Accuracy of teledermatology for nonpigmented neoplasms

Background Studies of teledermatology utilizing the standard reference of histopathology are lacking. Objective To compare accuracy of store-and-forward teledermatology for non-pigmented neoplasms with in-person dermatology. Methods This study was a repeated-measures equivalence trial involving veterans with non-pigmented skin neoplasms. Each lesion was evaluated by an in-person dermatologist and a teledermatologist; both generated a primary diagnosis, up to two differential diagnoses, and management plan. The primary outcome was aggregated diagnostic accuracy (percent correct matches of any chosen diagnosis with histopathology). Secondary outcomes included management plan accuracy (percent correct matches with expert panel management plan). Additional analyses included evaluation of the incremental effect of using polarized light dermatoscopy in addition to standard macro images, and evaluating benign and malignant lesion subgroups separately. Results Most of the 728 participants were male (97.8%) and Caucasian (98.9%). The aggregated diagnostic accuracy (primary outcome) of teledermatology (macro images) was not equivalent (95% confidence interval [CI] for difference within +/−10%) and was inferior (95% CI lower bound