Research Paper Volume 16, Issue 1 pp 820—843

Figure 4. Construction and validation of the lactate score in ovarian cancer. (A) ROC curve of different models constructed by two algorithms LASSO and RF. (B) LASSO coefficient profiles of the 230 LRGs, which come from the pathways named lactate metabolic process, abnormal lactate dehydrogenase levels, increased CSF lactate, increased serum lactate, increased circulating lactate dehydrogenase concentration. (C) The risk model consists of 17 genes and their coefficient. (D) The KM plot of high, middle and low lactate score patients in training set. (E) The KM plot of high, middle and low lactate score patients in GSE14764 dataset. (F) The KM plot of high, middle and low lactate score patients in GSE140082 dataset. (G) ROC curve of the risk model in training set. (H) ROC curve of the risk model in GSE14764 dataset; (I) ROC curve of the risk model in GSE140082 dataset. ROC: receiver operating characteristic; LASSO: least absolute shrinkage and selection operator; RF: random forest. LRGs: lactate-related genes.