Aggressive surgery could overcome the extent of initial peritoneal dissemination for advanced ovarian, fallopian tube, and peritoneal carcinoma

We examined whether the extent of initial peritoneal dissemination affected the prognosis of patients with advanced ovarian, fallopian tube, and peritoneal carcinoma when initially disseminated lesions > 1 cm in diameter were removed, regardless of the timing of aggressive cytoreductive surgery. The extent of peritoneal dissemination was assessed by the peritoneal cancer index (PCI) at initial laparotomy in 186 consecutive patients with stage IIIC/IV. Sixty patients underwent primary debulking surgery and 109 patients underwent neoadjuvant chemotherapy followed by interval debulking surgery. Seventeen patients could not undergo debulking surgery because of disease progression during neoadjuvant chemotherapy. The median initial PCI were 17. Upper abdominal surgery and bowel resection were performed in 149 (80%) and 171 patients (92%), respectively. Residual disease ≤ 1 cm after surgery was achieved in 164 patients (89%). The initial PCI was not significantly associated with progression-free survival (PFS; p = 0.13) and overall survival (OS; p = 0.09). No residual disease and a high-complexity surgery significantly prolonged PFS (p < 0.01 and p = 0.02, respectively) and OS (p < 0.01 and p ≤ 0.01, respectively). The extent of initial peritoneal dissemination did not affect the prognosis when initially disseminated lesions > 1 cm were resected.

tumor becomes invisible after NACT 20 . The previous study observed that the median progression-free survival (PFS) was longer in patients who underwent aggressive surgery with resection of lesions measuring > 1 cm before NACT than in those who underwent resection of only visible lesions during IDS 20 . Therefore, we concluded that initially disseminated tumors that could not be resected during IDS contributed to poor prognosis in patients with a high PCI score.
We hypothesized that aggressive surgery with resection of the initial > 1 cm dissemination would overcome the high PCI score by selecting PDS or NACT followed by IDS depending on whether cytoreduction to no residual disease is achievable at initial laparotomy before starting treatment. In this study, we examined whether PCI affected the prognosis of patients with International Federation of Gynecology and Obstetrics (FIGO) stage IIIC/ IV ovarian, fallopian tube, and peritoneal carcinoma when initially disseminated lesions > 1 cm were resected, regardless of the timing of cytoreductive surgery.

Results
Patients' characteristics and correlation with the initial PCI ( Table 1). The median initial PCI was 17 (minimum 2, maximum 31, first to third quartile range 10 to 22). The initial PCI was significantly associated with primary site, histology, timing of cytoreductive surgery, residual disease, and surgical complexity score. The initial PCI was not associated with performance status and FIGO stage.
Sixty patients (32%) underwent PDS. One hundred and nine patients underwent IDS after NACT. Seventeen patients did not undergo IDS because of disease progression during NACT (no debulking surgery). The median initial PCI of the patients who underwent PDS, IDS, and no debulking surgery were 8, 19, and 22, respectively. Surgical procedure and outcome. Of the 186 patients, 149 patients (80%) underwent upper abdominal procedure, including right diaphragm resection (n = 145) and splenectomy with or without distal pancreatectomy (n = 98), and 171 patients (92%) underwent bowel resection, including rectosigmoid colon (n = 163), large bowel other than the rectosigmoid colon (n = 90), and small bowel (n = 20). The surgical complexity score had a median value of 12 and was significantly higher in the patients who underwent IDS than in those who underwent PDS (13 vs 10, p < 0.001).

Survival analysis.
After a median follow-up duration of 40 months (range, 0.8-127 months), the median PFS for the entire cohort was 28 months, and the median overall survival (OS) was not reached.
In the Kaplan-Meier analysis (Fig. 2), PFS and OS did not significantly differ among the initial PCI groups, which were classified into three groups based on the PCI scores: 1-10 (n = 49), 11-20 (n = 81), 21-39 (n = 56) (p = 0.18 and p = 0.39, respectively). Kaplan-Meier curves were drawn for subgroup analyses of patients who underwent PDS, for those who underwent IDS, and for those who received chemotherapy without IDS (no debulking surgery). We observed that the PCI was not associated with PFS and OS in the subgroups of patients who underwent PDS and IDS (Fig. 3a,b,d,e). In contrast, the PCI was associated with PFS and OS in the subgroup of patients who received chemotherapy without IDS (Fig. 3c,f). The PDS, IDS, and no debulking surgery groups had a median PFS of 33, 28, and 6 months, respectively, and a median OS of not reached, 72, and 8 months, respectively.
Kaplan-Meier curves were drawn for subgroup analyses of patients without and with residual disease (those with lesions measuring 0.1-1 cm and also those with lesions measuring > 1 cm). We observed that the PCI was not associated with PFS and OS in patients without residual disease (Fig. 4a-d). In contrast, the PCI was associated with OS in patients with residual disease (Fig. 4e,f).
Multivariate Cox regression analysis (

