Our study demonstrated an OS advantage of intraperitoneal cisplatin/paclitaxel chemotherapy without bevacizumab in women with advanced ovarian cancer compared with the combination of intravenous carboplatin/paclitaxel chemotherapy and bevacizumab (log-rank test, p = 0.02, Table 1, Fig. 1). After adjusting with the variable "suboptimal debulking", there still existed a trend of better OS in the group of women who received intraperitoneal chemotherapy without bevacizumab (adjusted HR = 0.32, p = 0.065, Table 2).
A phase II trial found that women with newly diagnosed ovarian cancer who received intraperitoneal chemotherapy without bevacizumab seems to have a superior OS, compared with those who received intraperitoneal chemotherapy plus bevacizumab (median OS: 79.7 versus 68 months) [18]. In addition, bevacizumab was reported to have no survival benefit in the GOG 218 and ICON 7 studies [11, 12]. In the most recent Japanese phase II/III iPOCC trial, intraperitoneal carboplatin in combination with dose-dense paclitaxel without bevacizumab improved progression-free survival versus an intravenous regimen in women with advanced ovarian cancer, regardless of the residual tumor size after initial surgery [19]. Taken together with our study, intraperitoneal chemotherapy without bevacizumab seems to be a viable option for treating advanced ovarian cancer in the era of bevacizumab.
Intraperitoneal chemotherapy provides pharmacologic advantage by directly exposing tumor to a greater concentration of chemotherapeutic drugs [20], whereas bevacizumab causes a reduction in tumor vascularization and angiogenesis [9–10]. The reason why our IP group had a better survival benefit is unknown (Table 2). However, antiangiogenic therapy could prune tumor vessels excessively, rather than normalizing them, and thus decreases the delivery of chemotherapeutic agents [21]. Tumor angiogenesis can rapidly rebound when discontinuing vascular endothelial growth factor inhibition [22, 23]. The above phenomenon might be a feasible explanation to justify our finding of a superior OS in the IP group, compared with the IVB group.
The possibility of intraperitoneal bevacizumab therapy has been explored by several investigators, but mainly limited to the palliative treatment of malignant ascites [24–26]. Intraperitoneal administration of bevacizumab was reported to be safe and effective, but there is no standard recommended dose or frequency for intraperitoneal bevacizumab. To the best of our knowledge, no comparison between intraperitoneal administration and intravenous injection of bevacizumab has ever been made in clinical setting.
Intraperitoneal chemotherapy with cisplatin 75 mg/m2 was listed as an alternative to cisplatin 100 mg/m2 in the NCCN guideline. However, our study found that the use of cisplatin 100 mg/m2 was an independent predictor for a better OS (adjusted HR = 0.14, 95% CI = 0.02 to 0.87, p = 0.03, Table 3), compared with other regimens (including cisplatin 75 mg/m2 and carboplatin). The GOG 172 study used an intraperitoneal cisplatin dose of 100 mg/m2 and showed a superiority of intraperitoneal chemotherapy [6]. Nonetheless, the cisplatin dose was modified from 100 mg/m2 to 75 mg/m2 in the GOG 252 study, and there was no survival benefit between the intraperitoneal and intravenous chemotherapy groups [7]. In addition, Markman et al. found that intraperitoneal cisplatin resulted in a higher overall response rate, compared with intraperitoneal carboplatin [27]. Similarly, Eoh et al. reported that the intraperitoneal 100 mg/m2 group had a better OS, compared with intraperitoneal carboplatin [28]. Besides, Bouchard-Fortier et al. reported 4-year OS of 78% and 58% for the intraperitoneal cisplatin and intraperitoneal carboplatin groups, respectively (p = 0.107) [29]. Thus, taken together with our study, 100 mg/m2 cisplatin should be recommended as the agent of choice for intraperitoneal chemotherapy, instead of 75 mg/m2 cisplatin or carboplatin.
The number of chemotherapy cycles had a pronounced effect on the OS of women with advanced disease who received either intraperitoneal or intravenous chemotherapy with bevacizumab. Both Ting et al. and Yen et al. reported that at least five cycles were needed to effectively prolong the survival rate in women treated with intraperitoneal chemotherapy [14, 30]. Similarly, from our ROC analysis, the number of intraperitoneal or intravenous chemotherapy cycles ≥ 6 was the optimum cutoff value to predict OS (Fig. 2).
We acknowledge that the clinical evidence of this study is limited due to its retrospective, limited sample size and nonrandomized nature. Besides, a higher proportion of women in the IP group had optimally debulked surgery. However, we used a multivariable analysis and included the variable "suboptimal debulking" to adjust the effect of the IP group (Table 2). In addition, bevacizumab 7.5 mg/kg was used in this study, instead of 15 mg/kg.