Lung Cancer Segmentation With Transfer Learning: Usefulness of a Pretrained Model Constructed From an Artificial Dataset Generated Using a Generative Adversarial Network

Automatic Segmentation and Alignment of Uterine Shapes from 3D Ultrasound Data

BACKGROUND

The uterus is the most important organ in the female reproductive system. Its shape plays a critical role in fertility and pregnancy outcomes. Advances in medical imaging, such as 3D ultrasound, have significantly improved the exploration of the female genital tract, thereby enhancing gynecological healthcare. Despite well-documented data for organs like the liver and heart, large-scale studies on the uterus are lacking. Existing classifications, such as VCUAM and ESHRE/ESGE, provide different definitions for normal uterine shapes but are not based on real-world measurements. Moreover, the lack of comprehensive datasets significantly hinders research in this area. Our research, part of the larger NURSE study, aims to fill this gap by establishing the shape of a normal uterus using real-world 3D vaginal ultrasound scans. This will facilitate research into uterine shape abnormalities associated with infertility and recurrent miscarriages.

METHODS

We developed an automated system for the segmentation and alignment of uterine shapes from 3D ultrasound data, which consists of two steps: automatic segmentation of the uteri in 3D ultrasound scans using deep learning techniques, and alignment of the resulting shapes with standard geometrical approaches, enabling the extraction of the normal shape for future analysis. The system was trained and validated on a comprehensive dataset of 3D ultrasound images from multiple medical centers. Its performance was evaluated by comparing the automated results with manual annotations provided by expert clinicians.

RESULTS

The presented approach demonstrated high accuracy in segmenting and aligning uterine shapes from 3D ultrasound data. The segmentation achieved an average Dice similarity coefficient (DSC) of 0.90. Our method for aligning uterine shapes showed minimal translation and rotation errors compared to traditional methods, with the preliminary average shape exhibiting characteristics consistent with expert findings of a normal uterus.

CONCLUSION

We have presented an approach to automatically segment and align uterine shapes from 3D ultrasound data. We trained a deep learning nnU-Net model that achieved high accuracy and proposed an alignment method using a combination of standard geometrical techniques. Additionally, we have created a publicly available dataset of 3D transvaginal ultrasound volumes with manual annotations of uterine cavities to support further research and development in this field. The dataset and the trained models are available at https://github.com/UL-FRI-LGM/UterUS.

Intelligent Stroke Disease Prediction Model Using Deep Learning Approaches

Stroke is a high morbidity and mortality disease that poses a serious threat to people's health. Early recognition of the various warning signs of stroke is necessary so that timely clinical intervention can help reduce the severity of stroke. Deep neural networks have powerful feature representation capabilities and can automatically learn discriminant features from large amounts of data. This paper uses a range of physiological characteristic parameters and collaborates with deep neural networks, such as the Wasserstein generative adversarial networks with gradient penalty and regression network, to construct a stroke prediction model. Firstly, to address the problem of imbalance between positive and negative samples in the stroke public data set, we performed positive sample data augmentation and utilized WGAN-GP to generate stroke data with high fidelity and used it for the training of the prediction network model. Then, the relationship between observable physiological characteristic parameters and the predicted risk of suffering a stroke was modeled as a nonlinear mapping transformation, and a stroke prediction model based on a deep regression network was designed. Finally, the proposed method is compared with commonly used machine learning-based classification algorithms such as decision tree, random forest, support vector machine, and artificial neural networks. The prediction results of the proposed method are optimal in the comprehensive measurement index F. Further ablation experiments also show that the designed prediction model has certain robustness and can effectively predict stroke diseases.

Survey of Transfer Learning Approaches in the Machine Learning of Digital Health Sensing Data

Machine learning and digital health sensing data have led to numerous research achievements aimed at improving digital health technology. However, using machine learning in digital health poses challenges related to data availability, such as incomplete, unstructured, and fragmented data, as well as issues related to data privacy, security, and data format standardization. Furthermore, there is a risk of bias and discrimination in machine learning models. Thus, developing an accurate prediction model from scratch can be an expensive and complicated task that often requires extensive experiments and complex computations. Transfer learning methods have emerged as a feasible solution to address these issues by transferring knowledge from a previously trained task to develop high-performance prediction models for a new task. This survey paper provides a comprehensive study of the effectiveness of transfer learning for digital health applications to enhance the accuracy and efficiency of diagnoses and prognoses, as well as to improve healthcare services. The first part of this survey paper presents and discusses the most common digital health sensing technologies as valuable data resources for machine learning applications, including transfer learning. The second part discusses the meaning of transfer learning, clarifying the categories and types of knowledge transfer. It also explains transfer learning methods and strategies, and their role in addressing the challenges in developing accurate machine learning models, specifically on digital health sensing data. These methods include feature extraction, fine-tuning, domain adaptation, multitask learning, federated learning, and few-/single-/zero-shot learning. This survey paper highlights the key features of each transfer learning method and strategy, and discusses the limitations and challenges of using transfer learning for digital health applications. Overall, this paper is a comprehensive survey of transfer learning methods on digital health sensing data which aims to inspire researchers to gain knowledge of transfer learning approaches and their applications in digital health, enhance the current transfer learning approaches in digital health, develop new transfer learning strategies to overcome the current limitations, and apply them to a variety of digital health technologies.

