Iris nevus growth into melanoma: analysis of 1611 consecutive eyes: the ABCDEF guide

Abstract

Purpose: To determine clinical features predictive of growth of iris nevus into melanoma.

Design: Retrospective, comparative case series.

Participants: A total of 1611 consecutive patients referred to an ocular oncology center with iris nevus.

Intervention: Observation and photographic documentation.

Main outcome measures: Growth into melanoma.

Results: The mean age at referral for iris nevus was 51 years (median, 54; range, <1-94 years). At presentation, the mean tumor basal diameter was 3 mm (median, 3 mm; range, <1-12 mm) and mean tumor thickness was 0.8 mm (median, 0.5 mm; range, 0-5 mm). All patients were initially diagnosed with benign iris nevus. Growth of iris nevus to melanoma was confirmed in 2% of eyes (n = 27) over a mean follow-up of 68 months (median, 46 months; range, 3-465 months). By Kaplan-Meier estimates, iris nevus growth to melanoma occurred in <1%, 3%, 4%, 8%, and 11% at 1, 5, 10, 15, and 20 years, respectively. Factors predictive of iris nevus growth to melanoma by multivariable analysis included age ≤ 40 years at presentation (hazard ratio [HR], 3), episode of hyphema (HR, 9), 4:00 to 9:00 clock hour location of tumor (HR, 9), diffuse tumor (involving entire iris surface) (HR, 14), ectropion uveae (HR, 4), and feathery tumor margins (HR, 3). Additional important factors by univariable analysis included tumor seeding on the iris or in the anterior chamber angle, feeder vessels, and nodule formation. These factors can be remembered using the mnemonic ABCDEF, representing A = age young, B = blood, C = clock hour inferior, D = diffuse, E = ectropion, and F = feathery margin.

Conclusions: In an analysis of 1611 cases of iris nevus referred for evaluation at an ocular oncology center, growth into melanoma occurred in 8% by 15 years. Risk factors for growth, identified by ABCDEF included Age young, Blood (hyphema), Clock hour inferior, Diffuse configuration, Ectropion uveae, and Feathery tumor margin.