical image processing systems. VESCL provides an intuitive interface for drawing variable-width parametric curves along vessels in 2D images. It includes highly optimized localized filtering to automatically fit drawn curves to the nearest vessel centerline and automatically determine the varying vessel width along each curve. To support a variety of segmentation paradigms, VESCL can export multiple segmentation representations including binary segmentations, occupancy maps, and distance fields. VESCL provides sub-pixel resolution for vessel centerlines and vessel widths. It is optimized to segment small vessels with single- or sub-pixel widths that are visible to the human eye but hard to segment automatically via conventional filters. When tested on neurovascular digital subtraction angiography (DSA), VESCL's intuitive hand-drawn input with automatic curve fitting increased the speed of fully manual segmentation by 22× over conventional methods and by 3× over the best publicly available computer-assisted manual segmentation method. Accuracy was shown to be within the range of inter-operator variability of gold standard manually segmented data from a publicly available dataset of neurovascular DSA images as measured using Dice scores. Preliminary tests showed similar improvements for segmenting DSA of coronary arteries and RGB images of retinal arteries. VESCL is an open-source C++ library for contouring vessels in 2D images which can be used to reduce the tedious, labor-intensive process of manually generating gold-standard segmentations for training, testing, and comparing automatic segmentation methods.

Estimation of Fractal Dimension and Segmentation of Brain Tumor with Parallel Features Aggregation Network

The accurate recognition of a brain tumor (BT) is crucial for accurate diagnosis, intervention planning, and the evaluation of post-intervention outcomes. Conventional methods of manually identifying and delineating BTs are inefficient, prone to error, and time-consuming. Subjective methods for BT recognition are biased because of the diffuse and irregular nature of BTs, along with varying enhancement patterns and the coexistence of different tumor components. Hence, the development of an automated diagnostic system for BTs is vital for mitigating subjective bias and achieving speedy and effective BT segmentation. Recently developed deep learning (DL)-based methods have replaced subjective methods; however, these DL-based methods still have a low performance, showing room for improvement, and are limited to heterogeneous dataset analysis. Herein, we propose a DL-based parallel features aggregation network (PFA-Net) for the robust segmentation of three different regions in a BT scan, and we perform a heterogeneous dataset analysis to validate its generality. The parallel features aggregation (PFA) module exploits the local radiomic contextual spatial features of BTs at low, intermediate, and high levels for different types of tumors and aggregates them in a parallel fashion. To enhance the diagnostic capabilities of the proposed segmentation framework, we introduced the fractal dimension estimation into our system, seamlessly combined as an end-to-end task to gain insights into the complexity and irregularity of structures, thereby characterizing the intricate morphology of BTs. The proposed PFA-Net achieves the Dice scores (DSs) of 87.54%, 93.42%, and 91.02%, for the enhancing tumor region, whole tumor region, and tumor core region, respectively, with the multimodal brain tumor segmentation (BraTS)-2020 open database, surpassing the performance of existing state-of-the-art methods. Additionally, PFA-Net is validated with another open database of brain tumor progression and achieves a DS of 64.58% for heterogeneous dataset analysis, surpassing the performance of existing state-of-the-art methods.

GETNet: Group Normalization Shuffle and Enhanced Channel Self-Attention Network Based on VT-UNet for Brain Tumor Segmentation.

Currently, brain tumors are extremely harmful and prevalent. Deep learning technologies, including CNNs, UNet, and Transformer, have been applied in brain tumor segmentation for many years and have achieved some success. However, traditional CNNs and UNet capture insufficient global information, and Transformer cannot provide sufficient local information. Fusing the global information from Transformer with the local information of convolutions is an important step toward improving brain tumor segmentation. We propose the Group Normalization Shuffle and Enhanced Channel Self-Attention Network (GETNet), a network combining the pure Transformer structure with convolution operations based on VT-UNet, which considers both global and local information. The network includes the proposed group normalization shuffle block (GNS) and enhanced channel self-attention block (ECSA). The GNS is used after the VT Encoder Block and before the downsampling block to improve information extraction. An ECSA module is added to the bottleneck layer to utilize the characteristics of the detailed features in the bottom layer effectively. We also conducted experiments on the BraTS2021 dataset to demonstrate the performance of our network. The Dice coefficient (Dice) score results show that the values for the regions of the whole tumor (WT), tumor core (TC), and enhancing tumor (ET) were 91.77, 86.03, and 83.64, respectively. The results show that the proposed model achieves state-of-the-art performance compared with more than eleven benchmarks.

