Deep learning approach to skin layers segmentation in inflammatory dermatoses

Artificial Intelligence: A Snapshot of Its Application in Chronic Inflammatory and Autoimmune Skin Diseases

Immuno-correlated dermatological pathologies refer to skin disorders that are closely associated with immune system dysfunction or abnormal immune responses. Advancements in the field of artificial intelligence (AI) have shown promise in enhancing the diagnosis, management, and assessment of immuno-correlated dermatological pathologies. This intersection of dermatology and immunology plays a pivotal role in comprehending and addressing complex skin disorders with immune system involvement. The paper explores the knowledge known so far and the evolution and achievements of AI in diagnosis; discusses segmentation and the classification of medical images; and reviews existing challenges, in immunological-related skin diseases. From our review, the role of AI has emerged, especially in the analysis of images for both diagnostic and severity assessment purposes. Furthermore, the possibility of predicting patients' response to therapies is emerging, in order to create tailored therapies.

A deep learning fusion network trained with clinical and high-frequency ultrasound images in the multi-classification of skin diseases in comparison with dermatologists: a prospective and multicenter study

Background

Clinical appearance and high-frequency ultrasound (HFUS) are indispensable for diagnosing skin diseases by providing internal and external information. However, their complex combination brings challenges for primary care physicians and dermatologists. Thus, we developed a deep multimodal fusion network (DMFN) model combining analysis of clinical close-up and HFUS images for binary and multiclass classification in skin diseases.

Methods

Between Jan 10, 2017, and Dec 31, 2020, the DMFN model was trained and validated using 1269 close-ups and 11,852 HFUS images from 1351 skin lesions. The monomodal convolutional neural network (CNN) model was trained and validated with the same close-up images for comparison. Subsequently, we did a prospective and multicenter study in China. Both CNN models were tested prospectively on 422 cases from 4 hospitals and compared with the results from human raters (general practitioners, general dermatologists, and dermatologists specialized in HFUS). The performance of binary classification (benign vs. malignant) and multiclass classification (the specific diagnoses of 17 types of skin diseases) measured by the area under the receiver operating characteristic curve (AUC) were evaluated. This study is registered with www.chictr.org.cn (ChiCTR2300074765).

Findings

The performance of the DMFN model (AUC, 0.876) was superior to that of the monomodal CNN model (AUC, 0.697) in the binary classification (P = 0.0063), which was also better than that of the general practitioner (AUC, 0.651, P = 0.0025) and general dermatologists (AUC, 0.838; P = 0.0038). By integrating close-up and HFUS images, the DMFN model attained an almost identical performance in comparison to dermatologists (AUC, 0.876 vs. AUC, 0.891; P = 0.0080). For the multiclass classification, the DMFN model (AUC, 0.707) exhibited superior prediction performance compared with general dermatologists (AUC, 0.514; P = 0.0043) and dermatologists specialized in HFUS (AUC, 0.640; P = 0.0083), respectively. Compared to dermatologists specialized in HFUS, the DMFN model showed better or comparable performance in diagnosing 9 of the 17 skin diseases.

Interpretation

The DMFN model combining analysis of clinical close-up and HFUS images exhibited satisfactory performance in the binary and multiclass classification compared with the dermatologists. It may be a valuable tool for general dermatologists and primary care providers.

Funding

This work was supported in part by the National Natural Science Foundation of China and the Clinical research project of Shanghai Skin Disease Hospital.

