scholarly journals The Development of a Skin Cancer Classification System for Pigmented Skin Lesions Using Deep Learning

Biomolecules ◽  
2020 ◽  
Vol 10 (8) ◽  
pp. 1123 ◽  
Author(s):  
Shunichi Jinnai ◽  
Naoya Yamazaki ◽  
Yuichiro Hirano ◽  
Yohei Sugawara ◽  
Yuichiro Ohe ◽  
...  
Keyword(s):  
Skin Cancer ◽  
Malignant Tumors ◽  
Skin Lesions ◽  
Training Dataset ◽  
Clinical Image ◽  
Benign Tumors ◽  
Clinical Images ◽  
Predictive Values ◽  
Test Dataset ◽  
Pigmented Skin Lesions

Recent studies have demonstrated the usefulness of convolutional neural networks (CNNs) to classify images of melanoma, with accuracies comparable to those achieved by dermatologists. However, the performance of a CNN trained with only clinical images of a pigmented skin lesion in a clinical image classification task, in competition with dermatologists, has not been reported to date. In this study, we extracted 5846 clinical images of pigmented skin lesions from 3551 patients. Pigmented skin lesions included malignant tumors (malignant melanoma and basal cell carcinoma) and benign tumors (nevus, seborrhoeic keratosis, senile lentigo, and hematoma/hemangioma). We created the test dataset by randomly selecting 666 patients out of them and picking one image per patient, and created the training dataset by giving bounding-box annotations to the rest of the images (4732 images, 2885 patients). Subsequently, we trained a faster, region-based CNN (FRCNN) with the training dataset and checked the performance of the model on the test dataset. In addition, ten board-certified dermatologists (BCDs) and ten dermatologic trainees (TRNs) took the same tests, and we compared their diagnostic accuracy with FRCNN. For six-class classification, the accuracy of FRCNN was 86.2%, and that of the BCDs and TRNs was 79.5% (p = 0.0081) and 75.1% (p < 0.00001), respectively. For two-class classification (benign or malignant), the accuracy, sensitivity, and specificity were 91.5%, 83.3%, and 94.5% by FRCNN; 86.6%, 86.3%, and 86.6% by BCD; and 85.3%, 83.5%, and 85.9% by TRN, respectively. False positive rates and positive predictive values were 5.5% and 84.7% by FRCNN, 13.4% and 70.5% by BCD, and 14.1% and 68.5% by TRN, respectively. We compared the classification performance of FRCNN with 20 dermatologists. As a result, the classification accuracy of FRCNN was better than that of the dermatologists. In the future, we plan to implement this system in society and have it used by the general public, in order to improve the prognosis of skin cancer.

scholarly journals SKIN LESION CLASSIFICATION FROM DERMOSCOPY AND CLINICAL IMAGES WITH A DEEP LEARNING APPROACH

2021 ◽  
Vol 9 (10) ◽  
pp. 1294-1300
Author(s):  
Aigli Korfiati ◽  
◽  
Giorgos Livanos ◽  
Christos Konstandinou ◽  
Sophia Georgiou ◽  
...  
Keyword(s):  
Deep Learning ◽  
Skin Cancer ◽  
Medical Image ◽  
Skin Lesions ◽  
Learning Approaches ◽  
Early Screening ◽  
Clinical Images ◽  
Cad Systems ◽  
Learning Classifier ◽  
Aided Diagnosis

Computer-aided diagnosis (CAD) systems based on deep learning approaches are now feasible due to the availability of big data and the availability of powerful computational resources.The medical image-based CAD systems are of great interest in numerous diseases, but especially for skin cancer diagnosis, deep learning models have been mostly developed for dermoscopy images. Models for clinical images are few, mainly due to the unavailability of big volumes of relevant data. However, CAD systems able to classify skin lesions from clinical images would be of great valueboth for the population and clinicians as an initial early screening of lesions that would leadpatients to visiting a dermatologist in case of suspicious lesions. This is even more pronounced in areas where there is lack of dermoscopy instruments. Thus, in this paper, we aimed to build a classifier based on bothdermoscopy and clinical images able to discriminate skin cancer from skin lesions. The classification is made among three benign and two malignant categories, which include Nevus, Benign but not nevus, Benign but suspicious for malignancy, Melanoma and Non-Melanocytic Carcinoma.The proposed deep learning classifier achieves an Area Under Curve ranging between 0.75 and 0.9 for the five examined categories.

