Radical surgery versus standard surgery for primary cytoreduction of bulky stage IIIC and IV ovarian cancer: an observational study

The aim of this study was to evaluate the survival benefit of radical surgery with additional extensive upper abdominal procedures (EUAS) for the treatment of stage IIIC and IV ovarian cancer with bulky upper abdominal disease (UAD). An observational study was conducted between 2009 and 2012 involving two different surgical teams. Team A was composed of the “believers” in EUAS and Team B the “non-believers” in EUAS. Patients were divided into a radical surgery group (EUAS group) or a standard surgery group (non-EUAS group) according to whether or not they had received EUAS. All patients underwent primary cytoreductive surgery with the goal of optimal debulking (≤1 cm); this was reviewed in the pelvis, middle abdomen, and upper abdomen. The baseline for the two groups was optimal cytoreduction in both the pelvis and middle abdomen. Progression-free survival (PFS) was evaluated. Radical surgery was performed in 70.7 % (82/116) and 12.7 % (30/237) of the patients by Teams A and B, respectively. The study groups had similar clinicopathologic characteristics. The median PFS and OS were significantly improved in the radical surgery group, compared with standard surgery groups (PFS: 19.5 vs. 13.3 months, HR: 0.61; 95 % CI: 0.46–0.80, P < 0.001; OS: not reached vs. 39.3 months, HR: 0.47; 95 % CI: 0.30–0.72, P < 0.001). Positive predictors of complete cytoreduction were treatment with neoadjuvant chemotherapy, improved American Society of Anesthesiologists performance status, and the absence of bowel mesenteric carcinomatosis. Radical surgery lengthens the PFS and overall survival times of ovarian cancer patients with bulky UAD. However, a well-designed randomized trial is needed to confirm the present results.


Background
Epithelial ovarian cancer (EOC) is the most lethal of all gynecological cancers [1]. The goal of primary cytoreduction for advanced EOC is advocated to be no visible residual disease,which has been confirmed in several studies, but only less than 30 % of women with bulky upper abdominal disease (UAD) can achieve complete cytoreduction [2,3]. Thus, it still remains controversial as to whether or not patients with bulky UAD can benefit from upper abdominal procedures (EUAS). It has been suggested that upper abdominal procedures should only be performed when complete or optimal cytoreduction is attainable [4][5][6].
In China, only a few surgeons are willing to undertake EUAS because most lack the relevant surgical skills, or there is tension between patients and physicians regarding the invasiveness of the treatment. Consequently, to date, there have been no Chinese studies in this area [7]. Most of the surgeons tend to accept neoadjuvant chemotherapy followed by surgery as the standard approach, which is in line with the result of EORTC 55971 study reported in 2010 [8].
Herein, we define radical surgery as the EUAS procedures complementing an optimal cytoreduction within the middle abdomen and the pelvis. These procedures include diaphragmatic peritonectomy, hepatic resection, splenectomy, distal pancreatectomy, cholecystectomy, and the resection of tumor on the surface of upper abdominal organs. Standard surgery is defined as the optimal surgical outcome achieved in both the middle abdomen and the pelvis (including small and/or large bowel resections), and the subsequent attempt to resect tumor nodes measuring ≥1 cm in the upper abdomen. An exploratory study was conducted to compare the survival after radical surgery with standard surgery in patients with bulky stage IIIC and IV ovarian cancer.

Patients
A single institute observational study was conducted between 2009 and 2012. Patients who were diagnosed with stage IIIC or IV epithelial ovarian cancer with bulky UAD were identified and the data were collected retrospectively, which was approved by Fudan University Shanghai Cancer Center Institutional Review Board (SCCIRB-090371-2). The residual disease was evaluated in the pelvis, middle abdomen, and upper abdomen, respectively. In the standard surgery (non-EUAS surgery) group, patients with residual disease measuring ˃1 cm in the pelvis and middle abdomen were excluded.
Medical records were abstracted for the following data: age at primary cytoreduction; International Federation of Gynecology and Obstetrics (FIGO) stage; histology; tumor grade; Eastern Cooperative Oncology Group (ECOG) performance status; American Society of Anesthesiologists (ASA) score; preoperative serum cancer antigen (CA125); ascites and extent of disease (categorized as solitary, localized, and carcinomatosis defined as diffused involving of the peritoneal surfaces) at primary surgery; cytoreductive procedures; residual disease after primary cytoreduction; surgeons involved; type of frontline chemotherapy; estimated blood loss; intraoperative blood transfusion; operative time; intensive care unit (ICU) stay; length of hospitalization; postoperative 30day morbidity and mortality; progression-free survival (PFS), follow-up time; and survival time. The computed tomography (CT) or magnetic resonance imaging (MRI) scan was reviewed by the radiologist (CY) in all possible patients to re-evaluate the upper abdominal disease.

