Dermatologists' perceptions on the utility and limitations of teledermatology after examining 55,000 lesions.

INTRODUCTION: Few studies have assessed the perception of teledermatologists about the utility and limitations of teledermatology, especially to diagnose a broad range of skin diseases. This study aimed to evaluate dermatologists' confidence in teledermatology, its utility and limitations for dermatological conditions in primary care. METHODS: An analytical study that used a survey for dermatologists who diagnosed 30,916 patients with 55,012 lesions through teledermatology during a 1-year project in São Paulo, Brazil. RESULTS: Dermatologists found teledermatology useful for triage and diagnosis, especially for xerotic eczema, pigmentary disorders and superficial infections. Their confidence in teledermatology was statistically higher by the end of the project (p = 0.0012). Limitations included some technical issues and the impossibility to suggest how soon the patient should be assisted face-to-face by a dermatologist. The most treatable group of diseases by teledermatology was superficial infections (92%). The use of dermoscopy images would significantly increase the confidence to treat atypical naevi and malignant tumours (p < 0.0001 and p = 0.0003 respectively). Follow-ups by teledermatology or feedback from primary-care physicians would be desirable, according to the dermatologists. DISCUSSION: We found it interesting that dermatologists became increasingly confident in teledermatology after the project and how they classified teledermatology as useful for triage, diagnosis and even treatment of most types of skin conditions followed at primary care. Dermoscopy should definitely be added to the photographs, especially for malignant tumours and atypical naevi. Most of the technical limitations found could be solved with a few improvements in the software/platform.

Computer algorithms show potential for improving dermatologists' accuracy to diagnose cutaneous melanoma: Results of the International Skin Imaging Collaboration 2017.

BACKGROUND: Computer vision has promise in image-based cutaneous melanoma diagnosis but clinical utility is uncertain. OBJECTIVE: To determine if computer algorithms from an international melanoma detection challenge can improve dermatologists' accuracy in diagnosing melanoma. METHODS: In this cross-sectional study, we used 150 dermoscopy images (50 melanomas, 50 nevi, 50 seborrheic keratoses) from the test dataset of a melanoma detection challenge, along with algorithm results from 23 teams. Eight dermatologists and 9 dermatology residents classified dermoscopic lesion images in an online reader study and provided their confidence level. RESULTS: The top-ranked computer algorithm had an area under the receiver operating characteristic curve of 0.87, which was higher than that of the dermatologists (0.74) and residents (0.66) (P < .001 for all comparisons). At the dermatologists' overall sensitivity in classification of 76.0%, the algorithm had a superior specificity (85.0% vs. 72.6%, P = .001). Imputation of computer algorithm classifications into dermatologist evaluations with low confidence ratings (26.6% of evaluations) increased dermatologist sensitivity from 76.0% to 80.8% and specificity from 72.6% to 72.8%. LIMITATIONS: Artificial study setting lacking the full spectrum of skin lesions as well as clinical metadata. CONCLUSION: Accumulating evidence suggests that deep neural networks can classify skin images of melanoma and its benign mimickers with high accuracy and potentially improve human performance.