HISTOPATHOLOGICAL SPECTRUM OF SKIN LESIONS: A TWO YEAR RETROSPECTIVE STUDY.

2021 ◽  
pp. 26-29
Author(s):  
Shruti Shemawat ◽  
Sakshi Apurva ◽  
D.P Soni ◽  
Saurabh Soni
Keyword(s):  
Squamous Cell Carcinoma ◽  
Cell Carcinoma ◽  
Relative Frequency ◽  
Malignant Tumors ◽  
Skin Lesions ◽  
Benign Tumors ◽  
Histopathological Diagnosis ◽  
Neoplastic Lesions ◽  
Skin Biopsies ◽  
Age And Sex

INTRODUCTION: The skin being largest organ of the body has vast spectrum of disorders which can be difcult to diagnose correctly solely on the basis of clinical features. Hence histopathological examination is necessary to categorise skin lesions. The aim was to study relative frequency of various skin lesions and distribution of these lesions according to age and sex. METHODS: This is a retrospective descriptive hospital based study. The skin biopsies samples which came in the duration of two years from January 2019 to December 2020 at the Department of Pathology, Sardar Patel Medical college and associated group of hospitals, Bikaner, Rajasthan were taken in this study. All skin biopsies that showed denite histopathological diagnosis were included. After proper xing and staining procedures these lesions were examined under light microscopy and categorized as non-neoplastic and neoplastic. Relative frequency of various lesions, distribution of lesions according to age and sex was analyzed. The data collected was tabulated, interpreted and compared with other similar studies. RESULTS: Out of 346 patients, incidence of neoplastic lesions 259 (74.9%) were higher than non-neoplastic lesions 87(25.1%). Males were affected more compared to females with male to female ratio 1.45:1. Non-neoplastic lesions were mostly caused because of infectious etiologies among which leprosy was the most common infection. Keratinocytic tumors 99(52.2%) constituted most common type of neoplastic lesion. Benign tumors 191(73.7%) outnumbered malignant tumors 68(26.3%). The cases of benign tumors were seen more in younger population while that of malignant tumors were seen in older age groups. Among the keratinocytic type of malignant skin tumors squamous cell carcinoma (63.5%) was the most common variant which was followed by basal cell carcinoma 19(36.5%). Male predominance was observed in both squamous cell carcinoma and basal cell carcinoma. CONCLUSION: A wide heterogenesity of skin lesions was observed in the present study . These skin lesions were mostly affecting age group of 10-30 years. Inspite of extensive programmes and research, leprosy and tuberculosis remains a rampant cause of infectious non-neoplastic skin lesions. Sometimes ignorance by patient for a very small appearing skin lesions becomes life threatening. Hence early clinician consultation with proper examination and accurate histopathological diagnosis becomes the mainstay in early treatment and recovery.

scholarly journals Diagnostic accuracy of computed tomography to identify adenomas among adrenal incidentalomas in an endocrinological population

2018 ◽  
Vol 178 (5) ◽  
pp. 439-446 ◽  
Author(s):  
M Marty ◽  
D Gaye ◽  
P Perez ◽  
C Auder ◽  
M L Nunes ◽  
...  
Keyword(s):  
Computed Tomography ◽  
Malignant Tumors ◽  
Correlation Coefficients ◽  
University Hospital ◽  
Benign Tumors ◽  
Adrenal Incidentalomas ◽  
Reference Standard ◽  
Quantitative Ct ◽  
Endocrine Society ◽  
Predictive Values