MRI-based automatic identification and segmentation of extrahepatic cholangiocarcinoma using deep learning network

BACKGROUND

Accurate identification of extrahepatic cholangiocarcinoma (ECC) from an image is challenging because of the small size and complex background structure. Therefore, considering the limitation of manual delineation, it's necessary to develop automated identification and segmentation methods for ECC. The aim of this study was to develop a deep learning approach for automatic identification and segmentation of ECC using MRI.

METHODS

We recruited 137 ECC patients from our hospital as the main dataset (C1) and an additional 40 patients from other hospitals as the external validation set (C2). All patients underwent axial T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and diffusion-weighted imaging (DWI). Manual delineations were performed and served as the ground truth. Next, we used 3D VB-Net to establish single-mode automatic identification and segmentation models based on T1WI (model 1), T2WI (model 2), and DWI (model 3) in the training cohort (80% of C1), and compared them with the combined model (model 4). Subsequently, the generalization capability of the best models was evaluated using the testing set (20% of C1) and the external validation set (C2). Finally, the performance of the developed models was further evaluated.

RESULTS

Model 3 showed the best identification performance in the training, testing, and external validation cohorts with success rates of 0.980, 0.786, and 0.725, respectively. Furthermore, model 3 yielded an average Dice similarity coefficient (DSC) of 0.922, 0.495, and 0.466 to segment ECC automatically in the training, testing, and external validation cohorts, respectively.

CONCLUSION

The DWI-based model performed better in automatically identifying and segmenting ECC compared to T1WI and T2WI, which may guide clinical decisions and help determine prognosis.

Deep learning image segmentation approaches for malignant bone lesions: a systematic review and meta-analysis

Introduction

Image segmentation is an important process for quantifying characteristics of malignant bone lesions, but this task is challenging and laborious for radiologists. Deep learning has shown promise in automating image segmentation in radiology, including for malignant bone lesions. The purpose of this review is to investigate deep learning-based image segmentation methods for malignant bone lesions on Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and Positron-Emission Tomography/CT (PET/CT).

Method

The literature search of deep learning-based image segmentation of malignant bony lesions on CT and MRI was conducted in PubMed, Embase, Web of Science, and Scopus electronic databases following the guidelines of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). A total of 41 original articles published between February 2017 and March 2023 were included in the review.

Results

The majority of papers studied MRI, followed by CT, PET/CT, and PET/MRI. There was relatively even distribution of papers studying primary vs. secondary malignancies, as well as utilizing 3-dimensional vs. 2-dimensional data. Many papers utilize custom built models as a modification or variation of U-Net. The most common metric for evaluation was the dice similarity coefficient (DSC). Most models achieved a DSC above 0.6, with medians for all imaging modalities between 0.85-0.9.

Discussion

Deep learning methods show promising ability to segment malignant osseous lesions on CT, MRI, and PET/CT. Some strategies which are commonly applied to help improve performance include data augmentation, utilization of large public datasets, preprocessing including denoising and cropping, and U-Net architecture modification. Future directions include overcoming dataset and annotation homogeneity and generalizing for clinical applicability.

CheSS: Chest X-Ray Pre-trained Model via Self-supervised Contrastive Learning

Training deep learning models on medical images heavily depends on experts' expensive and laborious manual labels. In addition, these images, labels, and even models themselves are not widely publicly accessible and suffer from various kinds of bias and imbalances. In this paper, chest X-ray pre-trained model via self-supervised contrastive learning (CheSS) was proposed to learn models with various representations in chest radiographs (CXRs). Our contribution is a publicly accessible pretrained model trained with a 4.8-M CXR dataset using self-supervised learning with a contrastive learning and its validation with various kinds of downstream tasks including classification on the 6-class diseases in internal dataset, diseases classification in CheXpert, bone suppression, and nodule generation. When compared to a scratch model, on the 6-class classification test dataset, we achieved 28.5% increase in accuracy. On the CheXpert dataset, we achieved 1.3% increase in mean area under the receiver operating characteristic curve on the full dataset and 11.4% increase only using 1% data in stress test manner. On bone suppression with perceptual loss, we achieved improvement in peak signal to noise ratio from 34.99 to 37.77, structural similarity index measure from 0.976 to 0.977, and root-square-mean error from 4.410 to 3.301 when compared to ImageNet pretrained model. Finally, on nodule generation, we achieved improvement in Fréchet inception distance from 24.06 to 17.07. Our study showed the decent transferability of CheSS weights. CheSS weights can help researchers overcome data imbalance, data shortage, and inaccessibility of medical image datasets. CheSS weight is available at https://github.com/mi2rl/CheSS .