RDLR: A Robust Deep Learning-Based Image Registration Method for Pediatric Retinal Images.

Retinal diseases stand as a primary cause of childhood blindness. Analyzing the progression of these diseases requires close attention to lesion morphology and spatial information. Standard image registration methods fail to accurately reconstruct pediatric fundus images containing significant distortion and blurring. To address this challenge, we proposed a robust deep learning-based image registration method (RDLR). The method consisted of two modules: registration module (RM) and panoramic view module (PVM). RM effectively integrated global and local feature information and learned prior information related to the orientation of images. PVM was capable of reconstructing spatial information in panoramic images. Furthermore, as the registration model was trained on over 280,000 pediatric fundus images, we introduced a registration annotation automatic generation process coupled with a quality control module to ensure the reliability of training data. We compared the performance of RDLR to the other methods, including conventional registration pipeline (CRP), voxel morph (WM), generalizable image matcher (GIM), and self-supervised techniques (SS). RDLR achieved significantly higher registration accuracy (average Dice score of 0.948) than the other methods (ranging from 0.491 to 0.802). The resulting panoramic retinal maps reconstructed by RDLR also demonstrated substantially higher fidelity (average Dice score of 0.960) compared to the other methods (ranging from 0.720 to 0.783). Overall, the proposed method addressed key challenges in pediatric retinal imaging, providing an effective solution to enhance disease diagnosis. Our source code is available at https://github.com/wuwusky/RobustDeepLeraningRegistration .

Mask the Unknown: Assessing Different Strategies to Handle Weak Annotations in the MICCAI2023 Mediastinal Lymph Node Quantification Challenge

Pathological lymph node delineation is crucial in cancer diagnosis, progression assessment, and treatment planning. The MICCAI 2023 Lymph Node Quantification Challenge published the first public dataset for pathological lymph node segmentation in the mediastinum. As lymph node annotations are expensive, the challenge was formed as a weakly supervised learning task, where only a subset of all lymph nodes in the training set have been annotated. For the challenge submission, multiple methods for training on these weakly supervised data were explored, including noisy label training, loss masking of unlabeled data, and an approach that integrated the TotalSegmentator toolbox as a form of pseudo labeling in order to reduce the number of unknown voxels. Furthermore, multiple public TCIA datasets were incorporated into the training to improve the performance of the deep learning model. Our submitted model achieved a Dice score of 0.628 and an average symmetric surface distance of 5.8~mm on the challenge test set. With our submitted model, we accomplished the third rank in the MICCAI2023 LNQ challenge. A finding of our analysis was that the integration of all visible, including non-pathological lymph nodes improved the overall segmentation performance on pathological lymph nodes of the test set. Furthermore, segmentation models trained only on clinically enlarged lymph nodes, as given in the challenge scenario, could not generalize to smaller pathological lymph nodes. The code and model for the challenge submission are available at <a href='https://gitlab.lrz.de/compai/MediastinalLymphNodeSegmentation'>https://gitlab.lrz.de/compai/MediastinalLymphNodeSegmentation</a>

Partial Image Active Annotation (PIAA): An Efficient Active Learning Technique Using Edge Information in Limited Data Scenarios

AbstractActive learning (AL) algorithms are increasingly being used to train models with limited data for annotation tasks. However, the selection of data for AL is a complex issue due to the restricted information on unseen data. To tackle this problem, a technique we refer to as Partial Image Active Annotation (PIAA) employs the edge information of unseen images as prior knowledge to gauge uncertainty. This uncertainty is determined by examining the divergence and entropy in model predictions across edges. The resulting measure is then applied to choose superpixels from input images for active annotation. We demonstrate the effectiveness of PIAA in multi-class Optical Coherence Tomography (OCT) segmentation tasks, attaining a Dice score comparable to state-of-the-art OCT segmentation algorithms trained with extensive annotated data. Concurrently, we successfully reduce annotation label costs to 12%, 2.3%, and 3%, respectively, across three publicly accessible datasets (Duke, AROI, and UMN).

