In the mobile healthcare scenario, the efficient deployment of lightweight image classification models on edge devices can significantly enhance the accessibility and real-time performance of medical services, providing reliable technical support for scenarios such as scarce medical resources in remote areas, real-time diagnosis on mobile terminals, and remote image analysis. Aiming at the problems such as insufficient cross-domain adaptability and inadequate feature extraction of lightweight models in the task of medical image classification, this paper proposes a lightweight dual-stream multi-scale feature fusion medical adaptation classification network based on guided enhancement (DMF-MobileMamba). This network adopts a parallel dual stream architecture, combining the local texture extraction capability of Convolutional Neural Network (CNN) with the global remote dependency modeling advantage of the improved lightweight multi-scale adapter Mamba module, and achieving heterogeneous feature complementarity through decoupling design. The multi-scale attention modulation fusion module (MSA-Fusion) is used to dynamically and weighted fuse local and global features; Innovatively proposed the Cross-level Guided Enhancement Attention Module (CLGE), which utilizes shallow high-resolution details to dynamically correct deep semantic biases and alleviate the representation mismatch problem between model levels. Experiments show that DMF-MobileMamba only requires 4.039 million(M) parameters and 2.438 Giga Floating-point Operations Per Second(GFLOPS). On six medical datasets, its classification accuracy is significantly better than that of mainstream advanced lightweight models, and it achieves a real-time inference speed of 134.64 millisecond(ms) on mobile devices. It provides high-precision and low-cost solutions for resource-constrained scenarios.

Skin lesion classification is a core task in automated dermatological analysis. Compared to single modalities, multi-modal data can provide complementary information, accelerating the progress of dermatological classification research. However, multi-modal integration faces challenges of insufficient feature interaction and modality heterogeneity. In this study, we propose a multi-modal fusion framework, CDMFuse, for multi-modal skin lesion classification. CDMFuse driven by two novel modules: the Sum-Difference Feature Interaction Module (SDF) and the Patch-Metadata Cross-Attention Fusion Module (PMCA). The SDF module addresses the insufficient integration of dermoscopic and clinical images in existing methods, which typically fuse features only at late stages, leading to information loss. By introducing sum and difference branches, SDF hierarchically models complementary spatial-channel interactions: the sum branch enhances shared structural features (e.g., lesion boundaries), while the difference branch captures modality-specific details (e.g., microvascular patterns in dermoscopy). The PMCA module tackles the heterogeneity gap between image modalities and metadata, where direct concatenation amplifies redundancy and noise. PMCA employs a cross-attention mechanism to dynamically align image patches with metadata embeddings, resolving structural mismatches. It further integrates content-aware multi-scale convolutions to adaptively adjust receptive fields. The effectiveness of our approach is validated on the publicly available Derm7pt multi-modal skin lesion dataset, achieving an average accuracy of 77.69

BackgroundPsoriasis is a common autoimmune skin disease with high morbidity and associated complications, characterized by epidermal hyperplasia and cutaneous infiltration of immune cells. The role of S100A4, a key antimicrobial peptide, is highly expressed in psoriatic skin and has aroused considerable interest in recent years, yet its specific function and associated molecular mechanisms remain elusive.MethodsUsing CRISPR/Cas9 to generate S100a4 gene knockout mice, imiquimod was continuously applied to the back skin to induce the psoriasis disease model. Single-cell RNA sequencing (scRNA-seq) was employed to investigate changes in epidermal cell composition and gene expression profiles in mice subjected to different treatments. Multiple bioinformatics analyses were conducted to elucidate the biological role of S100A4 in psoriasis pathogenesis.ResultsWe observed that S100a4 knockout mice exhibited significant pathological improvement in psoriasis-like lesions, including reduced inflammatory cell infiltration and decreased epidermal hyperplasia. The results of scRNA-seq revealed that after S100a4 knockout, the pathologically relevant keratinocyte subpopulation was significantly reduced, and the related tumor necrosis factor (TNF) and interleukin-17 (IL-17) signal transducers and activator of keratinization were downregulated. Moreover, S100a4 depletion moderated the abnormal proliferation and differentiation dynamics of keratinocytes. Additionally, Klf9-mediated transcriptional dysregulation of Krt15 in keratinocytes was identified as a key driver of hyperkeratosis, while S100a4 deficiency contributed to restoring cellular homeostasis in this process.ConclusionsOur findings suggest a potential pathogenic role for S100A4 in psoriasis and highlight previously uncharacterized cell-specific transcriptional landscapes and regulatory mechanisms. Our results provide novel insights into the complex pathology of psoriasis and could offer important clues for the development of new targeted therapeutic strategies.