High-frequency ultrasound in anti-aging skin therapy monitoring

Over the last few decades, high-frequency ultrasound has found multiple applications in various diagnostic fields. The fast development of this imaging technique opens up new diagnostic paths in dermatology, allergology, cosmetology, and aesthetic medicine. In this paper, being the first in this area, we discuss the usability of HFUS in anti-aging skin therapy assessment. The fully automated algorithm combining high-quality image selection and entry echo layer segmentation steps followed by the dermal parameters estimation enables qualitative and quantitative evaluation of the effectiveness of anti-aging products. Considering the parameters of subcutaneous layers, the proposed framework provides a reliable tool for TCA-peel therapy assessment; however, it can be successfully applied to other skin-condition-related problems. In this randomized controlled clinical trial, forty-six postmenopausal women were randomly assigned to the experimental and control groups. Women were treated four times at one-week intervals and applied skin cream daily between visits. The three month follow-up study enables measurement of the long-term effect of the therapy. According to the results, the TCA-based therapy increased epidermal (entry echo layer) thickness, indicating that the thinning process has slowed down and the skin's condition has improved. An interesting outcome is the obtained growth in the intensity of the upper dermis in the experimental group, which might suggest a reduced photo-aging effect of TCA-peel and increased water content. The same conclusions connected with the anti-aging effect of TCA-peel can be drawn by observing the parameters describing the contribution of low and medium-intensity pixels in the upper dermis. The decreased share of low-intensity pixels and increased share of medium-intensity pixels in the upper dermis suggest a significant increase in local protein synthesis.

High-frequency ultrasound biomicroscopy findings of the skin of dogs with atopic dermatitis

BACKGROUND

The high-frequency ultrasonographic appearance of skin of dogs with atopic dermatitis (cAD) has not been described.

OBJECTIVES

To compare high-frequency ultrasonographic findings among lesional, macroscopically nonlesional skin of dogs with cAD, and the macroscopically nonlesional skin of healthy dogs. Additionally, to determine whether there is any correlation between the ultrasonographic findings in lesional skin and local Canine Atopic Dermatitis Extent and Severity Index, 4th iteration (CADESI-04) or its domains (erythema, lichenification, excoriations/alopecia). As a secondary aim, six cAD dogs were re-evaluated after management intervention.

ANIMALS

Twenty dogs with cAD (six were re-examined after treatment) and six healthy dogs.

MATERIALS AND METHODS

In all dogs, ultrasonographic examination was performed on the same 10 skin sites, using a 50 MHz transducer. Wrinkling of skin surface, presence/width of subepidermal low echogenic band, hypoechogenicity of dermis and thickness of the skin were evaluated and scored/measured blindly.

RESULTS

Dermal hypoechogenicity was more common and severe in lesional compared to macroscopically nonlesional skin of dogs with cAD. In lesional skin, presence/severity of wrinkling of skin surface and of dermal hypoechogenicity were positively correlated with presence/severity of lichenification, while severity of dermal hypoechogenicity was positively correlated with local CADESI-04. A positive correlation between the change in skin thickness and the change in the severity of erythema during treatment was noted.

CONCLUSIONS AND CLINICAL RELEVANCE

High-frequency ultrasound biomicroscopy may be useful for the evaluation of skin of dogs with cAD and for evaluating the progression of skin lesions during treatment.

Multi-frequency acoustic hologram generation with a physics-enhanced deep neural network

Here, a physics-enhanced multi-frequency acoustic hologram deep neural network (PhysNet_MFAH) method is proposed for designing multi-frequency acoustic holograms, which is built by incorporating multiple physical models that represent the physical processes of acoustic waves propagation for a set of design frequencies into a deep neural network. It is demonstrated that one needs only to feed a set of frequency-specific target patterns into the network, the proposed PhysNet_MFAH method can automatically, accurately, and rapidly generate a high-quality multi-frequency acoustic hologram for holographic rendering of different target acoustic fields in the same or distinct regions of the target plane when driven at different frequencies. Remarkably, it is also demonstrated that the proposed PhysNet_MFAH method can achieve a higher quality of the reconstructed acoustic intensity fields than the existing optimization methods IASA and DS for designing multi-frequency acoustic holograms at a relatively fast-computational speed. Furthermore, the performance dependencies of the proposed PhysNet_MFAH method on different design parameters are established, which provide insight into the performance of the reconstructed acoustic intensity fields when subject to different design conditions of the proposed PhysNet_MFAH method. We believe that the proposed PhysNet_MFAH method can facilitate many potential applications of acoustic holograms, ranging from dynamic particle manipulation to volumetric display.

A deep learning-based approach toward differentiating scalp psoriasis and seborrheic dermatitis from dermoscopic images

OBJECTIVES

This study aims to develop a new diagnostic method for discriminating scalp psoriasis and seborrheic dermatitis based on a deep learning (DL) model, which uses the dermatoscopic image as input and achieved higher accuracy than dermatologists trained with dermoscopy.

