Propensity of skin diseases to manifest in a variety of forms, lack and maldistribution of qualified dermatologists, and exigency of timely and accurate diagnosis call for automated Computer-Aided Diagnosis (CAD). This study aims at extending previous works on CAD for dermatology by exploring the potential of Deep Learning to classify hundreds of skin diseases, improving classification performance, and utilizing disease taxonomy. We trained state-of-the-art Deep Neural Networks on two of the largest publicly available skin image datasets, namely DermNet and ISIC Archive, and also leveraged disease taxonomy, where available, to improve classification performance of these models. On DermNet we establish new state-of-the-art with 80% accuracy and 98% Area Under the Curve (AUC) for classification of 23 diseases. We also set precedence for classifying all 622 unique sub-classes in this dataset and achieved 67% accuracy and 98% AUC. On ISIC Archive we classified all 7 diseases with 93% average accuracy and 99% AUC. This study shows that Deep Learning has great potential to classify a vast array of skin diseases with near-human accuracy and far better reproducibility. It can have a promising role in practical real-time skin disease diagnosis by assisting physicians in large-scale screening using clinical or dermoscopic images.

Rosacea is one of the most common dermatoses of adults. Recent studies have improved our understanding of the pathophysiology of rosacea. Current concepts suggest that known clinical trigger factors of rosacea such as UV radiation, heat, cold, stress, spicy food, and microbes modulate Toll-like receptor signaling, induce reactive oxygen species, as well as enhance antimicrobial peptide and neuropeptide production. Downstream of these events cytokines and chemokines orchestrate an inflammatory response that leads to the recruitment and activation of distinct leukocyte subsets and induces the characteristic histopathological features of rosacea. Here we summarize the current knowledge of the cytokine and chemokine network in rosacea and propose pathways that may be of therapeutic interest.

The novel keratinocyte-specific chemokine CCL27 plays a critical role in the organization of skin-associated immune responses by regulating T cell homing under homeostatic and inflammatory conditions. Here we demonstrate that human keratinocyte-derived skin tumors may evade T cell-mediated antitumor immune responses by down-regulating the expression of CCL27 through the activation of epidermal growth factor receptor (EGFR)–Ras–MAPK-signaling pathways. Compared with healthy skin, CCL27 mRNA and protein expression was progressively lost in transformed keratinocytes of actinic keratoses and basal and squamous cell carcinomas.In vivo, precancerous skin lesions as well as cutaneous carcinomas showed significantly elevated levels of phosphorylated ERK compared with normal skin, suggesting the activation of EGFR–Ras signaling pathways in keratinocyte-derived malignancies.In vitro, exogenous stimulation of the EGFR–Ras signaling pathway through EGF or transfection of the dominant-active form of the Ras oncogene (H-RasV12) suppressed whereas an EGFR tyrosine kinase inhibitor increased CCL27 mRNA and protein production in keratinocytes. In mice, neutralization of CCL27 led to decreased leukocyte recruitment to cutaneous tumor sites and significantly enhanced primary tumor growth. Collectively, our data identify a mechanism of skin tumors to evade host antitumor immune responses.