Utilizing deep learning to process multi-modal information, including clinical images, dermatoscopy images, and patient metadata, for multi-modal skin lesion diagnosis (MSLD) aligns with modern medical dermatological diagnostic methods. A crucial task in achieving MSLD is to fully leverage multi-modal information, which previous works often failed to accomplish. To address this challenge, in this paper, we propose a novel network, MSfuseNet, for multi-modal skin disease classification. Our method primarily consists of four modules: (1) First, we employ a Coordinate-Spatial Attention Fusion Module to align the two types of images in both coordinate and spatial dimensions, combining this module at different stages with intermediate fusion strategies to reduce information loss in the model. (2) Then, we utilize a MIX Module to facilitate the transformation of shallow local features into deep global features within the model, thereby enhancing the model’s modeling capabilities and robustness. (3) We adopt a Double-Modality Cross-Attention Fusion Module, employing cross-attention mechanisms for global modeling of image features. (4) Finally, we employ a Triple-Modality Fusion Module to aggregate textual features and image features, achieving a full integration of multi-modal information. We have validated the effectiveness of our approach on a public dataset named Derm7pt and a dataset we collected named XJU-MMSD. Compared to state-of-the-art methods, our method achieved the highest average accuracy of 77.73

Skin cancer is one of the most serious threats to human health among skin lesions. Computer-aided diagnosis methods can assist patients in identifying and detecting skin lesion types early, thereby enabling corresponding treatments. In this paper, we propose a dual-branch neural network model Conformer with Residual Cosine Similarity Attention and Bidirectional Convolutional fusion strategy, named RCSABC-Conformer. The core of this network structure comprises three parts: a Convolutional Neural Network (CNN) branch with Residual Cosine Similarity Attention (RCSA), a Transformer branch, and a Feature Couple Unit with Bidirectional Convolutional strategy (BC-FCU). The RCSA module calculates the cosine similarity value between the feature map generated by the convolutional operation and the feature map of the residual edge to assess whether their semantic information is similar. The semantic information of similar parts is weighted by exponential normalization to enhance the network's memory of similar features of the same type of skin lesion. The BC-FCU module interactively fuses local features and global representations of skin lesion images with different resolutions in the two branches. Specifically, when the global representations is integrated into local features, we introduce a new bidirectional convolution strategy to extract the feature map from both forward and backward directions, and then select the element with the smaller feature value from the two directions to fuse into local features. In this way, we can minimize the interference of the artifact features extracted by the Transformer branch on the CNN branch. In addition, taking advantage of the Transformer branch's capacity to construct global representations, our model can learn contextual semantic information of normal skin and lesion areas to enhance model robustness. We conducted experiments on three datasets, consisting of clinical and dermoscopic skin lesion images, as well as a hybrid of both. The experimental results show that RCSABC-Conformer outperforms both advanced and classical classification methods in terms of classification accuracy across all three datasets, without requiring an increase in the number of parameters and computational complexity. Compared with the baseline model, the classification accuracy of our proposed method improves by 2.40%, 5.39%, and 4.44% on the three datasets, respectively. To the best of our knowledge, this is the first study to apply an interactive fusion dual-branch network for multi-disease classification on different modalities of skin lesion databases. Code will be available at https://github.com/AlenLi817/RCSABC-Conformer.