Odontogenic cystic lesion segmentation on cone-beam CT using an auto-adapting multi-scaled UNet.

Precise segmentation of Odontogenic Cystic Lesions (OCLs) from dental Cone-Beam Computed Tomography (CBCT) is critical for effective dental diagnosis. Although supervised learning methods have shown practical diagnostic results in segmenting various diseases, their ability to segment OCLs covering different sub-class varieties has not been extensively investigated. In this study, we propose a new supervised learning method termed OCL-Net that combines a Multi-Scaled U-Net model, along with an Auto-Adapting mechanism trained with a combined supervised loss. Anonymous CBCT images were collected retrospectively from one hospital. To assess the ability of our model to improve the diagnostic efficiency of maxillofacial surgeons, we conducted a diagnostic assessment where 7 clinicians were included to perform the diagnostic process with and without the assistance of auto-segmentation masks. We collected 300 anonymous CBCT images which were manually annotated for segmentation masks. Extensive experiments demonstrate the effectiveness of our OCL-Net for CBCT OCLs segmentation, achieving an overall Dice score of 88.84%, an IoU score of 81.23%, and an AUC score of 92.37%. Through our diagnostic assessment, we found that when clinicians were assisted with segmentation labels from OCL-Net, their average diagnostic accuracy increased from 53.21% to 55.71%, while the average time spent significantly decreased from 101s to 47s (P<0.05). The findings demonstrate the potential of our approach as a robust auto-segmentation system on OCLs in CBCT images, while the segmented masks can be used to further improve OCLs dental diagnostic efficiency.

URCA: Uncertainty-based region clipping algorithm for semi-supervised medical image segmentation

Background and objectiveTraining convolutional neural networks based on large amount of labeled data has made great progress in the field of image segmentation. However, in medical image segmentation tasks, annotating the data is expensive and time-consuming because pixel-level annotation requires experts in the relevant field. Currently, the combination of consistent regularization and pseudo labeling-based semi-supervised methods has shown good performance in image segmentation. However, in the training process, a portion of low-confidence pseudo labels are generated by the model. And the semi-supervised segmentation method still has the problem of distribution bias between labeled and unlabeled data. The objective of this study is to address the challenges of semi-supervised learning and improve the segmentation accuracy of semi-supervised models on medical images. MethodsTo address these issues, we propose an Uncertainty-based Region Clipping Algorithm for semi-supervised medical image segmentation, which consists of two main modules. A module is introduced to compute the uncertainty of two sub-networks predictions with diversity using Monte Carlo Dropout, allowing the model to gradually learn from more reliable targets. To retain model diversity, we use different loss functions for different branches and use Non-Maximum Suppression in one of the branches. The other module is proposed to generate new samples by masking the low-confidence pixels in the original image based on uncertainty information. New samples are fed into the model to facilitate the model to generate pseudo labels with high confidence and enlarge the training data distribution. ResultsComprehensive experiments on the combination of two benchmarks ACDC and BraTS2019 show that our proposed model outperforms state-of-the-art methods in terms of Dice, HD95 and ASD. The results reach an average Dice score of 87.86 % and a HD95 score of 4.214 mm on ACDC dataset. For the brain tumor segmentation, the results reach an average Dice score of 84.79 % and a HD score of 10.13 mm. ConclusionsOur proposed method improves the accuracy of semi-supervised medical image segmentation. Extensive experiments on two public medical image datasets including 2D and 3D modalities demonstrate the superiority of our model. The code is available at: https://github.com/QuintinDong/URCA.