Discussion
This study showed that PCI did not affect the prognosis of patients with FIGO stage IIIC/IV ovarian, fallopian tube, and peritoneal carcinoma when initially disseminated lesions > 1 cm in diameter were resected, regardless of the timing of cytoreductive surgery. No residual disease after surgery and high-surgical complexity score affected the prognosis. In other words, aggressive surgery could overcome the extent of peritoneal dissemination when  www.nature.com/scientificreports/ patients with a high PCI score undergo high-complexity surgery by selecting PDS or IDS, depending on whether cytoreduction to no residual disease is achievable at initial laparotomy before starting treatment. www.nature.com/scientificreports/ In this study, the timing of cytoreductive surgery was determined at initial laparotomy on the based on the feasibility of achieving no residual disease. As a result, the number of the patients who did not underwent PDS was more than twice as large as the number of the patients underwent PDS. Contrary to our study, the timing of cytoreductive surgery was usually primary at many institutions where aggressive cytoreductive surgery was performed [16][17][18][19] . However, studies have shown that the PCI score was negatively associated with the completeness of cytoreduction and residual disease after PDS 10 . Moreover, the rate of complications with PDS was higher than that with IDS 18,24 . Conversely, the use of NACT/IDS enabled the surgeons to accomplish a safe and highcomplexity surgery in patients with high extent of peritoneal dissemination because NACT reduces the massive ascites and the tumor volume, improves the general condition, and decreases the difficulty of complex surgery. In this study, high-complexity surgery was performed on 71% of the entire cohort; this led to high rate (80%) of complete cytoreduction rate. Our results showed that not only PDS but also selective use of NACT/IDS allowed many patients to receive complete cytoreduction safely, regardless of the initial PCI.
In this study, the surgical complexity score of the patients who underwent IDS was higher than those of the patients who underwent PDS because initial PCI of the patients who underwent IDS was higher than those of the patients who underwent PDS. On the contrary, several studies showed that the surgical complexity and the rate of extra-gynecologic surgery were lower with IDS than with PDS 18,25,26 . The difference between this study and others lies in the surgical policy for IDS. In the present study, all disseminated tumors > 1 cm in diameter identified at initial laparotomy was removed even if the initially disseminated tumor was no longer visible at the time of IDS after NACT. In other words, at IDS, we performed the surgical procedures that would be performed to achieve residual tumor < 1 cm if PDS were to be performed. A drawback of the conventional IDS is that the tumors > 1 cm diameter identified at initial laparotomy would not be resected during IDS if they become invisible due to a good response to NACT. This may lead to a risk of early recurrence or relapse, similar to residual disease > 1 cm diameter after PDS. Our previous study showed that microscopic disease remains present especially in the rectosigmoid colon, transverse mesentery, greater omentum, right diaphragm, paracolic gutters, and vesicouterine pouch even if the tumors become invisible during IDS 20 . Similar to ours, Lim et al. reported that detached scars had residual cancer cells that assumedly included cancer stem cells 27,28 . The surgical removal of all the tumors > 1 cm identified at initial laparotomy could overcome tumor biology and tumor burden.
Unlike gastrointestinal carcinoma, another reason that PCI did not affect prognosis in this study is the high response rate to chemotherapy for ovarian cancer. In this study, 87% patients did not have disease progression during NACT and could undergo IDS. In other reports, the response rates to chemotherapy in FIGO stage III/IV ovarian carcinoma is higher in colorectal carcinoma (73-90 vs 15-62%) 18,[29][30][31][32] . For colorectal cancer, patients with high score of PCI were reported to have poor prognosis 29,33-36 because PCI directly affect the rate of no residual disease after surgery, frequency of postoperative complications, and prognosis 29,33-36 . Similar results have been reported for gastric and appendix cancers. www.nature.com/scientificreports/ This study had some limitations such as the single-institution analysis, the small number of patients, and a higher percentage of IDS cases compared with that of PDS cases. The selection criteria for performing PDS or IDS varies among institutions. Notably, 42 of 56 patients with a high PCI score (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39) were included in the NACT/ IDS group; therefore, it is reasonable to conclude that in addition to aggressive surgery, chemotherapy may have affected the prognosis. In this study, 73% of patients with a high PCI score who could not undergo PDS underwent high-complexity surgery following the administration of NACT. Therefore, optimal patient selection for PDS or IDS can enable high-complexity surgery with favorable prognosis even in patients with a high PCI score.
In conclusion, the extent of initial peritoneal dissemination did not affect the prognosis for patients with FIGO stage IIIC/IV ovarian, fallopian, and peritoneal carcinoma when initially disseminated lesions > 1 cm was resected, regardless of the timing of cytoreductive surgery. No residual disease after surgery and a high-surgical complexity led to favorable prognosis. Aggressive cytoreductive surgery with selective use of IDS could overcome the extent of peritoneal dissemination.

