Deep learning for MRI lesion segmentation in rectal cancer.

Rectal cancer (RC) is a globally prevalent malignant tumor, presenting significant challenges in its management and treatment. Currently, magnetic resonance imaging (MRI) offers superior soft tissue contrast and radiation-free effects for RC patients, making it the most widely used and effective detection method. In early screening, radiologists rely on patients' medical radiology characteristics and their extensive clinical experience for diagnosis. However, diagnostic accuracy may be hindered by factors such as limited expertise, visual fatigue, and image clarity issues, resulting in misdiagnosis or missed diagnosis. Moreover, the distribution of surrounding organs in RC is extensive with some organs having similar shapes to the tumor but unclear boundaries; these complexities greatly impede doctors' ability to diagnose RC accurately. With recent advancements in artificial intelligence, machine learning techniques like deep learning (DL) have demonstrated immense potential and broad prospects in medical image analysis. The emergence of this approach has significantly enhanced research capabilities in medical image classification, detection, and segmentation fields with particular emphasis on medical image segmentation. This review aims to discuss the developmental process of DL segmentation algorithms along with their application progress in lesion segmentation from MRI images of RC to provide theoretical guidance and support for further advancements in this field.

TG-Net: Using text prompts for improved skin lesion segmentation.

Automatic skin segmentation is an efficient method for the early diagnosis of skin cancer, which can minimize the missed detection rate and treat early skin cancer in time. However, significant variations in texture, size, shape, the position of lesions, and obscure boundaries in dermoscopy images make it extremely challenging to accurately locate and segment lesions. To address these challenges, we propose a novel framework named TG-Net, which exploits textual diagnostic information to guide the segmentation of dermoscopic images. Specifically, TG-Net adopts a dual-stream encoder-decoder architecture. The dual-stream encoder comprises Res2Net for extracting image features and our proposed text attention (TA) block for extracting textual features. Through hierarchical guidance, textual features are embedded into the process of image feature extraction. Additionally, we devise a multi-level fusion (MLF) module to merge higher-level features and generate a global feature map as guidance for subsequent steps. In the decoding stage of the network, local features and the global feature map are utilized in three multi-scale reverse attention modules (MSRA) to produce the final segmentation results. We conduct extensive experiments on three publicly accessible datasets, namely ISIC 2017, HAM10000, and PH2. Experimental results demonstrate that TG-Net outperforms state-of-the-art methods, validating the reliability of our method. Source code is available at https://github.com/ukeLin/TG-Net.

Nonproliferative diabetic retinopathy dataset(NDRD): A database for diabetic retinopathy screening research and deep learning evaluation.

OBJECTIVES: In this article, we provide a database of nonproliferative diabetes retinopathy, which focuses on early diabetes retinopathy with hard exudation, and further explore its clinical application in disease recognition. METHODS: We collect the photos of nonproliferative diabetes retinopathy taken by Optos Panoramic 200 laser scanning ophthalmoscope, filter out the pictures with poor quality, and label the hard exudative lesions in the images under the guidance of professional medical personnel. To validate the effectiveness of the datasets, five deep learning models are used to perform learning predictions on the datasets. Furthermore, we evaluate the performance of the model using evaluation metrics. RESULTS: Nonproliferative diabetes retinopathy is smaller than proliferative retinopathy and more difficult to identify. The existing segmentation models have poor lesion segmentation performance, while the intersection over union (IOU) value for deep lesion segmentation of models targeting small lesions can reach 66.12%, which is higher than ordinary lesion segmentation models, but there is still a lot of room for improvement. CONCLUSION: The segmentation of small hard exudative lesions is more challenging than that of large hard exudative lesions. More targeted datasets are needed for model training. Compared with the previous diabetes retina datasets, the NDRD dataset pays more attention to micro lesions.

A joint model for lesion segmentation and classification of MS and NMOSD.

