Ficantly increased mortality at univariate survival analysis, with surviving patients dropping to 22 after 9 months; at the same time point, 60 of the patients with low suPAR urinary levels were still alive. The association between high urinary suPAR/creatinine levels and poorer survival was not evident among resected patients at univariate analysis, possibly due to the small sample size (n = 21) of this patients group. However, the link between high urinary suPAR levels and worse prognosis was confirmed by multivariate survival analysis: Cox regression including both resected and unresected patients identified a high suPAR/ creatinine <a href="">CYM5442</a> level as an independent predictor of poor prognosis (odds ratio = 2.10). The indication by multivariate survival analysis that urinary suPAR/creatinine gives additional prognostic information independent of that given by tumour stage is further supported by proportions analysis that showed the absence of correlation between tumour stage and suPAR/ creatinine. In fact, patients with resectable (stage I and II), locally advanced (stage III) or metastatic (stage IV) tumours showed a similar proportion of cases displaying high suPAR/creatinine urinary levels (p = 0.17).Sorio et al. BMC Cancer 2011, 11:448 7 ofFigure 3 Univariate survival analysis of pancreatic adenocarcinoma patients. Kaplan-meier analysis of pancreatic adenocarcinoma patients based on surgery and suPAR/creatinine levels; curves compared by Mantel-Cox log-rank test, brackets illustrate 95 CI of survival. A) survival difference between resected patients bearing levels of suPAR/creatinine higher or lower than 9.1 ng/mg (95th percentile of healthy donors) is not significant (p = 0.46). B) Difference in survival between non resected patients according <a href="" title=View Abstract(s)">PubMed ID:</a> to suPAR/creatinine levels (p = 0.034). <a href="" title=View Abstract(s)">PubMed ID:</a> C) Survival difference between resected and non-resected patients without considering suPAR/creatinine levels (p = 0.0001). D) Survival of resected vs. nonresected patients with either low or high suPAR/creatinine (log-rank test p < 0.0001, log-rank test for trend between curves p = 0.014).Our work showed that half of PDAC patients had markedly elevated suPAR/creatinine values and poorer prognosis, while the other half of PDAC patients had values comparable to those of chronic pancreatitis patients and healthy subjects. Interestingly, previous immunohistochemical studies reported that the level of uPAR in primary pancreatic adenocarcinoma was higher than in normal pancreas and tended to be associated with a worse prognosis [32-34]. In particular, Hildenbrand et al showed that virtually all (48 of 50) primary carcinomas overexpress uPAR compared to normal pancreatic tissue by immunohistochemistry, and that 52 of them had uPAR gene amplification by FISH analysis. This latter feature correlated with worse prognosis and with a higher protein concentration in the tumour as quantitatively measured by ELISA on tissue lysates;conversely, tumours without gene amplification and chronic pancreatitis samples had comparably lower concentrations of uPAR [34]. Compared to its measurement on the tumour tissue, the measurement of suPAR in urine provides at least three advantages. First, the measurement does not require tumour biopsies, and can thus be performed on both resectable subjects, before and after surgery, and on unresected patients. Second, the measurements can be taken repeatedly and may be used in the follow-up of.