MSU-Net: Multi-Scale U-Net for 2D Medical Image Segmentation

A review of deep learning-based reconstruction methods for accelerated MRI using spatiotemporal and multi-contrast redundancies

Accelerated magnetic resonance imaging (MRI) has played an essential role in reducing data acquisition time for MRI. Acceleration can be achieved by acquiring fewer data points in k-space, which results in various artifacts in the image domain. Conventional reconstruction methods have resolved the artifacts by utilizing multi-coil information, but with limited robustness. Recently, numerous deep learning-based reconstruction methods have been developed, enabling outstanding reconstruction performances with higher acceleration. Advances in hardware and developments of specialized network architectures have produced such achievements. Besides, MRI signals contain various redundant information including multi-coil redundancy, multi-contrast redundancy, and spatiotemporal redundancy. Utilization of the redundant information combined with deep learning approaches allow not only higher acceleration, but also well-preserved details in the reconstructed images. Consequently, this review paper introduces the basic concepts of deep learning and conventional accelerated MRI reconstruction methods, followed by review of recent deep learning-based reconstruction methods that exploit various redundancies. Lastly, the paper concludes by discussing the challenges, limitations, and potential directions of future developments.

Next-Gen Medical Imaging: U-Net Evolution and the Rise of Transformers

The advancement of medical imaging has profoundly impacted our understanding of the human body and various diseases. It has led to the continuous refinement of related technologies over many years. Despite these advancements, several challenges persist in the development of medical imaging, including data shortages characterized by low contrast, high noise levels, and limited image resolution. The U-Net architecture has significantly evolved to address these challenges, becoming a staple in medical imaging due to its effective performance and numerous updated versions. However, the emergence of Transformer-based models marks a new era in deep learning for medical imaging. These models and their variants promise substantial progress, necessitating a comparative analysis to comprehend recent advancements. This review begins by exploring the fundamental U-Net architecture and its variants, then examines the limitations encountered during its evolution. It then introduces the Transformer-based self-attention mechanism and investigates how modern models incorporate positional information. The review emphasizes the revolutionary potential of Transformer-based techniques, discusses their limitations, and outlines potential avenues for future research.

CapNet: An Automatic Attention-Based with Mixer Model for Cardiovascular Magnetic Resonance Image Segmentation

Deep neural networks have shown excellent performance in medical image segmentation, especially for cardiac images. Transformer-based models, though having advantages over convolutional neural networks due to the ability of long-range dependence learning, still have shortcomings such as having a large number of parameters and and high computational cost. Additionally, for better results, they are often pretrained on a larger data, thus requiring large memory size and increasing resource expenses. In this study, we propose a new lightweight but efficient model, namely CapNet, based on convolutions and mixing modules for cardiac segmentation from magnetic resonance images (MRI) that can be trained from scratch with a small amount of parameters. To handle varying sizes and shapes which often occur in cardiac systolic and diastolic phases, we propose attention modules for pooling, spatial, and channel information. We also propose a novel loss called the Tversky Shape Power Distance function based on the shape dissimilarity between labels and predictions that shows promising performances compared to other losses. Experiments on three public datasets including ACDC benchmark, Sunnybrook data, and MS-CMR challenge are conducted and compared with other state of the arts (SOTA). For binary segmentation, the proposed CapNet obtained the Dice similarity coefficient (DSC) of 94% and 95.93% for respectively the Endocardium and Epicardium regions with Sunnybrook dataset, 94.49% for Endocardium, and 96.82% for Epicardium with the ACDC data. Regarding the multiclass case, the average DSC by CapNet is 93.05% for the ACDC data; and the DSC scores for the MS-CMR are 94.59%, 92.22%, and 93.99% for respectively the bSSFP, T2-SPAIR, and LGE sequences of the MS-CMR. Moreover, the statistical significance analysis tests with p-value < 0.05 compared with transformer-based methods and some CNN-based approaches demonstrated that the CapNet, though having fewer training parameters, is statistically significant. The promising evaluation metrics show comparative results in both Dice and IoU indices compared to SOTA CNN-based and Transformer-based architectures.

