A serum lipidomic strategy revealed potential lipid biomarkers for early-stage cervical cancer

Cervical cancer (CC) is a common tumor of women worldwide. Here, we conducted a non-targeted lipidomic study to discover novel lipid biomarkers for early-stage CC. The lipidomic analysis of 71 samples in discovery set and 72 samples in validation set were performed by coupling ultra-high-pressure liquid chromatography (UHPLC) with quadrupole time-of-flight tandem mass spectrometry (Q-TOF-MS). Lipids with variable importance (VIP) values greater than 1, adj. p < 0.05 (the adjusted p value obtained from false discovery rate correction) and fold change (FC) higher than 1.5 were reserved as potential biomarkers. Subsequently, receiver operating characteristic (ROC) curve and binary logistic regression were implemented to assess the diagnostic potential of these biomarkers and to acquire the best biomarker combination. A lipid biomarker panel, including phosphatidylcholine (PC, PC 14:0/18:2) and phosphatidylethanolamine (PE, PE 15:1e/22:6 and PE 16:1e/18:2), was established. This panel was effective in distinguishing between CC and non-CC (squamous intraepithelial lesions [SIL] and healthy controls) within the area under the ROC curve (AUC), sensitivity, and specificity reaching 0.966, 0.952, and 0.860 for discovery set and 0.961, 0.920, and 0.915 for external validation set. Furthermore, this panel was also capable of discriminating early-stage CC from SIL with AUC, sensitivity, and specificity reaching 0.946, 0.952, and 0.800 for discovery set and 0.956, 0.960, and 0.815 for external validation set. The combination of PC 14:0/18:2, PE 15:1e/22:6, and PE 16:1e/18:2 could serve as a promising serum biomarker for discriminating early-stage CC from SIL and healthy subjects.