Introduction: Multiple sclerosis (MS) and neuromyelitis optic spectrum disorder (NMOSD) are mimic autoimmune diseases of the central nervous system with a very high disability rate. Their clinical symptoms and imaging findings are similar, making it difficult to diagnose and differentiate. Existing research typically employs the T2-weighted fluid-attenuated inversion recovery (T2-FLAIR) MRI imaging technique to focus on a single task in MS and NMOSD lesion segmentation or disease classification, while ignoring the collaboration between the tasks. Methods: To make full use of the correlation between lesion segmentation and disease classification tasks of MS and NMOSD, so as to improve the accuracy and speed of the recognition and diagnosis of MS and NMOSD, a joint model is proposed in this study. The joint model primarily comprises three components: an information-sharing subnetwork, a lesion segmentation subnetwork, and a disease classification subnetwork. Among them, the information-sharing subnetwork adopts a dualbranch structure composed of a convolution module and a Swin Transformer module to extract local and global features, respectively. These features are then input into the lesion segmentation subnetwork and disease classification subnetwork to obtain results for both tasks simultaneously. In addition, to further enhance the mutual guidance between the tasks, this study proposes two information interaction methods: a lesion guidance module and a crosstask loss function. Furthermore, the lesion location maps provide interpretability for the diagnosis process of the deep learning model. Results: The joint model achieved a Dice similarity coefficient (DSC) of 74.87% on the lesion segmentation task and accuracy (ACC) of 92.36% on the disease classification task, demonstrating its superior performance. By setting up ablation experiments, the effectiveness of information sharing and interaction between tasks is verified. Discussion: The results show that the joint model can effectively improve the performance of the two tasks.

GLGFormer: Global Local Guidance Network for Mucosal Lesion Segmentation in Gastrointestinal Endoscopy Images.

Automatic mucosal lesion segmentation is a critical component in computer-aided clinical support systems for endoscopic image analysis. Image segmentation networks currently rely mainly on convolutional neural networks (CNNs) and Transformers, which have demonstrated strong performance in various applications. However, they cannot cope with blurred lesion boundaries and lesions of different scales in gastrointestinal endoscopy images. To address these challenges, we propose a new Transformer-based network, named GLGFormer, for the task of mucosal lesion segmentation. Specifically, we design the global guidance module to guide single-scale features patch-wise, enabling them to incorporate global information from the global map without information loss. Furthermore, a partial decoder is employed to fuse these enhanced single-scale features, achieving single-scale to multi-scale enhancement. Additionally, the local guidance module is designed to refocus attention on the neighboring patch, thus enhancing local features and refining lesion boundary segmentation. We conduct experiments on a private atrophic gastritis segmentation dataset and four public gastrointestinal polyp segmentation datasets. Compared to the current lesion segmentation networks, our proposed GLGFormer demonstrates outstanding learning and generalization capabilities. On the public dataset ClinicDB, GLGFormer achieved a mean intersection over union (mIoU) of 91.0% and a mean dice coefficient (mDice) of 95.0%. On the private dataset Gastritis-Seg, GLGFormer achieved an mIoU of 90.6% and an mDice of 94.6%.

Classification of cervical lesions based on multimodal features fusion.

Cervical cancer is a severe threat to women's health worldwide with a long cancerous cycle and a clear etiology, making early screening vital for the prevention and treatment. Based on the dataset provided by the Obstetrics and Gynecology Hospital of Fudan University, a four-category classification model for cervical lesions including Normal, low-grade squamous intraepithelial lesion (LSIL), high-grade squamous intraepithelial lesion (HSIL) and cancer (Ca) is developed. Considering the dataset characteristics, to fully utilize the research data and ensure the dataset size, the model inputs include original and acetic colposcopy images, lesion segmentation masks, human papillomavirus (HPV), thinprep cytologic test (TCT) and age, but exclude iodine images that have a significant overlap with lesions under acetic images. Firstly, the change information between original and acetic images is introduced by calculating the acetowhite opacity to mine the correlation between the acetowhite thickness and lesion grades. Secondly, the lesion segmentation masks are utilized to introduce prior knowledge of lesion location and shape into the classification model. Lastly, a cross-modal feature fusion module based on the self-attention mechanism is utilized to fuse image information with clinical text information, revealing the features correlation. Based on the dataset used in this study, the proposed model is comprehensively compared with five excellent models over the past three years, demonstrating that the proposed model has superior classification performance and a better balance between performance and complexity. The modules ablation experiments further prove that each proposed improved module can independently improve the model performance.

Improving assessment of lesions in longitudinal CT scans: a bi-institutional reader study on an AI-assisted registration and volumetric segmentation workflow.