MADR-Net: multi-level attention dilated residual neural network for segmentation of medical images

Medical image segmentation has made a significant contribution towards delivering affordable healthcare by facilitating the automatic identification of anatomical structures and other regions of interest. Although convolution neural networks have become prominent in the field of medical image segmentation, they suffer from certain limitations. In this study, we present a reliable framework for producing performant outcomes for the segmentation of pathological structures of 2D medical images. Our framework consists of a novel deep learning architecture, called deep multi-level attention dilated residual neural network (MADR-Net), designed to improve the performance of medical image segmentation. MADR-Net uses a U-Net encoder/decoder backbone in combination with multi-level residual blocks and atrous pyramid scene parsing pooling. To improve the segmentation results, channel-spatial attention blocks were added in the skip connection to capture both the global and local features and superseded the bottleneck layer with an ASPP block. Furthermore, we introduce a hybrid loss function that has an excellent convergence property and enhances the performance of the medical image segmentation task. We extensively validated the proposed MADR-Net on four typical yet challenging medical image segmentation tasks: (1) Left ventricle, left atrium, and myocardial wall segmentation from Echocardiogram images in the CAMUS dataset, (2) Skin cancer segmentation from dermoscopy images in ISIC 2017 dataset, (3) Electron microscopy in FIB-SEM dataset, and (4) Fluid attenuated inversion recovery abnormality from MR images in LGG segmentation dataset. The proposed algorithm yielded significant results when compared to state-of-the-art architectures such as U-Net, Residual U-Net, and Attention U-Net. The proposed MADR-Net consistently outperformed the classical U-Net by 5.43%, 3.43%, and 3.92% relative improvement in terms of dice coefficient, respectively, for electron microscopy, dermoscopy, and MRI. The experimental results demonstrate superior performance on single and multi-class datasets and that the proposed MADR-Net can be utilized as a baseline for the assessment of cross-dataset and segmentation tasks.

Automatic Segmentation of Membranous Glottal Gap Area with U-Net-Based Architecture

BACKGROUND

While videostroboscopy is recognized as the most popular approach for investigating vocal fold function, evaluating the numerical values, such as the membranous glottal gap area, remains too time consuming for clinical applications.

METHODS

We used a total of 2507 videostroboscopy images from 137 patients and developed five U-Net-based deep-learning image segmentation models for automatic masking of the membranous glottal gap area. To further validate the models, we used another 410 images from 41 different patients.

RESULTS

During development, all five models exhibited acceptable and similar metrics. While the VGG19 U-Net had a long inference time of 1654 ms, the other four models had more practical inference times, ranging from 16 to 138 ms. During further validation, Efficient U-Net demonstrated the highest intersection over union of 0.8455, the highest Dice coefficient of 0.9163, and the lowest Hausdorff distance of 1.5626. The normalized membranous glottal gap area index was also calculated and validated. Efficient U-Net and VGG19 U-Net exhibited the lowest mean squared errors (3.5476 and 3.3842) and the lowest mean absolute errors (1.8835 and 1.8396).

CONCLUSIONS

Automatic segmentation of the membranous glottal gap area can be achieved through U-net-based architecture. Considering the segmentation quality and speed, Efficient U-Net is a reasonable choice for this task, while the other four models remain valuable competitors. The models' masked area enables possible calculation of the normalized membranous glottal gap area and analysis of the glottal area waveform, revealing promising clinical applications for this model.

LEVEL OF EVIDENCE

NA Laryngoscope, 134:2835-2843, 2024.

MRUNet-3D: A multi-stride residual 3D UNet for lung nodule segmentation

Obtaining an accurate segmentation of the pulmonary nodules in computed tomography (CT) images is challenging. This is due to: (1) the heterogeneous nature of the lung nodules; (2) comparable visual characteristics between the nodules and their surroundings. A robust multi-scale feature extraction mechanism that can effectively obtain multi-scale representations at a granular level can improve segmentation accuracy. As the most commonly used network in lung nodule segmentation, UNet, its variants, and other image segmentation methods lack this robust feature extraction mechanism. In this study, we propose a multi-stride residual 3D UNet (MRUNet-3D) to improve the segmentation accuracy of lung nodules in CT images. It incorporates a multi-slide Res2Net block (MSR), which replaces the simple sequence of convolution layers in each encoder stage to effectively extract multi-scale features at a granular level from different receptive fields and resolutions while conserving the strengths of 3D UNet. The proposed method has been extensively evaluated on the publicly available LUNA16 dataset. Experimental results show that it achieves competitive segmentation performance with an average dice similarity coefficient of 83.47 % and an average surface distance of 0.35 mm on the dataset. More notably, our method has proven to be robust to the heterogeneity of lung nodules. It has also proven to perform better at segmenting small lung nodules. Ablation studies have shown that the proposed MSR and RFIA modules are fundamental to improving the performance of the proposed model.

