Lesion identification and malignancy prediction from clinical dermatological images.

We consider machine-learning-based lesion identification and malignancy prediction from clinical dermatological images, which can be indistinctly acquired via smartphone or dermoscopy capture. Additionally, we do not assume that images contain single lesions, thus the framework supports both focal or wide-field images. Specifically, we propose a two-stage approach in which we first identify all lesions present in the image regardless of sub-type or likelihood of malignancy, then it estimates their likelihood of malignancy, and through aggregation, it also generates an image-level likelihood of malignancy that can be used for high-level screening processes. Further, we consider augmenting the proposed approach with clinical covariates (from electronic health records) and publicly available data (the ISIC dataset). Comprehensive experiments validated on an independent test dataset demonstrate that (1) the proposed approach outperforms alternative model architectures; (2) the model based on images outperforms a pure clinical model by a large margin, and the combination of images and clinical data does not significantly improves over the image-only model; and (3) the proposed framework offers comparable performance in terms of malignancy classification relative to three board certified dermatologists with different levels of experience.

A shallow deep learning approach to classify skin cancer using down-scaling method to minimize time and space complexity.

The complex feature characteristics and low contrast of cancer lesions, a high degree of inter-class resemblance between malignant and benign lesions, and the presence of various artifacts including hairs make automated melanoma recognition in dermoscopy images quite challenging. To date, various computer-aided solutions have been proposed to identify and classify skin cancer. In this paper, a deep learning model with a shallow architecture is proposed to classify the lesions into benign and malignant. To achieve effective training while limiting overfitting problems due to limited training data, image preprocessing and data augmentation processes are introduced. After this, the 'box blur' down-scaling method is employed, which adds efficiency to our study by reducing the overall training time and space complexity significantly. Our proposed shallow convolutional neural network (SCNN_12) model is trained and evaluated on the Kaggle skin cancer data ISIC archive which was augmented to 16485 images by implementing different augmentation techniques. The model was able to achieve an accuracy of 98.87% with optimizer Adam and a learning rate of 0.001. In this regard, parameter and hyper-parameters of the model are determined by performing ablation studies. To assert no occurrence of overfitting, experiments are carried out exploring k-fold cross-validation and different dataset split ratios. Furthermore, to affirm the robustness the model is evaluated on noisy data to examine the performance when the image quality gets corrupted.This research corroborates that effective training for medical image analysis, addressing training time and space complexity, is possible even with a lightweighted network using a limited amount of training data.

Metastatic melanoma in the breast and axilla: A case report.

Metastatic melanoma of the breast is rare, and demonstrates nonspecific imaging findings which may overlap with both benign and malignant pathology.1-3 Immunohistochemical stains are important to confirm the diagnosis, particularly combining S100, a sensitive marker for melanoma, with more specific tumor markers such as Melan-A and HMB-45, and lack of cytokeratin staining.4-7 We present a case of a 64-year-old female who presented for diagnostic imaging of a palpable abnormality in her right breast, with medical history notable for previously excised cutaneous melanoma, recent COVID-19 vaccination, and significant family history of breast cancer. Diagnostic mammogram of the right breast demonstrated a circumscribed mass in the lower inner quadrant corresponding to the area of palpable concern, as well as an additional non-palpable circumscribed mass in the lower inner quadrant. Targeted right breast ultrasound demonstrated corresponding circumscribed cystic versus solid masses as well as a morphologically abnormal right axillary lymph node. Pathologic results after tissue sampling of the two right breast masses and right axillary lymph node all yielded metastatic melanoma.

PET/CT variants and pitfalls in malignant melanoma.

