Deep learning technologies for skin cancer detection have been dramatically advanced based on high resolution dermoscopic images. The low-cost approach based on clinical skin images of low resolution is promising but remains technically challenging due to their undermined image quality. In this paper, we propose a coarse-to-fine efficient super resolution transformer (CF-ESRT) network to reconstruct the dermoscopy-level high resolution skin image from a low resolution clinical image. By connecting the refinement network to the original super resolution transformers and applying perceptual and gradient losses, our framework noticeably improves the finer texture details of skin lesions in the super resolution (SR) images, and is effective to elevate the perceptual quality of the SR images. Quantitative and qualitative evaluations show that our method outperforms ESRT the basis model as well as the other state-of-the-art SR models.

Breast cancer remains a significant health concern worldwide, necessitating early detection and treatment. The rapid appearance or a change in the shape of skin lesion on the breast or just of the nipple-areola complex(NAC), especially if it is unilateral needs to be evaluated. Several primary or secondary tumours such as inflammatory skin lesions can affect this body area. Ultrasound, mammography, or magnetic resonance imaging (MRI) are the most used non-invasive diagnostic techniques for breast cancer, but in recent years reflectance confocal microscopy (RCM) and optical coherence tomography have emerged as additional tools to diagnose cases characterized by cutaneous changes.The introduction of new techniques such as reflectance confocal microscopy (RCM), may open new perspectives in the non-invasive diagnosis of mammary paget disease (MPD). RCM is a non-invasive diagnostic technique that allows high-resolution images of the skin at microscopic level in real time. It is non-invasive, painless and repeatable, making it a valuable tool in the early diagnosis and monitoring of skin diseases. We describe an atypical clinical presentation of mammary paget disease early diagnosed by reflectance confocal microscopy evaluation.

Super-high (x400) magnification dermoscopy (D400) is a new non-invasive imaging technique that has been shown to add information for the differential diagnosis of melanocytic lesions in a pilot study. Our study aimed to confirm if D400 can add details for the discrimination of clinically atypical nevus and melanoma. This is a retrospective observational, multicentric study enrolling patients who received an x20 (D20) and x400 (D400) magnification dermoscopic examination of pigmented skin lesions. Dermoscopic images were retrospectively evaluated by three observers for the presence/absence of nine D20 and twenty D400 dermoscopic features defined in the previous pilot study. Univariate and multivariate analyses were carried out to find predictors of benign and malignant behaviors. At D20 atypical pigment network, blue-whitish veil, atypical vascular pattern, regression, and homogenous pattern were more frequent in melanoma than in nevi (p<0.001). At D400 melanoma showed more frequently than benign lesions melanocytes with an irregular arrangement and irregular in shape and size (p<0.001). Network with edged papillae was more frequent in benign lesions than melanomas (p<0.001). Our study confirms that D400 can identify melanocytes with an irregular arrangement and irregular in shape and size in melanomas, adding information to the conventional D20 examination.

Cutaneous metastases from clear cell renal carcinoma (ccRC) are uncommon and often indicate a poor prognosis. These metastases typically occur on the scalp, face, and trunk, and they can be difficult to diagnose due to their resemblance to benign dermatological tumors. We report the case of a 56-year-old patient with a history of ccRC (TNM stage 4) who was referred to our Dermatology department with two rapidly enlarging, painful lesions on the left jawline and scalp, which had developed one month and one week earlier, respectively. On examination, the lesions appeared as well-defined, round to oval plaques with a central ulceration and a peripheral red rim. Dermoscopic examination revealed a structureless pink to orange pattern, atypical central vessels, and irregular linear vessels in a corona-like arrangement. Despite the patient's stable oncological treatment for six months, pain management had recently included paracetamol, tramadol, and NSAIDs. Given the recent medication changes and the clinical presentation, differential diagnoses initially included drug-induced eruptions such as erythema multiforme or fixed drug eruption. However, a biopsy of the facial lesion revealed immunohistochemical positivity for CD10, CAIX, and PAX8, confirming the diagnosis of metastatic ccRC with sarcomatoid differentiation. Unfortunately, despite continued targeted therapies and palliative care, the patient’s condition deteriorated rapidly, leading to death two months later. This case highlights the aggressive progression and poor prognosis associated with cutaneous metastases from renal cell carcinoma, particularly once the skin is involved.

Skin cancer is one of the most prevalent cancers worldwide, with an increasing incidence. Skin cancer is typically classified as melanoma or non-melanoma skin cancer. Melanoma, despite being much less common than basal or squamous cell carcinomas, is the most deadly, with nearly 8,300 Americans expected to die from melanoma each year. Biopsies are currently the gold standard in diagnosing melanoma; however, they can be invasive and expensive and inaccessible to lower-income individuals. Currently, suspicious lesions are triaged with image-based technologies, such as dermatoscopy and confocal microscopy. While these techniques are useful, there is wide inter-user variability and minimal training for dermatology residents on how to properly use these devices. The use of artificial intelligence (AI)-based technologies in dermatology has emerged in recent years to assist in the diagnosis of melanoma that may be more accessible to all patients and more accurate than current methods of screening. This review explores the current status of the application of AI-based algorithms in the detection of melanoma, underscoring its potential to aid dermatologists in clinical practice. We will specifically focus on AI’s application in clinical imaging, dermoscopic evaluation, algorithms that can distinguish melanoma from non-melanoma skin cancers, and in-vivo skin imaging devices.