PURPOSE: AI-assisted techniques for lesion registration and segmentation have the potential to make CT-based tumor follow-up assessment faster and less reader-dependent. However, empirical evidence on the advantages of AI-assisted volumetric segmentation for lymph node and soft tissue metastases in follow-up CT scans is lacking. The aim of this study was to assess the efficiency, quality, and inter-reader variability of an AI-assisted workflow for volumetric segmentation of lymph node and soft tissue metastases in follow-up CT scans. Three hypotheses were tested: (H1) Assessment time for follow-up lesion segmentation is reduced using an AI-assisted workflow. (H2) The quality of the AI-assisted segmentation is non-inferior to the quality of fully manual segmentation. (H3) The inter-reader variability of the resulting segmentations is reduced with AI assistance. MATERIALS AND METHODS: The study retrospectively analyzed 126 lymph nodes and 135 soft tissue metastases from 55 patients with stage IV melanoma. Three radiologists from two institutions performed both AI-assisted and manual segmentation, and the results were statistically analyzed and compared to a manual segmentation reference standard. RESULTS: AI-assisted segmentation reduced user interaction time significantly by 33% (222 s vs. 336 s), achieved similar Dice scores (0.80-0.84 vs. 0.81-0.82) and decreased inter-reader variability (median Dice 0.85-1.0 vs. 0.80-0.82; ICC 0.84 vs. 0.80), compared to manual segmentation. CONCLUSION: The findings of this study support the use of AI-assisted registration and volumetric segmentation for lymph node and soft tissue metastases in follow-up CT scans. The AI-assisted workflow achieved significant time savings, similar segmentation quality, and reduced inter-reader variability compared to manual segmentation.

Enhancing skin lesion segmentation with a fusion of convolutional neural networks and transformer models.

Accurate segmentation is crucial in diagnosing and analyzing skin lesions. However, automatic segmentation of skin lesions is extremely challenging because of their variable sizes, uneven color distributions, irregular shapes, hair occlusions, and blurred boundaries. Owing to the limited range of convolutional networks receptive fields, shallow convolution cannot extract the global features of images and thus has limited segmentation performance. Because medical image datasets are small in scale, the use of excessively deep networks could cause overfitting and increase computational complexity. Although transformer networks can focus on extracting global information, they cannot extract sufficient local information and accurately segment detailed lesion features. In this study, we designed a dual-branch encoder that combines a convolution neural network (CNN) and a transformer. The CNN branch of the encoder comprises four layers, which learn the local features of images through layer-wise downsampling. The transformer branch also comprises four layers, enabling the learning of global image information through attention mechanisms. The feature fusion module in the network integrates local features and global information, emphasizes important channel features through the channel attention mechanism, and filters irrelevant feature expressions. The information exchange between the decoder and encoder is finally achieved through skip connections to supplement the information lost during the sampling process, thereby enhancing segmentation accuracy. The data used in this paper are from four public datasets, including images of melanoma, basal cell tumor, fibroma, and benign nevus. Because of the limited size of the image data, we enhanced them using methods such as random horizontal flipping, random vertical flipping, random brightness enhancement, random contrast enhancement, and rotation. The segmentation accuracy is evaluated through intersection over union and duration, integrity, commitment, and effort indicators, reaching 87.7 % and 93.21 %, 82.05 % and 89.19 %, 86.81 % and 92.72 %, and 92.79 % and 96.21 %, respectively, on the ISIC 2016, ISIC 2017, ISIC 2018, and PH2 datasets, respectively (code: https://github.com/hyjane/CCT-Net).

Cross-modality transfer learning with knowledge infusion for diabetic retinopathy grading.

