Evaluation of dermoscopic algorithm for seborrhoeic keratosis: a prospective study in 412 patients

Abstract

Background: Although most seborrhoeic keratoses can be easily diagnosed with dermoscopy, some lesions mimic a number of other diseases and lack typical dermoscopic criteria. Validation of the diagnostic algorithm and further understanding about dermoscopy of seborrhoeic keratosis is necessary.

Objective: To evaluate the dermoscopic criteria, define dermoscopic patterns and optimize the diagnostic algorithm of seborrhoeic keratosis in a large series of cases.

Patients and methods: An unselected consecutive series of 416 clinically suspected seborrhoeic keratoses from 412 patients were prospectively included. All the lesions were imaged with dermoscopy, analysed for dermoscopic patterns and features. An independent blinded histopathological diagnosis as well as dermoscopic diagnosis was made for each lesion.

Results: A total of 416 lesions were analysed. Using histopathological diagnoses as the gold standard, the two-step algorithm achieved a sensitivity of 79.1, a specificity of 78.3% (P < 0.01) and a κ index of 0.223. An optimized algorithm that combined lack of blue-grey or blue-white colour, sharp demarcation, mica-like structure and yellowish colour with the two-step algorithm demonstrated a sensitivity of 95.7%, a specificity of 78.3% (P < 0.001) and a κ index of 0.594. Pattern analysis revealed 12 dermoscopic patterns. The elevated lesions mainly presented fissures/ridges pattern, exophytic papillary pattern, crypts pattern and mica-like pattern; the flat lesions mainly presented structureless pattern and coral-like pattern.

Conclusion: Seborrhoeic keratosis may present a variety of dermoscopic patterns and features. The combination of lack of blue-grey or blue-white colour, sharp demarcation, mica-like structure and yellowish colour with the two-step algorithm could improve its diagnostic accuracy.