Comparison of neoadjuvant treatment and surgery first for resectable or borderline resectable pancreatic carcinoma: A systematic review and network meta-analysis of randomized controlled trials.

Background: Current treatment recommendations for resectable or borderline pancreatic carcinoma support upfront surgery and adjuvant therapy. However, neoadjuvant therapy (NT) seems to increase prognosis of pancreatic carcinoma and come to everyone's attention gradually. Randomized controlled trials offering comparison with the NT are lacking and optimal neoadjuvant treatment regimen still remains uncertain. This study aims to compare both treatment strategies for resectable or borderline resectable pancreatic cancer.

Methods: The PRISMA checklist was used as a guide to systematically review relevant peer-reviewed literature reporting primary data analysis. We searched PubMed, Medline, EMBASE, Cochrane Datebase and related reviews for randomized controlled trials comparing neoadjuvant therapy with surgery first for resectable or borderline resectable pancreatic carcinoma. We estimated relative hazard ratios (HRs) for median overall survival and ratios risks (RRs) for microscopically complete (R0) resection among different neoadjuvant regimens and major complications. We assessed the effects of neoadjuvant therapy on R0 resection rate and median overall survival with Bayesian analysis.

Results: Thirteen eligible articles were included. Eight studies performed comparison neoadjuvant therapy with surgery first, and R0 resection rate was recorded in seven studies. Compared with surgery first, neoadjuvant therapy did increase the R0 resection rate (RR = 1.53, I2 = 0%, P< 0.00001), there was a certain possibility that gemcitabine + cisplatin (Gem+Cis) + Radiotherapy was the most favorable in terms of the fact that there was no significant difference concerning the results from the individual studies. In direct comparison, four studies were included and estimated that Neoadjuvant therapy improved mOS compared with upfront surgery (HR 0.68, 95% CI 0.58-0.92; P = 0.012; I2 = 15%), after Bayesian analysis it seemed that regimen with Cisplatin/ Epirubicin then Gemcitabine/ Capecitabine (PEXG) was most likely the best with a relatively small sample size. The rate of major surgical complications was available for six studies and ranged from 11% to 56% with neoadjuvant therapy and 11% to 45% with surgery first. There was no significant difference between neoadjuvant therapy and surgery first, also with a high heterogeneity (RR = 0.96, 95%CI = 0.65-1.43; P = 0.85; I2 = 46%).

Conclusion: In conclusion neoadjuvant therapy might offer benefit over up-front surgery. Neoadjuvant therapy increased the R0 resection rate with gemcitabine + cisplatin + Radiotherapy that was the most favorable and improved mOS with Cisplatin/ Epirubicin then Gemcitabine/ Capecitabine (PEXG) that was most likely the best.