Malignant Skin Lesions Research Articles

To Refer or Not to Refer in Teledermoscopy: Retrospective Study.

Challenges remain for general practitioners (GPs) in diagnosing (pre)malignant and benign skin lesions. Teledermoscopy (TDsc) supports GPs in diagnosing these skin lesions guided by teledermatologists' (TDs) diagnosis and advice and prevents unnecessary referrals to dermatology care. However, the impact of the availability of TDsc on GPs' self-reported referral decisions to dermatology care before and after the TDsc consultation is unknown. The objective of this study is to assess and compare the initial self-reported referral decisions of GPs before TDsc versus their final self-reported referral decisions after TDsc for skin lesions diagnosed by the TD as (pre)malignant or benign. TDsc consultations requested by GPs in daily practice between July 2015 and June 2020 with a TD assessment and diagnosis were extracted from a nationwide Dutch telemedicine database. Based on GP self-administered questions, the GPs' referral decisions before and their final referral decision after TDsc consultation were assessed for (pre)malignant and benign TD diagnoses. GP self-administered questions and TD diagnoses were evaluated for 6364 TDsc consultations (9.3% malignant, 8.8% premalignant, and 81.9% benign skin lesions). In half of the TDsc consultations, GPs adjusted their initial referral decision after TD advice and TD diagnosis. Initially, GPs did not have the intention to refer 67 (56.8%) of 118 patients with a malignant TD diagnosis and 26 (16.0%) of 162 patients with a premalignant TD diagnosis but then decided to refer these patients after the TDsc consultation. Furthermore, GPs adjusted their decision from referral to nonreferral for 2534 (74.9%) benign skin lesions (including 676 seborrheic keratosis and 131 vascular lesions). GPs adjusted their referral decision in 52% (n=3306) of the TDsc consultations after the TD assessment. The availability of TDsc is thus of added value and assists GPs in their (non)referral for patients with skin lesions to dermatology care. TDsc resulted in referrals of patients with (pre)malignant skin lesions that GPs would not have referred directly to the dermatologist. TDsc also led to a reduction of unnecessary referrals of patients with low complex benign skin lesions (eg, seborrheic keratosis and vascular lesions).

35126 Education to improve general medical doctors’ dermatology referrals

General hospital doctors often find it challenging to correctly diagnose and appropriately refer common benign and malignant skin lesions. A series of dermatology images of melanoma, nonmelanoma malignancies, actinic keratosis and benign lesions was presented in a quiz format to general medical doctors in a model 3 hospital. Dermatology experience, confidence in lesion recognition and decision for dermatology referral was surveyed pre-teaching. A teaching session with follow-up quiz and self-assessed confidence of lesion recognition and diagnosis was surveyed. Comparison pre and post quiz scores were analyzed. Twenty-four medical doctors ranging from intern to consultant participated in this study. 96% (n = 23) had no prior dermatology experience. Following teaching, 61% (n = 14) expressed feeling confident in their diagnosis of presented images, compared with 17% (n = 4) preteaching. The follow-up quiz showed a global 24% increase in lesion recognition. Correct referral of actinic keratosis improved from 79% to 100%, acral melanoma improved from 75% to 90%, melanoma referral improved from 77% to 93% and BCC referral improved from 88% to 100%. 91% (n = 21) found the teaching session helpful for future practice. The teaching session improved the ability to correctly identify features of common benign, premalignant and malignant skin lesions. The follow-up quiz demonstrated improvement in appropriate referrals for premalignant and malignant lesions. Regular dermatology teaching with images and self-assessment quizzes should improve the confidence of general hospital doctors to correctly recognize and refer to dermatology services. This study demonstrates the need for dermatology teaching to general medical doctors to improve the accurate diagnosis and referral of common skin lesions.

SkinNet-16: A deep learning approach to identify benign and malignant skin lesions.