Background: Ultra-wide-field (UWF) fundus photography represents an emerging retinal imaging technique offering a broader field of view, thus enhancing its utility in screening and diagnosing various eye diseases, notably diabetic retinopathy (DR). However, the application of computer-aided diagnosis for DR using UWF images confronts two major challenges. The first challenge arises from the limited availability of labeled UWF data, making it daunting to train diagnostic models due to the high cost associated with manual annotation of medical images. Secondly, existing models' performance requires enhancement due to the absence of prior knowledge to guide the learning process. Purpose: By leveraging extensively annotated datasets within the field, which encompass large-scale, high-quality color fundus image datasets annotated at either image-level or pixel-level, our objective is to transfer knowledge from these datasets to our target domain through unsupervised domain adaptation. Methods: Our approach presents a robust model for assessing the severity of diabetic retinopathy (DR) by leveraging unsupervised lesion-aware domain adaptation in ultra-wide-field (UWF) images. Furthermore, to harness the wealth of detailed annotations in publicly available color fundus image datasets, we integrate an adversarial lesion map generator. This generator supplements the grading model by incorporating auxiliary lesion information, drawing inspiration from the clinical methodology of evaluating DR severity by identifying and quantifying associated lesions. Results: We conducted both quantitative and qualitative evaluations of our proposed method. In particular, among the six representative DR grading methods, our approach achieved an accuracy (ACC) of 68.18% and a precision (pre) of 67.43%. Additionally, we conducted extensive experiments in ablation studies to validate the effectiveness of each component of our proposed method. Conclusion: In conclusion, our method not only improves the accuracy of DR grading, but also enhances the interpretability of the results, providing clinicians with a reliable DR grading scheme.

Improving multiple sclerosis lesion segmentation across clinical sites: A federated learning approach with noise-resilient training.

Accurately measuring the evolution of Multiple Sclerosis (MS) with magnetic resonance imaging (MRI) critically informs understanding of disease progression and helps to direct therapeutic strategy. Deep learning models have shown promise for automatically segmenting MS lesions, but the scarcity of accurately annotated data hinders progress in this area. Obtaining sufficient data from a single clinical site is challenging and does not address the heterogeneous need for model robustness. Conversely, the collection of data from multiple sites introduces data privacy concerns and potential label noise due to varying annotation standards. To address this dilemma, we explore the use of the federated learning framework while considering label noise. Our approach enables collaboration among multiple clinical sites without compromising data privacy under a federated learning paradigm that incorporates a noise-robust training strategy based on label correction. Specifically, we introduce a Decoupled Hard Label Correction (DHLC) strategy that considers the imbalanced distribution and fuzzy boundaries of MS lesions, enabling the correction of false annotations based on prediction confidence. We also introduce a Centrally Enhanced Label Correction (CELC) strategy, which leverages the aggregated central model as a correction teacher for all sites, enhancing the reliability of the correction process. Extensive experiments conducted on two multi-site datasets demonstrate the effectiveness and robustness of our proposed methods, indicating their potential for clinical applications in multi-site collaborations to train better deep learning models with lower cost in data collection and annotation.

LST-AI: A deep learning ensemble for accurate MS lesion segmentation.

Automated segmentation of brain white matter lesions is crucial for both clinical assessment and scientific research in multiple sclerosis (MS). Over a decade ago, we introduced an engineered lesion segmentation tool, LST. While recent lesion segmentation approaches have leveraged artificial intelligence (AI), they often remain proprietary and difficult to adopt. As an open-source tool, we present LST-AI, an advanced deep learning-based extension of LST that consists of an ensemble of three 3D U-Nets. LST-AI explicitly addresses the imbalance between white matter (WM) lesions and non-lesioned WM. It employs a composite loss function incorporating binary cross-entropy and Tversky loss to improve segmentation of the highly heterogeneous MS lesions. We train the network ensemble on 491 MS pairs of T1-weighted and FLAIR images, collected in-house from a 3T MRI scanner, and expert neuroradiologists manually segmented the utilized lesion maps for training. LST-AI also includes a lesion location annotation tool, labeling lesions as periventricular, infratentorial, and juxtacortical according to the 2017 McDonald criteria, and, additionally, as subcortical. We conduct evaluations on 103 test cases consisting of publicly available data using the Anima segmentation validation tools and compare LST-AI with several publicly available lesion segmentation models. Our empirical analysis shows that LST-AI achieves superior performance compared to existing methods. Its Dice and F1 scores exceeded 0.62, outperforming LST, SAMSEG (Sequence Adaptive Multimodal SEGmentation), and the popular nnUNet framework, which all scored below 0.56. Notably, LST-AI demonstrated exceptional performance on the MSSEG-1 challenge dataset, an international WM lesion segmentation challenge, with a Dice score of 0.65 and an F1 score of 0.63-surpassing all other competing models at the time of the challenge. With increasing lesion volume, the lesion detection rate rapidly increased with a detection rate of >75% for lesions with a volume between 10 mm3 and 100 mm3. Given its higher segmentation performance, we recommend that research groups currently using LST transition to LST-AI. To facilitate broad adoption, we are releasing LST-AI as an open-source model, available as a command-line tool, dockerized container, or Python script, enabling diverse applications across multiple platforms.

