Can we use video images to triage pigmented lesions?

Abstract

Background: Advances in telemedicine permit consultations where the doctor and patient are at different sites.

Objectives: To determine whether this technology could be used to triage referrals of pigmented lesions to a dermatology out-patient clinic, and thereby assist in managing the ever increasing number of lesions being referred.

Methods: When patients attended clinic, a clinical diagnosis of their pigmented lesion was made and recorded. A still image of each lesion was subsequently taken from a video camera using a PC card, and stored. These images were subsequently viewed in conjunction with the general practitioner's referral information and designated as 'warrants referral' or 'does not warrant referral'. For each lesion this decision was compared with the clinical diagnosis made during the live consultation in the clinic (the 'gold standard'). Clinical diagnoses designated as warranting referral were malignant melanoma, basal cell carcinoma, keratoacanthoma, atypical naevus and pyogenic granuloma (due to the potential clinical confusion with amelanotic melanoma). Lesions that were not considered to warrant referral included benign melanocytic naevus, seborrhoeic keratosis, dermatofibroma, congenital naevus, solar lentigo, actinic keratosis and various other benign conditions.

Results: In total, 819 lesions were evaluated, resulting in a mean sensitivity of 81% and specificity of 73% for the technique.

Conclusions: We feel that the overall sensitivity of 81% is encouraging as regards the use of such a technique as a triage tool, but that the inability to examine the whole patient or palpate the lesions is a major drawback in the safe triage of patients with pigmented lesions.