Skin Tumor Images Research Articles

Deep convolutional neural network with fusion strategy for skin cancer recognition: model development and validation

We aimed to develop an accurate and efficient skin cancer classification system using deep-learning technology with a relatively small dataset of clinical images. We proposed a novel skin cancer classification method, SkinFLNet, which utilizes model fusion and lifelong learning technologies. The SkinFLNet's deep convolutional neural networks were trained using a dataset of 1215 clinical images of skin tumors diagnosed at Taichung and Taipei Veterans General Hospital between 2015 and 2020. The dataset comprised five categories: benign nevus, seborrheic keratosis, basal cell carcinoma, squamous cell carcinoma, and malignant melanoma. The SkinFLNet's performance was evaluated using 463 clinical images between January and December 2021. SkinFLNet achieved an overall classification accuracy of 85%, precision of 85%, recall of 82%, F-score of 82%, sensitivity of 82%, and specificity of 93%, outperforming other deep convolutional neural network models. We also compared SkinFLNet's performance with that of three board-certified dermatologists, and the average overall performance of SkinFLNet was comparable to, or even better than, the dermatologists. Our study presents an efficient skin cancer classification system utilizing model fusion and lifelong learning technologies that can be trained on a relatively small dataset. This system can potentially improve skin cancer screening accuracy in clinical practice.

Use of Artificial Intelligence as a Predictor of the Response to Treatment in Alopecia Areata

Background Artificial intelligence (AI) has emerged in dermatology with some studies focusing on skin disorders such as skin cancer, atopic dermatitis, psoriasis, and onychomycosis. Alopecia areata (AA) is a dermatological disease whose prevalence is 0.7%-3% in the United States, and is characterized by oval areas of nonscarring hair loss of the scalp or body without evident clinical variables to predict its response to the treatment. Nonetheless, some studies suggest a predictive value of trichoscopic features in the evaluation of treatment responses. Assuming that black dots, broken hairs, exclamation marks, and tapered hairs are markers of negative predictive value of the treatment response, while yellow dots are markers of no response to treatment according to recent studies, the absence of these trichoscopic features could indicate favorable disease evolution without treatment or even predict its response. Nonetheless, no studies have reportedly evaluated the role of AI in AA on the basis of trichoscopic features. Objective This study aimed to develop an AI algorithm to predict, using trichoscopic images, those patients diagnosed with AA with a better disease evolution. Methods In total, 80 trichoscopic images were included and classified in those with or without features of negative prognosis. Using a data augmentation technique, they were multiplied to 179 images to train an AI algorithm, as previously carried out with dermoscopic images of skin tumors with a favorable response. Subsequently, 82 new images of AA were presented to the algorithm, and the algorithm classified these patients as responders and non-responders; this process was reviewed by an expert trichologist observer and presented a concordance higher than 90% with the algorithm identifying structures described previously. Evolution of the cases was followed up to truly determine their response to treatment and, therefore, to assess the predictive value of the algorithm. Results In total, 32 of 40 (80%) images of patients predicted as nonresponders scarcely showed response to the treatment, while 34 of 42 (81%) images of those predicted as responders showed a favorable response to the treatment. Conclusions The development of an AI algorithm or tool could be useful to predict AA evolution and its response to treatment. However, further research is needed, including larger sample images or trained algorithms, by using images previously classified in accordance with the disease evolution and not with trichoscopic features. Conflicts of Interest None declared.

NURBS functional network approach for automatic image segmentation of macroscopic medical images in melanoma detection

Image processing techniques are becoming standard technology in many medical specialities, such as dermatology, where they are a key tool for the early detection and diagnosis of melanoma and other skin cancers and tumors. A previous paper by the authors presented at SOCO 2020 conference introduced a new method for image segmentation of skin images through functional networks. The method performs well but it relies on a semi-automatic approach involving a combination of manual and automatic operations. This paper aims at making image segmentation of macroscopic skin images a fully automatic process. To this purpose, the present work extends our previous paper with five new relevant contributions: (1) a filtering strategy for removal of noise, hair and other artifacts; (2) two morphological operators for image enhancement; (3) a clustering-based binary classifier to separate the skin tumor from the image background; (4) a smoothing and discretization process to obtain the border points; and (5) a curve reconstruction method from the border points with NURBS using a new type of functional network particularly tailored for this task. This new method is applied to two different benchmarks, comprised respectively of four and two macroscopic medical images of skin tumors. The visual and numerical results show that the method performs very well, yielding segmented images which are suitable for clinical practice. This method is a significant step towards the future development of a fully automatic approach for the whole medical image analysis pipeline of skin images, including diagnosis and classification.

