Large-scale independent validation of the nuclear factor-kappa B p65 prognostic biomarker in prostate cancer

Over the last decade, we and others have uncovered a robust association between the nuclear localisation of nuclear factor-kappa B (NF-κB) p65, prostate cancer (PCa) aggressiveness and biochemical recurrence (BCR). Our goal was to validate these results in a large independent cohort of PCa patients who underwent radical prostatectomy.

A set of 1826 fully annotated prostate cancers treated by radical prostatectomy were analysed in a tissue microarray (TMA) format for NF-κB p65 immunohistochemistry-based protein expression. We performed standard Cox proportional hazard regression models for follow-up data, bootstrap procedure for model internal validation, Harrell's concordance index for model discrimination and graphical assessment of predicted versus actual outcomes for model calibration.

We observed a significant association between an increase in the nuclear frequency of NF-κB p65 and Gleason score (P<0.001), overall BCR (P<0.001) and development of metastases (P=0.001). NF-κB was found to be an independent predictor of BCR (P<0.001, Cox regression). However its contribution to the predictive accuracy of a multivariate model, which included preoperative PSA, Gleason score, extraprostatic extension, lymph node invasion, seminal vesicle involvement and surgical margin status, was modest.

Our study offers validating results linking NF-κB p65 with disease progression using a large cohort of European men. However, the contribution of NF-κB to a post-surgical predictive model appears modest. Further validating work should focus on evaluating the contribution of NF-κB p65 in pre-treatment models.