Purpose: We sought to compare the sensitivity, specificity, accuracy, and interobserver agreement of the two most commonly used classification systems for conjunctival melanocytic intraepithelial lesions with the new World Health Organization (WHO) classification. Design: Retrospective case series and evaluation of classification systems. Methods: We reviewed the pathology and medical records of all patients who underwent a primary biopsy procedure for conjunctival primary acquired melanosis (PAM) at Wills Eye Hospital between 1974 and 2002 who had ≥36 months of follow-up. Data collected included age, sex, clinical findings, recurrence, and progression to melanoma. Twelve ophthalmic pathologists analyzed scanned hematoxylin and eosin–stained virtual microscopic slides using 3 classification systems: PAM, conjunctival melanocytic intraepithelial neoplasia, and the WHO 4th edition classification of conjunctival melanocytic intraepithelial lesions. Observer agreement, sensitivity, specificity, and diagnostic accuracy of each classification system were assessed. Results: There were 64 patients who underwent 83 primary excisions with cryotherapy for conjunctival PAM who had adequate tissue for histopathologic evaluation. The interobserver agreement in distinction between the low- and high-grade lesions was 76% for PAM, 67% for conjunctival melanocytic intraepithelial neoplasia, and 81% for WHO classification system. Low-grade lesions provided the greatest interpretative challenge with all 3 classification systems. The 3 classification systems had comparable accuracy of 81%-83% in their ability to identify lesions with potential for recurrence. Conclusions: This study highlights the comparable strengths and limitations of the 3 classification systems for conjunctival melanocytic intraepithelial lesions and suggests that the simplified WHO classification scheme is appropriate for evaluation of these lesions.
ASJC Scopus subject areas