METHODS

A total of 1,358 pictures (obtained from 617 patients) with pathological and diagnostic confirmed skin diseases (508 psoriases, 850 seborrheic dermatitides) were randomly allocated into the training, validation, and testing datasets (1,088/134/136) in this study. A DL model concerning dermatoscopic images was established using the transfer learning technique and trained for diagnosing two diseases.

RESULTS

The developed DL model exhibits good sensitivity, specificity, and Area Under Curve (AUC) (96.1, 88.2, and 0.922%, respectively), it outperformed all dermatologists in the diagnosis of scalp psoriasis and seborrheic dermatitis when compared to five dermatologists with various levels of experience. Furthermore, non-proficient doctors with the assistance of the DL model can achieve comparable diagnostic performance to dermatologists proficient in dermoscopy. One dermatology graduate student and two general practitioners significantly improved their diagnostic performance, where their AUC values increased from 0.600, 0.537, and 0.575 to 0.849, 0.778, and 0.788, respectively, and their diagnosis consistency was also improved as the kappa values went from 0.191, 0.071, and 0.143 to 0.679, 0.550, and 0.568, respectively. DL enjoys favorable computational efficiency and requires few computational resources, making it easy to deploy in hospitals.

CONCLUSIONS

The developed DL model has favorable performance in discriminating two skin diseases and can improve the diagnosis, clinical decision-making, and treatment of dermatologists in primary hospitals.

Computer-Aided Diagnosis Methods for High-Frequency Ultrasound Data Analysis: A Review

Over the last few decades, computer-aided diagnosis systems have become a part of clinical practice. They have the potential to assist clinicians in daily diagnostic tasks. The image processing techniques are fast, repeatable, and robust, which helps physicians to detect, classify, segment, and measure various structures. The recent rapid development of computer methods for high-frequency ultrasound image analysis opens up new diagnostic paths in dermatology, allergology, cosmetology, and aesthetic medicine. This paper, being the first in this area, presents a research overview of high-frequency ultrasound image processing techniques, which have the potential to be a part of computer-aided diagnosis systems. The reviewed methods are categorized concerning the application, utilized ultrasound device, and image data-processing type. We present the bridge between diagnostic needs and already developed solutions and discuss their limitations and future directions in high-frequency ultrasound image analysis. A search was conducted of the technical literature from 2005 to September 2022, and in total, 31 studies describing image processing methods were reviewed. The quantitative and qualitative analysis included 39 algorithms, which were selected as the most effective in this field. They were completed by 20 medical papers and define the needs and opportunities for high-frequency ultrasound application and CAD development.

Deep learning for real-time auxiliary diagnosis of pancreatic cancer in endoscopic ultrasonography

In recent year, many deep learning have been playing an important role in the detection of cancers. This study aimed to real-timely differentiate a pancreatic cancer (PC) or a non-pancreatic cancer (NPC) lesion via endoscopic ultrasonography (EUS) image. A total of 1213 EUS images from 157 patients (99 male, 58 female) with pancreatic disease were used for training, validation and test groups. Before model training, regions of interest (ROIs) were manually drawn to mark the PC and NPC lesions using Labelimage software. Yolov5m was used as the algorithm model to automatically distinguish the presence of pancreatic lesion. After training the model based on EUS images using YOLOv5, the parameters achieved convergence within 300 rounds (GIoU Loss: 0.01532, Objectness Loss: 0.01247, precision: 0.713 and recall: 0.825). For the validation group, the mAP0.5 was 0.831, and mAP@.5:.95 was 0.512. In addition, the receiver operating characteristic (ROC) curve analysis showed this model seemed to have a trend of more AUC of 0.85 (0.665 to 0.956) than the area under the curve (AUC) of 0.838 (0.65 to 0.949) generated by physicians using EUS detection without puncture, although pairwise comparison of ROC curves showed that the AUC between the two groups was not significant (z= 0.15, p = 0.8804). This study suggested that the YOLOv5m would generate attractive results and allow for the real-time decision support for distinction of a PC or a NPC lesion.