Skin cancer these days have become quite a common occurrence especially in certain geographic areas such as Oceania. Early detection of such cancer with high accuracy is of utmost importance, and studies have shown that deep learning- based intelligent approaches to address this concern have been fruitful. In this research, we present a novel deep learning- based classifier that has shown promise in classifying this type of cancer on a relevant preprocessed dataset having important features pre-identified through an effective feature extraction method.Skin cancer in modern times has become one of the most ubiquitous types of cancer. Accurate identification of cancerous skin lesions is of vital importance in treating this malady. In this research, we employed a deep learning approach to identify benign and malignant skin lesions. The initial dataset was obtained from Kaggle before several preprocessing steps for hair and background removal, image enhancement, selection of the region of interest (ROI), region-based segmentation, morphological gradient, and feature extraction were performed, resulting in histopathological images data with 20 input features based on geometrical and textural features. A principle component analysis (PCA)-based feature extraction technique was put into action to reduce the dimensionality to 10 input features. Subsequently, we applied our deep learning classifier, SkinNet-16, to detect the cancerous lesion accurately at a very early stage. The highest accuracy was obtained with the Adamax optimizer with a learning rate of 0.006 from the neural network-based model developed in this study. The model also delivered an impressive accuracy of approximately 99.19%.

Assessing the impact of color blindness on the ability of identifying benign and malignant skin lesions by naked-eye examination.

Color vision deficiency describes the inability to distinguish certain shades of color. The aim of this study was to assess the impact of having color vision deficiency on the accuracy of distinguishing benign and malignant skin lesions by naked-eye examination. This was a cross-sectional study conducted during the period August 2020 to February 2021. We randomly selected a total of 20 nevi and 20 melanoma images from an open access image database. The 40 images were divided into four sets of images, each set contained 5 benign and 5 malignant skin lesion images simulated as if they were seen by a protanope physician, deuteranope physician, tritanope physician, and a set of images presented without simulation. In an online survey, students who were in their final year of medical school or had newly graduated were asked to diagnose each image as benign or malignant. A total of 140 participants were included with a mean (SD) age of 24.88 (1.51). We found a significantly higher mean accuracy for non-simulated images compared to deuteranope simulated images (p< 0.001, mean difference = 11.07, 95% CI 8.40 to 13.74). We did not find a significant difference in accuracy classification for protanope simulated images (p = 0.066), nor for tritanope simulated images (p = 0.315). Classification accuracy for malignant lesions was higher than classification accuracy for benign lesions, with the highest difference belonging to deuteranope simulated images, with a difference in mean accuracy of classifying malignant lesions by 32.2 (95% CI 27.0 to 37.6). Deuteranope participants (i.e., green color deficiency) had a significantly lower accuracy of distinguishing pigmented skin lesions as benign or malignant, an impact not found for other color vision deficiencies, which was mainly for misdiagnosing benign lesions as malignant.

Assessing individual patients' knowledge of benign versus malignant skin lesions in the dermatology clinic population.

Background:Skin cancer education targeted to patients’ needs is a goal of practicing dermatologists. Data regarding dermatology patients’ baseline knowledge regarding skin cancer could aid clinicians in tailoring education efforts.Objective:To help quantify existing patients’ existing visual recognition of skin cancer and common benign lesions, with the goal of helping to provide more targeted and meaningful education to patients.Methods:Two hundred forty-four adult patients from the dermatology clinics at University of Oklahoma and Loyola University Chicago were surveyed using digital images and questions regarding personal and family history of skin cancer, sun protection practices and sun protection knowledge.Results:Of the 244 subjects, 43% percent had a positive personal history of skin cancer, 40% had a positive family history. Scores differed minimally by personal history of skin cancer (p = .37) but differed more markedly by family history of skin cancer (p = .02).Limitations:Lack of generalizability to the general public, age range of subjects.Conclusions:There are knowledge gaps within the dermatology patient population regarding common benign and malignant skin lesions.

Common benign and malignant tumours of the aging skin: Characteristics and treatment options

Abstract The incidence of all types of malignant skin tumours, including both melanoma and non-melanoma types, has increased in recent decades, while basal cell carcinoma is the most common human malignancy in the Caucasian race. The aging of the skin is associated with an increase in both benign and malignant tumours. As the population ages and life expectancy extends, mostly in developed countries, dermatologists are likely to face growing numbers of patients seeking therapy for such abnormalities. It is primarily UV irradiation that is responsible for the development of skin cancers, although there are other risk factors, including air pollution and X-ray irradiation. Seborrhoeic keratosis, solar lentigo and other benign lesions, despite their harmless nature, may cause distress to patients, such as itching or aesthetic issues. This review article summarises the features of the most common benign and malignant lesions of aging skin.