The Transformer architecture has been widely applied in the field of image segmentation due to its powerful ability to capture long-range dependencies. However, its ability to capture local features is relatively weak and it requires a large amount of data for training. Medical image segmentation tasks, on the other hand, demand high requirements for local features and are often applied to small datasets. Therefore, existing Transformer networks show a significant decrease in performance when applied directly to this task. To address these issues, we have designed a new medical image segmentation architecture called CT-Net. It effectively extracts local and global representations using an asymmetric asynchronous branch parallel structure, while reducing unnecessary computational costs. In addition, we propose a high-density information fusion strategy that efficiently fuses the features of two branches using a fusion module of only 0.05M. This strategy ensures high portability and provides conditions for directly applying transfer learning to solve dataset dependency issues. Finally, we have designed a parameter-adjustable multi-perceptive loss function for this architecture to optimize the training process from both pixel-level and global perspectives. We have tested this network on 5 different tasks with 9 datasets, and compared to SwinUNet, CT-Net improves the IoU by 7.3% and 1.8% on Glas and MoNuSeg datasets respectively. Moreover, compared to SwinUNet, the average DSC on the Synapse dataset is improved by 3.5%.

Semi-supervised learning (SSL) provides methods to improve model performance through unlabeled samples. In medical image analysis, the challenges of multi-category classification and imbalance learning must be addressed effectively. Pseudo labeling is not specifically designed for multi-category and category imbalance problems. In this paper, we propose the Growth Threshold for Pseudo Labeling (GTPL) and Pseudo Label Dropout (PLD), which can be used separately or in combination. GTPL changes the threshold value of each category by combining the confidence of labeled and unlabeled samples. PLD alleviates the category imbalance by randomly discarding some of the pseudo labels. We apply GTPL and PLD to FixMatch and CoMatch and effectively improve their semi-supervised classification performance. We validate the effectiveness of our approach in skin lesion diagnosis on two long-tailed distributions of public medical images on the ISIC 2018 and ISIC 2019 challenge datasets, obtaining AUCs of 89.19%, 92.71%, 94.71%, and 94.76%, respectively, on four scales of labeled data from ISIC 2018.

BACKGROUND:The size variation, complex semantic environment and high similarity in medical images often prevent deep learning models from achieving good performance.PURPOSE:To overcome these problems and improve the model segmentation performance and generalizability.METHODS:We propose the key class feature reconstruction module (KCRM), which ranks channel weights and selects key features (KFs) that contribute more to the segmentation results for each class. Meanwhile, KCRM reconstructs all local features to establish the dependence relationship from local features to KFs. In addition, we propose the spatial gating module (SGM), which employs KFs to generate two spatial maps to suppress irrelevant regions, strengthening the ability to locate semantic objects. Finally, we enable the model to adapt to size variations by diversifying the receptive field.RESULTS:We integrate these modules into class key feature extraction and fusion network (CKFFNet) and validate its performance on three public medical datasets: CHAOS, UW-Madison, and ISIC2017. The experimental results show that our method achieves better segmentation results and generalizability than those of mainstream methods.CONCLUSION:Through quantitative and qualitative research, the proposed module improves the segmentation results and enhances the model generalizability, making it suitable for application and expansion.

Recent research on semi-supervised learning (SSL) is mainly based on the method of consistency regularization, which relies on data augmentation in the specific domain. Pseudolabelling is a more general method that has no such restrictions but is limited by noisy training. Medical datasets are a particular domain that exhibit a long-tail distribution. Combining these two limitations, we focus on the widespread use of weak augmentation to generate pseudolabels. We propose FixMatch-LS and a variant FixMatch-LS-v2 for medical image classification. First, we introduce label smoothing to change the pseudolabel threshold, which reduces the influence of noisy pseudolabels. In addition, pseudolabelling should be matched with consistency. A suitable consistency can constrain pseudolabelling to improve the quality of the pseudolabels. We validate our framework on skin lesion diagnoses from the ISIC 2018 and ISIC 2019 challenges, obtaining AUCs of 91.63%, 93.70%, 94.46%, and 95.44% on the four proportions of labelled data from ISIC 2018.

