Serum Inflammatory Biomarkers Contribute to the Prognosis Prediction in High-Grade Glioma

Abstract

Background: To evaluate the prognostic value of serum inflammatory biomarkers and develop a risk stratification model for high-grade glioma (HGG) patients based on clinical, laboratory, radiological, and pathological factors.

Materials and methods: A retrospective study of 199 patients with HGG was conducted. Patients were divided into a training cohort (n = 120) and a validation cohort (n = 79). The effects of potential associated factors on the overall survival (OS) time were investigated and the benefits of serum inflammatory biomarkers in improving predictive performance was assessed. Univariable and multivariable Cox regression analyses, the least absolute shrinkage and selection operator (LASSO) regression analysis, and support vector machines (SVM) were used to select variables for the final nomogram model.

Results: After multivariable Cox, LASSO, and SVM analysis, in addition to 3 other clinico-pathologic factors, platelet-to-lymphocyte ratio (PLR) >144.4 (hazard ratio [HR], 2.05; 95% confidence interval [CI], 1.25-3.38; P = 0.005) were left for constructing the predictive model. The model with PLR exhibited a better predictive performance than that without them in both cohorts. The nomogram based on the model showed an excellent ability of discrimination in the entire cohort (C-index, 0.747; 95%CI, 0.706-0.788). The calibration curves showed good consistency between the predicted and observed survival probability.

Conclusion: Our study confirmed the prognostic value of serum inflammatory biomarkers including PLR and established a comprehensive scoring system for the OS prediction in HGG patients.

Keywords: LASSO; SVM; glioma; nomogram; prognosis; serum inflammatory biomarker.

Figure 1

Least absolute shrinkage and selection operator (LASSO) regression analysis and support vector machines (SVM) was applied to further identify prognostic factors in the training cohort. (A) LASSO regression analysis showed that the 7 variables were all left in the LASSO model based on the partial likelihood deviance vs log (λ). The right dotted vertical line was drawn at the optimal value of λ by one standard error of the minimum criteria. (B) SVM showed that the model consists of the top 4 variables almost reached the lowest value of Root Mean Square Error (RMSE) based on 10-fold cross-validation. The blue curve represents the different value of RMSE based on models consists of different variables. Lower values of RMSE represents better consistency between prediction and actuality.

Figure 2

Decision curve analyses (DCA) of ModelA and ModelB at 1, 2, and 3 years after surgery in the training cohort (A) and 1, 2, and 3 years after surgery in the validation cohort (B). The y-axis represents the net benefit and the x-axis represents the corresponding risk threshold. The blue line represents that all patients die during the follow-up. The purple line represents that no patients die during the follow-up. In the most points of risk threshold, ModelB (green line) showed more benefits in predicting survival status than ModelA (red line).

Figure 3

Integrated Discrimination Improvements (IDI) and Net Reclassification Index (NRI) of ModelB comparing to ModelA at (A) 1 year, (B) 2 years, and (C) 3 years after surgery in the training cohort and (D) 1 year (E), 2 years, and (F) 3 years after surgery in the validation cohort. the red areas were greater than blue areas and the median value of NRI and IDI were all greater than zero, indicating that the predictive ability of ModelB may be better than ModelA.

Figure 4

Time-dependent receiver operating characteristic (ROC) curve of ModelA and ModelB in the training (A) and validation cohort (B). The y-axis represents the area under curve (AUC) and the x-axis represents the follow-up time. In the most points of follow-up time, the AUC value of ModelB (green line) was higher than ModelA (red line).

Figure 5

The nomogram for predicting 1-, 2-, and 3-year survival rates of high-grade glioma patients. For each variable, draw a straight line up to the Points axis to calculate the point. After summing the points and locating it on the Total Points axis, draw a straight line down to the 1-year survival, 2-year survival, and 3-year survival axis to determine the probability of surviving for 1, 2, and 3 years. PLR, platelet-to-lymphocyte ratio; IDH, isocitrate dehydrogenase.

Figure 6

Decision curve analyses (DCA), time-dependent receiver operating characteristic (ROC) curve and calibration curves of the nomogram. (A) DCA of the nomogram at 1, 2, and 3 years after surgery. The y-axis represents the net benefit and the x-axis represents the corresponding risk threshold. The blue line represents that all patients die during the follow-up. The purple line represents that no patients die during the follow-up. The red line represents the net benefits of nomogram at different risk threshold. (B) The predictive value of the nomogram at different points of follow-up after surgery. (C–E) The calibration curves of the nomogram to predict (C) 1-year, (D) 2-year, and (E) 3-year survival rates. The y-axis represents actual survival and the x-axis represents the predicted survival probability based on nomogram. The gray oblique line represents the ideal prediction and the red line represents the performance of the nomogram. Close fit to the grey oblique line indicates the consistency between the predicted and observed survival probability.