A deep learning model for generating [18F]FDG PET Images from early-phase [18F]Florbetapir and [18F]Flutemetamol PET images

IntroductionAmyloid-β (Aβ) plaques is a significant hallmark of Alzheimer's disease (AD), detectable via amyloid-PET imaging. The Fluorine-18-Fluorodeoxyglucose ([18F]FDG) PET scan tracks cerebral glucose metabolism, correlated with synaptic dysfunction and disease progression and is complementary for AD diagnosis. Dual-scan acquisitions of amyloid PET allows the possibility to use early-phase amyloid-PET as a biomarker for neurodegeneration, proven to have a good correlation to [18F]FDG PET. The aim of this study was to evaluate the added value of synthesizing the later from the former through deep learning (DL), aiming at reducing the number of PET scans, radiation dose, and discomfort to patients.MethodsA total of 166 subjects including cognitively unimpaired individuals (N = 72), subjects with mild cognitive impairment (N = 73) and dementia (N = 21) were included in this study. All underwent T1-weighted MRI, dual-phase amyloid PET scans using either Fluorine-18 Florbetapir ([18F]FBP) or Fluorine-18 Flutemetamol ([18F]FMM), and an [18F]FDG PET scan. Two transformer-based DL models called SwinUNETR were trained separately to synthesize the [18F]FDG from early phase [18F]FBP and [18F]FMM (eFBP/eFMM). A clinical similarity score (1: no similarity to 3: similar) was assessed to compare the imaging information obtained by synthesized [18F]FDG as well as eFBP/eFMM to actual [18F]FDG. Quantitative evaluations include region wise correlation and single-subject voxel-wise analyses in comparison with a reference [18F]FDG PET healthy control database. Dice coefficients were calculated to quantify the whole-brain spatial overlap between hypometabolic ([18F]FDG PET) and hypoperfused (eFBP/eFMM) binary maps at the single-subject level as well as between [18F]FDG PET and synthetic [18F]FDG PET hypometabolic binary maps.ResultsThe clinical evaluation showed that, in comparison to eFBP/eFMM (average of clinical similarity score (CSS) = 1.53), the synthetic [18F]FDG images are quite similar to the actual [18F]FDG images (average of CSS = 2.7) in terms of preserving clinically relevant uptake patterns. The single-subject voxel-wise analyses showed that at the group level, the Dice scores improved by around 13% and 5% when using the DL approach for eFBP and eFMM, respectively. The correlation analysis results indicated a relatively strong correlation between eFBP/eFMM and [18F]FDG (eFBP: slope = 0.77, R2 = 0.61, P-value < 0.0001); eFMM: slope = 0.77, R2 = 0.61, P-value < 0.0001). This correlation improved for synthetic [18F]FDG (synthetic [18F]FDG generated from eFBP (slope = 1.00, R2 = 0.68, P-value < 0.0001), eFMM (slope = 0.93, R2 = 0.72, P-value < 0.0001)).ConclusionWe proposed a DL model for generating the [18F]FDG from eFBP/eFMM PET images. This method may be used as an alternative for multiple radiotracer scanning in research and clinical settings allowing to adopt the currently validated [18F]FDG PET normal reference databases for data analysis.

Cellpose as a reliable method for single-cell segmentation of autofluorescence microscopy images.

Autofluorescence microscopy uses intrinsic sources of molecular contrast to provide cellular-level information without extrinsic labels. However, traditional cell segmentation tools are often optimized for high signal-to-noise ratio (SNR) images, such as fluorescently labeled cells, and unsurprisingly perform poorly on low SNR autofluorescence images. Therefore, new cell segmentation tools are needed for autofluorescence microscopy. Cellpose is a deep learning network that is generalizable across diverse cell microscopy images and automatically segments single cells to improve throughput and reduce inter-human biases. This study aims to validate Cellpose for autofluorescence imaging, specifically from multiphoton intensity images of NAD(P)H. Manually segmented nuclear masks of NAD(P)H images were used to train new Cellpose models. These models were applied to PANC-1 cells treated with metabolic inhibitors and patient-derived cancer organoids (across 9 patients) treated with chemotherapies. These datasets include co-registered fluorescence lifetime imaging microscopy (FLIM) of NAD(P)H and FAD, so fluorescence decay parameters and the optical redox ratio (ORR) were compared between masks generated by the new Cellpose model and manual segmentation. The Dice score between repeated manually segmented masks was significantly lower than that of repeated Cellpose masks (p<0.0001) indicating greater reproducibility between Cellpose masks. There was also a high correlation (R 2 >0.9) between Cellpose and manually segmented masks for the ORR, mean NAD(P)H lifetime, and mean FAD lifetime across 2D and 3D cell culture treatment conditions. Masks generated from Cellpose and manual segmentation also maintain similar means, variances, and effect sizes between treatments for the ORR and FLIM parameters. Overall, Cellpose provides a fast, reliable, reproducible, and accurate method to segment single cells in autofluorescence microscopy images such that functional changes in cells are accurately captured in both 2D and 3D culture.