DermSynth3D: Synthesis of in-the-wild annotated dermatology images.

In recent years, deep learning (DL) has shown great potential in the field of dermatological image analysis. However, existing datasets in this domain have significant limitations, including a small number of image samples, limited disease conditions, insufficient annotations, and non-standardized image acquisitions. To address these shortcomings, we propose a novel framework called DermSynth3D. DermSynth3D blends skin disease patterns onto 3D textured meshes of human subjects using a differentiable renderer and generates 2D images from various camera viewpoints under chosen lighting conditions in diverse background scenes. Our method adheres to top-down rules that constrain the blending and rendering process to create 2D images with skin conditions that mimic in-the-wild acquisitions, ensuring more meaningful results. The framework generates photo-realistic 2D dermatological images and the corresponding dense annotations for semantic segmentation of the skin, skin conditions, body parts, bounding boxes around lesions, depth maps, and other 3D scene parameters, such as camera position and lighting conditions. DermSynth3D allows for the creation of custom datasets for various dermatology tasks. We demonstrate the effectiveness of data generated using DermSynth3D by training DL models on synthetic data and evaluating them on various dermatology tasks using real 2D dermatological images. We make our code publicly available at https://github.com/sfu-mial/DermSynth3D.

Deep learning models for ischemic stroke lesion segmentation in medical images: A survey.

This paper provides a comprehensive review of deep learning models for ischemic stroke lesion segmentation in medical images. Ischemic stroke is a severe neurological disease and a leading cause of death and disability worldwide. Accurate segmentation of stroke lesions in medical images such as MRI and CT scans is crucial for diagnosis, treatment planning and prognosis. This paper first introduces common imaging modalities used for stroke diagnosis, discussing their capabilities in imaging lesions at different disease stages from the acute to chronic stage. It then reviews three major public benchmark datasets for evaluating stroke segmentation algorithms: ATLAS, ISLES and AISD, highlighting their key characteristics. The paper proceeds to provide an overview of foundational deep learning architectures for medical image segmentation, including CNN-based and transformer-based models. It summarizes recent innovations in adapting these architectures to the task of stroke lesion segmentation across the three datasets, analyzing their motivations, modifications and results. A survey of loss functions and data augmentations employed for this task is also included. The paper discusses various aspects related to stroke segmentation tasks, including prior knowledge, small lesions, and multimodal fusion, and then concludes by outlining promising future research directions. Overall, this comprehensive review covers critical technical developments in the field to support continued progress in automated stroke lesion segmentation.

SUTrans-NET: a hybrid transformer approach to skin lesion segmentation.

Melanoma is a malignant skin tumor that threatens human life and health. Early detection is essential for effective treatment. However, the low contrast between melanoma lesions and normal skin and the irregularity in size and shape make skin lesions difficult to detect with the naked eye in the early stages, making the task of skin lesion segmentation challenging. Traditional encoder-decoder built with U-shaped networks using convolutional neural network (CNN) networks have limitations in establishing long-term dependencies and global contextual connections, while the Transformer architecture is limited in its application to small medical datasets. To address these issues, we propose a new skin lesion segmentation network, SUTrans-NET, which combines CNN and Transformer in a parallel fashion to form a dual encoder, where both CNN and Transformer branches perform dynamic interactive fusion of image information in each layer. At the same time, we introduce our designed multi-grouping module SpatialGroupAttention (SGA) to complement the spatial and texture information of the Transformer branch, and utilize the Focus idea of YOLOV5 to construct the Patch Embedding module in the Transformer to prevent the loss of pixel accuracy. In addition, we design a decoder with full-scale information fusion capability to fully fuse shallow and deep features at different stages of the encoder. The effectiveness of our method is demonstrated on the ISIC 2016, ISIC 2017, ISIC 2018 and PH2 datasets and its advantages over existing methods are verified.