Discrimination of cancerous from benign pigmented skin lesions based on multispectral autofluorescence lifetime imaging dermoscopy and machine learning.

.SignificanceAccurate early diagnosis of malignant skin lesions is critical in providing adequate and timely treatment; unfortunately, initial clinical evaluation of similar-looking benign and malignant skin lesions can result in missed diagnosis of malignant lesions and unnecessary biopsy of benign ones.AimTo develop and validate a label-free and objective image-guided strategy for the clinical evaluation of suspicious pigmented skin lesions based on multispectral autofluorescence lifetime imaging (maFLIM) dermoscopy.ApproachWe tested the hypothesis that maFLIM-derived autofluorescence global features can be used in machine-learning (ML) models to discriminate malignant from benign pigmented skin lesions. Clinical widefield maFLIM dermoscopy imaging of 41 benign and 19 malignant pigmented skin lesions from 30 patients were acquired prior to tissue biopsy sampling. Three different pools of global image-level maFLIM features were extracted: multispectral intensity, time-domain biexponential, and frequency-domain phasor features. The classification potential of each feature pool to discriminate benign versus malignant pigmented skin lesions was evaluated by training quadratic discriminant analysis (QDA) classification models and applying a leave-one-patient-out cross-validation strategy.ResultsClassification performance estimates obtained after unbiased feature selection were as follows: 68% sensitivity and 80% specificity with the phasor feature pool, 84% sensitivity, and 71% specificity with the biexponential feature pool, and 84% sensitivity and 32% specificity with the intensity feature pool. Ensemble combinations of QDA models trained with phasor and biexponential features yielded sensitivity of 84% and specificity of 90%, outperforming all other models considered.ConclusionsSimple classification ML models based on time-resolved (biexponential and phasor) autofluorescence global features extracted from maFLIM dermoscopy images have the potential to provide objective discrimination of malignant from benign pigmented lesions. ML-assisted maFLIM dermoscopy could potentially assist with the clinical evaluation of suspicious lesions and the identification of those patients benefiting the most from biopsy examination.

Skin Lesion Classification Using Collective Intelligence of Multiple Neural Networks.

Skin lesion detection and analysis are very important because skin cancer must be found in its early stages and treated immediately. Once installed in the body, skin cancer can easily spread to other body parts. Early detection would represent a very important aspect since, by ensuring correct treatment, it could be curable. Thus, by taking all these issues into consideration, there is a need for highly accurate computer-aided systems to assist medical staff in the early detection of malignant skin lesions. In this paper, we propose a skin lesion classification system based on deep learning techniques and collective intelligence, which involves multiple convolutional neural networks, trained on the HAM10000 dataset, which is able to predict seven skin lesions including melanoma. The convolutional neural networks experimentally chosen, considering their performances, to implement the collective intelligence-based system for this purpose are: AlexNet, GoogLeNet, GoogLeNet-Places365, MobileNet-V2, Xception, ResNet-50, ResNet-101, InceptionResNet-V2 and DenseNet201. We then analyzed the performances of each of the above-mentioned convolutional neural networks to obtain a weight matrix whose elements are weights associated with neural networks and classes of lesions. Based on this matrix, a new decision matrix was used to build the multi-network ensemble system (Collective Intelligence-based System), combining each of individual neural network decision into a decision fusion module (Collective Decision Block). This module would then have the responsibility to take a final and more accurate decision related to the prediction based on the associated weights of each network output. The validation accuracy of the proposed system is about 3 percent better than that of the best performing individual network.