Recent research on semi-supervised learning (SSL) is mainly based on the method of consistency regularization, which relies on domain-specific data augmentation. Pseudo-labeling is a more general method that has no such restrictions but performs limited by noisy training. We combine both approaches and focus on generating pseudo-labels using domain-independent weak augmentation. In this article, we propose ReFixMatch-LS and apply it to the classification of medical images. First, we reduce the impact of noisy artificial labels by label smoothing and consistent regularization. Then, by recording high-confidence pseudo-labels generated from each epoch during training, we reuse the generated pseudo-labels to train the model in the subsequent epochs. ReFixMatch-LS effectively increases the number of pseudo-labels and improves the model performance. We validate the effectiveness of ReFixMatch-LS on skin lesion diagnosis in the ISIC 2018 and ISIC 2019 challenge datasets, obtaining AUCs of 91.54%, 93.68%, 94.55%, and 95.47% on the four proportions of labeled data from ISIC 2018.

患者,男,31岁.右侧面部出现红色斑块1个月,逐渐形成肿块改变.发病前有被羊角擦伤史.皮损组织病理示:表皮渗出结痂,表皮呈假上皮瘤样增生、棘层呈灶状水肿,真皮乳头高度水肿,真皮层较密集以淋巴细胞、中性粒细胞、嗜酸粒细胞及组织细胞和多核巨细胞为主的炎细胞浸润.皮损组织真菌培养、分枝杆菌快速培养和抗酸杆菌染色均阴性.局部分泌物细菌培养示:金黄色葡萄球菌(MSSA).诊断:面部慢性感染性肉芽肿(多重耐药性金黄色葡萄球菌引起).予口服莫西沙星治疗7天后痊愈.

BACKGROUND:Semi-supervised learning (SSL) can effectively use information from unlabeled data to improve model performance, which has great significance in medical imaging tasks. Pseudo-labeling is a classical SSL method that uses a model to predict unlabeled samples and selects the prediction with the highest confidence level as the pseudo-labels and then uses the generated pseudo-labels to train the model. Most of the current pseudo-label-based SSL algorithms use predefined fixed thresholds for all classes to select unlabeled data.PURPOSE:However, data imbalance is a common problem in medical image tasks, where the use of fixed threshold to generate pseudo-labels ignores different classes of learning status and learning difficulties. The aim of this study is to develop an algorithm to solve this problem.METHODS:In this work, we propose Multi-Curriculum Pseudo-Labeling (MCPL), which evaluates the learning status of the model for each class at each epoch and automatically adjusts the thresholds for each class. We apply MCPL to FixMatch and propose a new SSL framework for medical image classification, which we call the improved algorithm FaxMatch. To mitigate the impact of incorrect pseudo-labels on the model, we use label smoothing (LS) strategy to generate soft labels (SL) for pseudo-labels.RESULTS:We have conducted extensive experiments to evaluate our method on two public benchmark medical image classification datasets: the ISIC 2018 skin lesion analysis and COVID-CT datasets. Experimental results show that our method outperforms fully supervised baseline, which uses only labeled data to train the model. Moreover, our method also outperforms other state-of-the-art methods.CONCLUSIONS:We propose MCPL and construct a semi-supervised medical image classification framework to reduce the reliance of the model on a large number of labeled images and reduce the manual workload of labeling medical image data.

Background: Contrastive learning has achieved remarkable success in representation learning via self-supervision in massive unlabeled data. However, as to classification tasks, there is still a certain gap between the performance of self-supervised learning and fully supervised learning. Purpose: To make the feature extractor learn the feature representation of input samples according to the induced hierarchy and enhance the performance of classifier in fully supervised setting. Methods: This study proposes a novel method, namely multi-hierarchy contrastive learning with pareto optimality (MHC-PO), to promote the overall performance of skin lesion multi-classification. In this work, we introduce a multi-hierarchy contrastive learning framework that adapts the contrastive loss to induced hierarchy structure, to enable the model to learn the feature representation via induced hierarchy. The pareto optimality method is applied to balance the common trade-off issue between the performance of feature extractor and classifier, making fully use of learned features to enhance the performance of classifiers. Results: A series of experiments were conducted on two skin lesion datasets, one is the public dataset HAM10000, and the other is the private dataset XJUSL. Experimental analysis reveals that the proposed MHC-PO surpasses the EfficientNet and other typical models in the performance of all evaluation metrics used in this study. Conclusions: The results from the experiments reflect that the proposed method enhances the clustering of the representation space and improves the overall performance of multi-classification. In addition, MHC-PO is not only promising to be applied in the medical field, but also to other domains with hierarchical data.

