Identification of MUM1 as a prognostic immunohistochemical marker in follicular lymphoma using computerized image analysis.
Detection of MUM1+ cells in follicular lymphoma (FL) tissues was previously found to be associated with poor prognosis in a single report, whereas the usefulness of Ki-67 immunostaining remains debated. Our goal was to establish whether these markers have predictive value for patients with FL. We analyzed MUM1 and Ki-67 expression using immunohistochemistry in biopsy samples from 434 patients from the PRIMA randomized trial. The MUM1 prognostic value was then validated in a cohort of 138 patients from the FL2000 randomized trial, using the optimal cutoff value obtained from the PRIMA cohort. The surface of positive staining was quantified using computerized image analysis. In the PRIMA cohort, both high levels of MUM1 positivity (cutoff value of 0.80%) and high levels of Ki-67 positivity (cutoff value of 10.25%) were significantly associated with a shorter progression-free survival (PFS) (P = .004 and P = .007 for MUM1 and Ki-67, respectively). In a multivariate Cox proportional hazards regression model, only MUM1 retained a statistical significance (hazards ratio 1.56; 95% confidence interval, 1.02-2.37; P = .038) after adjustment for the maintenance arm of treatment and the follicular lymphoma international prognostic index score. In the FL2000 cohort, high levels of MUM1 positivity were significantly associated to a shorter PFS (P = .004) and to a trend toward a shorter overall survival (P = .043). This remained significant using a multivariate Cox regression model after adjustment for the follicular lymphoma international prognostic index and the treatment arm for PFS (P = .016). These results show that MUM1 is a strong and robust predictive immunohistochemical marker in patients with FL.Lire l‘article