Definitions
There were two different surgical teams: Team A was the "believers" and "deeds over words" regarding EUAS procedures, who attempted to resect any macroscopic disease; Team B was "non-believers" in EUAS.
PFS was defined as the time from initial treatment to the diagnosis of the first recurrence or last follow-up, whichever came first. Overall survival (OS) was defined as the time from initial treatment to death or last follow-up. Recurrence was diagnosed by one or more of the following: physical examination; elevated CA-125 levels as defined by the Gynecologic Oncology Intergroup [9]; and radiological imaging.
The abdominal tumor site (pelvis, middle abdominal, and upper abdominal disease) at primary cytoreduction was defined as previously described [10]. Optimal cytoreduction was defined as residual disease measuring ≤1 cm, but the cut-off points of 0 cm and 0.5 cm were also used to evaluate the impact on survival.

Statistical analysis
Statistical analysis was performed using the SPSS software package for Windows (version 16.0). The Chisquare or Mann-Whitney U tests were used to identify differences in the baseline level between the two groups. Median survival was evaluated using the Kaplan-Meier method and differences were determined using the logrank test. The Cox proportional hazards regression model was used to identify prognostic factors. Logistic regression analysis was conducted to detect the predictors of complete cytoreduction. A P-value of <0.05 was considered as being statistically significant.

Baseline and patient characteristics
Three hundred and fifty-three patients were included in this observational study. Of these patients, 112 received radical surgery including EUAS procedures, and 241 received standard surgery. In radical surgery group, Team A did 82 cases (73.2 %) radical surgery; and Team B did 30 (26.8 %) radical surgery. Altogether Team A did surgery in 116 patients, and Team B did surgery in 237 patients. Patient baseline characteristics as well as the imaging findings regarding the disease within the upper abdomen were similar between the two groups (Tables 1  and 2). In the radical surgery group, eight (7.1 %) patients were upstaged for pleural metastasis and 21 (18.8 %) were diagnosed with stage IV ovarian cancer in the EUAS group as compared with 26 (10.8 %) patients in the non-EUAS group. In the radical surgery group, the optimal cytoreduction achieved in the pelvis and middle abdomen was 98.2 % and 97.3 %, respectively; all patients in the standard surgery group underwent optimal cytoreduction. However, more patients in the radical surgery group had microscopic residual disease, both in the pelvis (82.1 % versus 53.5 %) and the middle abdomen (53.6 % versus 28.6 %) when compared with those patients in the control group.

Surgical outcomes
There were significant differences between radical surgery involving EUAS and standard surgery in terms of estimated blood loss, intraoperative blood transfusion, operative time, ICU stay, and length of hospitalization (Table 3). In the EUAS group, optimal cytoreduction was performed in 107 patients (95.5 %), and in 76 patients (67.9 %) complete cytoreduction was achieved in the upper abdomen. However, no patients achieved complete cytoreduction in the control arm, and only 43.6 % received optimal surgery.
The extensive upper abdominal procedures performed in the radical surgery group included diaphragm peritonectomy, full-thickness diaphragm resection, resection of the lesser omentum, splenectomy, liver resection, distal       Table 6). The positive predictors of complete cytoreduction were as follows: undergoing neoadjuvant chemotherapy; better ASA performance status; and the absence of bowel mesenteric carcinomatosis (P = 0.006, P = 0.014 and P = 0.026, respectively; Table 7).
In the standard surgery group, one patient died at 30 days after surgery because of a bowel obstruction, abdominal hemorrhage, and disease progression. However, there was no surgery-related death in the radical surgery group. The rate of Memorial Sloan-Kettering Cancer Centre (MSKCC) Grade III/IV complications in the two groups were 8.0 % (9/112) and 2.9 % (7/241), respectively. Thoracentesisor chest tube placement was more  Logistic regression analysis b Variable "surgical team" was the most significant determinant of surgical outcomes, but it was not included into this model c It is a variable found in surgery, but interestingly, it is a predictor of complete cytoreduction not the predictor of small residual disease of 0.5 cm, so it was included in the Logistic regression model  Table 9 Morbidity and mortality by surgeons  Table 8). There was no significant difference in morbidity and mortality between Teams A and B (Table 9).

