A biobank analysis of prognostic biomarkers of the systemic inflammatory response in patients presenting with malignancy of undefined primary origin.

BACKGROUND: Survival prediction in patients presenting with malignancy of undefined primary origin (MUO) is challenging, with a lack of validated prognostic tools. Biomarkers of the systemic inflammatory response independently predict survival in other cancer types, but their role in MUO is unclear. The aim of this study was to assess biomarkers of the systemic inflammatory response in patients presenting with MUO. PATIENTS AND METHODS: A biobank of 1049 patients presenting with MUO referred to a regional oncology service in Scotland was analysed. Key inflammatory biomarkers (white cell count, neutrophil count and C-reactive protein combined with albumin [to give the modified Glasgow Prognostic Score {mGPS}]) were examined. The relationship between these and survival was examined using Kaplan-Meier and Cox regression methods. RESULTS: Data were available for 1049 patients. Median survival was 4.3 months (interquartile range: 1.7-16.0 months). On multivariate analysis mGPS was independently associated with survival and stratified survival from 13.6 months (mGPS: 0) to 2.3 months (mGPS: 2) (p < 0.001). The mGPS was predictive of survival on multivariate analysis in patients found to have a non-cancer diagnosis (p = 0.034), an identified primary cancer (0.002), cancer of unknown primary (CUP) (p = 0.011), those for whom biopsy was not done (MUO) (p = 0.036), those found to have an identified primary cancer (0.002) and even those found to have a non-cancer diagnosis (p = 0.034) after further detailed investigations. In patients with CUP mGPS predicted survival regardless of the recognised clinicopathological prognostic subgroup (p < 0.001). CONCLUSIONS: The results of the present study demonstrate that biomarkers of the systemic inflammatory response are reliable prognostic factors in patients presenting with MUO. These simple, objective, routine clinical tests may inform clinical management.