MSDEnet: Multi-scale detail enhanced network based on human visual system for medical image segmentation

In medical image segmentation, accuracy is commonly high for tasks involving clear boundary partitioning features, as seen in the segmentation of X-ray images. However, for objects with less obvious boundary partitioning features, such as skin regions with similar color textures or CT images of adjacent organs with similar Hounsfield value ranges, segmentation accuracy significantly decreases. Inspired by the human visual system, we proposed the multi-scale detail enhanced network. Firstly, we designed a detail enhanced module to enhance the contrast between central and peripheral receptive field information using the superposition of two asymmetric convolutions in different directions and a standard convolution. Then, we expanded the scale of the module into a multi-scale detail enhanced module. The difference between central and peripheral information at different scales makes the network more sensitive to changes in details, resulting in more accurate segmentation. In order to reduce the impact of redundant information on segmentation results and increase the effective receptive field, we proposed the channel multi-scale module, adapted from the Res2net module. This creates independent parallel multi-scale branches within a single residual structure, increasing the utilization of redundant information and the effective receptive field at the channel level. We conducted experiments on four different datasets, and our method outperformed the common medical image segmentation algorithms currently being used. Additionally, we carried out detailed ablation experiments to confirm the effectiveness of each module.

HIMS-Net: Horizontal-vertical interaction and multiple side-outputs network for cyst segmentation in jaw images

Jaw cysts are mainly caused by abnormal tooth development, chronic oral inflammation, or jaw damage, which may lead to facial swelling, deformity, tooth loss, and other symptoms. Due to the diversity and complexity of cyst images, deep-learning algorithms still face many difficulties and challenges. In response to these problems, we present a horizontal-vertical interaction and multiple side-outputs network for cyst segmentation in jaw images. First, the horizontal-vertical interaction mechanism facilitates complex communication paths in the vertical and horizontal dimensions, and it has the ability to capture a wide range of context dependencies. Second, the feature-fused unit is introduced to adjust the network's receptive field, which enhances the ability of acquiring multi-scale context information. Third, the multiple side-outputs strategy intelligently combines feature maps to generate more accurate and detailed change maps. Finally, experiments were carried out on the self-established jaw cyst dataset and compared with different specialist physicians to evaluate its clinical usability. The research results indicate that the Matthews correlation coefficient (Mcc), Dice, and Jaccard of HIMS-Net were 93.61, 93.66 and 88.10% respectively, which may contribute to rapid and accurate diagnosis in clinical practice.

CSAM: A 2.5D Cross-Slice Attention Module for Anisotropic Volumetric Medical Image Segmentation

A large portion of volumetric medical data, especially magnetic resonance imaging (MRI) data, is anisotropic, as the through-plane resolution is typically much lower than the in-plane resolution. Both 3D and purely 2D deep learning-based segmentation methods are deficient in dealing with such volumetric data since the performance of 3D methods suffers when confronting anisotropic data, and 2D methods disregard crucial volumetric information. Insufficient work has been done on 2.5D methods, in which 2D convolution is mainly used in concert with volumetric information. These models focus on learning the relationship across slices, but typically have many parameters to train. We offer a Cross-Slice Attention Module (CSAM) with minimal trainable parameters, which captures information across all the slices in the volume by applying semantic, positional, and slice attention on deep feature maps at different scales. Our extensive experiments using different network architectures and tasks demonstrate the usefulness and generalizability of CSAM. Associated code is available at https://github.com/aL3x-O-o-Hung/CSAM.

Efficient semi-supervised semantic segmentation of electron microscopy cancer images with sparse annotations

Electron microscopy (EM) enables imaging at a resolution of nanometers and can shed light on how cancer evolves to develop resistance to therapy. Acquiring these images has become a routine task.However, analyzing them is now a bottleneck, as manual structure identification is very time-consuming and can take up to several months for a single sample. Deep learning approaches offer a suitable solution to speed up the analysis. In this work, we present a study of several state-of-the-art deep learning models for the task of segmenting nuclei and nucleoli in volumes from tumor biopsies. We compared previous results obtained with the ResUNet architecture to the more recent UNet++, FracTALResNet, SenFormer, and CEECNet models. In addition, we explored the utilization of unlabeled images through semi-supervised learning with Cross Pseudo Supervision. We have trained and evaluated all of the models on sparse manual labels from three fully annotated in-house datasets that we have made available on demand, demonstrating improvements in terms of 3D Dice score. From the analysis of these results, we drew conclusions on the relative gains of using more complex models, and semi-supervised learning as well as the next steps for the mitigation of the manual segmentation bottleneck.