Discriminative deep learning based benignity/malignancy diagnosis of dermatologic ultrasound skin lesions with pretrained artificial intelligence architecture.

Deep-learning algorithms (DLAs) have been used in artificial intelligence aided ultrasonography diagnosis of thyroid and breast lesions. However, its use has not been described in the case of dermatologic ultrasound lesions. Our purpose was to train a DLA to discriminate benign form malignant lesions in dermatologic ultrasound images. We trained a prebuilt neural network architecture (EfficientNet B4) in a commercial artificial intelligence platform (Peltarion, Stockholm, Sweden) with 235 color Doppler images of both benign and malignant ultrasound images of 235 excised and histologically confirmed skin lesions (84.3% training, 15.7% validation). An additional 35 test images were used for testing the algorithm discrimination for correct benign/malignant diagnosis. One dermatologist with more than 5 years of experience in dermatologic ultrasound blindly evaluated the same 35 test images for malignancy or benignity. EfficientNet B4 trained dermatologic ultrasound algorithm sensitivity; specificity; predictive positive values, and predicted negative values for validation algorithm were 0.8, 0.86, 0.86, and 0.8, respectively for malignancy diagnosis. When tested with 35 previously unevaluated images sets, the algorithm´s accuracy for correct benign/malignant diagnosis was 77.1%, not statistically significantly different from the dermatologist's evaluation (74.1%). An adequately trained algorithm, even with a limited number of images, is at least as accurate as a dermatologic-ultrasound experienced dermatologist in the evaluation of benignity/malignancy of ultrasound skin tumor images devoid of clinical data.

Deep learning-based, computer-aided classifier developed with dermoscopic images shows comparable performance to 164 dermatologists in cutaneous disease diagnosis in the Chinese population.

BackgroundDiagnoses of Skin diseases are frequently delayed in China due to lack of dermatologists. A deep learning-based diagnosis supporting system can facilitate pre-screening patients to prioritize dermatologists’ efforts. We aimed to evaluate the classification sensitivity and specificity of deep learning models to classify skin tumors and psoriasis for Chinese population with a modest number of dermoscopic images.MethodsWe developed a convolutional neural network (CNN) based on two datasets from a consecutive series of patients who underwent the dermoscopy in the clinic of the Department of Dermatology, Peking Union Medical College Hospital, between 2016 and 2018, prospectively. In order to evaluate the feasibility of the algorithm, we used two datasets. Dataset I consisted of 7192 dermoscopic images for a multi-class model to differentiate three most common skin tumors and other diseases. Dataset II consisted of 3115 dermoscopic images for a two-class model to classify psoriasis from other inflammatory diseases. We compared the performance of CNN with 164 dermatologists in a reader study with 130 dermoscopic images. The experts’ consensus was used as the reference standard except for the cases of basal cell carcinoma (BCC), which were all confirmed by histopathology.ResultsThe accuracies of multi-class and two-class models were 81.49% ± 0.88% and 77.02% ± 1.81%, respectively. In the reader study, for the multi-class tasks, the diagnosis sensitivity and specificity of 164 dermatologists were 0.770 and 0.962 for BCC, 0.807 and 0.897 for melanocytic nevus, 0.624 and 0.976 for seborrheic keratosis, 0.939 and 0.875 for the “others” group, respectively; the diagnosis sensitivity and specificity of multi-class CNN were 0.800 and 1.000 for BCC, 0.800 and 0.840 for melanocytic nevus, 0.850 and 0.940 for seborrheic keratosis, 0.750 and 0.940 for the “others” group, respectively. For the two-class tasks, the sensitivity and specificity of dermatologists and CNN for classifying psoriasis were 0.872 and 0.838, 1.000 and 0.605, respectively. Both the dermatologists and CNN achieved at least moderate consistency with the reference standard, and there was no significant difference in Kappa coefficients between them (P > 0.05).ConclusionsThe performance of CNN developed with relatively modest number of dermoscopic images of skin tumors and psoriasis for Chinese population is comparable with 164 dermatologists. These two models could be used for screening in patients suspected with skin tumors and psoriasis respectively in primary care hospital.

