The impact of peritoneal lavage cytology in biliary tract cancer (KHBO1701): Kansai Hepato‐Biliary Oncology Group

Abstract Background Only few studies in literature have analyzed the clinical effects of peritoneal lavage status in biliary tract cancers. Aim We aimed to assess the effect of cytology‐positive peritoneal lavage on survival for patients with biliary tract cancer who underwent curative resection. Methods The KHBO1701 study was a multi‐institutional retrospective study that assessed the clinical effects of peritoneal lavage cytology in biliary tract cancers. Using clinicopathological data from 11 Japanese institutions, we compared long‐term outcomes between patients with cytology‐positive and cytology‐negative peritoneal lavage. Results Of 169 patients who underwent curative resection, 164 were cytology‐negative, and five were cytology‐positive. The incidence of portal invasion and preoperative carbohydrate antigen 19‐9 levels were higher in the cytology‐positive group than in the cytology‐negative group. The incidence of peritoneal metastatic recurrence was also higher, and overall survival tended to be worse in the cytology‐positive group. In contrast, recurrence‐free survival was similar between the cytology‐negative and cytology‐positive groups. Conclusions The positive status of peritoneal lavage cytology could moderately affect the survival of patients with biliary tract cancers. Given that surgical resection is the only curative treatment option, it may be acceptable to resect biliary tract cancers without other non‐curative factors, regardless of peritoneal lavage cytology status.

in the cytology-negative group. The incidence of peritoneal metastatic recurrence was also higher, and overall survival tended to be worse in the cytology-positive group. In contrast, recurrence-free survival was similar between the cytologynegative and cytology-positive groups.
Conclusions: The positive status of peritoneal lavage cytology could moderately affect the survival of patients with biliary tract cancers. Given that surgical resection is the only curative treatment option, it may be acceptable to resect biliary tract cancers without other non-curative factors, regardless of peritoneal lavage cytology status. and ampullary region cancer (AmpCa), are intractable diseases with a dismal prognosis. 1 Radical resection without residual tumor may be the only option for a potential cure. 2,3 The presence of cancer cells in peritoneal lavage is a predictor of subsequent peritoneal dissemination of tumors, and cytology-positive peritoneal lavage (CY+) affects the survival of patients with gastric 4 and pancreatic cancers. 5 For patients with advanced gastric cancer, Japanese guidelines recommend staging laparoscopy to detect peritoneal dissemination, including CY+. 6 A change in the cytology result from positive to negative after neoadjuvant chemotherapy has been reported to be associated with improved survival. 4 Although CY+ is associated with poor prognosis in pancreatic cancer, 5,7-11 resection could improve the outcomes of patients with CY+. 11 However, only few studies with a small number of patients have analyzed the clinical effect of CY on BTC, 12,13 and the significance of CY+ in BTC remains unknown. The major concern regarding CY+ for BTC is whether surgical resection for these tumors is justified in the absence of other non-curative factors.
Thus, we conducted this multi-institutional retrospective study to assess the effect of CY+ on BTC.

| MATERIALS AND METHODS
We conducted this multi-institutional retrospective study to compare the outcomes between patients with BTC having CY+ and negative CY (CY−) who underwent curative resection. The study was approved by the institutional review board of each institution (protocol number in Yamaguchi University Hospital, which was the leading institution of this study: H29-094) and was conducted according to the ethical standards of the 2013 Declaration of Helsinki. This clinical trial has been registered at https://www.umin.ac.jp/icdr/index-j.html (identifier: UMIN000029888). Informed consent was waived because this study was a retrospective cohort study.

| Patients and study design
Clinicopathological data of patients with BTCs who underwent curative resection from January 2013 to January 2016 were collected from 11 institutions in Japan. These data were obtained from the clinical records in each institute, and the anonymized data were sent to an independent data center-the Osaka International Cancer Institute.
Clinical data included age, sex, preoperative therapy, adjuvant therapy, preoperative serum carbohydrate antigen 19-9 (CA19-9) level, tumor location, and operative procedure. Pathological data included T and N status according to the tumor-node-metastasis Classification of Malignant Tumours eighth edition by Union for International Cancer Control, 14 histological type, surgical margin status, and CY. CY status was not considered as a factor of residual tumor status (R). Data management was performed at the independent data center.
The collected data were carefully analyzed, and patients who underwent non-curative resection were excluded from this study.
These included seven patients with distant metastasis, including five with paraaortic lymph node metastasis, and 30 patients who underwent microscopic non-curative resection (R1), including patients with positive biliary margin (n = 22) and/or patients with cancer cells in exfoliative margin (n = 11). Forty-eight patients with AmpCa were also excluded because the outcome of these patients was significantly better than that of patients with the other three types of cancer ( Figures S1 and S2), and none of these patients had CY+.
The clinicopathological data of patients with AmpCa are shown in Table S1.

