Remote cutaneous confocal microscopy: A multicentric prospective study evaluating diagnostic accuracy for melanoma and keratinocyte carcinoma in tertiary settings.

Abstract

Cutaneous confocal microscopy (CCM) facilitates in vivo visualization of skin at a cellular level. Use of a "store and forward" approach for remote-CCM interpretation (remote-CCM) across multiple sites has not been tested and may increase access to noninvasive diagnosis. To test the diagnostic accuracy and safety of remote-CCM. We prospectively recruited lesions selected for biopsy for skin malignancy across 5 Australian tertiary dermatology centers. CCM, clinical and dermatoscopy images were acquired prebiopsy and accessed by a cloud-based platform for interpretation by CCM readers. CCM diagnosis was compared with histopathology results. Among the 201 lesions included, melanoma was the most common malignancy (34/72, 47.2%). Of the 89 lesions (44.8%) potentially "saved" from biopsy, 80 (90%) were truly benign lesions and 9 (10.1%) were missed malignant lesions of melanoma in situ (n = 7) and squamous cell carcinoma (SCC) (n = 2). No invasive melanomas were missed. Sensitivity of remote-CCM for detection of malignancy was 89% (95% CI, 79%-95%) and specificity was 64% (95% CI, 55%-73%). The study recruited from high-risk populations and excluded lesions that were not biopsied. Remote-CCM has comparable accuracy to bedside CCM and safely reduces unnecessary biopsies. Potential SCCs are not appropriate for remote-CCM. Follow-up of borderline melanocytic lesions is recommended.