Medical Images Segmentation for Lung Cancer Diagnosis Based on Deep Learning Architectures

Lung cancer presents one of the leading causes of mortalities for people around the world. Lung image analysis and segmentation are one of the primary steps used for early diagnosis of cancer. Handcrafted medical imaging segmentation presents a very time-consuming task for radiation oncologists. To address this problem, we propose in this work to develop a full and entire system used for early diagnosis of lung cancer in CT scan imaging. The proposed lung cancer diagnosis system is composed of two main parts: the first part is used for segmentation developed on top of the UNETR network, and the second part is a classification part used to classify the output segmentation part, either benign or malignant, developed on top of the self-supervised network. The proposed system presents a powerful tool for early diagnosing and combatting lung cancer using 3D-input CT scan data. Extensive experiments have been performed to contribute to better segmentation and classification results. Training and testing experiments have been performed using the Decathlon dataset. Experimental results have been conducted to new state-of-the-art performances: segmentation accuracy of 97.83%, and 98.77% as classification accuracy. The proposed system presents a new powerful tool to use for early diagnosing and combatting lung cancer using 3D-input CT scan data.

A transformer-based generative adversarial network for brain tumor segmentation

Brain tumor segmentation remains a challenge in medical image segmentation tasks. With the application of transformer in various computer vision tasks, transformer blocks show the capability of learning long-distance dependency in global space, which is complementary to CNNs. In this paper, we proposed a novel transformer-based generative adversarial network to automatically segment brain tumors with multi-modalities MRI. Our architecture consists of a generator and a discriminator, which is trained in min-max game progress. The generator is based on a typical "U-shaped" encoder-decoder architecture, whose bottom layer is composed of transformer blocks with Resnet. Besides, the generator is trained with deep supervision technology. The discriminator we designed is a CNN-based network with multi-scale L 1 loss, which is proved to be effective for medical semantic image segmentation. To validate the effectiveness of our method, we conducted exclusive experiments on BRATS2015 dataset, achieving comparable or better performance than previous state-of-the-art methods. On additional datasets, including BRATS2018 and BRATS2020, experimental results prove that our technique is capable of generalizing successfully.

Intelligent tuberculosis activity assessment system based on an ensemble of neural networks

This article proposes a novel approach to assess the degree of activity of pulmonary tuberculosis by active tuberculoma foci. It includes the development of a new method for processing lung CT images using an ensemble of deep convolutional neural networks using such special algorithms: an optimized algorithm for preliminary segmentation and selection of informative scans, a new algorithm for refining segmented masks to improve the final accuracy, an efficient fuzzy inference system for more weighted activity assessment. The approach also includes the use of medical classification of disease activity based on densitometric measures of tuberculomas. The selection and markup of the training sample images were performed manually by qualified pulmonologists from a base of approximately 9,000 CT lung scans of patients who had been enrolled in the dispensary for 15 years. The first basic step of the proposed approach is the developed algorithm for preprocessing CT lung scans. It consists in segmentation of intrapulmonary regions, which contain vessels, bronchi, lung walls to detect complex cases of ingrown tuberculomas. To minimize computational cost, the proposed approach includes a new method for selecting informative lung scans, i.e., those that potentially contain tuberculomas. The main processing step is binary segmentation of tuberculomas, which is proposed to be performed optimally by a certain ensemble of neural networks. Optimization of the ensemble size and its composition is achieved by using an algorithm for calculating individual contributions. A modification of this algorithm using new effective heuristic metrics has been proposed which improves the performance of the algorithm for this problem. A special algorithm was developed for post-processing of tuberculoma masks obtained during the segmentation step. The goal of this step is to refine the calculated mask for the physical placement of the tuberculoma. The algorithm consists in cleaning the mask from noisy formations on the scan, as well as expanding the mask area to maximize the capture of the tuberculoma location area. A simplified fuzzy inference system was developed to provide a more accurate final calculation of the degree of disease activity, which reflects data from current medical studies. The accuracy of the system was also tested on a test sample of independent patients, showing more than 96% correct calculations of disease activity, confirming the effectiveness and feasibility of introducing the system into clinical practice.