Deep learning automatic semantic segmentation of glioblastoma multiforme regions on multimodal magnetic resonance images.

In patients having naïve glioblastoma multiforme (GBM), this study aims to assess the efficacy of Deep Learning algorithms in automating the segmentation of brain magnetic resonance (MR) images to accurately determine 3D masks for 4 distinct regions: enhanced tumor, peritumoral edema, non-enhanced/necrotic tumor, and total tumor. A 3D U-Net neural network algorithm was developed for semantic segmentation of GBM. The training dataset was manually delineated by a group of expert neuroradiologists on MR images from the Brain Tumor Segmentation Challenge 2021 (BraTS2021) image repository, as ground truth labels for diverse glioma (GBM and low-grade glioma) subregions across four MR sequences (T1w, T1w-contrast enhanced, T2w, and FLAIR) in 1251 patients. The in-house test was performed on 50GBM patients from our cohort (PerProGlio project). By exploring various hyperparameters, the network's performance was optimized, and the most optimal parameter configuration was identified. The assessment of the optimized network's performance utilized Dice scores, precision, and sensitivity metrics. Our adaptation of the 3D U-net with additional residual blocks demonstrated reliable performance on both the BraTS2021 dataset and the in-house PerProGlio cohort, employing only T1w-ce sequences for enhancement and non-enhanced/necrotic tumor models and T1w-ce + T2w + FLAIR for peritumoral edema and total tumor. The mean Dice scores (training and test) were 0.89 and 0.75; 0.75 and 0.64; 0.79 and 0.71; and 0.60 and 0.55, for total tumor, edema, enhanced tumor, and non-enhanced/necrotic tumor, respectively. The results underscore the high precision with which our network can effectively segment GBM tumors and their distinct subregions. The level of accuracy achieved agrees with the coefficients recorded in previous GBM studies. In particular, our approach allows model specialization for each of the different tumor subregions employing only those MR sequences that provide value for segmentation.

Geometric deep learning for enhanced quantitative analysis of microstructures in X-ray computed tomography data

Quantitative microstructural analysis of XCT 3D images is key for quality assurance of materials and components. In this paper we implement a Graph Convolutional Neural Network (GCNN) architecture to segment a complex Al-Si Metal Matrix composite XCT volume (3D image). We train the model on a synthetic dataset and we assess its performance on both synthetic and experimental, manually-labeled, datasets. Our simple GCNN shows a comparable performance, measured via the Dice score, to more standard machine learning methods, but uses a greatly reduced number of parameters (less than 1/10 of parameters), features low training time, and needs little hardware resources. Our GCNN thus achieves a cost-effective reliable segmentation.

An automated two-stage approach to kidney and tumor segmentation in CT imaging.

The incidence of kidney tumors is progressively increasing each year. The precision of segmentation for kidney tumors is crucial for diagnosis and treatment. To enhance accuracy and reduce manual involvement, propose a deep learning-based method for the automatic segmentation of kidneys and kidney tumors in CT images. The proposed method comprises two parts: object detection and segmentation. We first use a model to detect the position of the kidney, then narrow the segmentation range, and finally use an attentional recurrent residual convolutional network for segmentation. Our model achieved a kidney dice score of 0.951 and a tumor dice score of 0.895 on the KiTS19 dataset. Experimental results show that our model significantly improves the accuracy of kidney and kidney tumor segmentation and outperforms other advanced methods. The proposed method provides an efficient and automatic solution for accurately segmenting kidneys and renal tumors on CT images. Additionally, this study can assist radiologists in assessing patients' conditions and making informed treatment decisions.