18F-FDG PET/CT plays an increasingly pivotal role in the staging and post-treatment monitoring of high-risk melanoma patients, augmented by the introduction of therapies, including tyrosine kinase inhibitors (TKI) and immune checkpoint inhibitors (ICIs), that have novel modes of action that challenge conventional response assessment. Simultaneously, technological advances have been regularly released, including advanced reconstruction algorithms, digital PET and motion correction, which have allowed the PET community to detect ever-smaller cancer lesions, improving diagnostic performance in the context of indications previously viewed as limitations, such as detection of in-transit disease and confirmation of the nature of small pulmonary metastases apparent on CT.This review will provide advice regarding melanoma-related PET protocols and will focus on variants encountered during the imaging of melanoma patients. Emphasis will be made on pitfalls related to non-malignant diseases and treatment-related findings that may confound accurate interpretation unless recognized. The latter include signs of immune activation and immune-related adverse events (irAEs). Technology-related pitfalls are also discussed, since while new PET technologies improve detection of small lesions, these may also induce false-positive cases and require a learning curve to be observed. In these times of the COVID 19 pandemic, cases illustrating lessons learned from COVID 19 or vaccination-related pitfalls will also be described.

Non-melanoma skin cancer diagnosis: a comparison between dermoscopic and smartphone images by unified visual and sonification deep learning algorithms.

PURPOSE: Non-melanoma skin cancer (NMSC) is the most frequent keratinocyte-origin skin tumor. It is confirmed that dermoscopy of NMSC confers a diagnostic advantage as compared to visual face-to-face assessment. COVID-19 restrictions diagnostics by telemedicine photos, which are analogous to visual inspection, displaced part of in-person visits. This study evaluated by a dual convolutional neural network (CNN) performance metrics in dermoscopic (DI) versus smartphone-captured images (SI) and tested if artificial intelligence narrows the proclaimed gap in diagnostic accuracy. METHODS: A CNN that receives a raw image and predicts malignancy, overlaid by a second independent CNN which processes a sonification (image-to-sound mapping) of the original image, were combined into a unified malignancy classifier. All images were histopathology-verified in a comparison between NMSC and benign skin lesions excised as suspected NMSCs. Study criteria outcomes were sensitivity and specificity for the unified output. RESULTS: Images acquired by DI (n = 132 NMSC, n = 33 benign) were compared to SI (n = 170 NMSC, n = 28 benign). DI and SI analysis metrics resulted in an area under the curve (AUC) of the receiver operator characteristic curve of 0.911 and 0.821, respectively. Accuracy was increased by DI (0.88; CI 81.9-92.4) as compared to SI (0.75; CI 68.1-80.6, p < 0.005). Sensitivity of DI was higher than SI (95.3%, CI 90.4-98.3 vs 75.3%, CI 68.1-81.6, p < 0.001), but not specificity (p = NS). CONCLUSION: Telemedicine use of smartphone images might result in a substantial decrease in diagnostic performance as compared to dermoscopy, which needs to be considered by both healthcare providers and patients.

Developing a Recognition System for Diagnosing Melanoma Skin Lesions Using Artificial Intelligence Algorithms.

In recent years, computerized biomedical imaging and analysis have become extremely promising, more interesting, and highly beneficial. They provide remarkable information in the diagnoses of skin lesions. There have been developments in modern diagnostic systems that can help detect melanoma in its early stages to save the lives of many people. There is also a significant growth in the design of computer-aided diagnosis (CAD) systems using advanced artificial intelligence. The purpose of the present research is to develop a system to diagnose skin cancer, one that will lead to a high level of detection of the skin cancer. The proposed system was developed using deep learning and traditional artificial intelligence machine learning algorithms. The dermoscopy images were collected from the PH2 and ISIC 2018 in order to examine the diagnose system. The developed system is divided into feature-based and deep leaning. The feature-based system was developed based on feature-extracting methods. In order to segment the lesion from dermoscopy images, the active contour method was proposed. These skin lesions were processed using hybrid feature extractions, namely, the Local Binary Pattern (LBP) and Gray Level Co-occurrence Matrix (GLCM) methods to extract the texture features. The obtained features were then processed using the artificial neural network (ANNs) algorithm. In the second system, the convolutional neural network (CNNs) algorithm was applied for the efficient classification of skin diseases; the CNNs were pretrained using large AlexNet and ResNet50 transfer learning models. The experimental results show that the proposed method outperformed the state-of-art methods for HP2 and ISIC 2018 datasets. Standard evaluation metrics like accuracy, specificity, sensitivity, precision, recall, and F-score were employed to evaluate the results of the two proposed systems. The ANN model achieved the highest accuracy for PH2 (97.50%) and ISIC 2018 (98.35%) compared with the CNN model. The evaluation and comparison, proposed systems for classification and detection of melanoma are presented.