Deep convolutional neural network (DCNN) models have been widely used to diagnose skin lesions, and some of them have achieved diagnostic results comparable to or even better than dermatologists. Most publicly available skin lesion datasets used to train DCNN were dermoscopic images. Expensive dermoscopic equipment is rarely available in rural clinics or small hospitals in remote areas. Therefore, it is of great significance to rely on clinical images for computer-aided diagnosis of skin lesions. This paper proposes an improved dual-branch fusion network called CR-Conformer. It integrates a DCNN branch that can effectively extract local features and a Transformer branch that can extract global features to capture more valuable features in clinical skin lesion images. In addition, we improved the DCNN branch to extract enhanced features in four directions through the convolutional rotation operation, further improving the classification performance of clinical skin lesion images. To verify the effectiveness of our proposed method, we conducted comprehensive tests on a private dataset named XJUSL, which contains ten types of clinical skin lesions. The test results indicate that our proposed method reduced the number of parameters by 11.17 M and improved the accuracy of clinical skin lesion image classification by 1.08

Objective To analyze the dermoscopic, reflectance confocal microscopic and histopathological features of 60 cases of atypical scabies. Methods Sixty cases of atypical scabies were analyzed in terms of demographic data, disease duration, and clinical, dermoscopic and reflectance confocal microscopic features. Skin histopathology was performed in 25 patients. Results A larger portion of the patients were aged 18-65 years (40.00%). Retirees, students and migrant workers accounted for 36.67% (22 cases), 20.00% (12 cases) and 18.33% (11 cases), respectively. The average disease duration was 38.12±22.56 days. Twenty-seven patients (45.00%) had a history of contact with patients with scabies. The commonly involved body sites were the palms (52 cases, 86.67%), abdomen (49 cases, 81.67%), wrists (44 cases, 73.33%) and vulva (38 cases, 63.33%). The most common skin rashes were papules (55 cases, 91.67%). Most patients had been misdiagnosed more than twice. The misdiagnoses included pruritus (17 cases, 28.33%), atopic dermatitis (16 cases, 26.67%) and eczema (14 cases, 23.33%). Over 60% of the misdiagnosed patients were treated with glucocorticoids (38 cases, 63.33%). Dermoscopic features included white burrows with triangular structures in brown or dark brown color at the end of burrows (48 cases) and gray-edge lines (42 cases). The typical mite body, burrows and feces were observed in 55 cases under a reflectance confocal microscope. Histopathologically, mites or eggs were seen in the stratum corneum in 17 out of 25 cases, and acute or subacute dermatitis with spongy edema were observed in 12 cases. Dermal infiltrates of eosinophils and lymphocytes were seen in 18 and 7 cases, respectively. Conclusions Atypical scabies displays divergent clinical features. Dermoscopy, reflectance confocal microscopy and histopathology are helpful in the diagnosis of atypical scabies.

BACKGROUND:The results of medical image segmentation can provide reliable evidence for clinical diagnosis and treatment. The U-Net proposed previously has been widely used in the field of medical image segmentation. Its encoder extracts semantic features of different scales at different stages, but does not carry out special processing for semantic features of each scale.OBJECTIVE:To improve the feature expression ability and segmentation performance of U-Net, we proposed a feature supplement and optimization U-Net (FSOU-Net).METHODS:First, we put forward the view that semantic features of different scales should be treated differently. Based on this view, we classify the semantic features automatically extracted by encoders into two categories: shallow semantic features and deep semantic features. Then, we propose the shallow feature supplement module (SFSM), which obtains fine-grained semantic features through up-sampling to supplement the shallow semantic information. Finally, we propose the deep feature optimization module (DFOM), which uses the expansive convolution of different receptive fields to obtain multi-scale features and then performs multi-scale feature fusion to optimize the deep semantic information.RESULTS:The proposed model is experimented on three medical image segmentation public datasets, and the experimental results prove the correctness of the proposed idea. The segmentation performance of the model is higher than the advanced models for medical image segmentation. Compared with baseline network U-NET, the main index of Dice index is 0.75% higher on the RITE dataset, 2.3% higher on the Kvasir-SEG dataset, and 0.24% higher on the GlaS dataset.CONCLUSIONS:The proposed method can greatly improve the feature representation ability and segmentation performance of the model.

