Predictive value of LDH kinetics in bevacizumab treatment and survival of patients with advanced NSCLC

Background: The combination of bevacizumab and chemotherapy is still one of the standard treatments for advanced non-small-cell lung cancer (NSCLC) patients in the new era of targeted therapy. Although a high level of baseline lactate dehydrogenase (LDH) was found to predict survival benefit from bevacizumab in patients with metastatic colorectal cancer, the predictive value of serum level of LDH in NSCLC patients treated with bevacizumab has not been investigated yet. Moreover, dynamic evaluation of serum level of LDH changes may be more informative and promising in predicting patients' prognosis. We thus sought to analyze LDH kinetics and evaluate its predictive role in the response and survival of advanced NSCLC patients treated with bevacizumab. Method: We retrospectively collected and analyzed a total of 161 advanced NSCLC patients who had undergone treatment with bevacizumab. Univariate and multivariate logistic regression analyses of serum level of LDH were used for response analyses, and Cox models for both overall survival (OS) and progression-free survival analyses (PFS). Longitudinal analysis of LDH was performed using a mixed-effect regression model. Results: On multivariate Cox models, increase of serum level of LDI-I after 4 cycles with bevacizumab (INC4) treatment was shown to be the independent risk factor for OS (hazard ratio =2.17, 95% Cl: 1.21-3.90, P=0.009), and the serum level of LDH after 2 cycles (LDH2) and the increase of LDH after 6 cycles with bevacizumab (INC6) treatment were the predictive factors for PFS (hazard ratio =2.33, 95% CI: 1.38-3.93, P=0.001; hazard ratio =1.96, 95% CI: 1.27-3.03, P=0.002, respectively). Patients with increase of serum level of LDH after 2 cycles of treatment with bevacizumab (INC2) (odds ratio =3.75, 95% CI: 1.83-7.68, P<0.001) were more likely to attain stable disease/progressive disease on multivariate logistic regression analyses, while patients with complete response (CR)/partial response (PR) experienced a reduction of serum level of LDH every 2 cycles (Coef =-0.076, std error =0.017, P<0.001) over time. Conclusion: Dynamic changes of LDH were superior to baseline LDH in predicting prognosis of NSCLC patients treated with bevacizumab. Serum level of LDI-I reducing over time was a potential biomarker for patients to achieve good clinical response (CR/PR) to bevacizumab.