New Apps Drive Health Care Innovation, Access to Care.

With the ubiquitous nature of smartphones, apps are a regular part of our day-to-day lives. They are also becoming a larger presence in health care, where they have the ability to expand access to care, help people monitor health changes, provide support for people living with chronic conditions, and coordinate communication between patients and their doctors. From detecting skin cancer to helping people with diabetes, new apps aim to change how people think about their health.

COVID-19 Vaccine-Induced Radiation Recall Phenomenon.

Radiation recall phenomenon (RRP) is an uncommon, late occurring, acute inflammatory skin reaction that emerges in localized areas coincident with previously irradiated radiation therapy (RT) treatment fields. RRP has been known to be triggered by a number of chemotherapy agents. To the best of our knowledge, this report is the first description of RRP after administration of the Pfizer-BioNTech vaccine for COVID-19, or any other currently available vaccine against COVID-19. Acute skin reactions were observed in 2 RT patients with differing timelines of RT and vaccinations. In both cases however, the RRP presented within days of the patient receiving the second dose of vaccine. For each RT course, the treatment planning dosimetry of the radiation fields was compared with the area of the observable RRP. RRP developed within the borders of treatment fields where prescription dose constraints were prioritized over skin sparing. Our observation is currently limited to 2 patients. The actual incidence of RRP in conjunction with Pfizer-BioNTech vaccine or any other vaccine against COVID-19 is unknown. For patients with cancer being treated with radiation with significant dose to skin, consideration should be given to the probability of RRP side effects from vaccinations against COVID-19.

Direct-to-consumer, store-and-forward teledermatology with dermoscopy using the pharmacist as patient point-of-contact.

OBJECTIVE(S): To evaluate the frequency of nonmelanoma skin cancer (NMSC), NMSC precursors, and melanoma on a store-and-forward dermatology model featuring the pharmacist as the patient's point-of-contact. The secondary objective was to define lesion changes and symptoms perceived by patients (clinical prediction rules by nonexpert observers) that can be predictive of malignity. METHODS: A cross-sectional study of teledermatology consultation was performed. All patients who underwent a teledermatology consultation between September 2018 and March 2020 were included. A patient could have more than 1 lesion per consultation. The object of the study was a defined dermatologic lesion. The differences between the variables were analyzed using a univariate model based on the chi-square test for independent qualitative variables and Fisher exact test in cases when the expected values in any of the cells of a contingency table were less than 5. Statistical significance was set at P < 0.05 (2-tailed). RESULTS: A total of 225 lesions in 218 patients were considered for this study; 53.8% (n = 121) of the lesions were classified as benign, 16.4% (n = 37) as dubious, 23.1% (n = 52) as NMSC precursors, 5.8% (n = 13) as NMSC, and 0.9% (n = 2) as melanomas. Of the reported clinical lesion changes, spontaneous pain, pruritus, surface texture changes, color changes, or form changes had no statistically significant relationship with the diagnostic group, whereas the presence of spontaneous bleeding (P = 0.015) and size changes (P = 0.026) were more frequently observed in the "dubious lesion" and "of oncological relevance lesion" groups. CONCLUSION: This "direct-to-consumer," store-and-forward teledermatology with dermoscopy model featuring the pharmacist as the patient's point-of-contact is useful for the diagnosis of melanoma, NMSC, and NMSC precursors when backed by a robust dermatology service.

Direct patient-to-physician teledermatology: Not a flash in the pan(demic).

Dermatology is a clinical and visual discipline, which makes it the quintessential medical specialty for spot diagnosis and telemedicine. The COVID-19 pandemic has led to an unprecedented worldwide renaissance of teledermatology (TD). It has helped deliver high-quality medical care, while protecting the medical personnel and vulnerable patients from potential infection. Examining a patient from a distance through digital photography has many drawbacks, including lack of physical touch, difficulties in performing full body examinations, and several legal and ethical issues. We summarize have summarized the more common pitfalls and highlight the key aspects of direct patient-to-physician TD. Basic practical advice includes the use of TD for obtaining patient history, examining patient-captured photographs for inflammatory skin disease, and skin cancer screening.