Automated intracranial vessel segmentation of 4D flow MRI data in patients with atherosclerotic stenosis using a convolutional neural network.

Intracranial 4D flow MRI enables quantitative assessment of hemodynamics in patients with intracranial atherosclerotic disease (ICAD). However, quantitative assessments are still challenging due to the time-consuming vessel segmentation, especially in the presence of stenoses, which can often result in user variability. To improve the reproducibility and robustness as well as to accelerate data analysis, we developed an accurate, fully automated segmentation for stenosed intracranial vessels using deep learning. 154 dual-VENC 4D flow MRI scans (68 ICAD patients with stenosis, 86 healthy controls) were retrospectively selected. Manual segmentations were used as ground truth for training. For automated segmentation, deep learning was performed using a 3D U-Net. 20 randomly selected cases (10 controls, 10 patients) were separated and solely used for testing. Cross-sectional areas and flow parameters were determined in the Circle of Willis (CoW) and the sinuses. Furthermore, the flow conservation error was calculated. For statistical comparisons, Dice scores (DS), Hausdorff distance (HD), average symmetrical surface distance (ASSD), Bland-Altman analyses, and interclass correlations were computed using the manual segmentations from two independent observers as reference. Finally, three stenosis cases were analyzed in more detail by comparing the 4D flow-based segmentations with segmentations from black blood vessel wall imaging (VWI). Training of the network took approximately 10 h and the average automated segmentation time was 2.2 ± 1.0 s. No significant differences in segmentation performance relative to two independent observers were observed. For the controls, mean DS was 0.85 ± 0.03 for the CoW and 0.86 ± 0.06 for the sinuses. Mean HD was 7.2 ± 1.5 mm (CoW) and 6.6 ± 3.7 mm (sinuses). Mean ASSD was 0.15 ± 0.04 mm (CoW) and 0.22 ± 0.17 mm (sinuses). For the patients, the mean DS was 0.85 ± 0.04 (CoW) and 0.82 ± 0.07 (sinuses), the HD was 8.4 ± 3.1 mm (CoW) and 5.7 ± 1.9 mm (sinuses) and the mean ASSD was 0.22 ± 0.10 mm (CoW) and 0.22 ± 0.11 mm (sinuses). Small bias and limits of agreement were observed in both cohorts for the flow parameters. The assessment of the cross-sectional lumen areas in stenosed vessels revealed very good agreement (ICC: 0.93) with the VWI segmentation but a consistent overestimation (bias ± LOA: 28.1 ± 13.9%). Deep learning was successfully applied for fully automated segmentation of stenosed intracranial vasculatures using 4D flow MRI data. The statistical analysis of segmentation and flow metrics demonstrated very good agreement between the CNN and manual segmentation and good performance in stenosed vessels. To further improve the performance and generalization, more ICAD segmentations as well as other intracranial vascular pathologies will be considered in the future.

Coronary Vessel Segmentation in X-ray Angiography Images Using Edge-Based Tracking Method

Automated extraction of coronary arteries is an essential process in the diagnosis of treatment for coronary artery disease (CAD) with computer assistance. Accurately outlining the coronary artery is difficult when using X-ray coronary angiography (XCA) because of the low signal-to-noise ratio and the presence of interfering background structures. In this paper, a new approach for segmenting vessels in angiograms is presented, specifically designed to tackle the difficulties arising from non-uniform illumination, artifacts, and noise present in angiographic images. The proposed method employs an edge-based tracking tool to generate an initial probability map for segmentation. A segmentation method based on coronary vessel tracking is presented for finding the border and centerline of the vessel. The proposed method is designed based on two main components: preprocessing and tracking. In the preprocessing stage, a guided filter and edge-sharpening algorithms are used to enhance the features of the original image. In the tracking stage, an initial point is selected, and using the Gaussian property, a semi-circle operator is applied to track the line perpendicular to the vessel. The proposed method demonstrated remarkable performance in terms of sensitivity and specificity, achieving values of 86.93 and 99.61, respectively. Additionally, the method achieved an accuracy rate of 97.81. Notably, the proposed method outperformed existing state-of-the-art segmentation methods, as indicated by its higher dice score. These impressive results signify a significant advancement in the field of vessel segmentation, highlighting the effectiveness and superiority of the proposed approach.