Methods
Patient selection. This study was approved by the Institutional Review Board of Chiba University Graduate School of Medicine. All methods were carried out in accordance with relevant guidelines and regulations. Patients with FIGO 2014 stage IIIC/IV ovarian, fallopian tube, and peritoneal carcinoma who were consecutively treated at Chiba University Hospital from January 2008 to December 2017 were included. Written informed consent was obtained from all patients before surgery. Of the 208 patients during the study period, 22 patients who did not undergo exploratory laparotomy before NACT because of poor general condition (performance status ≥ 3 or ileus) and/or those aged ≥ 80 years, or who did not be evaluated in whole abdominal cavity due to adhesion were excluded. Therefore, this study included 186 patients in whom the extent of peritoneal dissemination upon initial laparotomy was assessed.
Selection of Primary or Interval debulking surgery. The treatment protocol used in this study is described in our previous reports 20, 21 . In brief, the selection of primary or interval debulking surgery was decided during the initial laparotomy which was performed as early as possible after a patient's first visit. The patients was selected for IDS when disseminated tumor burden explored at initial laparotomy was too high to achieve no residual disease; gastrectomy, resection of the hepatic hilum or head of the pancreas, massive intestinal resection, or total colectomy was required; and/or massive ascites caused coagulopathy. At that time, patients underwent ovarian, fallopian tube, or omental biopsies, followed by NACT. Patients other than those mentioned above were selected for primary debulking surgery including upper abdominal surgery and bowel resection, to achieve no residual disease after surgery.
The exploration of initial peritoneal dissemination. The extent of peritoneal dissemination was assessed during initial laparotomy according to the peritoneal cancer index (PCI) 1,2 . The initial PCI groups were classified into three groups based on the PCI scores: 1-10, 11-20, 21-39. When the patients were selected for IDS, the margins of the disseminated tumor (> 1 cm) were marked with a non-absorbable, 3-0 black silk suture after diagnostic biopsy of ovary, fallopian tube or omentum was performed 20,21 . Interval debulking surgery. The timing of IDS was described at our previous reports 20,21 . An implantable port system placed in the abdominal cavity at initial laparotomy was used for collection of peritoneal washing cytology, which was performed every 3 to 4 weeks during NACT. IDS was performed when the peritoneal washing cytology was negative and/or the serum CA-125 had decreased to 15 IU/mL or the serum CA-125 level stopped decreasing when the peritoneal washing cytology remained positive. If the disease progressed during NACT, we did not perform IDS. During IDS, the regions in which the disseminated tumors > 1 cm at the initial laparotomy was removed even when initial tumors were invisible, using the non-absorbable suture marked at initial laparotomy as landmarks.
The complexity of surgery performed. The complexity of surgery performed were scored according to the surgical complexity score 22,23 . The surgical complexity score was classified into three groups: the score of 1-3 was low, 4-7 was moderate, and 8-18 was high 22,23 . Chemotherapy. For the adjuvant chemotherapy after PDS, six cycles of weekly paclitaxel (60-80 mg/m 2 ) and carboplatin (AUC 2-3) for 3 weeks were administered. For the NACT, weekly paclitaxel and carboplatin were administered until the conditions mentioned above 20 . The median number of NACT cycles was 5. For adjuvant chemotherapy after IDS, six cycles of gemcitabine (500 mg/m 2 ) and irinotecan (50 mg/m 2 ) on days 1 and 8 every 3 weeks were administered. After December 2013, upon its approval for use by the Japanese public insurance, bevacizumab (15 mg/kg) every 21 days for 21 cycles was administered as the first-line therapy.
Statistical analysis. The association between the initial PCI and the clinical factors were analyzed using the Mann-Whitney U-test and Kruskal-Wallis H-test. Age and initial PCI score were entered as continuous variables. The association between the initial PCI and survival was analyzed by the Kaplan-Meier. Kaplan-Meier curves were drawn for the entire cohort and for subgroups of patients who underwent PDS, for those who underwent IDS, and for those who received chemotherapy without IDS. Kaplan-Meier curves were also drawn for subgroups of patients without and with residual disease (those with lesions measuring 0.1-1 cm and also those with lesions measuring > 1 cm). Univariate and multivariate Cox proportional hazards regression analyses Scientific Reports | (2020) 10:21307 | https://doi.org/10.1038/s41598-020-78296-0