Adding Iris Lesion Thickness to ABCDEF Criteria Improves Melanoma Prediction Accuracy

In this study, biopsy-confirmed iris melanomas were, on average, significantly larger (3.20mm vs. 1.80mm) and thicker (1.40mm vs. 1.00mm) than non-melanomas. Photo: Christine Sindt, OD. Click image to enlarge. The mnemonic “ABCDEF” is often used by eyecare providers to help identify predictors of melanoma in patients presenting with an iris lesion. Each letter represents a high-risk clinical feature: Age ≤40, Blood in the anterior chamber, Clock hour inferior, Diffuse configuration, Ectropion uveae and Feathery margins. The diagnostic performance of this tool was assessed in a recent study presented on Tuesday at ARVO 2026 in Denver. Its authors found that ectropion uveae was the single strongest ABCDEF predictor of melanoma and that measuring tumor thickness substantially improved diagnostic discrimination.The retrospective cohort study reviewed adult patients (age ≥18) with documented pigmented iris lesions referred to a tertiary academic center. Non-pigmented lesions and patients under 18 were excluded. Clinical data captured included the ABCDEF features, intrinsic (tumor) vessels, extent of clock-hour involvement (≥three clock hours), angle seeding and tumor dimensions. Of 262 lesions, 13.4% were biopsy-confirmed iris melanomas. Melanomas were larger (median largest diameter: 3.20mm vs. 1.80mm for benign lesions) and thicker (median thickness: 1.40mm vs. 1.00mm). On univariate analysis, several findings were far more common in melanomas than benign lesions: intrinsic tumor vessels (82.8% vs. 12.4%), involvement of ≥three clock hours (75.0% vs. 18.7%), angle seeding (86.2% vs. 26.5%) and ectropion uveae (65.6% vs. 31.5%).In multivariable analysis restricted to ABCDEF features, ectropion uveae remained the only independent predictor (odds ratio [OR]: 3.31). When tumor thickness was added to the model, both thickness (OR: 2.76 per mm increase) and ectropion uveae (OR: 3.71) were independently associated with melanoma. Adding thickness improved the diagnostic area-under-the-curve from 0.726 to 0.869.Based on these findings, the study authors concluded in their abstract, “Although intrinsic lesion vascularity, angle seeding and extensive clock hour features are strong univariate indicators, incorporating thickness with ABCDEF criteria provides a more accurate framework for clinical risk stratification in pigmented iris lesions.”Original abstracts ©2026 Association for Research in Vision and Ophthalmology. Click here for the source. Pellack D, Grimes K, Pouw AE. Iris Lesion Characteristics: Differentiating Pigmentation from Melanoma. ARVO 2026 annual meeting. This article was developed by the editorial staff in conjunction with experts in the field. In the process, AI may have been among the editorial tools used to meet the goals of human editors, who approved all content.