Development of a deep learning based image processing tool for enhanced organoid analysis

Contrary to 2D cells, 3D organoid structures are composed of diverse cell types and exhibit morphologies of various sizes. Although researchers frequently monitor morphological changes, analyzing every structure with the naked eye is difficult. Given that deep learning (DL) has been used for 2D cell image segmentation, a trained DL model may assist researchers in organoid image recognition and analysis. In this study, we developed OrgaExtractor, an easy-to-use DL model based on multi-scale U-Net, to perform accurate segmentation of organoids of various sizes. OrgaExtractor achieved an average dice similarity coefficient of 0.853 from a post-processed output, which was finalized with noise removal. Correlation between CellTiter-Glo assay results and daily measured organoid images shows that OrgaExtractor can reflect the actual organoid culture conditions. The OrgaExtractor data can be used to determine the best time point for organoid subculture on the bench and to maintain organoids in the long term.

PMJAF-Net: Pyramidal multi-scale joint attention and adaptive fusion network for explainable skin lesion segmentation

Traditional convolutional neural networks have achieved remarkable success in skin lesion segmentation. However, the successive pooling operations and convolutional spans reduce the feature resolution and hinder the dense prediction for spatial information, resulting in blurred boundaries, low accuracy and poor interpretability for irregular lesion segmentation under low contrast. To solve the above issues, a pyramidal multi-scale joint attention and adaptive fusion network for explainable (PMJAF-Net) skin lesion segmentation is proposed. Firstly, an adaptive spatial attention module is designed to establish the long-term correlation between pixels, enrich the global and local contextual information, and refine the detailed features. Subsequently, an efficient pyramidal multi-scale channel attention module is proposed to capture the multi-scale information and edge features by using the pyramidal module. Meanwhile, a channel attention module is devised to establish the long-term correlation between channels and highlight the most related feature channels to capture the multi-scale key information on each channel. Thereafter, a multi-scale adaptive fusion attention module is put forward to efficiently fuse the scale features at different decoding stages. Finally, a novel hybrid loss function based on region salient features and boundary quality is presented to guide the network to learn from map-level, patch-level and pixel-level and to accurately predict the lesion regions with clear boundaries. In addition, visualizing attention weight maps are utilized to visually enhance the interpretability of our proposed model. Comprehensive experiments are conducted on four public skin lesion datasets, and the results demonstrate that the proposed network outperforms the state-of-the-art methods, with the segmentation assessment evaluation metrics Dice, JI, and ACC improved to 92.65%, 87.86% and 96.26%, respectively.

Multiscale deep learning framework captures systemic immune features in lymph nodes predictive of triple negative breast cancer outcome in large-scale studies

The suggestion that the systemic immune response in lymph nodes (LNs) conveys prognostic value for triple-negative breast cancer (TNBC) patients has not previously been investigated in large cohorts. We used a deep learning (DL) framework to quantify morphological features in haematoxylin and eosin-stained LNs on digitised whole slide images. From 345 breast cancer patients, 5,228 axillary LNs, cancer-free and involved, were assessed. Generalisable multiscale DL frameworks were developed to capture and quantify germinal centres (GCs) and sinuses. Cox regression proportional hazard models tested the association between smuLymphNet-captured GC and sinus quantifications and distant metastasis-free survival (DMFS). smuLymphNet achieved a Dice coefficient of 0.86 and 0.74 for capturing GCs and sinuses, respectively, and was comparable to an interpathologist Dice coefficient of 0.66 (GC) and 0.60 (sinus). smuLymphNet-captured sinuses were increased in LNs harbouring GCs (p < 0.001). smuLymphNet-captured GCs retained clinical relevance in LN-positive TNBC patients whose cancer-free LNs had on average ≥2 GCs, had longer DMFS (hazard ratio [HR] = 0.28, p = 0.02) and extended GCs' prognostic value to LN-negative TNBC patients (HR = 0.14, p = 0.002). Enlarged smuLymphNet-captured sinuses in involved LNs were associated with superior DMFS in LN-positive TNBC patients in a cohort from Guy's Hospital (multivariate HR = 0.39, p = 0.039) and with distant recurrence-free survival in 95 LN-positive TNBC patients of the Dutch-N4plus trial (HR = 0.44, p = 0.024). Heuristic scoring of subcapsular sinuses in LNs of LN-positive Tianjin TNBC patients (n = 85) cross-validated the association of enlarged sinuses with shorter DMFS (involved LNs: HR = 0.33, p = 0.029 and cancer-free LNs: HR = 0.21 p = 0.01). Morphological LN features reflective of cancer-associated responses are robustly quantifiable by smuLymphNet. Our findings further strengthen the value of assessment of LN properties beyond the detection of metastatic deposits for prognostication of TNBC patients. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Predicting Recurrence in Pancreatic Ductal Adenocarcinoma after Radical Surgery Using an AX-Unet Pancreas Segmentation Model and Dynamic Nomogram