Context The recent recommendations of the European Endocrine Society states that the performance of computed tomography (CT) to characterize ‘true' adrenal incidentalomas (AIs) remains debatable. Objective To determine relevant thresholds for usual CT parameters for the diagnosis of benign tumors using robust reference standard among a large series of ‘true’ AIs recruited in an endocrinological setting. Design Retrospective study of 253 AIs in 233 consecutive patients explored in a single university hospital: 183 adenomas, 33 pheochromocytomas, 23 adrenocortical carcinomas, 5 other malignant tumors and 9 other benign tumors. Reference standard was histopathology in 118 AIs, biological diagnosis of pheochromocytoma in 2 AIs and size stability after at least 1 year of follow-up in 133 AIs. Methods Sensitivity, specificity and positive and negative predictive values were estimated for various thresholds of size, unenhanced attenuation (UA), relative and absolute wash-out (RPW, APW) of contrast media. 197 scans were reviewed independently in a blinded fashion by two expert radiologists to assess inter-observer reproducibility of measurements. Results Criteria associated with a 100% positive predictive value for the diagnosis of benign AI were: a combination of size and UA: 30 mm and 20 HU or 40 mm and 15 HU, respectively; RPW >53%; and APW >78%. Non-adenomatous AIs with rapid contrast wash-out were exclusively benign pseudocysts and pheochromocytomas, suggesting that classical thresholds of 60% and 40% for APW and RPW, respectively, can be safely used for patients with normal metanephrine values. Inter-observer reproducibility of all parameters was excellent (intra-class correlation coefficients: 0.96–0.99). Conclusions Our study, the largest conducted in AIs recruited in an endocrinological setting, suggests safe thresholds for quantitative CT parameters to avoid false diagnoses of benignity.

The Advantages of Dermoscopy in the Diagnosis of Acral Melanoma from Other Podiatric Lesions: A Literature Review

2016 ◽  
Vol 106 (sp1) ◽  
pp. 3-3
Author(s):  
Mansi Patel ◽  
HyunJi Boo ◽  
Suganthi Kandasamy ◽  
Dhagash Patel ◽  
Anthony Iorio
Keyword(s):  
English Language ◽  
Malignant Tumors ◽  
Correct Diagnosis ◽  
Skin Lesions ◽  
Exclusion Criteria ◽  
Dermoepidermal Junction ◽  
Diabetic Ulcer ◽  
Predictive Values ◽  
Acral Melanoma

INTRODUCTION AND OBJECTIVES: Melanoma is one of the most common primary malignant tumors arising in the lower extremity. It is crucial to diagnose melanoma as quickly and as efficiently as possible for a better prognosis. The use of dermoscopy is helpful in diagnosing such conditions. Dermoscopy is a non-invasive, in-vivo technique primarily used in the examination of pigmented skin lesions. This procedure allows the visualization of subsurface skin structures in the epidermis, dermoepidermal junction, and upper dermis - structures not visible to the naked eye. This poster presents the advantages of dermoscopy in the field of podiatry by assessing the dermoscopic criteria with positive predictive values for distinguishing acral melanoma from acral nevus. Additionally it analyzes cases of melanomas misdiagnosed as a plantar-pigmented wart and a diabetic ulcer. METHODS: The authors used PubMed to perform an English language literature search. The exclusion criteria included articles older than 10 years. Inclusion criteria consisted of research done on humans and the terms dermoscopy and foot lesions or melanoma. After retrieving a total of 140 articles, 14 articles were found to meet both the inclusion and exclusion criteria. RESULTS: Qualitative analysis of relevant articles demonstrates that the detection of malignant dermoscopic patterns enhances quick and correct diagnosis. CONCLUSIONS: The use of dermoscopy is slowly evolving in podiatry. It aims to minimize the amount of biopsies taken, thereby decreasing the risk of creating an ulcer and reducing the patients exposure to anesthesia. While controversy remains over sensitivity and specificity of using a dermatoscope alone to diagnose pedal lesions, particular attention should be paid to the accuracy of diagnosing a lesion when dermoscopy is used in conjunction with a biopsy. With the continued usage of a dermatoscope along with experience and expertise in the field, the need for a biopsy could eventually be eliminated.