Simulation of acquisition shifts in T2 weighted fluid-attenuated inversion recovery magnetic resonance images to stress test artificial intelligence segmentation networks.

Purpose: To provide a simulation framework for routine neuroimaging test data, which allows for "stress testing" of deep segmentation networks against acquisition shifts that commonly occur in clinical practice for T2 weighted (T2w) fluid-attenuated inversion recovery magnetic resonance imaging protocols. Approach: The approach simulates "acquisition shift derivatives" of MR images based on MR signal equations. Experiments comprise the validation of the simulated images by real MR scans and example stress tests on state-of-the-art multiple sclerosis lesion segmentation networks to explore a generic model function to describe the F1 score in dependence of the contrast-affecting sequence parameters echo time (TE) and inversion time (TI). Results: The differences between real and simulated images range up to 19% in gray and white matter for extreme parameter settings. For the segmentation networks under test, the F1 score dependency on TE and TI can be well described by quadratic model functions (R2>0.9). The coefficients of the model functions indicate that changes of TE have more influence on the model performance than TI. Conclusions: We show that these deviations are in the range of values as may be caused by erroneous or individual differences in relaxation times as described by literature. The coefficients of the F1 model function allow for a quantitative comparison of the influences of TE and TI. Limitations arise mainly from tissues with a low baseline signal (like cerebrospinal fluid) and when the protocol contains contrast-affecting measures that cannot be modeled due to missing information in the DICOM header.

Adaptive Feature Medical Segmentation Network: an adaptable deep learning paradigm for high-performance 3D brain lesion segmentation in medical imaging.

Introduction: In neurological diagnostics, accurate detection and segmentation of brain lesions is crucial. Identifying these lesions is challenging due to its complex morphology, especially when using traditional methods. Conventional methods are either computationally demanding with a marginal impact/enhancement or sacrifice fine details for computational efficiency. Therefore, balancing performance and precision in compute-intensive medical imaging remains a hot research topic. Methods: We introduce a novel encoder-decoder network architecture named the Adaptive Feature Medical Segmentation Network (AFMS-Net) with two encoder variants: the Single Adaptive Encoder Block (SAEB) and the Dual Adaptive Encoder Block (DAEB). A squeeze-and-excite mechanism is employed in SAEB to identify significant data while disregarding peripheral details. This approach is best suited for scenarios requiring quick and efficient segmentation, with an emphasis on identifying key lesion areas. In contrast, the DAEB utilizes an advanced channel spatial attention strategy for fine-grained delineation and multiple-class classifications. Additionally, both architectures incorporate a Segmentation Path (SegPath) module between the encoder and decoder, refining segmentation, enhancing feature extraction, and improving model performance and stability. Results: AFMS-Net demonstrates exceptional performance across several notable datasets, including BRATs 2021, ATLAS 2021, and ISLES 2022. Its design aims to construct a lightweight architecture capable of handling complex segmentation challenges with high precision. Discussion: The proposed AFMS-Net addresses the critical balance issue between performance and computational efficiency in the segmentation of brain lesions. By introducing two tailored encoder variants, the network adapts to varying requirements of speed and feature. This approach not only advances the state-of-the-art in lesion segmentation but also provides a scalable framework for future research in medical image processing.

[Breast cancer lesion segmentation based on co-learning feature fusion and Transformer].

The PET/CT imaging technology combining positron emission tomography (PET) and computed tomography (CT) is the most advanced imaging examination method currently, and is mainly used for tumor screening, differential diagnosis of benign and malignant tumors, staging and grading. This paper proposes a method for breast cancer lesion segmentation based on PET/CT bimodal images, and designs a dual-path U-Net framework, which mainly includes three modules: encoder module, feature fusion module and decoder module. Among them, the encoder module uses traditional convolution for feature extraction of single mode image; The feature fusion module adopts collaborative learning feature fusion technology and uses Transformer to extract the global features of the fusion image; The decoder module mainly uses multi-layer perceptron to achieve lesion segmentation. This experiment uses actual clinical PET/CT data to evaluate the effectiveness of the algorithm. The experimental results show that the accuracy, recall and accuracy of breast cancer lesion segmentation are 95.67%, 97.58% and 96.16%, respectively, which are better than the baseline algorithm. Therefore, it proves the rationality of the single and bimodal feature extraction method combining convolution and Transformer in the experimental design of this article, and provides reference for feature extraction methods for tasks such as multimodal medical image segmentation or classification.