A Comparison of CT-Based Pancreatic Segmentation Deep Learning Models

Rationale and ObjectivesPancreas segmentation accuracy at CT is critical for the identification of pancreatic pathologies and is essential for the development of imaging biomarkers. Our objective was to benchmark the performance of five high-performing pancreas segmentation models across multiple metrics stratified by scan and patient/pancreatic characteristics that may affect segmentation performance. Materials and MethodsIn this retrospective study, PubMed and ArXiv searches were conducted to identify pancreas segmentation models which were then evaluated on a set of annotated imaging datasets. Results (Dice score, Hausdorff distance [HD], average surface distance [ASD]) were stratified by contrast status and quartiles of peri-pancreatic attenuation (5mm region around pancreas). Multivariate regression was performed to identify imaging characteristics and biomarkers (n=9) that were significantly associated with Dice score. ResultsFive pancreas segmentation models were identified: Abdomen Atlas [AAUNet, AASwin, trained on 8448 scans], TotalSegmentator [TS, 1204 scans], nnUNetv1 [MSD-nnUNet, 282 scans], and a U-Net based model for predicting diabetes [DM-UNet, 427 scans]. These were evaluated on 352 CT scans (30 females, 25 males, age 58±7 years old [±1SD], 327 age unknown) from 2000-2023. Overall, TS, AAUNet, and AASwin were the best performers, Dice=80±11%, 79±16%, and 77±18% respectively (pairwise Sidak test not-significantly different). AASwin and MSD-nnUNet performed worse (for all metrics) on non-contrast scans (vs contrast, P<.001). The worst performer was DM-UNet (Dice=67±16%). All algorithms except TS showed lower Dice scores with increasing peri-pancreatic attenuation (P<.01). Multivariate regression showed non-contrast scans, (P<.001; MSD-nnUNet), smaller pancreatic length (P=.005, MSD-nnUNet), and height (P=.003, DM-UNet) were associated with lower Dice scores. ConclusionThe convolutional neural network-based models trained on a diverse set of scans performed best (TS, AAUnet, and AASwin). TS performed equivalently to AAUnet and AASwin with only 13% of the training set size (8488 vs 1204 scans). Though trained on the same dataset, a transformer network (AASwin) had poorer performance on non-contrast scans whereas its convolutional network counterpart (AAUNet) did not. This study highlights how aggregate assessment metrics of pancreatic segmentation algorithms seen in other literature are not enough to capture differential performance across common patient and scanning characteristics in clinical populations.

From quantitative metrics to clinical success: assessing the utility of deep learning for tumor segmentation in breast surgery.

Preventing positive margins is essential for ensuring favorable patient outcomes following breast-conserving surgery (BCS). Deep learning has the potential to enable this by automatically contouring the tumor and guiding resection in real time. However, evaluation of such models with respect to pathology outcomes is necessary for their successful translation into clinical practice. Sixteen deep learning models based on established architectures in the literature are trained on 7318 ultrasound images from 33 patients. Models are ranked by an expert based on their contours generated from images in our test set. Generated contours from each model are also analyzed using recorded cautery trajectories of five navigated BCS cases to predict margin status. Predicted margins are compared with pathology reports. The best-performing model using both quantitative evaluation and our visual ranking framework achieved a mean Dice score of 0.959. Quantitative metrics are positively associated with expert visual rankings. However, the predictive value of generated contours was limited with a sensitivity of 0.750 and a specificity of 0.433 when tested against pathology reports. We present a clinical evaluation of deep learning models trained for intraoperative tumor segmentation in breast-conserving surgery. We demonstrate that automatic contouring is limited in predicting pathology margins despite achieving high performance on quantitative metrics.