This study aims to investigate the reliability of radiomic features extracted from contrast-enhanced computer tomography (CT) by AX-Unet, a pancreas segmentation model, to analyse the recurrence of pancreatic ductal adenocarcinoma (PDAC) after radical surgery. In this study, we trained an AX-Unet model to extract the radiomic features from preoperative contrast-enhanced CT images on a training set of 205 PDAC patients. Then we evaluated the segmentation ability of AX-Unet and the relationship between radiomic features and clinical characteristics on an independent testing set of 64 patients with clear prognoses. The lasso regression analysis was used to screen for variables of interest affecting patients' post-operative recurrence, and the Cox proportional risk model regression analysis was used to screen for risk factors and create a nomogram prediction model. The proposed model achieved an accuracy of 85.9% for pancreas segmentation, meeting the requirements of most clinical applications. Radiomic features were found to be significantly correlated with clinical characteristics such as lymph node metastasis, resectability status, and abnormally elevated serum carbohydrate antigen 19-9 (CA 19-9) levels. Specifically, variance and entropy were associated with the recurrence rate (p < 0.05). The AUC for the nomogram predicting whether the patient recurred after surgery was 0.92 (95% CI: 0.78-0.99) and the C index was 0.62 (95% CI: 0.48-0.78). The AX-Unet pancreas segmentation model shows promise in analysing recurrence risk factors after radical surgery for PDAC. Additionally, our findings suggest that a dynamic nomogram model based on AX-Unet can provide pancreatic oncologists with more accurate prognostic assessments for their patients.

A Multitask Approach for Automated Detection and Segmentation of Thyroid Nodules in Ultrasound Images

An increase in the incidence and diagnosis of thyroid nodules and thyroid cancer underscores the need for a better approach to nodule detection and risk stratification in ultrasound (US) images that can reduce healthcare costs, patient discomfort, and unnecessary invasive procedures. However, variability in ultrasound technique and interpretation makes the diagnostic process partially subjective. Therefore, an automated approach that detects and segments nodules could improve performance on downstream tasks, such as risk stratification.Current deep learning architectures for segmentation are typically semi-automated because they are evaluated solely on images known to have nodules and do not assess ability to identify suspicious images. However, the proposed multitask approach both detects suspicious images and segments potential nodules; this allows for a clinically translatable model that aptly parallels the workflow for thyroid nodule assessment. The multitask approach is centered on an anomaly detection (AD) module that can be integrated with any U-Net architecture variant to improve image-level nodule detection. Ultrasound studies were acquired from 280 patients at UCLA Health, totaling 9,888 images, and annotated by collaborating radiologists. Of the evaluated models, a multi-scale UNet (MSUNet) with AD achieved the highest F1 score of 0.829 and image-wide Dice similarity coefficient of 0.782 on our hold-out test set. Furthermore, models were evaluated on two external validations datasets to demonstrate generalizability and robustness to data variability. Ultimately, the proposed architecture is an automated multitask method that expands on previous methods by successfully both detecting and segmenting nodules in ultrasound.

Prognostic Significance of Tumor-Infiltrating Lymphocytes Determined Using LinkNet on Colorectal Cancer Pathology Images

PURPOSE

Tumor-infiltrating lymphocytes (TILs) have a significant prognostic value in cancers. However, very few automated, deep learning-based TIL scoring algorithms have been developed for colorectal cancer (CRC).

MATERIALS AND METHODS

We developed an automated, multiscale LinkNet workflow for quantifying TILs at the cellular level in CRC tumors using H&E-stained images from the Lizard data set with annotations of lymphocytes. The predictive performance of the automatic TIL scores (TILsLink) for disease progression and overall survival (OS) was evaluated using two international data sets, including 554 patients with CRC from The Cancer Genome Atlas (TCGA) and 1,130 patients with CRC from Molecular and Cellular Oncology (MCO).

RESULTS

The LinkNet model provided outstanding precision (0.9508), recall (0.9185), and overall F1 score (0.9347). Clear continuous TIL-hazard relationships were observed between TILsLink and the risk of disease progression or death in both TCGA and MCO cohorts. Both univariate and multivariate Cox regression analyses for the TCGA data demonstrated that patients with high TIL abundance had a significant (approximately 75%) reduction in risk for disease progression. In both the MCO and TCGA cohorts, the TIL-high group was significantly associated with improved OS in univariate analysis (30% and 54% reduction in risk, respectively). The favorable effects of high TIL levels were consistently observed in different subgroups (classified according to known risk factors).