Machine Learning in Melanoma Diagnosis. Limitations About to be Overcome

BackgroundAutomated image classification is a promising branch of machine learning (ML) useful for skin cancer diagnosis, but little has been determined about its limitations for general usability in current clinical practice. ObjectiveTo determine limitations in the selection of skin cancer images for ML analysis, particularly in melanoma. MethodsRetrospective cohort study design, including 2,849 consecutive high-quality dermoscopy images of skin tumors from 2010 to 2014, for evaluation by a ML system. Each dermoscopy image was assorted according to its eligibility for ML analysis. ResultsOf the 2,849 images chosen from our database, 968 (34%) met the inclusion criteria for analysis by the ML system. Only 64.7% of nevi and 36.6% of melanoma met the inclusion criteria. Of the 528 melanomas, 335 (63.4%) were excluded. An absence of normal surrounding skin (40.5% of all melanomas from our database) and absence of pigmentation (14.2%) were the most common reasons for exclusion from ML analysis. DiscussionOnly 36.6% of our melanomas were admissible for analysis by state-of-the-art ML systems. We conclude that future ML systems should be trained on larger datasets which include relevant non-ideal images from lesions evaluated in real clinical practice. Fortunately, many of these limitations are being overcome by the scientific community as recent works show.

Uso del aprendizaje automático en el diagnóstico del melanoma. Limitaciones por superar

BackgroundAutomated image classification is a promising branch of machine learning (ML) useful for skin cancer diagnosis, but little has been determined about its limitations for general usability in current clinical practice. ObjectiveTo determine limitations in the selection of skin cancer images for ML analysis, particularly in melanoma. MethodsRetrospective cohort study design, including 2,849 consecutive high-quality dermoscopy images of skin tumors from 2010 to 2014, for evaluation by a ML system. Each dermoscopy image was assorted according to its eligibility for ML analysis. ResultsOf the 2,849 images chosen from our database, 968 (34%) met the inclusion criteria for analysis by the ML system. Only 64.7% of nevi and 36.6% of melanoma met the inclusion criteria. Of the 528 melanomas, 335 (63.4%) were excluded. An absence of normal surrounding skin (40.5% of all melanomas from our database) and absence of pigmentation (14.2%) were the most common reasons for exclusion from ML analysis. DiscussionOnly 36.6% of our melanomas were admissible for analysis by state-of-the-art ML systems. We conclude that future ML systems should be trained on larger datasets which include relevant non-ideal images from lesions evaluated in real clinical practice. Fortunately, many of these limitations are being overcome by the scientific community as recent works show.

Automated Analysis of the Pigment Network in Dermatoscopic Images of Melanocytic Skin Tumors

A method for recognition of the pigment network lines in dermatoscopic images of skin tumors is presented. The method provides calculation of characteristics of the pigment network lines and imaging of the obtained results. Experimental assessment of the effectiveness of the proposed method showed it to be promising for use in melanoma recognition systems.

Usage of Skin Tumor Images on the Internet for Personal Computer Based Automated Cognition

Usage of Skin Tumor Images on the Internet for Personal Computer Based Automated Cognition

Differential optical spectropolarimetric imaging system assisted by liquid crystal devices for skin imaging

Skin cancer diagnosis depends not only on histopathological examination but also on visual inspection before and after the excision of suspected lesion. Neoplasm is accompanied with changes in birefringence of collagen, pleomorphicity, and hyperchromatic state of epithelial nuclei. These phenomena can be measured by spectral and polarization changes of light backscattered by the examined tissue. A new differential spectropolarimetric system is proposed using liquid crystal devices, one as a tunable filter and the other as a polarization rotator, both operating at wide spectral ranges from the visible to the near-infrared. Since collagen's fibrils texture orientation depends on its location in the skin and since it is not well organized, our system scans the bipolarization states by continuously rotating the linearly polarized light incident on a skin lesion, and collecting differential contrasts between sequenced images when simultaneously averaging the statistical readout of a video camera. This noninvasive method emphasizes areas on skin where the neoplasm, or tumor, minimizes the statistical polarization change of the scattered light from the lesion. The module can be considered as an assistant tool for epiluminescence microscopy. Images of skin tumors were captured in vivo before the patients having their surgery and compared to histopathological results.