Use of an AI-powered visualization platform to support precision clinical decision-making in patients with breast cancer undergoing neoadjuvant chemotherapy.

e12618 Background: Chemotherapy to reduce cancer burden and tumor size ("downstaging") is widely used in neoadjuvant treatment of early-stage breast cancer (ESBC) and can have a dramatic impact on selection of further treatment courses, including surgery. The resulting decision-making process between physicians and cancer patients requires detailed and effective communication of the patient's disease status, and clear presentation of this information in Tumor Board settings. Artificial intelligence (AI) -powered visualization platforms offer the potential to aid in evaluating breast tumor location and morphology, and accordingly optimize treatment. Additionally, these tools offer enhanced communication with patients regarding future treatment options, particularly serving as clinical decision support (CDS) tools. Methods: We assessed the performance of an AI platform that employs dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to render three-dimensional (3D) visualizations of breast tumors and 5 additional chest tissues on an independent validation dataset comprise of 100 ESBC cases with ground-truth tumor labels vetted by 2 breast-specialized radiologists. The platform provides 3D depictions of the tumor and its multi-tissue context, as well as functionalities for quantifying tumor morphology, tumor-to-landmark structure distances, excision volumes, and approximate surgical margins. We subsequently evaluated the tool's utility in a real-world setting to understand how the tool impacts clinical decision-making and patient communication, specifically around surgical planning in the context of neoadjuvant chemotherapy (NACT). Results: The platform generated accurate 3D tumor maps, as evidenced by a median Dice score of 75.6% and a median Hausdorff distances of 15.1 mm relative to the radiologist validation set. The automatic longest tumor dimension calculations were linearly correlated with radiologist direct measurements (Pearson correlation coefficient = 0.78) and had a median absolute difference of 6.3 mm. We found strong overlap between the platform’s ground-truth convex hulls, as shown by a median Dice score of 81.5%. When used by a skilled breast-specialized surgeon, the tool enabled rapid and intuitive patient-physician communication of the location and extent of disease, and potential degrees of downstaging to breast-conserving surgery. Conclusions: Detailed 3D tumor and surrounding multi-tissue depictions offer both qualitative and quantitative comprehension of cancer topology and may aid in formulating an optimal treatment approach for breast cancer treatment. We show that detailed depictions enhance communication between physicians and patients and may aid in streamlining decision-making.

GAIR-U-Net: 3D guided attention inception residual u-net for brain tumor segmentation using multimodal MRI images

Deep learning technologies have led to substantial breakthroughs in the field of biomedical image analysis. Accurate brain tumor segmentation is an essential aspect of treatment planning. Radiologists agree that manual segmentation is a difficult and time-consuming task that frequently causes delays in the diagnosing process. While U-Net-based methods have been widely used for brain tumor segmentation, many challenges persist, particularly when dealing with tumors of varying sizes, locations, and shapes. Additionally, segmenting tumor regions with structures requires a comprehensive model, which can increase computational complexity and potentially cause gradient vanishing issues. This study presents a novel method called 3D Guided Attention-based deep Inception Residual U-Net (GAIR-U-Net) to address these challenges. This model combines attention mechanisms, an inception module, and residual blocks with dilated convolution to enhance feature representation and spatial context understanding. The backbone of the model is the U-Net model, which leverages the power of inception and residual connections to capture intricate patterns and hierarchical features while expanding the model’s width in three-dimensional space without significantly increasing computational complexity. The attention mechanisms play a role in focusing on important regions and areas while downgrading irrelevant details. The dilated convolutions in the network help in learning both local and global information, improving accuracy and adaptability in segmenting tumors. All the experiments in this study were carried out on multimodal MRI scans that include (T1-weighted, T1-ce, T2-weighted, and FLAIR sequences) from the BraTS 2020 dataset. The presented model is trained and tested on the same dataset, which exhibited promising performance compared to previous methods. On the BraTS 2020 validation dataset, the proposed model obtained a dice score of 0.8796, 0.8634, and 0.8441 for whole tumor (WT), tumor core (TC), and enhancing tumor (ET), respectively. These results demonstrate the model’s efficacy in precisely segmenting brain tumors across various modalities. Comparative analyses underscore the model’s versatility in handling tumor shape variations, size, and location, making it a promising solution for clinical applications.