Inflammation as a validated prognostic determinant in carcinoma of unknown primary site

Carcinoma of unknown primary (CUP) is a clinical presentation with a poor prognosis. Inflammation-based prognostic systems are stage-independent prognostic predictors in various malignancies. We aimed to assess the accuracy of the modified Glasgow Prognostic Score (mGPS), neutrophil/lymphocyte ratio (NLR) and platelet/lymphocyte ratio (PLR) as objective prognostic models in CUP.We derived inflammatory scores in 60 consecutive CUP referrals to the Imperial College oncology unit between 1996 and 2011. Patient demographics, treatment and staging data and full blood profiles were collected. An independent cohort of 179 patients presenting to the Taipei Veterens Hospital between 2000 and 2009 were used as a 'validation' data set. Uni- and multivariate survival analysis was used to predict the overall survival (OS).Sixty patients were included: median age 61 (range: 33-86); 51% men; median OS 5.9 months (0.7-42.9); 88% with distant metastases. On univariate analysis NLR >5 (P=0.04) and mGPS (score 1-2) (P=0.03) correlated with OS. Multivariate analysis demonstrated significant hazard ratios for NLR; 2.02 (CI 1.0-4.1) (P=0.04) and mGPS; 1.52 (CI 1.0-2.3) (P=0.03). These findings were reinforced by analysis of the validation data.NLR and mGPS are independent, externally validated prognostic markers in CUP, with superior objectivity compared with performance status.