Multiphoton Laser Scanning Microscopy—A Novel Diagnostic Method for Superficial Skin Cancers

The increasing incidence of skin cancer and the importance of early diagnosis is a challenge, which requires the development of reliable, cost-effective, noninvasive, diagnostic techniques. Several such methods based on optical imaging techniques are available and currently being investigated. A novel method in this field is multiphoton laser scanning microscopy (MPLSM). This technique is based on the nonlinear process of 2-photon excitation of endogenous fluorophores, which can be used to acquire horizontal optical sectioning of intact biological tissue samples. When studying human skin, MPLSM provides high-resolution fluorescence imaging, allowing visualization of cellular and subcellular structures of the epidermis and upper dermis. This review covers the application of MPLSM as a diagnostic tool for superficial skin cancers, such as basal cell carcinomas, squamous cell carcinoma in situ, and melanomas. MPLSM has also been applied in other research areas related to skin, which will be mentioned briefly. The morphologic features observed in MPLSM images of skin tumors are comparable to traditional histopathology. Safety issues, limitations, and further improvements are discussed. Although further investigations are required to make MPLSM a mainstream clinical diagnostic tool, MPLSM has the potential of becoming a noninvasive, bedside, histopathologic technique for the diagnosis of superficial skin cancers.

The real-time, three-dimensional analyses of benign and malignant skin tumors by confocal laser scanning microscopy

In obtain images of skin tumors non-invasively with real-time, confocal laser scanning microscope (CLSM) is introduced. Reconstructed images of given horizontal sections were converted into three-dimensions using the data set of a large number of tomograms in the horizontal directions. To develop the multiplaner reconstruction images of skin tumors in vertical directions and three-dimensionally reconstructed images of tumors will be obtained from the continuously collected horizontal image data sets. Three-dimensional analyses of the skin tumors from reconstructed images of the CLSM scanning have provided the information as to their physiological characteristics as well as the extent of deep invasion in real-time with non-invasive manner. High performance three-dimensional conversion software was effective in displaying three-dimensional construction of skin tumors. The CLSM scanning images followed by three-dimensional reconstruction using them can provide the real-time and non-invasive diagnoses of skin tumors and analyze the radial growth phase of tumors and the three-dimensional growth characteristics.

Border detection on digitized skin tumor images.

A radial search technique is presented for detecting skin tumor borders in clinical dermatology images. First, it includes two rounds of radial search based on the same tumor center. The first-round search is independent, and the second-round search is knowledge-based tracking. Then a rescan with a new center is used to solve the blind-spot problem. The algorithm is tested on model images with excellent performance, and on 300 real clinical images with a satisfactory result.

Performance of AI methods in detecting melanoma

This research has shown that features extracted from color skin tumor images by computer vision methods can be reliable discriminators of malignant tumors from benign ones. Reliability was demonstrated by the monotonically increasing success ratios with increasing training set size and by the small standard deviations from the mean success rates. An average success rate of 70 percent in diagnosing melanoma was attained for a training set size of 60 percent. The presence or absence of atypical moles in the training and test sets was shown to have a dramatic impact on the effectiveness of the generated classification rules. This was the case with both AIM and lst-Class, and indicates a high potential for success if a method can be found for discriminating between atypical moles and melanoma.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Neural network diagnosis of malignant melanoma from color images.