scholarly journals Dermoscopy Improves Accuracy of Primary Care Physicians to Triage Lesions Suggestive of Skin Cancer

2006 ◽  
Vol 24 (12) ◽  
pp. 1877-1882 ◽  
Author(s):  
Giuseppe Argenziano ◽  
Susana Puig ◽  
Iris Zalaudek ◽  
Francesco Sera ◽  
Rosamaria Corona ◽  
...  
Keyword(s):  
Primary Care ◽  
Skin Cancer ◽  
Primary Care Physicians ◽  
Randomized Study ◽  
Skin Lesions ◽  
Skin Cancer Screening ◽  
Predictive Values ◽  
Naked Eye ◽  
Pigmented Lesions ◽  
Malignant Skin Tumors

Purpose Primary care physicians (PCPs) constitute an appropriate target for new interventions and educational campaigns designed to increase skin cancer screening and prevention. The aim of this randomized study was to determine whether the adjunct of dermoscopy to the standard clinical examination improves the accuracy of PCPs to triage lesions suggestive of skin cancer. Patients and Methods PCPs in Barcelona, Spain, and Naples, Italy, were given a 1-day training course in skin cancer detection and dermoscopic evaluation, and were randomly assigned to the dermoscopy evaluation arm or naked-eye evaluation arm. During a 16-month period, 73 physicians evaluated 2,522 patients with skin lesions who attended their clinics and scored individual lesions as benign or suggestive of skin cancer. All patients were re-evaluated by expert dermatologists at clinics for pigmented lesions. Referral accuracy of both PCP groups was calculated by their scores, which were compared to those tabulated for dermatologists. Results Referral sensitivity, specificity, and positive and negative predictive values were 54.1%, 71.3%, 11.3%, and 95.8%, respectively, in the naked-eye arm, and 79.2%, 71.8%, 16.1%, and 98.1%, respectively, in the dermoscopy arm. Significant differences were found in terms of sensitivity and negative predictive value (P = .002 and P = .004, respectively). Histopathologic examination of equivocal lesions revealed 23 malignant skin tumors missed by PCPs performing naked-eye observation and only six by PCPs using dermoscopy (P = .002). Conclusion The use of dermoscopy improves the ability of PCPs to triage lesions suggestive of skin cancer without increasing the number of unnecessary expert consultations.

scholarly journals Calprotectin and Calgranulin C as Biomarkers of Pancreatic Tumors: Baseline Levels and Level Changes after Surgery

2019 ◽  
Vol 2019 ◽  
pp. 1-11
Author(s):  
Michal Holub ◽  
Eva Bartáková ◽  
Alžběta Stráníková ◽  
Eva Koblihová ◽  
Simona Arientová ◽  
...  
Keyword(s):  
Postoperative Period ◽  
Malignant Tumors ◽  
Serum Levels ◽  
Pancreatic Tissue ◽  
Pancreatic Tumors ◽  
Major Surgery ◽  
Benign Tumors ◽  
Baseline Serum ◽  
Predictive Values ◽  
Potential Biomarkers

