Supplementary Material for: Serum Neuroendocrine Markers Predict Therapy Outcome of Patients with Metastatic Castration-Resistant Prostate Cancer: A Meta-Analysis
datasetposted on 05.12.2018 by Liu Y., Zhao S., Wang J., Zhu Z., Luo L., Li E., Tang F., Zhao Z.
Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
Purpose: To evaluate whether serum neuroendocrine markers could effectively predict treatment outcomes in patients with metastatic castration-resistant prostate cancer (mCRPC). Methods: The PubMed, Cochrane Library and Embase databases were sought to identify eligible studies concerning serum neuroendocrine markers and the prognosis of post-treatment mCRPC from inception to April 2018. The association between serum neuroendocrine markers, that is, chromogranin A (CgA) and neurone-specific enolase (NSE), levels and the prognosis of post-treatment mCRPC were summarized using a random-effects model and hazard ratio (HR) with 95% CI Sensitivity analyses were conducted to assess potential bias. Results: A total of 234 participants are included in this meta-analysis (mean age = 71.3 years) from 6 studies. The pooled results show that higher markers’ levels at baseline in patients were associated with unfavorable overall survival (OS; univariate analysis: HR 3.775, 95% CI 1.469–9.698, p = 0.006; multivariate analysis: HR 3.838, 95% CI 1.774–8.304, p = 0.001), and a similar situation was observed in progression-free survival (PFS; univariate analysis: HR 2.785, 95% CI 1.315–5.898, p = 0.007; multivariate analysis: HR 1.266, 95% CI 1.017–1.577, p = 0.035). Estimates of the total effects were generally consistent in the sensitivity analysis. Publication bias was observed when performing the univariate analysis of PFS, and we have the explanation accordingly. Conclusions: The results of this pooled analysis confirm serum neuroendocrine markers could be the effective predictor of treatment outcome in patients with mCRPC. In addition, a combination of CgA and NSE is more valuable to predict worse OS. Further randomized case-control trials are required to validate this relationship.