患者女, 64岁, 因全身丘疹、溃疡伴瘙痒7个月就诊。2020年2月患者无诱因前胸部出现散在红斑、丘疹, 米粒大小, 伴轻度瘙痒。后逐渐累及头面、颈、背、腹部及左侧大腿, 间歇自行涂抹软膏(具体不详), 经数日皮疹可消退并遗留色素沉着。2020年7月新发皮疹明显增多, 躯干、头面部均受累, 背部瘙痒剧烈, 部分皮疹中央轻度糜烂并结痂, 自行口服抗过敏药物, 病情未见改善。当地医院诊断"大疱性疾病可能", 给予抗过敏、调节免疫治疗, 用药数周效果不理想, 故至本院就诊。门诊以"类天疱疮?"收住入院。患者近期体力、体重未见明显变化。家族中无类似病史。既往体健, 否认糖尿病、心脏病、高血压等病史, 否认药物过敏史, 否认既往皮肤疾病史, 否认特殊物质接触史。体检:一般情况可, 各系统检查未见明显异常。皮肤科检查:头面部、躯干可见米粒至黄豆大小脐状丘疹、斑丘疹, 色泽暗红, 边缘清晰, 部分中央溃烂, 上覆褐色痂皮, 不易去除, 部分皮疹消退后遗留色素沉着, 未见大疱及水疱(图1A、1B)。实验室检查:总三碘甲状腺原氨酸0.72 ng/ml(参考值:0.8 ～ 2.0 ng/ml, 下同), 总甲状腺素4.78 μg/dl(5.1 ～ 14.1 μg/dl), 其他甲状腺功能指标未见异常。血脂系列:甘油三酯1.99 mmol/L(0.48 ～ 1.7 mmol/L), 极低密度脂蛋白0.91 mmol/L(0.18 ～ 0.78 mmol/L), 余未见异常。血尿粪常规、血糖、肝肾功能、电解质、体液免疫(补体C3、补体C4、IgG、IgA、IgM)、自身免疫性抗体(抗核抗体、抗双链DNA抗体、抗可提取性核抗原抗体)等均未见明显异常。肿瘤标志物均阴性。2020-09-21行间接免疫荧光(indirect immunofluorence, IIF)法检测抗棘细胞桥粒抗体阳性(1∶100)、抗桥粒芯糖蛋白(desmoglein, Dsg)1抗体186.00 Ru/ml(<20 Ru/ml);2020-09-28行IIF检测抗棘细胞桥粒抗体阴性, 抗Dsg1抗体>200.00 Ru/ml, 抗表皮基底膜带抗体、抗Dsg3抗体、抗大疱性类天疱疮(bullous pemphigoid, BP)180抗体、抗BP230抗体均阴性。头颅、肺部、腹盆腔计算机断层扫描(CT)均未见异常。

Skin is the first line of defence of the human body. Accurate skin lesions image segmentation is essential for skin disease diagnosis and treatment planning. In this work, we create a new multi-scale channel attention module - MS-CA, which can display more accurate and relevant feature channels on multiple scales. And a network segmentation model for multi-scale channel attention(MS-CA) is proposed. The network model embeds the MS-CA module into two different types of benchmark networks, and modifies the two types of benchmark network models to obtain an image segmentation model suitable for skin diseases. Specifically, by extending the encoder and decoder structures of the LCA-Net network respectively, the MS-CA segmentation model of LCA-Net is proposed, and experiments are carried out on the ISIC2017 and ISIC2018 datasets, and the Dice and accuracy evaluation indicators are improved from 85.33% to 86.71 %, 96.07% increased to 97.23% and the MS-CA segmentation model of TransUNet was proposed. The Jaccard and Dice coefficients were increased from 66.21% to 68.80%, respectively. 79.67% increased to 81.52%. Compared with the existing methods, the MS-CA attention network segmentation model achieves better segmentation results, and it can be used as a new skin disease segmentation network as a computer-aided diagnosis tool.