Pancreatic tumors and their surgical resection are associated with significant morbidity and mortality, and the biomarkers currently used for these conditions have limited sensitivity and specificity. Because calprotectin and calgranulin C serum levels have been demonstrated to be potential biomarkers of certain cancers and complications of major surgery, the levels of both proteins were tested in the current study in patients with benign and malignant pancreatic tumors that were surgically removed. The baseline serum levels and kinetics of calprotectin and calgranulin C during the 7-day postoperative period were evaluated with immunoassays in 98 adult patients who underwent pancreatic surgery. The baseline serum levels of calprotectin and calgranulin C in patients with malignant (n=84) and benign tumors (n=14) were significantly higher (p<0.01) when compared to those in the healthy controls (n=26). The serum levels of both proteins were also significantly (p<0.05) higher in patients with benign tumors than in those with malignant tumors. After surgery, the serum levels of calprotectin and calgranulin C were significantly (p<0.01) higher than their baseline values, and this elevation persisted throughout the seven days of the follow-up period. Interestingly, starting on day 1 of the postoperative period, the serum levels of both proteins were significantly (p<0.05) higher in the 37 patients who developed postoperative pancreatic fistulas (POPFs) than in the patients who had uneventful recoveries (n=61). Moreover, the serum levels of calprotectin and calgranulin C demonstrated a significant predictive value for the development of POPF; the predictive values of these two proteins were better than those of the serum level of C-reactive protein and the white blood cell count. Taken together, the results of this study suggest that calprotectin and calgranulin C serum levels are potential biomarkers for pancreatic tumors, surgical injury to the pancreatic tissue and the development of POPFs.

Use of an app in skin cancer screening in pomeran immigrant communities in the state of Espírito Santo: An observational study (Preprint)

2021 ◽  
Author(s):  
Patrícia Henriques Lyra Frasson ◽  
Amanda Da Silva Salomão ◽  
Gustavo Ribeiro Lima ◽  
Renato A. Krohling ◽  
André G. C. Pacheco ◽  
...  
Keyword(s):  
Skin Cancer ◽  
Histopathological Examination ◽  
Demographic Data ◽  
Skin Lesions ◽  
Analysis Software ◽  
Assistance Program ◽  
Clinical Images ◽  
Adequate Treatment ◽  
Espírito Santo ◽  
Espirito Santo

BACKGROUND Cancer occupies the second leading cause of death in the world, behind only cardiovascular diseases. Among cancers, skin cancer is the most frequent in Brazil and worldwide. The University Extension Program entitled Dermatological Assistance Program to Pomeranian Farmers in Espírito Santo (PAD) of the Federal University of Espírito Santo, has been promoting prevention, diagnosis, and adequate treatment in the Pomeranian population of Espírito Santo since 1986, through joint efforts of medical care. The result of these joint efforts represents, on average, 300 doctor’s appointments, 500 to 900 cryotherapies, and 100 surgeries per county where the visits occur, promoted once a month. OBJECTIVE Currently, there is no set of public cancer-related data in the literature that provides information about clinical images and medical histories of affected patients. In partnership with the Engineering and Computer Science Sectors, the Dermatological Analysis Software (SADE) was created to store data and images of the skin lesions of patients operated by the program. The goal of this study was to evaluate the scenario of skin cancer in the communities served by the PAD, based on data stored at SADE between 2018 and 2019, totaling 2,935 visits during this period. METHODS This is a retrospective study carried out from the database collected via the SADE platform (Dermatological Analysis Software). The data is collected using the smartphone application, which connects to the local internet server to store the data. The application was developed using a specific type of Deep Learning model known as Convolutional Neural Networks (CNN). This model is trained using clinical images and patient demographic data collected using the software described above. RESULTS In view of the neoplastic lesions, 1,201 lesions were removed, which after histopathological examination showed 593 basal cell carcinomas (BCC), 95 squamous cell carcinomas (SCC), 81 melanomas and 48 associated BCC and SCC. CONCLUSIONS The results highlight the potential of the software to compose an application in the future that will help doctors in remote locations who have no training in dermatology, optimizing the referral of patients with suspected skin lesions to the specialist, aiming at a faster treatment and suggesting differential diagnoses that may not have been considered. The application of technological instruments in the identification of cancer is an increasingly current reality and its use is already widely used.

scholarly journals Categorization of Common Pigmented Skin Lesions (CPSL) using Multi-Deep Features and Support Vector Machine