Multicenter prospective clinical trial of molecular genetic markers for non-invasive differential diagnosis of benign and malignant melanocytic skin lesions.

e21581 Background: Improving the accuracy of non-invasive diagnosis of skin tumors is becoming an urgent problem, given the clear increase in the incidence of melanoma in many countries. Dermoscopy and examination by an experienced dermatologist can reduce the NNE to 5–10 per melanoma (M) detected, however, achieving even this, clearly suboptimal, rate requires a long specialist training. Methods: We report results of prospective nonrandomized trial evaluating sensitivity and specificity of RNA profiling of cytological samples from adhesive transparent patches (ATP) applied to skin lesions. Routine pathology reports were obtained for all surgery samples. RNA extracted from ATP specimens (N = 126) was tested for mRNA expression of PRAME and LINC00518 genes. GAPDH was used as a reference gene. Sample was considered marker-positive if either PRAME or LINC00518 expression was detected. Results: Between June 2021 and November 2021, 126 pts undegoing excisional biopsy of skin lesions were included in the study at 5 centers. On pathology, invasive M was detected in 49 (38.9%) patients (mean Breslow thickness was 2.41 mm [95%CI 1,44 to 3.37]), M in situ in 6 (4.8%), non-melanoma skin cancer in 4 (3.2%) and dysplastic nevus in 56 (44.4%). In the day of preplanned excision ATP was applied to the target lesion for 5–6 min and sent to the central lab. Amplifiable mRNA was found in all 126 ATP samples. 82 samples of 126 (65,1%) tested marker-positive. Sensitivity of RNA profiling for detection of skin cancers reached 88.14% (95% CI 77.07% – 95.09%), while specificity was only 53.85% (41.03% – 66.30%), with PPV of 63.41% (56.74% – 69.61%), NPV of 83.33% (70.65% – 91.22%) and accuracy of 70.16% (61.29% – 78.04%). Among the seven false negative results there were 3 M in situ and 4 invasive M. Conclusions: RNA profiling of skin ATP specimens is feasible and has high sensitivity but rather low specificity. Randomized studies are needed to evaluate if this technique can add anything to dermoscopy or other noninvasive diagnostic modalities. Clinical trial information: NCT04353050. [Table: see text]

HOW WE DO IT: the Keystone flap for large skin defects: our experience with 30 consecutive cases

Background The keystone perforator-based fasciocutaneous island flap is a method of locoregional skin defect reconstruction, consisting of two conjoined V to Y advancement flaps. In this article, we report a series of 30 consecutive cases in various body parts with the aim to demonstrate the versatility of the keystone flap in the reconstruction of various skin defects after removal of a suspected malignant skin lesion and evaluate our clinical experience while maintaining oncological safety and good aesthetic results. Methods Patients who underwent keystone flap reconstruction between January 2016 and January 2021 were retrospectively reviewed. Their demographics, operative details, oncological data and postoperative details were obtained from the institution’s internal electronic patient records system. Results The majority of the skin defects were the result of an oncologic resection (97%, n = 29) and were located in the pretibial region (47%, n = 14). In five cases (17%) there was a need for a double keystone flap to cover the skin defect. One or more risk factors (hypertension, smoking, anticoagulation, …) were identified in 22 patients (73%). Minor complications (partial wound dehiscence, effusion or surgical site infection) occurred in ten cases (33%) which were resolved with conservative wound management or systemic antibiotics. There were no major postoperative flap-related complications such as flap failure or need for re-surgery. Conclusion In our experience, the keystone island flap is an effective and reliable technique to attain a single stage tension-free closure of large skin defects. Due to its low technical complexity and high success rate, we consider it a feasible alternative to other reconstructive flaps and grafts for direct closure of secondary defects.

Ultrasonography in dermatologic surgery: revealing the unseen for improved surgical planning

Ultrasonography (US) is a modern, in vivo imaging method, which is increasingly being used in dermatology as a complementary tool to clinical examination and dermoscopy. At higher frequencies (15MHz and above), US is an established method for assessing benign and malignant skin lesions, locoregional staging, monitoring the therapeutic efficacy in various inflammatory skin conditions, and patient follow-up. One field, which may increasingly benefit from performant imaging techniques such as US is dermatologic surgery. Preoperative imaging of cutaneous tumors, inflammatory skin conditions (hidradenitis suppurativa, abscesses, etc.), or nail pathology provide dermatologic surgeons with relevant information for an optimal surgical planning, identifying potential complex aspects which might require interdisciplinary approaches, herein sparing unnecessary surgical interventions and increasing patients' compliance. In this review, we discuss the increasing significance of US in the field of dermatologic surgery, as well as the spectrum of cutaneous pathology where sonography can aid in the preoperative setting to provide a more precise, individualized surgical planning for better counseling to our patients and improved surgical results.

