In accordance with the guidelines of Preferred Reporting Items for Systematic Reviews and Meta-Analyses and Meta-Analysis of Observational Studies in Epidemiology[11, 12], we performed this systematic review and meta-analysis which has been registered at the International Prospective Register of Systematic Reviews(number CRD42020207903)
- Search strategy and inclusion criteria
The search was formulated in PubMed, Embase, Cochrane Central Register of Controlled Trials, Web of Science, ClinicalTrials.gov, and China National Knowledge Infrastructure using the same subject headings and keywords. Search terms included (neuroendocrine tumor OR neuroendocrine neoplasm OR neuroendocrine carcinoma OR mixed neuroendocrine-non-neuroendocrine neoplasm OR mixed adenoneuroendocrine carcinoma OR carcinoid) AND (neoadjuvant OR preoperative) AND (gastrointestinal OR gastroenteropancreatic OR digestive OR gut OR pancreas).
- Data extraction
Two authors (Chen and Zhu) independently extracted information from eligible studies using standardized forms. Baseline clinical characteristics (author, year, country, and study design, included number, age, genders), endpoint-related variables (HR for OS and PFS, 3-Year OS,5-Year OS, 3-Year PFS,5-Year PFS), complications, and pre- and post-operative TNM grades were extracted. All discrepancies between the reviewers were resolved by consulting a third reviewer.
- Inclusion Criteria and Exclusion Criteria
The inclusion criteria for clinical articles were (1) being a definitive pathologic diagnosis of GEP-NETs, (2) patients without metastases or with only liver metastases having undergone surgery and at least one preoperative neoadjuvant therapy (including chemotherapy, radiation therapy, chemoradiotherapy and PRRT therapy), (3) at least including one of HR or survival rate, and (4) having full text of studies available.
The exclusion criteria were (1) diagnoses not confirmed by pathology, (2) only endoscopic treatment, (3) originating outside the digestive system and metastasizing to the digestive tract, (4) cancers with neuroendocrine differentiation, (5) intraoperative chemotherapy, and (6) fewer than ten patients included.
- Quality assessment
Two reviewers independently assessed the eligibility and validity of each study via the Cochrane tool for assessing the risk of bias in randomized controlled trials (RCTs) and the Newcastle-Ottawa scale in observational studies. Newcastle-Ottawa scale measured quality in the three parameters of selection, comparability, and exposure/outcome as well as allocated a maximum of 4, 2, and 3 points respectively. High-quality studies scored greater than 7, moderate-quality studies scored between 5 and 7, and low-quality studies scored under 5. We conducted research bias detection on the articles included. The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) method serves as the guideline in the systematic review and meta-analysis. Each study was independently assessed by 2 authors, and any disagreement was to be discussed with a third reviewer and resolved in consensus.
- Statistical Analysis
This study primarily focused on the evaluation of the overall survival time, meanwhile covering the analysis of complications, PFS, 5-year OS and 5-year PFS.
For articles that presented Kaplan-Meier curves without Hazard Ratio or related survival information, we used software Engauge Digitizer (version 12.0) and SPSS (version 21) to extract variables in the coordinates of the Kaplan-Meier curves. Cochrane I2 and χ2 statistics were applied to estimate statistical heterogeneity, P <0 .05 and I2 > 50% being highly heterogeneous. Random-effects models were conducted in the case of I² >50%; otherwise, fixed-effects models were chosen. Statistical significance was defined at the level of 5% (P = 0.05). The Inverse variance method was employed in analyzing subgroups. Odds ratios (OR) were calculated for the analysis of complications. Egger linear regression was conducted to define publication bias. All statistical analysis was performed in the software STATA（version 14.0）and RevMan (version 5.3).