Abstract

326 Background: CA19.9, NLR and PLR have all been proposed as prognostic in pancreatic cancer. We analysed correlation between NLR, PLR and rate of change of CA19.9. Methods: A total of 63 metastatic pancreatic cancer patients were identified from our database and evaluated retrospectively for blood count, NLR, PLR and serial CA19.9 levels during treatment. Daily Rate of Change of CA19.9 levels were calculated for the first 90 days (DRC90) of the patient’s treatment. Kaplan-Meir curves, univariate and multivariate Cox-regression analyses were calculated to assess the effects of these 3 markers on overall survival. Results: In a univariate analysis, PLR > 240, NLR > 5 and DRC90 > 0.4% were all significantly associated with deceased overall survival. The Cox proportional hazards model showed that NLR < 5 (HR 0.475, 95% CI 0.259 to 0.873, P = 0.017), PLR < 240 (HR 0.444, 95% CI 0.229 to 0.861, P = 0.016), and a DRC90 < 0.4% (HR 0.294, 95% CI 0.102 to 0.851, P = 0.024) were independent predictors of good prognosis (22.6 months vs. 9.6 months, 22.3 months vs 12.4 months and 23.9 months vs. 9.3 months respectively). In multivariate analysis, only a DRC90 < 0.4% was independently associated with a longer survival (HR 0.239, 95% CI 0.076 to 0.752, P = 0.014). The formula (F) {PLR + (NLRxNLR) + (DRC90 x 100)} was predictive for survival, as patients with F > 190 (HR 3.295, 95% CI 1.232 to 8.807, P = 0.017), having a significantly lower survival rate than patients with F < 190 (25.1 months vs. 10.6 months, log-rank P = 0.009). Conclusions: These findings indicate the prognostic utility of the rate of CA 19.9 decline - measured as a standardised daily percentage change in value over 90 days. Our data validates daily rate of change of CA 19.9 over 90 days as an independent variable that correlates with prognosis, independent of PLR and NLR. We also identified a novel formula - PLR + (NLRxNLR) + (DRC90 x 100) - as being predictive for survival. We would like to increase the sample size to further validate our initial findings and investigate possible relationships in combining these variables for better prognostication in metastatic pancreatic cancer.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call