Discussion
The role of radical surgery with an additional upper abdominal cytoreduction It is well established that EUAS could increase the proportion of patients achieving optimal cytoreduction as well as complete cytoreduction [4,5,[11][12][13]. However, it is still unclear as to which patients would benefit from EUAS. Rodriguez and colleagues reviewed 2655 patients enrolled in the Gynecologic Oncology Group 182 trial from 2001 to 2004 who had achieved optimal cytoreduction (<1 cm) [4]. A total of 482 (18.1 %) patients received upper abdominal procedures. These authors reported that both the PFS (18.2 versus 14.8 months; P < 0.01) and OS (49.8 versus 43.7 months; P = 0.01) were higher in patients who did not require an upper abdominal procedure relative to patients who required upper abdominal procedures; they suggested that requiring upper abdominal surgery was an indicator of more extensive disease, which would have a negative impact on survival. When subgroup analysis was performed in patients with a high disease burden in the upper abdomen (n = 1636), it did not reveal an improved PFS (P = 0.43) in the 482 (29 %) patients who had undergone upper abdominal surgery relative to patients who had not, while the OS exhibited a modest improvement of 3.8 months. However, in the subset of patients who had received upper abdominal surgery, when considering postoperative residual disease, both the PFS (20.2 versus 13.7 months; P < 0.01) and the OS (54.6 versus 40.4 months; P < 0.01) were significantly prolonged in patients with complete cytoreduction (n = 141) relative to patients with minimal residual disease measuring <1 cm (n = 341). The authors concluded that in patients who required upper abdominal surgery, only the increased rate of complete cytoreduction could translate into prolonged OS [4]. However, regarding the MSKCC studies, their data demonstrated that EUAS significantly improved PFS and OS [13,14]. In addition, the most recent study from this center has shown that patients with macroscopic residual disease ≤1 cm who required EUAS did not have a worse OS than those who did not require EUAS (45 months versus 52 months; P = 0.56) [5].
Unlike previous studies [4][5][6][11][12][13][14], the current study was unique with respect to patient enrollment. First, it was an observational study and we only included patients from a single institution with bulky upper abdominal disease, who had achieved optimal primary resection in both the pelvis and middle abdomen in a recent period. Second, although patient selection bias still existed in the present study, the difference in the inherent tumor biology between the radical surgery group and the standard surgery group was balanced to a certain degree. Consequently, we could specifically analyze the role of EUAS and obtain results as to whether or not patients could obtain a survival benefit from EUAS. In the present study, a 6.2-month prolongation in the median PFS was observed in patients who had undergone additional EUAS. Although the data were immature, a significant difference in OS was also observed. Because of the EUAS procedures, eight patients were upstaged for pleural metastasis. Therefore, thoracic exploration in EUAS could obtain an exact FIGO staging in patients with a heavy diaphragm tumor burden, especially those with an untapped pleural effusion.
As diaphragmatic peritonectomy was performed most frequently regarding EUAS, pleural effusion was the most common morbidity [15]. However, symptomatic pleural effusion can be well managed through thoracentesis or chest tube placement. The other morbidities such as pulmonary embolism, deep venous thrombosis, bowel obstruction, abdominal infections, and wound infection were not significantly increased in the radical surgery group. In addition, there was no surgery-related death in the radical surgery group. Although EUAS procedures were complicated, radical surgery did not increase the morbidity and mortality when compared with standard surgery.
"Maximal" cytoreduction and residual disease in bulky ovarian cancer Upper abdominal surgical procedures have proved to be more practical than most surgeons had anticipated. In our cohort, 89.3 % of the patients in the radical surgery  group achieved residual disease measuring <0.5 cm and showed a survival advantage (Table 3). Additionally, the surgical outcome was accurately predicted by logistic regression analysis of CT/MRI scan data regarding the upper abdomen (Table 11). Moreover, residual disease in the upper abdomen was found to be one of the predictors of PFS (Table 10).