Malignant melanoma is the deadliest form of all skin cancers. Approximately 32,000 new cases of malignant melanoma were diagnosed in 1991 in the United States, with approximately 80% of patients expected to survive five years [1]. Fortunately, if detected early, even malignant melanoma may be treated successfully. Thus, in recent years, there has been rising interest in the automated detection and diagnosis of skin cancer, particularly malignant melanoma [2]. In this paper, we present a novel neural network approach for the automated separation of melanoma from three benign categories of tumors which exhibit melanoma-like characteristics. Our approach uses discriminant features, based on tumor shape and relative tumor color, that are supplied to an artificial neural network for classification of tumor images as malignant or benign. With this approach, for reasonably balanced training/testing sets, we are able to obtain above 80% correct classification of the malignant and benign tumors on real skin tumor images.

Automatic color segmentation algorithms-with application to skin tumor feature identification

Two color-image segmentation methods are described. The first is based on a spherical coordinate transform of original RGB data. The second is based on a mathematically optimal transform, the principal components transform (also known as eigenvector, discrete Karhunen-Loeve, or Hotelling transform). These algorithms are applied to the extraction from skin tumor images of various features such as tumor border, crust, hair scale, shiny areas, and ulcer. The results of this research will be used in the development of a computer vision system that will seve as the visual front-end of a medical expert system to automate visual feature identification for skin tumor evaluation.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Automatic detection of irregular borders in melanoma and other skin tumors

An irregularity index previously developed is applied to detect irregular borders automatically in skin tumor images, particularly malignant melanoma. The irregularity index is used to classify various tumor borders as irregular or regular. This procedure processes tumor images with borders automatically determined by a radial search algorithm previously described. Potential use of this algorithm in an in vivo skin cancer detection system and errors expected in the use of the algorithm are discussed.

An automatic color segmentation algorithm with application to identification of skin tumor borders

A principal components transform algorithm for automatic color segmentation of images is described. This color segmentation algorithm was used to find tumor borders in six different color spaces including the original red, green, and blue (RGB) color space of the digitized image, the intensity/hue/saturation (IHS) transform, the spherical transform, chromaticity coordinates, the CIE transform and the uniform color transform designated CIE-LUV. Five hundred skin tumor images were separated into a training set and a test set for comparison of the different color spaces. Automatic induction was applied to dynamically determine the number of colors for segmentation. Ninety-one percent of image variance was contained in the image component along the principal axis (also containing the most image information). When compared to a luminance radial search method, the principal components color segmentation border method performed equally well by one measure and 10% better by another measure, including more near border points outside the tumor. The spherical transform provides the highest success rate and the chromaticity transform the lowest error rate, although large variances in the data preclude definitive statistical comparisons.

Boundary detection in skin tumor images: An overall approach and a radial search algorithm

Although computerized boundary detection has been studied in depth, current general algorithms are not highly successful when applied to in vivo medical images where the borders are often not clearly defined and are sometimes difficult even for the human eye to detect. In these cases, domain-specific algorithms are necessary to achieve the required accuracy. This paper addresses the problem of automatic detection of tumor borders in digitized images of skin tumors. The complexity of boundary detection in this domain is such that algorithms based on single boundary determinants such as color, luminance, texture, or three-dimensional information are not capable of correctly identifying the boundary in all cases. Hence, an overall approach is discussed that will use confidence levels to combine results from several border detection algorithms based on the individual criteria mentioned. The development of one such algorithm (using a radial search method on luminance information) and its results are presented. With this approach, problems such as numerous false borders and unknown shape and size of the tumors are overcome.

Automatic color segmentation of images with application to detection of variegated coloring in skin tumors

A description is given of a computer vision system, developed to serve as the front-end of a medical expert system, that automates visual feature identification for skin tumor evaluation. The general approach is to create different software modules that detect the presence or absence of critical features. Image analysis with artificial intelligence (AI) techniques, such as the use of heuristics incorporated into image processing algorithms, is the primary approach. On a broad scale, this research addressed the problem of segmentation of a digital image based on color information. The algorithm that was developed to segment the image strictly on the basis of color information was shown to be a useful aid in the identification of tumor border, ulcer, and other features of interest. As a specific application example, the method was applied to 200 digitized skin tumor images to identify the feature called variegated coloring. Extensive background information is provided, and the development of the algorithm is described.