Binary dermoscopic classification inadequately stratifies urgent and convenience excision decisions in community dermatology practice.

Abstract

Dermoscopy is traditionally framed as a binary benign-malignant tool. Yet real-world practice uses multi-level triage: dermatologists allocate lesions to urgent excision, non-urgent excision, or surveillance based on dermoscopic atypia, contextual factors, and surveillance feasibility. Convenience excisions -low-suspicion lesions removed when monitoring is impractical - are a common, unacknowledged pattern. Dermoscopic literature reports only binary outcomes, ignoring this process. To evaluate whether clinical urgency classification captures melanoma risk beyond isolated dermoscopic assessment in community practice with limited surveillance capacity. We retrospectively analysed 371 consecutively excised melanocytic lesions. Two blinded dermatologists independently assigned four-level gestalt atypia scores. Referring dermatologists prospectively classified excisions as urgent or non-urgent, integrating morphological concern with surveillance constraints. Histopathology identified melanoma prevalence and number needed to excise (NNE). Multivariable regression evaluated predictors of melanoma. Melanoma prevalence was 23.2% (86 melanomas: 41 in situ, 45 invasive; NNE=4.3). Inter-rater agreement for morphology-only assessment was poor (κ=0.16), with weak histopathological correlation (ρ=0.23-0.29). Most excisions (75%) were non-urgent, predominantly convenience excisions, yet captured 44 of 86 melanomas (51%). For invasive melanoma alone, NNE rose to 8.2, still favourable versus the 1:10 benchmark. Age predicted melanoma-in situ diagnoses, but age-stratified analysis showed no disproportionate overdiagnosis. Clinical urgency -which integrates morphological concern with contextual factors and shares information with dermoscopic assessment- predicted melanoma (OR=3.05, p=0.002) beyond gestalt scores, age, and phenotypic complexity. Binary dermoscopic frameworks inadequately reflect real-world multi-factorial triage. Effective melanoma detection depends on integrating morphology with patient-specific risk and system-level factors beyond dermoscopy alone.