CONCLUSION

The proposed deep-learning workflow for automatic TIL quantification on the basis of LinkNet can be a useful tool for CRC. TILsLink is likely an independent risk factor for disease progression and carries predictive information of disease progression beyond the current clinical risk factors and biomarkers. The prognostic significance of TILsLink for OS is also evident.

CAT-Net: A Cross-Slice Attention Transformer Model for Prostate Zonal Segmentation in MRI

Prostate cancer is the second leading cause of cancer death among men in the United States. The diagnosis of prostate MRI often relies on accurate prostate zonal segmentation. However, state-of-the-art automatic segmentation methods often fail to produce well-contained volumetric segmentation of the prostate zones since certain slices of prostate MRI, such as base and apex slices, are harder to segment than other slices. This difficulty can be overcome by leveraging important multi-scale image-based information from adjacent slices, but current methods do not fully learn and exploit such cross-slice information. In this paper, we propose a novel cross-slice attention mechanism, which we use in a Transformer module to systematically learn cross-slice information at multiple scales. The module can be utilized in any existing deep-learning-based segmentation framework with skip connections. Experiments show that our cross-slice attention is able to capture cross-slice information significant for prostate zonal segmentation in order to improve the performance of current state-of-the-art methods. Cross-slice attention improves segmentation accuracy in the peripheral zones, such that segmentation results are consistent across all the prostate slices (apex, mid-gland, and base). The code for the proposed model is available at https://bit.ly/CAT-Net.

Application of an Improved U2-Net Model in Ultrasound Median Neural Image Segmentation

To investigate whether an improved U2-Net model could be used to segment the median nerve and improve segmentation performance, we performed a retrospective study with 402 nerve images from patients who visited Huashan Hospital from October 2018 to July 2020; 249 images were from patients with carpal tunnel syndrome, and 153 were from healthy volunteers. From these, 320 cases were selected as training sets, and 82 cases were selected as test sets. The improved U2-Net model was used to segment each image. Dice coefficients (Dice), pixel accuracy (PA), mean intersection over union (MIoU) and average Hausdorff distance (AVD) were used to evaluate segmentation performance. Results revealed that the Dice, MIoU, PA and AVD values of our improved U2-Net were 72.85%, 79.66%, 95.92% and 51.37 mm, respectively, which were comparable to the actual ground truth; the ground truth came from the labeling of clinicians. However, the Dice, MIoU, PA and AVD values of U-Net were 43.19%, 65.57%, 86.22% and 74.82 mm, and those of Res-U-Net were 58.65%, 72.53%, 88.98% and 57.30 mm. Overall, our data suggest our improved U2-Net model might be used for segmentation of ultrasound median neural images.

A survey on deep learning applied to medical images: from simple artificial neural networks to generative models

Deep learning techniques, in particular generative models, have taken on great importance in medical image analysis. This paper surveys fundamental deep learning concepts related to medical image generation. It provides concise overviews of studies which use some of the latest state-of-the-art models from last years applied to medical images of different injured body areas or organs that have a disease associated with (e.g., brain tumor and COVID-19 lungs pneumonia). The motivation for this study is to offer a comprehensive overview of artificial neural networks (NNs) and deep generative models in medical imaging, so more groups and authors that are not familiar with deep learning take into consideration its use in medicine works. We review the use of generative models, such as generative adversarial networks and variational autoencoders, as techniques to achieve semantic segmentation, data augmentation, and better classification algorithms, among other purposes. In addition, a collection of widely used public medical datasets containing magnetic resonance (MR) images, computed tomography (CT) scans, and common pictures is presented. Finally, we feature a summary of the current state of generative models in medical image including key features, current challenges, and future research paths.

MC-ViT: Multi-path cross-scale vision transformer for thymoma histopathology whole slide image typing

OBJECTIVES

Accurate histological typing plays an important role in diagnosing thymoma or thymic carcinoma (TC) and predicting the corresponding prognosis. In this paper, we develop and validate a deep learning-based thymoma typing method for hematoxylin & eosin (H&E)-stained whole slide images (WSIs), which provides useful histopathology information from patients to assist doctors for better diagnosing thymoma or TC.

