Treatment strategy and prognostic factors for Krukenberg tumors of gastric origin: report of a 10-year single-center experience from China

Background Gastric cancer patient with ovarian metastasis is common in clinical practice, but it is still uncertain whether surgical resection of ovarian metastasis could improve the outcome. This study aimed to explore the survival benefit of metastasectomy plus chemotherapy over chemotherapy alone in the treatment of Krukenberg tumors arising from gastric cancer and to identify prognostic factors. Results A total of 152 patients were identified, including 93 patients with synchronous ovarian metastasis and 59 patients with metachronous ovarian metastasis. Overall survival (OS) was significantly better in metastasectomy group relative to the non-metastasectomy group for patients with synchronous ovarian metastasis (19.0 months vs. 11.8 months; P < 0.001) and those with metachronous ovarian metastasis (24.6 months vs. 14.3 months; P = 0.02), respectively. Metastasectomy (hazard ration [HR] 0.486; 95% confidence interval [CI] 0.323–0.729; P < 0.001), peritoneal carcinomatosis (HR 1.934; 95% CI 1.230–3.049; P = 0.004), and expression status of ER-β (HR 0.404; 95% CI 0.251–0.648; P < 0.001) and PR (HR 0.496; 95% CI 0.301–0.817; P < 0.001) were independent predictors of OS. Methods All patients who were diagnosed with gastric cancer and ovarian metastases between January 2005 and December 2014 were included in the current study. Patients were subdivided according to treatment modality: the metastasectomy group (metastasectomy plus chemotherapy) and the non-metastasectomy group (chemotherapy alone). The clinicopathological features and the treatment records were reviewed in detail and their association with survival were analyzed. Conclusion Metastasectomy plus chemotherapy was associated with survival benefits in patients with Krukenberg tumors from gastric cancer. Metastasectomy, peritoneal carcinomatosis, and expression status of ER-β and PR were independent prognostic factors for survival.


INTRODUCTION
Gastric cancer is one of the most common cancers worldwide, and the incidence is particularly high in Asian countries, including China [1]. Metastasis and recurrence are the major causes for poor prognosis in gastric cancer. Ovarian metastasis (Krukenberg tumor) is usually seen in female patients, including synchronous metastasis, and metachronous metastasis after curative resection of gastric cancer [2]. The reported incidence of ovarian metastasis or Krukenberg tumor is approximately 0.3% to 6.7%, however, some autopsy studies have reported incidence rates of 33% to 41% [3,4]. Ovarian metastasis is associated with poor prognosis, and is one of the most important causes of treatment failure for gastric cancer in female patients [5]. Although systemic chemotherapy can provide symptom palliation and prolonged survival in patients with ovarian metastasis, the efficiency and survival time remain disappointing [2,6]. Several studies have explored the utility of metastasectomy for Krukenberg tumors in patients with gastric cancer [7,8] However, the role of ovarian metastasectomy is still under debate and is likely to benefit only a specific subset of patients [9]. So far, the optimal treatment strategy for Krukenberg tumors from gastric cancer had not been clearly established. This study was conducted to determine associations between metastasectomy of Krukenberg tumors, clinicopathological features, and survival outcome, and consequently provide optimal treatment strategy for these patients.

Patient characteristics
The median follow-up duration for all patients was 37.2 (range 2.5-71) months; median age at treatment onset was 43.4 (range 18-65) years, and mean size of metastatic ovarian tumors was 8.13 (range 2-20) cm.
Clinicopathologic characteristics of gastric cancer with synchronous or metachronous ovarian metastasis are listed in Tables 1 and 2, respectively. Of these, tumor location, differentiation, ascites, ER and PR expression, tumor markers, and TNM stage were similar for both groups of either synchronous or metachronous ovarian metastasis.

Correlation between expression of sex hormone receptors, clinicopathologic factors, and survival
The positive rate of ER-β and PR was 44.7% and 31.6%.However, ER-α showed no positivity in any gastric cancer tissue sample.
The positive/negative expression profiles of representative sex hormone receptors are shown in Figures  4 and 5. Uni-and multivariate analyses showed a positive correlation between expression of ER-β and PR and better survival. The average OS in ER-β-positive and -negative patients was 24.4 (95% CI 21.5-27.4) months and 12.6 (95% CI 10.8-14.4) months, respectively (P < 0.001; Figure  6). Average OS in PR-positive and -negative patients was 23.7 (95% CI 20.2-27.3) months and 15.4 (95% CI 13.2-17.5) months, respectively (P < 0.001; Figure 7). These results indicate that expression of ER-β or PR are favorably associated with better prognosis in patients with ovarian metastases from a primary gastric cancer.
Furthermore, we evaluated the relationship between the expression of ER-β, PR, and clinicopathologic factors, and found negative expression of ER-β to be significantly associated with peritoneal metastasis, signet-ring cells, and ascites; negative expression of PR was significantly associated with laterality as well as peritoneal metastasis [ Table 4]. On multivariate analysis, peritoneal metastasis was significantly associated with negative expression of ER-β and PR. In addition, laterality was also significantly associated with negative expression of PR.

