PRAME staining was associated with shorter survival in a cohort of people with primary uveal melanoma, with the clearest pattern appearing when researchers examined whole sections of tumour. Among the 83 whole-section cases, median overall survival was 13 months for PRAME-positive tumours and 52 months for PRAME-negative tumours. Median metastasis-free survival, the time before spread was recorded, was 5 months versus 20 months. The reported p values were 0.039 and less than 0.001, respectively.
PRAME is a protein measured here with immunohistochemistry, a laboratory staining test. A tumour counted as positive when nuclear staining appeared in more than 10% of its cells. Two observers scored the slides without knowing patient outcomes. The researchers drew on hospital records, pathology and molecular data, and registry follow-up; the report also described ethics approval and consent or permission for historic anonymised samples.
The specimen changed the picture
PRAME assessment was available for 383 cases tested on tissue microarrays and 83 tested on whole-tumour sections, with 57 cases assessed by both methods. The difference in positive results was large: 19 of 383 tissue-microarray cases, or 5%, were positive, compared with 33 of 83 whole-section cases, or 39.8%.
That difference was reflected in the survival analysis. In the tissue-microarray group, PRAME-positive tumours had median overall survival of 26 months, compared with 50 months for PRAME-negative tumours. Estimated five-year overall survival was 25% versus 42%, and the reported p value was 0.03. The metastasis result was less certain: median metastasis-free survival was 14 months versus 20 months, the five-year estimates were approximately 0% versus approximately 11%, and the difference was not statistically significant, with a reported p value of 0.376. Only 19 tissue-microarray tumours were PRAME-positive.
Whole-section PRAME was also associated with chromosome 8q gain. In the tissue-microarray cases, it was associated with spindle-cell type and loss of nuclear BAP1. The reported p values for those associations were 0.009, 0.013 and 0.038, respectively.
The combined markers sharpened the risk signal
The researchers also combined PRAME with BAP1, a second marker assessed in the cohort. In whole-section analysis, the combined groups differed in both overall and metastasis-free survival. Median overall survival was 52 months for tumours negative for both PRAME and BAP1, compared with 16 months for tumours positive for PRAME but negative for BAP1. Median metastasis-free survival was 29 months versus 6 months. The reported p values for the two comparisons were 0.021 and less than 0.001.
A similar pattern appeared after chromosome 3 status was added. The PRAME-positive, monosomy 3 group had median overall survival of 6 months, compared with 52 months for PRAME-negative, monosomy 3 tumours. Median metastasis-free survival was 6 months versus 29 months. Across the chromosome 3 categories, the reported p values were 0.008 for overall survival and less than 0.001 for metastasis-free survival. The PRAME-positive, monosomy 3 subgroup had the poorest reported prognosis, while metastasis-free survival for at least one other subgroup could not be estimated.
Whole-section PRAME positivity also remained associated with shorter metastasis-free survival in the reported multivariable Cox model. A hazard ratio is a measure comparing event rates over time; the study reported a value of 10.5, with a p value of 0.002.
The evidence has limits
These results are associations, not proof that PRAME causes metastasis or death. The study was a cohort built from clinical records, pathology and molecular data, and registry follow-up. The whole-section analysis covered 83 cases, and the combined biomarker analyses included sparse groups, with some estimates not available. Those features make the most detailed subgroup comparisons harder to interpret.
The way tissue was sampled may also have affected the result. The authors report that 95% of tissue-microarray-scored cases were PRAME-negative, while 21 tumours that were negative on the microarray were positive when whole sections were examined. That pattern suggests limited sampling can miss focal or uneven PRAME expression, helping explain why the whole-section positivity rate was higher.
The study therefore supports a possible role for whole-section PRAME staining in estimating risk in uveal melanoma, especially when read with BAP1 or chromosome 3 status. It does not establish that the stain changes treatment or improves survival. The evidence is a link between a marker and outcomes, not proof that the marker causes those outcomes.
Paper data and sources
Original title: PRAME predicting outcome in uveal melanoma
Authors: Owens S, Walsh N, Horgan N et al.
Journal/Repository: The journal of pathology. Clinical research
Status: Peer-reviewed
First online: 2026-10-14
DOI: Not available
Original paper · Full text