Abstract
Purpose :
The purpose of this study was to assess the relationship between the cytology cell type (spindle vs epithelioid vs mixed) obtained during fine needle aspiration biopies of uveal melanoma and the results of commercially available genetic testing (gene expression profiling (GEP), PRAME analysis) with a focus on mismatched patients, cases in which cytomorphology and genomics had opposite prognostic predictions for mortality.
Methods :
This was a retrospective review of 235 patients who underwent a fine needle aspiration biopsy for cytopathologic and genomic analyses of uveal melanoma. Only patients with 12 months or more of follow-up and a full data set for analysis were included. Clinical variables were reviewed including: patient demographic data; clinical tumor data (size of tumor, AJCC class); biopsy technique details (transscleral vs transvitreal vs transcorneal, number of passes, GEP class and discriminant value, Prame status, cytology findings). Statistical analyses were performed.
Results :
Metastasis-free survival (MFS) stratified by GEP and Prame was consistent with published data, with Kaplan-Meier analysis demonstrating a 95% 5 year MFS in GEP Class 1A patients; 80% in Class 1B patients, and 25% in Class 2 patients (p<0.01). Patients with spindle/ predominantly spindle cells had an 87% 3-year MFS, vs. epithelioid/predominantly epithelioid with 63% (p=0.03). Prame expression was associated with a higher rate of metastases at 3 years than Prame negativity (85% vs. 62%, p=0.06). 34 patients demonstrated a mismatch in metastatic disease prediction between cytomorphology and genomics. Among epithelioid/Class 1A or epithelioid/1B patients, 3 year MFS was 90%; among Spindle/Class 2 patients, 3-year MFS was 61%. Patients who were Epithelioid/Class 2 had a 3-year MFS rate of 48%, vs. Spindle/Class 2 patients with a 61% rate, suggesting that although GEP was a significantly better overall predictor of metastatic mortality, cytomorphology in mismatched cases accurately predicted a modified risk (p<0.01).
Conclusions :
In addition to identifying viable cells for genomic testing, experienced cytologic analysis by an ocular pathologist appears to modify the accuracy of metastatic mortality predictions in genomic/cytologic mismatched cases. Cytomorphologic features may reflect cells with intermediate genomic risk, and/or rare cases with genomic heterogeneity between two or more biopsy sites.
This is a 2020 ARVO Annual Meeting abstract.