High-Frequency Ultrasound Dataset for Deep Learning-Based Image Quality Assessment

This study aims at high-frequency ultrasound image quality assessment for computer-aided diagnosis of skin. In recent decades, high-frequency ultrasound imaging opened up new opportunities in dermatology, utilizing the most recent deep learning-based algorithms for automated image analysis. An individual dermatological examination contains either a single image, a couple of pictures, or an image series acquired during the probe movement. The estimated skin parameters might depend on the probe position, orientation, or acquisition setup. Consequently, the more images analyzed, the more precise the obtained measurements. Therefore, for the automated measurements, the best choice is to acquire the image series and then analyze its parameters statistically. However, besides the correctly received images, the resulting series contains plenty of non-informative data: Images with different artifacts, noise, or the images acquired for the time stamp when the ultrasound probe has no contact with the patient skin. All of them influence further analysis, leading to misclassification or incorrect image segmentation. Therefore, an automated image selection step is crucial. To meet this need, we collected and shared 17,425 high-frequency images of the facial skin from 516 measurements of 44 patients. Two experts annotated each image as correct or not. The proposed framework utilizes a deep convolutional neural network followed by a fuzzy reasoning system to assess the acquired data’s quality automatically. Different approaches to binary and multi-class image analysis, based on the VGG-16 model, were developed and compared. The best classification results reach 91.7% accuracy for the first, and 82.3% for the second analysis, respectively.

Automated segmentation of epidermis in high-frequency ultrasound of pathological skin using a cascade of DeepLab v3+ networks and fuzzy connectedness

This study proposes a novel, fully automated framework for epidermal layer segmentation in different skin diseases based on 75 MHz high-frequency ultrasound (HFUS) image data. A robust epidermis segmentation is a vital first step to detect changes in thickness, shape, and intensity and therefore support diagnosis and treatment monitoring in inflammatory and neoplastic skin lesions. Our framework links deep learning and fuzzy connectedness for image analysis. It consists of a cascade of two DeepLab v3+ models with a ResNet-50 backbone and a fuzzy connectedness analysis module for fine segmentation. Both deep models are pre-trained on the ImageNet dataset and subjected to transfer learning using our HFUS database of 580 images with atopic dermatitis, psoriasis and non-melanocytic skin tumors. The first deep model is used to detect the appropriate region of interest, while the second stands for the main segmentation procedure. We use the softmax layer of the latter twofold to prepare the input data for fuzzy connectedness analysis: as a reservoir of seed points and a direct contribution to the input image. In the experiments, we analyze different configurations of the framework, including region of interest detection, deep model backbones and training loss functions, or fuzzy connectedness analysis with parameter settings. We also use the Dice index and epidermis thickness to compare our results to state-of-the-art approaches. The Dice index of 0.919 yielded by our model over the entire dataset (and exceeding 0.93 in inflammatory diseases) proves its superiority over the other methods.

Deep Learning-Based High-Frequency Ultrasound Skin Image Classification with Multicriteria Model Evaluation

This study presents the first application of convolutional neural networks to high-frequency ultrasound skin image classification. This type of imaging opens up new opportunities in dermatology, showing inflammatory diseases such as atopic dermatitis, psoriasis, or skin lesions. We collected a database of 631 images with healthy skin and different skin pathologies to train and assess all stages of the methodology. The proposed framework starts with the segmentation of the epidermal layer using a DeepLab v3+ model with a pre-trained Xception backbone. We employ transfer learning to train the segmentation model for two purposes: to extract the region of interest for classification and to prepare the skin layer map for classification confidence estimation. For classification, we train five models in different input data modes and data augmentation setups. We also introduce a classification confidence level to evaluate the deep model’s reliability. The measure combines our skin layer map with the heatmap produced by the Grad-CAM technique designed to indicate image regions used by the deep model to make a classification decision. Moreover, we propose a multicriteria model evaluation measure to select the optimal model in terms of classification accuracy, confidence, and test dataset size. The experiments described in the paper show that the DenseNet-201 model fed with the extracted region of interest produces the most reliable and accurate results.