Pancreatic cancer has the worst prognosis among all cancers1. Cancer screening programs based on the analysis of body fluids can improve the survival time of patients, who are often diagnosed too late at an incurable stage2. Several studies have reported the dysregulation of lipid metabolism in tumor cells and tissues3, suggesting that the changes of blood lipidome may accompany tumor growth and progression. Analytical methods based on mass spectrometry (MS) using either direct infusion or chromatographic separation4 are convenient for high-throughput lipidomic profiling. Here we show that the comprehensive quantitation of a wide range of serum lipids reveals statistically significant differences between pancreatic cancer patients and healthy controls visualized by multivariate data analysis. Initial results for 364 human serum samples in the discovery phase were subsequently verified in the qualification phase on 554 samples measured by three independent laboratories, and finally on 830 samples from four blood collection sites in the verification phase. Concentrations suggestive of dysregulation of some very long chain sphingomyelins (SM 42:1, SM 41:1, SM 39:1, and SM 40:1), ceramides (Cer 41:1, and Cer 42:1), and (lyso)phosphatidylcholines (LPC 18:2) were recorded. Some lipid species indicated a potential as biomarkers of survival. The sensitivity and specificity to diagnose pancreatic cancer is over 90%, which outperforms CA 19-9, especially in early stage, and is comparable to established imaging diagnostic methods. The accuracy of lipidomic approach is not influenced by the cancer stage, analytical method, or blood collection site.
EP-1418 Initial results of carbon ion radiotherapy combined with S-1 for locally advanced pancreatic cancer
