Plasma lysophosphatidylcholine levels: potential biomarkers for colorectal cancer.
Journal of clinical oncology : official journal of the American Society of Clinical Oncology 2007 Jul 1; 25(19):2696-701
Zhao Z Z, Xiao Y Y, Elson P P, Tan H H, Plummer S SJ, Berk M M,
Department of Obstetrics and Gynecology, Indiana University Cancer Center, Indiana University School
PURPOSE Plasma levels of lysophospholipids were evaluated as potential biomarkers for colorectal cancer (CRC), where a highly reliable and minimally invasive blood test is lacking. PATIENTS AND METHODS Patients with CRC (n = 133) and control subjects (n = 125) were recruited through the Cleveland Clinic. Preoperative plasma samples were analyzed for lysophospholipid levels using liquid chromatography mass spectrometry in a blinded fashion. Participants were randomly divided in a 2:1 ratio into a "training set" (TS) and a "validation set" (VS). Logistic regression models were used in the TS to identify markers that best discriminated between CRC and controls. A cutoff point for the final discriminating model was developed using the receiver operating characteristic curve to achieve 95% specificity. All analyses were then independently validated in the VS. Results Plasma levels of several lysophosphatidylcholines (LPCs), including 18:1- and 18:2-LPC, were significantly decreased in CRC patients compared with controls (P < .001). A model based on total saturated LPC and the difference between the proportional amounts of 18:2-LPC and 18:1-LPC in the unsaturated LPC fraction was derived from the TS. This model achieved a sensitivity and specificity of 82% and 93%, respectively, in the VS. Overall, 118 (94%) of 125 control subjects and 113 (85%) of 133 CRC cases were correctly identified, including eight (89%) of nine CRC cases with stage T1 disease. CONCLUSION Percentage of 18:1-LPC or 18:2-LPC plasma levels compared with total saturated LPC levels, either individually or in combination, may represent potential biomarkers for CRC.