| Peritoneal lavage cytology
After laparotomy, the pelvic and/or subhepatic space was washed with 0.9% sodium chloride (10-200 mL), and the peritoneal washing fluid was collected for pathological examination. Smears were prepared using centrifuged deposits, stained with Papanicolaou and/or Giemsa staining, and examined by experienced pathologists. CY+ was defined as the presence of cancer cells in peritoneal lavage. The CY status results were obtained before the resection in some centers and T A B L E 1 Clinicopathological characteristics of patients

| Statistical analyses
Clinicopathological and survival data were compared between the patients with CY+ and CY−. Data are presented as medians and interquartile ranges. Continuous variables were analyzed using the Mann-Whitney U test, and categorical variables were analyzed using the chi-square test or Fisher's exact test, as appropriate. The Kaplan-Meier method was used to calculate the recurrence-free survival (RFS) and overall survival (OS), with differences being evaluated using the log-rank test. The cumulative incidence of peritoneal metastasis was estimated using the cumulative incidence function, taking into consideration the competing risk of death before peritoneal metastasis. The differences between the groups were compared using Gray's test.
Multivariate analysis to identify independent prognostic factors of OS was conducted using Cox proportional regression model. Several potential confounders reported as predictors for OS, including lymph node metastasis, differentiation, vascular invasion, combined vascular resection, 1 and modified Glasgow Prognostic Score (mGPS), were included in the model. 15 The Fine and Gray competing risks proportional hazards regression model was used to identify the independent predictors, accounting for the competing risk of death before peritoneal metastasis.
Variables with P < .10 in the univariate analysis were included in the model for peritoneal recurrence because the predictors for peritoneal recurrence have not been fully clarified. All tests were two-sided, and P < .05 was considered to be statistically significant. All statistical analyses were performed with R (The R Foundation for Statistical Computing, Vienna, Austria).

| RESULTS
One hundred and eight patients with ECC, 33 patients with GBC, and 28 patients with ICC were included in this study. Among 169 patients who underwent curative resection, five patients (3.0%) had a CY+ status, and 164 patients had a CY− status. Overall, the postoperative complication rate of more than Dindo-Clavian grade IIIa was 26.0%; however, the 90-day mortality rate was zero.

| Comparison of clinicopathological factors between the CY+ and CY− groups
Preoperative CA19-9 levels in the CY+ group were significantly higher than those in the CY− group (Table 1). Fifteen patients in the CY− group and one patient in the CY+ group received preoperative therapy, including gemcitabine plus radiation, 16 Table S2. None of the five patients underwent preoperative transhepatic biliary drainage, and one patient alone underwent preoperative percutaneous tumor biopsy in the CY+ group.
All five patients in the CY+ group had a moderately differentiated tumor. Although portal invasion was higher in the CY+ group than in the CY− group, the rate for combined portal and arterial resection was similar between the two groups.
To assess predictors for peritoneal recurrence, the presence of preoperative and postoperative therapy, vascular invasion, serosal invasion, mGPS, and CY+ were included in the model. Preoperative therapy (subdistribution HR 0.033, P = .002) and CY+ (subdistribution HR 4.251, P = .020) were identified as predictive factors of peritoneal recurrence (Table 4).

| DISCUSSION
In this study, we demonstrated that the survival of patients with BTC who underwent curative resection with CY+ tended to be worse than that of those with CY−. These results are partially inconsistent with those of previous studies 12,13 that found that the CY status did not affect patients' outcomes. This discrepancy may be due to the small number of cases in previous studies, and the accumulation of more CY+ cases than in our study could further clarify the difference in prognosis. In addition, our results showed that the cumulative incidence of peritoneal metastasis after surgery was higher in the CY+ group than in the CY− group.
The major concern regarding CY+ BTC is whether surgical resec- F I G U R E 3 Cumulative incidence of peritoneal metastasis and death before peritoneal metastasis in patients who underwent curative resection with cytology-positive peritoneal lavage (CY+), A, and cytology-negative peritoneal lavage (CY−), B. The cumulative incidence of peritoneal metastasis was higher in the CY+ group than in the CY− group (P = .034) strong prognostic factors were not significant, and mGPS, serosal invasion, and not well-differentiated types were prognostic factors in this study. The cause of the weak effect is unknown; however, our study included several cancer types, which may have led to this result.
We could no suggest the reason as to why the number of patients with CY+ in BTC was much smaller in this study (3.0%) than in previous reports (7.7% 12 and 9.8% 13 ). Although it is difficult to draw a definitive conclusion because of small number of CY+ and peritoneal metastatic recurrence in CY+ cases, our results implied a higher cumulative incidence of peritoneal metastasis in patients with CY+ after curative resection. Additionally, some studies showed that higher seeding metastasis occurs after percutaneous transhepatic biliary drainage and resection than after endoscopic biliary drainage for perihilar cholangiocarcinoma 30  cisplatin, and S-1 17 are promising preoperative chemotherapy, and postoperative therapy with capecitabine 33 or S-1 34 is also promising.
Moreover, hyperthermic intraperitoneal chemotherapy could be a choice for patients with CY+ at risk of developing peritoneal metastasis. 35 This study has some limitations. First, this was a retrospective study that included some bias in indication and method of peritoneal lavage cytology. Thus, further prospective studies defining indications and methods of peritoneal lavage cytology to assess the incidence and effect of CY+ on survival are needed. Second, the status of cytology was diagnosed in each institute; therefore, unexpected bias could have occurred.
Third, the small number of CY+ cases in this study was a limitation. However, the increase in peritoneal recurrence in patients with CY+ in our study could indicate the usefulness of peritoneal lavage cytology in BTC.
In conclusion, the positive status of peritoneal lavage cytology could moderately affect the survival of patients with BTC with increasing the incidence of peritoneal recurrence. Considering that surgical resection is the only potentially curative therapeutic option, it may be acceptable to resect BTCs without other non-curative factors, regardless of the CY status.