DISCUSSION
Cancers of the gastrointestinal tract are most likely to generate ovarian metastases in female patients and this is one of the most important causes of treatment failure [11]. The prognosis of patients with gastric cancer and metastasis to the ovaries has been reported to be poorer, as compared with prognosis in other primary gastrointestinal tumors [12]. In the past, chemotherapy was the main treatment for gastric cancer with ovarian metastasis, but the efficiency was disappointing, with a median survival time of 7 to 11 months [13]. Some retrospective study reports in recent years indicated that metastasectomy of Krukenberg tumors could improve prognosis of these patients [7,8,9]. However, an optimal therapeutic strategy for Krukenberg tumors from gastric cancer has not been fully established.  In the present study, we retrospectively analyzed outcomes of 152 patients with synchronous and metachronous ovarian metastasis. Metastasectomy, relative to non-metastasectomy treatment, resulted in significantly better OS in patients with synchronous (19.0 months vs. 13.5 months; P < 0.001) and metachronous (24.6 months vs. 14.3 months; P = 0.02) ovarian metastasis. The present series included a large study sample, with similar clinicopathologic characteristics for both study groups. Our results are in accordance with those reported in the earlier literature, therefore, we can conclude that metastasectomy should be recommended in patients with ovarian metastasis.
Analysis of prognostic factors can, moreover, facilitate identification of patients most likely to benefit from treatment. Metastasectomy, peritoneal carcinomatosis, and expression of ER-β and PR were independent predictors of OS. Krukenberg tumors are often associated with different degrees of peritoneal metastasis, which frequently induces ascites, intestinal obstruction, or hydronephrosis and seriously impair patient quality of life [14]. There is currently no standard treatment for ovarian metastasis when it presents with peritoneal metastases from gastric cancer, however, cytoreductive surgery in combination with chemotherapy could offer a survival advantage for these patients [15,16]. In our study, limited peritoneal metastases were resected in a metastasectomy, and this was followed by systemic chemotherapy. Subgroup analysis showed that the survival of the R0 resection group was superior to that of the non-R0 resection group. Therefore, surgical resection without gross residual disease may improve prognosis in patients with Krukenberg tumors. Chemotherapy is another main treatment for these patients, and improves quality of life and OS [17]. Several chemotherapeutic agents are effective against Krukenberg tumors with peritoneal metastasis, including fluoropyrimidine, platinum, taxanes, and epirubicin, either alone or in combination [18,19]. A European study reported encouraging results from the use of hyperthermic intraperitoneal chemotherapy (HIPEC) [5], but these need to be validated in future research.
Some studies have reported a correlation between expression of sex hormone receptors and the incidence and progression of gastric cancer [20]. It has been suggested that sex hormones play critical protective roles in female patients with gastric cancer [21,22]. Recently, a meta-analysis suggested that longer exposure to the effects of estrogen may potentially decrease the risk of gastric cancer [23]. However, the presence and   role of sex hormone receptors in Krukenberg tumors were undetermined. We evaluated the expression of ER-a, ER-β, and PR in gastric cancer specimens, and found that only ER-β and PR were detected. Our data showed that positive expression of ER-β or PR were favorably associated with better prognosis of gastric cancer patients with ovarian metastases. From our results and those reported in the related literature, ER and PR play an important role in the etiopathogenesis of Krukenberg tumors. However, some issues remain to be resolved. It is unclear whether targeted therapy of sex hormone receptors can lower the incidence of ovarian metastasis from ER-or PR-positive gastric cancer. Moreover, the possibility of ER or PR expression being used as a reliable indicator of prophylactic ovariectomy is worth exploring. Thus, future studies investigating the significance of ER or PR expression in carcinogenesis and tumor progression of gastric cancer and Krukenberg tumor are necessary. Our study had certain limitations because it was a retrospective analysis, and several confounding factors may have influenced our findings. However, the fact that it reports results from a large series of patients and presents the first evaluation of the role of sex hormone receptors in Krukenberg tumors from gastric cancer. The results are expected to provide useful information for formulating treatment strategies for those patients.
The present study demonstrated that the metastasectomy in combination with chemotherapy was associated with survival benefits in patients with synchronous or metachronous Krukenberg tumors from gastric cancer. Metastasectomy, peritoneal carcinomatosis, and ER-β and PR expression status were prognostic factors for survival. Well-designed prospective studies are needed to confirm these results, and will be important in developing optimal treatment strategies for Krukenberg tumors of gastric origin.

Detection of sex hormone receptors in gastric cancer specimens
The expression of estrogen receptor (ER)-α,ER-β, and progesterone receptor (PR) in gastric cancer, which may potentially be related to the ovarian metastasis, was detected by immunohistochemistry (IHC) using the Novolink Polymer Detection System as described previously [10]. The following primary antibodies were used: ER-α (ab37438, dilution 1:200; Abcam, Cambridge, UK), ER-β (ab288, dilution 1:100; Abcam), and PR (ab16661, dilution 1:100; Abcam). All specimens were independently evaluated by two pathologists who were blinded to study grouping. Samples where more than 10% of the tumor cells were stained were regarded as positive.

Ethics statement
The study was approved by the institutional ethics review board of Zhejiang Cancer Hospital. Written informed consent was obtained from all study participants. All study procedures were undertaken in accordance with the principles of the Declaration of Helsinki and relevant human research policies in China.

Statistical analysis
For continuous variables, two-tailed Student ttests were used to compare demographic and clinical characteristics between the study groups. For discrete variables, a chi-square test was used. Kaplan-Meier analysis was used to calculate survival, and log-rank test was performed to compare OS between treatment groups. Independent prognostic factors were determined by multivariate analysis using the Cox proportional hazards model. All statistical analyses were conducted on SPSS 19.0 for Windows (IBM Corporation, Armonk, NY, USA). A P value less than 0.05 was considered indicative of statistical significance.

Author contributions
Pengfei Yu, Litao Yang, Gaiguo Dai, Jieer Ying, Yian Du: Concept and design of study; Ling Huang, Guoping Cheng,:analysis and interpretation of data; Pengfei Yu, Ling Huang and Yian Du, Drafting the article; Pengfei Yu and Yian Du :Final approval of the version to be published.