2021 ◽  
Author(s):  
SANTI BEHERA ◽  
PRABIRA SETHY
Keyword(s):  
Support Vector Machine ◽  
Skin Cancer ◽  
Principal Component ◽  
Skin Lesions ◽  
Classification Model ◽  
Healthcare Sector ◽  
Support Vector ◽  
Svm Classifier ◽  
Deep Feature ◽  
Pigmented Skin Lesions

Abstract The skin is the main organ. It is approximately 8 pounds for the average adult. Our skin is a truly wonderful organ. It isolates us and shields our bodies from hazards. However, the skin is also vulnerable to damage and distracted from its original appearance; brown, black, or blue, or combinations of those colors, known as pigmented skin lesions. These common pigmented skin lesions (CPSL) are the leading factor of skin cancer, or can say these are the primary causes of skin cancer. In the healthcare sector, the categorization of CPSL is the main problem because of inaccurate outputs, overfitting, and higher computational costs. Hence, we proposed a classification model based on multi-deep feature and support vector machine (SVM) for the classification of CPSL. The proposed system comprises two phases: first, evaluate the 11 CNN model's performance in the deep feature extraction approach with SVM. Then, concatenate the top performed three CNN model's deep features and with the help of SVM to categorize the CPSL. In the second step, 8192 and 12288 features are obtained by combining binary and triple networks of 4096 features from the top performed CNN model. These features are also given to the SVM classifiers. The SVM results are also evaluated with principal component analysis (PCA) algorithm to the combined feature of 8192 and 12288. The highest results are obtained with 12288 features. The experimentation results, the combination of the deep feature of Alexnet, VGG16 & VGG19, achieved the highest accuracy of 91.7% using SVM classifier. As a result, the results show that the proposed methods are a useful tool for CPSL classification.

scholarly journals Clinico-histopathological correlation of pigmented skin lesions: A hospital based study at BPKIHS

2019 ◽  
Vol 9 (2) ◽  
pp. 1550-1554
Author(s):  
Isha Bohra ◽  
Punam Paudyal ◽  
Anju Pradhan ◽  
Dhan Kesar Khadka
Keyword(s):  
Malignant Tumors ◽  
Skin Lesions ◽  
Pathological Examination ◽  
Cross Sectional ◽  
Histopathological Correlation ◽  
Melanocytic Lesions ◽  
Benign Pathology ◽  
Seborrhoeic Keratosis ◽  
Pigmented Skin Lesions ◽  
Age Range

Background: Pigmented skin lesions refers to melanocytic as well as nonmelanocytic lesions. Pigmentation is not just a cosmetic deformity but can also reflect underlying benign pathology as nevi or malignant lesions as melanoma. With this study we intend to evaluate the spectrum of pigmented skin lesions and to correlate the clinical diagnosis with the histological diagnosis. Materials and Methods: This is a hospital based cross sectional descriptive study where clinicohistopathological evaluation of 46 cases of pigmented skin lesions were analyzed on paraffin embedded tissue sections for a duration of 1 year at the Department of Pathology, B. P. Koirala Institute of Health Sciences. Results: Out of the 46 cases evaluated there were 32 cases of melanocytic lesions comprising of benign melanocytic nevi (n=27), malignant melanoma (n=5) and 14 cases of nonmelanocytic lesions including basal cell carcinoma and seborrhoeic keratosis (5 cases each). Angiokeratoma (n=1), sebaceous hyperplasia (n=1), trichoepitheloma (n=1) and venous haemangioma (n=1). The age range was from 8-83 years with slight female predominance (52.2%) and the most common site involved was head and neck (58.7%). 76.1% of the patients belonged to the Terai region. Clinicohistopathological correlation showed positive correlation in 26 cases (56.5%) and negative correlation in 20 cases (43.5%). Conclusions: Pigmented skin lesions are common presenting problem, while majority are benign a small minority can be malignant. So, clinically pigmented skin lesions should be submitted for pathological examination in order not to miss a small percentage of malignant tumors and to differentiate melanocytic lesions from its nonmelanocytic mimickers.

Sign in / Sign up

Export Citation Format

Share Document