Elvia Mastrogiuseppe, Rumana N Hussain, Zakariah Ali, Sarah Coupland, Heinrich Heimann
A tailored sequential multimodal approach achieved satisfactory local disease control, with local recurrence being primarily driven by non-bulbar involvement and residual pigmentation after primary therapy, while metastatic spread was strongly associated with ulcerative features, followed by non-bulbar involvement and tumour thickness.
PURPOSE: To evaluate the outcomes of a stepwise multimodal treatment approach for conjunctival melanocytic lesions ranging from low-grade conjunctival melanocytic intraepithelial lesion (LG-CMIL) through high-grade CMIL (HG-CMIL) to invasive conjunctival melanoma (iCoM) and to propose a practical management framework.
DESIGN: Retrospective, single-centre cohort study.
SUBJECTS: Patients diagnosed with LG-CMIL, HG-CMIL, or iCoM and managed at the Liverpool Ocular Oncology Centre between January 2015 and April 2025 were included.
METHODS: A sequential stepwise strategy was employed, comprising surgical excision or incisional biopsy, topical chemotherapy (mitomycin C, 5-fluorouracil, or interferon alpha-2b), and adjuvant Ruthenium-106 plaque brachytherapy or proton beam radiotherapy when indicated, selected according to histopathological grade and clinical growth pattern (focal versus diffuse). Kaplan-Meier survival analysis and Cox proportional hazards regression identified prognostic factors for recurrence and metastasis.
MAIN OUTCOME MEASURES: Local recurrence was the primary endpoint; metastatic disease and residual pigmentation after primary therapy were the secondary endpoints.
RESULTS: Among 165 eyes, including 39 LG-CMIL, 39 HG-CMIL, and 87 iCoM cases, HG-CMIL demonstrated the highest recurrence rate (24.3%; 9 eyes) and the earliest recurrence events (median 8.2 months), whereas LG-CMIL showed the lowest recurrence risk (11.1%; 4 eyes). Metastatic disease occurred exclusively in iCoM (11.4%; 8 eyes), accounting for 16.7% of deaths in the cohort. Non-bulbar involvement (HR 2.76, p=0.014) and residual pigmentation (HR 2.77, p=0.024) independently predicted local recurrence, while ulcerative features (HR 15.08, p=0.001), non-bulbar involvement (HR 7.28, p=0.015) and increased tumour thickness (HR 2.15, p=0.030) were associated with metastatic spread.
CONCLUSIONS: A tailored sequential multimodal approach achieved satisfactory local disease control, with local recurrence being primarily driven by non-bulbar involvement and residual pigmentation after primary therapy, while metastatic spread was strongly associated with ulcerative features, followed by non-bulbar involvement and tumour thickness.