Masked autoencoders with generalizable self-distillation for skin lesion segmentation.

In the field of skin lesion image segmentation, accurate identification and partitioning of diseased regions is of vital importance for in-depth analysis of skin cancer. Self-supervised learning, i.e., MAE, has emerged as a potent force in the medical imaging domain, which autonomously learns and extracts latent features from unlabeled data, thereby yielding pre-trained models that greatly assist downstream tasks. To encourage pre-trained models to more comprehensively learn the global structural and local detail information inherent in dermoscopy images, we introduce a Teacher-Student architecture, named TEDMAE, by incorporating a self-distillation mechanism, it learns holistic image feature information to improve the generalizable global knowledge learning of the student MAE model. To make the image features learned by the model suitable for unknown test images, two optimization strategies are, Exterior Conversion Augmentation (EC) utilizes random convolutional kernels and linear interpolation to effectively transform the input image into one with the same shape but altered intensities and textures, while Dynamic Feature Generation (DF) employs a nonlinear attention mechanism for feature merging, enhancing the expressive power of the features, are proposed to enhance the generalizability of global features learned by the teacher model, thereby improving the overall generalization capability of the pre-trained models. Experimental results from the three public skin disease datasets, ISIC2019, ISIC2017, and PH 2 indicate that our proposed TEDMAE method outperforms several similar approaches. Specifically, TEDMAE demonstrated optimal segmentation and generalization performance on the ISIC2017 and PH 2 datasets, with Dice scores reaching 82.1% and 91.2%, respectively. The best Jaccard values were 72.6% and 84.5%, while the optimal HD95% values were 13.0% and 8.9%, respectively.

SEU2-Net: multi-scale U2-Net with SE attention mechanism for liver occupying lesion CT image segmentation.

Liver occupying lesions can profoundly impact an individual's health and well-being. To assist physicians in the diagnosis and treatment of abnormal areas in the liver, we propose a novel network named SEU2-Net by introducing the channel attention mechanism into U2-Net for accurate and automatic liver occupying lesion segmentation. We design the Residual U-block with Squeeze-and-Excitation (SE-RSU), which is to add the Squeeze-and-Excitation (SE) attention mechanism at the residual connections of the Residual U-blocks (RSU, the component unit of U2-Net). SEU2-Net not only retains the advantages of U2-Net in capturing contextual information at multiple scales, but can also adaptively recalibrate channel feature responses to emphasize useful feature information according to the channel attention mechanism. In addition, we present a new abdominal CT dataset for liver occupying lesion segmentation from Peking University First Hospital's clinical data (PUFH dataset). We evaluate the proposed method and compare it with eight deep learning networks on the PUFH and the Liver Tumor Segmentation Challenge (LiTS) datasets. The experimental results show that SEU2-Net has state-of-the-art performance and good robustness in liver occupying lesions segmentation.

Anomaly-guided weakly supervised lesion segmentation on retinal OCT images.

The availability of big data can transform the studies in biomedical research to generate greater scientific insights if expert labeling is available to facilitate supervised learning. However, data annotation can be labor-intensive and cost-prohibitive if pixel-level precision is required. Weakly supervised semantic segmentation (WSSS) with image-level labeling has emerged as a promising solution in medical imaging. However, most existing WSSS methods in the medical domain are designed for single-class segmentation per image, overlooking the complexities arising from the co-existence of multiple classes in a single image. Additionally, the multi-class WSSS methods from the natural image domain cannot produce comparable accuracy for medical images, given the challenge of substantial variation in lesion scales and occurrences. To address this issue, we propose a novel anomaly-guided mechanism (AGM) for multi-class segmentation in a single image on retinal optical coherence tomography (OCT) using only image-level labels. AGM leverages the anomaly detection and self-attention approach to integrate weak abnormal signals with global contextual information into the training process. Furthermore, we include an iterative refinement stage to guide the model to focus more on the potential lesions while suppressing less relevant regions. We validate the performance of our model with two public datasets and one challenging private dataset. Experimental results show that our approach achieves a new state-of-the-art performance in WSSS for lesion segmentation on OCT images.