Riccardo Pampena, Stefano Migliorati1, Giovanni Paolino, Michela Lai, Nicola Lippolis, Stefania Guida, Stefania Borsari, Sebastiano Pellerone, Sofia Maria Di Ciaccio, Elvira Moscarella, Giovanni Pellacani, Giuseppe Argenziano, Caterina Longo
ABSTRACT
Introduction: The diagnosis of fully amelanotic skin tumors is difficult on clinical and dermoscopic examination.
Objectives: We sought to identify an accurate and user-friendly dermoscopic algorithm to differenti-ate between benign and malignant pink lesions.
Methods: The database of 1 referral center was retrospectively reviewed for images of non-inflammatory fully amelanotic skin lesions. Two dermatologists jointly assessed a validation set of images for dermoscopic criteria and constructed a diagnostic algorithm, the 2-step-7-pink rule (2S-7PR). Two external clinicians, with different skills in dermoscopy and blinded to the final diagno-sis, separately evaluated images from the validation test sets using the prevalent criterion method and the new 2S-7PR algorithm.
Results: A total of 763 lesions from 652 patients were included in the validation set database, of which 68.3% were malignant and 31.7% were benign. Three suspicious dermoscopic criteria were included in the first step of the 2S-7PR: polymorphous or sharply focused vessels, scales or crusts, and erosions or ulcerations; and 4 non-suspicious criteria were included in the second: white collarette, white scar-like area, vascular lacunae, and necklace pinpoint vessels. High levels of specificity and sensitivity were calculated in the validation and test phases for both the expert and non-expert eval-uators, the former achieving higher levels of both sensitivity and specificity by employing the 2S-7PR compared to the prevalent method, and the latter only improved specificity.
Conclusions: The present study showed that an algorithm focused on a few reproducible and easily recognizable criteria could improve diagnostic accuracy in the management of amelanotic lesions.