Risks and benefits of dermatological machine learning health care applications-an overview and ethical analysis.

Visual data are particularly amenable for machine learning techniques. With clinical photography established for skin surveillance and documentation purposes as well as progress checks, dermatology is an ideal field for the development and application of emerging machine learning health care applications (ML-HCAs). To date, several ML-HCAs have detected malignant skin lesions on par with experts or found overlooked visual patterns that correlate with certain dermatological diseases. However, it is well established that ML-HCAs come with ethical and social implications. Currently, there is a lack of research that establishes model design, training, usage and regulation of such technologies sufficient to ensure ethically and socially responsible development and clinical translation, specifically within the field of dermatology. With this paper, we aim to give an overview of currently discussed ethical issues relating to dermatological ML-HCAs. On the basis of a thematic, keyword-based literature search, we performed an ethical analysis against established frameworks of biomedical ethics. We combined our results with current, relevant normative machine learning ethics literature to identify the status quo of the ethics of ML-HCAs in dermatology. We describe the benefits and risks of dermatological ML-HCAs that are currently being developed for clinical purposes. The potential benefits range from better patient outcomes to better knowledge accessibility to decreasing health care disparities, that is, standards of care between different population groups. The risks associated with ML-HCAs range from confidentiality issues to individual patient outcomes as well as the exacerbation of prevalent health care disparities. We discuss the practical implications for all stages of dermatological ML-HCA development. We found that ML-HCAs present stakeholder-specific risks for patients, health care professionals and society, which need to be considered separately. The discipline lacks sufficient biomedical ethics research that could standardize the approach to ML-HCA model design, training, use and regulation of such technologies.

Diagnostic Accuracy and Cost Savings Associated with Dermoscopy: An Economic Study.

Cutaneous malignancies are the most common overall cancer worldwide. Dermoscopy is widely used among various specialties to evaluate skin lesions and to differentiate benign from malignant lesions. Our objective was to evaluate the diagnostic capability and the economic impact of dermoscopy. The accuracy of diagnoses and economic costs of treating skin lesions were compared between two periods: 2001 to 2007 (prior the use of dermoscopy) and 2009 to 2011 (following implementation of dermoscopy). During the earlier period, 6,549 skin lesions were excised, of which 1,042 (15.9%) were malignant. During the later period, 2,578 lesions were excised, of which 610 (23.7%) were malignant. The potential savings estimated for the Israeli health care system are 6,500,000 USD. The use of dermoscopy increases sensitivity in diagnosing malignant skin lesions, enables diagnoses at an earlier stage, and has the potential for major cost savings for the health care system.

Xeroderma pigmentosum: an updated review.

BackgroundEarly recognition of xeroderma pigmentosum is important to minimize the complications arising from the harmful effects of exposure to ultraviolet radiation. This narrative review aims to familiarize physicians with the clinical features, diagnosis and management of xeroderma pigmentosum.MethodsA search was conducted in December 2021 in PubMed Clinical Queries using the key term “xeroderma pigmentosum”. The search strategy included all clinical trials, observational studies and reviews published within the past 10 years. The information retrieved from the search was used in the compilation of this article.ResultsXeroderma pigmentosum is a condition of abnormal DNA repair of ultraviolet radiation-induced and oxidative DNA damage, which leads to increased skin cancer susceptibility. Approximately 50% of patients with xeroderma pigmentosum have increased photosensitivity and certain types of xeroderma pigmentosum are more prone to ocular disease and progressive neurodegeneration depending on the causative mutation. The diagnosis should be suspected in patients with increased photosensitivity and characteristic cutaneous, ophthalmological and neurological findings. A definite diagnosis can be made by the identification of biallelic mutation in one of the causative genes. Strict and consistent sun avoidance and protection and early detection and treatment of premalignant and malignant skin lesions are the mainstays of management. Treatment options for actinic keratosis include cryotherapy, topical imiquimod, topical 5-fluorouracil, chemical peeling, excision, CO2 laser resurfacing, fractional/pulsed laser therapy, and photodynamic therapy. Cutaneous malignancy can be treated by photodynamic therapy, curettage and electrodesiccation, or surgical excision. Oral isotretinoin, oral niacinamide, topical imiquimod and topical fluorouracil can be used for the prevention of skin malignancy. Treatment options for poikiloderma include chemical peeling, dermabrasion and laser resurfacing. Methylcellulose eyedrops and soft ultraviolet-protective contact lenses may be used to keep the cornea moist and protect against the harmful effects of keratitis sicca. Investigational therapies include the use of T4 endonuclease-V liposome lotion and oral nicotinamide to reduce the rate of actinic keratoses and non-melanoma skin cancers and gene therapy for radical cure of this condition.ConclusionAlthough currently there is no cure for xeroderma pigmentosum, increased awareness and early diagnosis of the condition, followed by rigorous sun avoidance and protection and optimal management, can dramatically improve the quality of life and life expectancy.