DETECT-LC: A 3D Deep Learning and Textural Radiomics Computational Model for Lung Cancer Staging and Tumor Phenotyping Based on Computed Tomography Volumes

Lung Cancer is one of the primary causes of cancer-related deaths worldwide. Timely diagnosis and precise staging are pivotal for treatment planning, and thus can lead to increased survival rates. The application of advanced machine learning techniques helps in effective diagnosis and staging. In this study, a multistage neurobased computational model is proposed, DETECT-LC learning. DETECT-LC handles the challenge of choosing discriminative CT slices for constructing 3D volumes, using Haralick, histogram-based radiomics, and unsupervised clustering. ALT-CNN-DENSE Net architecture is introduced as part of DETECT-LC for voxel-based classification. DETECT-LC offers an automatic threshold-based segmentation approach instead of the manual procedure, to help mitigate this burden for radiologists and clinicians. Also, DETECT-LC presents a slice selection approach and a newly proposed relatively light weight 3D CNN architecture to improve existing studies performance. The proposed pipeline is employed for tumor phenotyping and staging. DETECT-LC performance is assessed through a range of experiments, in which DETECT-LC attains outstanding performance surpassing its counterparts in terms of accuracy, sensitivity, F1-score and Area under Curve (AuC). For histopathology classification, DETECT-LC average performance achieved an improvement of 20% in overall accuracy, 0.19 in sensitivity, 0.16 in F1-Score and 0.16 in AuC over the state of the art. A similar enhancement is reached for staging, where higher overall accuracy, sensitivity and F1-score are attained with differences of 8%, 0.08 and 0.14.

The value of AI in the Diagnosis, Treatment, and Prognosis of Malignant Lung Cancer

Malignant tumors is a serious public health threat. Among them, lung cancer, which has the highest fatality rate globally, has significantly endangered human health. With the development of artificial intelligence (AI) and its integration with medicine, AI research in malignant lung tumors has become critical. This article reviews the value of CAD, computer neural network deep learning, radiomics, molecular biomarkers, and digital pathology for the diagnosis, treatment, and prognosis of malignant lung tumors.

An Image-Based Benchmark Dataset and a Novel Object Detector for Water Surface Object Detection

Water surface object detection is one of the most significant tasks in autonomous driving and water surface vision applications. To date, existing public large-scale datasets collected from websites do not focus on specific scenarios. As a characteristic of these datasets, the quantity of the images and instances is also still at a low level. To accelerate the development of water surface autonomous driving, this paper proposes a large-scale, high-quality annotated benchmark dataset, named Water Surface Object Detection Dataset (WSODD), to benchmark different water surface object detection algorithms. The proposed dataset consists of 7,467 water surface images in different water environments, climate conditions, and shooting times. In addition, the dataset comprises a total of 14 common object categories and 21,911 instances. Simultaneously, more specific scenarios are focused on in WSODD. In order to find a straightforward architecture to provide good performance on WSODD, a new object detector, named CRB-Net, is proposed to serve as a baseline. In experiments, CRB-Net was compared with 16 state-of-the-art object detection methods and outperformed all of them in terms of detection precision. In this paper, we further discuss the effect of the dataset diversity (e.g., instance size, lighting conditions), training set size, and dataset details (e.g., method of categorization). Cross-dataset validation shows that WSODD significantly outperforms other relevant datasets and that the adaptability of CRB-Net is excellent.

Automated Final Lesion Segmentation in Posterior Circulation Acute Ischemic Stroke Using Deep Learning

Final lesion volume (FLV) is a surrogate outcome measure in anterior circulation stroke (ACS). In posterior circulation stroke (PCS), this relation is plausibly understudied due to a lack of methods that automatically quantify FLV. The applicability of deep learning approaches to PCS is limited due to its lower incidence compared to ACS. We evaluated strategies to develop a convolutional neural network (CNN) for PCS lesion segmentation by using image data from both ACS and PCS patients. We included follow-up non-contrast computed tomography scans of 1018 patients with ACS and 107 patients with PCS. To assess whether an ACS lesion segmentation generalizes to PCS, a CNN was trained on ACS data (ACS-CNN). Second, to evaluate the performance of only including PCS patients, a CNN was trained on PCS data. Third, to evaluate the performance when combining the datasets, a CNN was trained on both datasets. Finally, to evaluate the performance of transfer learning, the ACS-CNN was fine-tuned using PCS patients. The transfer learning strategy outperformed the other strategies in volume agreement with an intra-class correlation of 0.88 (95% CI: 0.83-0.92) vs. 0.55 to 0.83 and a lesion detection rate of 87% vs. 41-77 for the other strategies. Hence, transfer learning improved the FLV quantification and detection rate of PCS lesions compared to the other strategies.