Publication:

Confocal reflectance microscopy versus digital dermoscopy for the diagnosis of melanoma in equivocal lesions: a systematic review

The diagnosis of melanoma in clinically or dermoscopically equivocal melanocytic lesions is difficult and gives rise to many unnecessary excisions. This review compares in vivo reflectance confocal microscopy or confocal reflectance microscopy (RCM of CRM) and digital dermoscopy in the diagnosis of melanoma in equivocal lesions. We conducted a systematic review of PubMed, Embase, Cochrane Central and Web of Science (according to PRISMA) until 30 January 2026. Eligible studies were direct comparisons of RCM and digital dermoscopy in adults with equivocal lesions and had histopathology as the reference standard. Data extraction and quality appraisal using the quality assessment of diagnostic accuracy studies-2 (QUADAS-2) instrument were performed in duplicate. Where studies were homogenous enough for the statistical analysis, we used a bivariate random-effects model to pool sensitivity and specificity; or used a structured narrative synthesis otherwise. Fifteen (2007-2020) studies were included in which referral cohorts of equivocal lesions were included. In data pooled from the similar studies sensitivity of RCM was about 91% {[95% confidence intervals (CI) 88-94%]} with specificity almost at 69% (95% CI 62-76%). Digital dermoscopy demonstrated the same sensitivity, approximately 88% (95% CI 83 92) but significantly worse specificity, approximately 34% (95% CI 2841). Heterogeneity was high and was determined by referral setting, reader experience, device type and algorithm use. Studies consistently found that the number needed to excise was reduced from approximately 3.7 by dermoscopy alone to near 1.1 by RCM guided excision. For lesions that are still equivocal even after being clinically and assessed by dermoscopy, the RCM substantially improves specificity but at the cost of a high sensitivity. RCM is a useful second-level test in specialist environments but operator education, cost, and poor penetration depth limit the broader use of this modality.