Xeroderma Pigmentosum-Associated Childhood Interstitial Lung Disease.

Chromosomal breakage syndromes are a group of genetic disorders that are ascribable to the autosomal recessive mode of inheritance. Xeroderma pigmentosum is one of the chromosomal breakage syndromes which is due to the involvement of deformity in the deoxyribonucleic acid’s nucleotide excision repair. Xeroderma pigmentosum is a genetic disorder, which is autosomal recessive, heterogeneous, and more common in cases of consanguinity, caused by mutations in at least 10 genes and 9 complementation groups. The disorder is very rare. Patients experience photophobia and extreme photosensitivity and have pigmentary changes in ultraviolet light-exposed areas of the body with freckling, premalignant, and malignant skin lesions arising in keratinocytes soon after the fleeting exposure to sunlight. Patients are also oversensitive to environmental mutagens such as cigarette smoke and possibly to the widely used agricultural insecticide, diazinon. Progressive neurological abnormalities along with some rare complications are also noticed among these patients. Symptoms and thoracic high-resolution computed tomography are considered for diagnosis. Only corticosteroids can be given to limit the progression of the disease. Xeroderma pigmentosum-related interstitial lung disease is one of the rarest forms and we thereby report an interesting case.

Real-time high-resolution millimeter-wave imaging for in-vivo skin cancer diagnosis

High-resolution millimeter-wave imaging (HR-MMWI), with its high discrimination contrast and sufficient penetration depth, can potentially provide affordable tissue diagnostic information noninvasively. In this study, we evaluate the application of a real-time system of HR-MMWI for in-vivo skin cancer diagnosis. 136 benign and malignant skin lesions from 71 patients, including melanoma, basal cell carcinoma, squamous cell carcinoma, actinic keratosis, melanocytic nevi, angiokeratoma, dermatofibroma, solar lentigo, and seborrheic keratosis were measured. Lesions were classified using a 3-D principal component analysis followed by five classifiers including linear discriminant analysis (LDA), K-nearest neighbor (KNN) with different K-values, linear and Gaussian support vector machine (LSVM and GSVM) with different margin factors, and multilayer perception (MLP). Our results suggested that the best classification was achieved by using five PCA components followed by MLP with 97% sensitivity and 98% specificity. Our findings establish that real-time millimeter-wave imaging can be used to distinguish malignant tissues from benign skin lesions with high diagnostic accuracy comparable with clinical examination and other methods.

Heterogeneous Skin Phantoms for Experimental Validation of Microwave-Based Diagnostic Tools.

Considerable exploration has been done in recent years to exploit the reported inherent dielectric contrast between healthy and malignant tissues for a range of medical applications. In particular, microwave technologies have been investigated towards new diagnostic medical tools. To assess the performance and detection capabilities of such systems, tissue-mimicking phantoms are designed for controlled laboratory experiments. We here report phantoms developed to dielectrically represent malign skin lesions such as liposarcoma and nonsyndromic multiple basal cell carcinoma. Further, in order to provide a range of anatomically realistic scenarios, and provide meaningful comparison between different phantoms, cancer-mimicking lesions are inserted into two different types of skin phantoms with varying tumor–skin geometries. These configurations were measured with a microwave dielectric probe (0.5–26.5 GHz), yielding insight into factors that could affect the performance of diagnostic and detection tools.