METHODS

We propose a multi-path cross-scale vision transformer (MC-ViT), which first uses the cross attentive scale-aware transformer (CAST) to classify the pathological information related to thymoma, and then uses such pathological information priors to assist the WSIs transformer (WT) for thymoma typing. To make full use of the multi-scale (10×, 20×, and 40×) information inherent in a WSI, CAST not only employs parallel multi-path to capture different receptive field features from multi-scale WSI inputs, but also introduces the cross-correlation attention module (CAM) to aggregate multi-scale features to achieve cross-scale spatial information complementarity. After that, WT can effectively convert full-scale WSIs into 1D feature matrices with pathological information labels to improve the efficiency and accuracy of thymoma typing.

RESULTS

We construct a large-scale thymoma histopathology WSI (THW) dataset and annotate corresponding pathological information and thymoma typing labels. The proposed MC-ViT achieves the Top-1 accuracy of 0.939 and 0.951 in pathological information classification and thymoma typing, respectively. Moreover, the quantitative and statistical experiments on the THW dataset also demonstrate that our pipeline performs favorably against the existing classical convolutional neural networks, vision transformers, and deep learning-based medical image classification methods.

CONCLUSION

This paper demonstrates that comprehensively utilizing the pathological information contained in multi-scale WSIs is feasible for thymoma typing and achieves clinically acceptable performance. Specifically, the proposed MC-ViT can well predict pathological information classes as well as thymoma types, which show the application potential to the diagnosis of thymoma and TC and may assist doctors in improving diagnosis efficiency and accuracy.

A bi-directional deep learning architecture for lung nodule semantic segmentation

Lung nodules are abnormal growths and lesions may exist. Both lungs may have nodules. Most lung nodules are harmless (not cancerous/malignant). Pulmonary nodules are rare in lung cancer. X-rays and CT scans identify the lung nodules. Doctors may term the growth a lung spot, coin lesion, or shadow. It is necessary to obtain properly computed tomography (CT) scans of the lungs to get an accurate diagnosis and a good estimate of the severity of lung cancer. This study aims to design and evaluate a deep learning (DL) algorithm for identifying pulmonary nodules (PNs) using the LUNA-16 dataset and examine the prevalence of PNs using DB-Net. The paper states that a new, resource-efficient deep learning architecture is called for, and it has been given the name of DB-NET. When a physician orders a CT scan, they need to employ an accurate and efficient lung nodule segmentation method because they need to detect lung cancer at an early stage. However, segmentation of lung nodules is a difficult task because of the nodules' characteristics on the CT image as well as the nodules' concealed shape, visual quality, and context. The DB-NET model architecture is presented as a resource-efficient deep learning solution for handling the challenge at hand in this paper. Furthermore, it incorporates the Mish nonlinearity function and the mask class weights to improve segmentation effectiveness. In addition to the LUNA-16 dataset, which contained 1200 lung nodules collected during the LUNA-16 test, the LUNA-16 dataset was extensively used to train and assess the proposed model. The DB-NET architecture surpasses the existing U-NET model by a dice coefficient index of 88.89%, and it also achieves a similar level of accuracy to that of human experts.

Local Label Point Correction for Edge Detection of Overlapping Cervical Cells

Accurate labeling is essential for supervised deep learning methods. However, it is almost impossible to accurately and manually annotate thousands of images, which results in many labeling errors for most datasets. We proposes a local label point correction (LLPC) method to improve annotation quality for edge detection and image segmentation tasks. Our algorithm contains three steps: gradient-guided point correction, point interpolation, and local point smoothing. We correct the labels of object contours by moving the annotated points to the pixel gradient peaks. This can improve the edge localization accuracy, but it also causes unsmooth contours due to the interference of image noise. Therefore, we design a point smoothing method based on local linear fitting to smooth the corrected edge. To verify the effectiveness of our LLPC, we construct a largest overlapping cervical cell edge detection dataset (CCEDD) with higher precision label corrected by our label correction method. Our LLPC only needs to set three parameters, but yields 30–40% average precision improvement on multiple networks. The qualitative and quantitative experimental results show that our LLPC can improve the quality of manual labels and the accuracy of overlapping cell edge detection. We hope that our study will give a strong boost to the development of the label correction for edge detection and image segmentation. We will release the dataset and code at: https://github.com/nachifur/LLPC.

