Dear Editor

We read with great interests on the recent paper by Han L et al. “Prognostic value of circulating tumor cells in patients with pancreatic cancer: a meta-analysis” published online in Tumor Biology [1]. The investigators (Han L et al.) performed a meta-analysis of nine cohort studies to assess the prognostic value of circulating tumor cells (CTC) in patients with pancreatic cancer. They concluded that CTC-positive pancreatic cancer patients may have worser progression free survival (PFS) and overall survival (OS) than CTC-negative patients. Detection of CTC in peripheral blood may be a promising biomarker for the detection and prognosis of pancreatic cancer. It is the first comprehensive meta-analysis of all eligible studies concerning the prognostic role of CTC in patients with pancreatic cancer. Nevertheless, we have several queries which we would like to communicate with the investigators.

  1. 1.

    We think that there are some deficiencies in the literature search. The investigators just provided us some keywords and MeSH terms, while did not describe the search strategy report for databases in details, which played important roles in the meta-analysis. Manual searches were also not expressed clearly. The lack of a manual search protocol may be regarded as a drawback of this meta-analysis. If possible, we suggest that the investigators provide us a complete search protocol to strengthen the credibility of the meta-analysis.

  2. 2.

    The investigators conducted the Newcastle-Ottawa Scale (NOS) criteria to assess the methodological quality of the included studies [210]. They included the study by Uchikura K et al. [10], which NOS scale was just 5. Based on the NOS criteria, we are wondering whether this study could be included or not.

  3. 3.

    The investigators just extracted data to calculate hazard ratios (HR) for OS and PFS comparing CTC-positive to CTC-negative patients. However, they did not extract data to calculate odds ratios (OR) for objective response ratio (ORR) comparing CTC-positive to CTC-negative patients. To make the meta-analysis better, these correlations should be added in the meta-analysis.

  4. 4.

    The investigators clarified that “The random effect model (the DerSimonian Laird method) was conducted when there was a significant Q test with P < 0.10 or I 2 > 30 %.” Frequently, the heterogeneity was considered by P < 0.10 or I 2 > 50 %. We hope that the investigators could make some explanations for the discrepancy.

  5. 5.

    The investigators showed forest plots for the difference in OS and PFS between CTC-positive and CTC-negative pancreatic cancer patients just in one figure. If possible, we suggest that the forest plots for OS and PFS should be showed in two single figures.

  6. 6.

    The investigators did not strictly follow the meta-analysis of observational studies in epidemiology (MOOSE). As a minor suggestion, adding MOOSE checklist might make this meta-analysis better.

Thanks go to the investigators for their contributions to supplying us with an assessment of prognostic value of CTC in patients with pancreatic cancer. However, further detailed studies are still required to confirm these findings.