Skin lesion classification system using a K-nearest neighbor algorithm

One of the most critical steps in medical health is the proper diagnosis of the disease. Dermatology is one of the most volatile and challenging fields in terms of diagnosis. Dermatologists often require further testing, review of the patient’s history, and other data to ensure a proper diagnosis. Therefore, finding a method that can guarantee a proper trusted diagnosis quickly is essential. Several approaches have been developed over the years to facilitate the diagnosis based on machine learning. However, the developed systems lack certain properties, such as high accuracy. This study proposes a system developed in MATLAB that can identify skin lesions and classify them as normal or benign. The classification process is effectuated by implementing the K-nearest neighbor (KNN) approach to differentiate between normal skin and malignant skin lesions that imply pathology. KNN is used because it is time efficient and promises highly accurate results. The accuracy of the system reached 98% in classifying skin lesions.

Validation of a Market-Approved Artificial Intelligence Mobile Health App for Skin Cancer Screening: A Prospective Multicenter Diagnostic Accuracy Study

Background: Mobile health (mHealth) consumer applications (apps) have been integrated with deep learning for skin cancer risk assessments. However, prospective validation of these apps is lacking. Objectives: To identify the diagnostic accuracy of an app integrated with a convolutional neural network for the detection of premalignant and malignant skin lesions. Methods: We performed a prospective multicenter diagnostic accuracy study of a CE-marked mHealth app from January 1 until August 31, 2020, among adult patients with at least one suspicious skin lesion. Skin lesions were assessed by the app on an iOS or Android device after clinical diagnosis and before obtaining histopathology. The app outcome was compared to the histopathological diagnosis, or if not available, the clinical diagnosis by a dermatologist. The primary outcome was the sensitivity and specificity of the app to detect premalignant and malignant skin lesions. Subgroup analyses were conducted for different smartphone types, the lesion’s origin, indication for dermatological consultation, and lesion location. Results: In total, 785 lesions, including 418 suspicious and 367 benign control lesions, among 372 patients (50.8% women) with a median age of 71 years were included. The app performed at an overall 86.9% (95% CI 82.3–90.7) sensitivity and 70.4% (95% CI 66.2–74.3) specificity. The sensitivity was significantly higher on the iOS device compared to the Android device (91.0 vs. 83.0%; p = 0.02). Specificity calculated on benign control lesions was significantly higher than suspicious skin lesions (80.1 vs. 45.5%; p < 0.001). Sensitivity was higher in skin fold areas compared to smooth skin areas (92.9 vs. 84.2%; p = 0.01), while the specificity was higher for lesions in smooth skin areas (72.0 vs. 56.6%; p = 0.02). Conclusion: The diagnostic accuracy of the mHealth app is far from perfect, but is potentially promising to empower patients to self-assess skin lesions before consulting a health care professional. An additional prospective validation study, particularly for suspicious pigmented skin lesions, is warranted. Furthermore, studies investigating mHealth implementation in the lay population are needed to demonstrate the impact on health care systems.

Stacked Cross Validation with Deep Features

Detection of malignant skin lesions is important for early and accurate diagnosis of skin cancer. In this work, a hybrid method for malignant lesion detection from dermoscopy images is proposed. The method combines the feature extraction process of convolutional neural networks (CNN) with an ensemble learner called stacked cross-validation (CV). The features extracted by three different CNN architectures, namely, ResNet50, Xception, and VGG16 are used for training of four different baseline classifiers, which are support vector machines, k-nearest neighbors, artificial neural networks, and random forests. The stacked outputs of these classifiers are used to train a logistic regression model as a meta-classifier. The performance of the proposed method is compared with the baseline classifiers trained individually as well as AdaBoost classifier, another ensemble learner. Feature extraction with Xception architecture, outperforms all other benchmark models by achieving scores of 0.909, 0.896, 0.886, and 0.917 for accuracy, F1-score, sensitivity, and AUC, respectively.