Automatic and Efficient Prediction of Hematoma Expansion in Patients with Hypertensive Intracerebral Hemorrhage Using Deep Learning Based on CT Images

Patients with hypertensive intracerebral hemorrhage (ICH) have a high hematoma expansion (HE) incidence. Noninvasive prediction HE helps doctors take effective measures to prevent accidents. This study retrospectively analyzed 253 cases of hypertensive intraparenchymal hematoma. Baseline non-contrast-enhanced CT scans (NECTs) were collected at admission and compared with subsequent CTs to determine the presence of HE. An end-to-end deep learning method based on CT was proposed to automatically segment the hematoma region, region of interest (ROI) feature extraction, and HE prediction. A variety of algorithms were employed for comparison. U-Net with attention performs best in the task of segmenting hematomas, with the mean Intersection overUnion (mIoU) of 0.9025. ResNet-34 achieves the most robust generalization capability in HE prediction, with an area under the receiver operating characteristic curve (AUC) of 0.9267, an accuracy of 0.8827, and an F1 score of 0.8644. The proposed method is superior to other mainstream models, which will facilitate accurate, efficient, and automated HE prediction.

U-Net-Based Medical Image Segmentation

Deep learning has been extensively applied to segmentation in medical imaging. U-Net proposed in 2015 shows the advantages of accurate segmentation of small targets and its scalable network architecture. With the increasing requirements for the performance of segmentation in medical imaging in recent years, U-Net has been cited academically more than 2500 times. Many scholars have been constantly developing the U-Net architecture. This paper summarizes the medical image segmentation technologies based on the U-Net structure variants concerning their structure, innovation, efficiency, etc.; reviews and categorizes the related methodology; and introduces the loss functions, evaluation parameters, and modules commonly applied to segmentation in medical imaging, which will provide a good reference for the future research.

Automatic Segmentation of Novel Coronavirus Pneumonia Lesions in CT Images Utilizing Deep-Supervised Ensemble Learning Network

Background: The novel coronavirus disease 2019 (COVID-19) has been spread widely in the world, causing a huge threat to the living environment of people. Objective: Under CT imaging, the structure features of COVID-19 lesions are complicated and varied greatly in different cases. To accurately locate COVID-19 lesions and assist doctors to make the best diagnosis and treatment plan, a deep-supervised ensemble learning network is presented for COVID-19 lesion segmentation in CT images. Methods: Since a large number of COVID-19 CT images and the corresponding lesion annotations are difficult to obtain, a transfer learning strategy is employed to make up for the shortcoming and alleviate the overfitting problem. Based on the reality that traditional single deep learning framework is difficult to extract complicated and varied COVID-19 lesion features effectively that may cause some lesions to be undetected. To overcome the problem, a deep-supervised ensemble learning network is presented to combine with local and global features for COVID-19 lesion segmentation. Results: The performance of the proposed method was validated in experiments with a publicly available dataset. Compared with manual annotations, the proposed method acquired a high intersection over union (IoU) of 0.7279 and a low Hausdorff distance (H) of 92.4604. Conclusion: A deep-supervised ensemble learning network was presented for coronavirus pneumonia lesion segmentation in CT images. The effectiveness of the proposed method was verified by visual inspection and quantitative evaluation. Experimental results indicated that the proposed method has a good performance in COVID-19 lesion segmentation.

Multi-path cascaded U-net for vessel segmentation from fundus fluorescein angiography sequential images

BACKGROUND AND OBJECTIVE

Fundus fluorescein angiography (FFA) technique is widely used in the examination of retinal diseases. In analysis of FFA sequential images, accurate vessel segmentation is a prerequisite for quantification of vascular morphology. Current vessel segmentation methods concentrate mainly on color fundus images and they are limited in processing FFA sequential images with varying background and vessels.

METHODS

We proposed a multi-path cascaded U-net (MCU-net) architecture for vessel segmentation in FFA sequential images, which is capable of integrating vessel features from different image modes to improve segmentation accuracy. Firstly, two modes of synthetic FFA images that enhance details of small vessels and large vessels are prepared, and are then used together with the raw FFA image as inputs of the MCU-net. By fusion of vessel features from the three modes of FFA images, a vascular probability map is generated as output of MCU-net.

RESULTS

The proposed MCU-net was trained and tested on the public Duke dataset and our own dataset for FFA sequential images as well as on the DRIVE dataset for color fundus images. Results show that MCU-net outperforms current state-of-the-art methods in terms of F1-score, sensitivity and accuracy, and is able of reserving details such as thin vessels and vascular connections. It also shows good robustness in processing FFA images captured at different perfusion stages.

CONCLUSIONS

The proposed method can segment vessels from FFA sequential images with high accuracy and shows good robustness to FFA images in different perfusion stages. This method has potential applications in quantitative analysis of vascular morphology in FFA sequential images.