Although surgeons have made more efforts regarding how to best describe optimal cytoreduction in ovarian cancer, we believe that it is still not clear as to what constitutes "maximal" cytoreduction, or how to improve the surgical approach among the different teams, institutions, and countries. Complete cytoreduction is not equally applicable between neoadjuvant chemotherapy followed by interval cytoreductive surgery and primary cytoreductive surgery. Neoadjuvant chemotherapy can eliminate carcinomatosis in the peritoneum if it is chemosensitive, but its effects differ in patients undergoing "maximal" cytoreduction and peritonectomy. In the present study, Team B resected patients with tumor masses measuring >1 cm but left the peritoneal carcinomatosis in the upper abdomen; this was considered suboptimal by Team A. Thus, different interpretations of the term "maximal" can result in different surgical outcomes.
In the radical surgery group, we considered the residual disease as ≤0.5 cm when peritoneal carcinomatosis was cytoreduced to microscopic residual by electronic knife, while not by en-block peritonectomy. We did not find a survival difference between lesions measuring 0.1-0.5 cm and 0 cm when residual disease was evaluated in either the upper abdomen alone or the whole abdomen (Figs. 3 and 4a). The common pattern of tumor involvement in the diaphragmatic and other upper regions of the abdominal peritoneum was carcinomatosis (339/353; 96 %); consequently, resection of all tumor nodes to a microscopic residual size in these patients was usually impossible. In our series, the cut-off point of 0.5 cm for residual disease  is recommended for patients with bulky upper abdominal disease. This is because a biological complete cytoreduction is always rare when tumor cells have spread to the whole coelom (Figs. 4a, b; 5 and 6). However, a 41.1 % complete cytoreduction was achieved in the radical surgery group, so we should always keep in mind that complete gross cytoreductions are generally accepted, particularly in patients with less bulky UAD or less aggressive bowel carcinomatosis [3]. A longer operation time (≥4 h) did not improve the PFS, but shortened the OS, although not significantly (data not shown). Team A did 82 cases (73.2 %) radical surgery; and Team B only did 30 (26.8 %) radical surgery. There were 2 cases of relaparotomy for hemorrhage, 2 blood transfusion for hemorrhage, 1 intestinal fistula and 1 death in Team B, with no case in Team A. Therefore, EUAS performed by under-trained gynecologic surgeons is not recommended because it can result in more morbidity and mortality ( Table 9).
Because of non-randomization, the weakness of our study was potential selection bias. To minimize the bias of the study, the basic characters of the two groups  Fig. 6 Overall survival by residual disease in overall after primary cytoreductive surgery. a: OS by residual disease in overall after primary cytoreductive surgery; b: OS by residual disease in overall with a comparison of cut-off point R0.5 cm; c: OS by residual disease in overall with a comparison of cut-off point R0 cm were analyzed in Table 1, and no significant difference was found between two groups in the median age, primary tumor, histology, tumor grade, ECOG performance status, ASA status, CA125 level, Neoadjuvant chemotherapy (NAC), ascites, and bowel mesenteric carcinomatosis (p > 0.05). More patients with stage IV disease were in radical surgery group, as stage IV disease required more radical surgery during the operation (18.8 % vs. 10.8 %, p = 0.045). It is still not clear whether or not patients with stage IV disease benefit from radical surgery (Fig. 4d). However, the results of the current study provide evidence for designing a randomized clinical trial.

Conclusions
Extensive upper abdominal surgery lengthens the PFS and OS of ovarian cancer patients with bulky upper abdominal disease. Although these findings are based on short-term follow-up data, long-term follow-up is in progress. A welldesigned randomized trial is needed to confirm the present results.