Abstract
Aim: Hepatic resection has become the standard treatment for patients with primary or metastatic liver malignancies. The aim of our study was to evaluate the clinical outcome of hepatic resection in patients with advanced ovarian cancer (AOC). Patients and Methods: All patients undergoing hepatic resection for AOC in our institution between 11/1991 and 02/2007 were evaluated by a validated intraoperative documentation tool. Results: Seventy patients were evaluated (median age=59 years; range=29-76 years). Forty-one (58.6%) patients underwent liver resection; 29 patients had unresectable disease. Additional multivisceral procedures performed were: colic resection (51.4%), small bowel resection (32.9%), gastric resection (5.7%), pancreatic resection (4.3%), splenectomy (5.7%). The median survival of patients with R0 resection was 42 months (95% confidence interval (CI)=17-66 months), 4 months for R1, 6 months (95% CI=0-11 months) for R2, and 5 months (95% CI=0-9 months) for those without liver resection. In multivariate analysis, postoperative residual tumor mass was the strongest predictor of survival. Conclusion: Our data indicate that complete macroscopical tumor resection remains the strongest predictor of survival in patients with liver metastases from AOC.
The established standard for newly-diagnosed advanced-stage ovarian cancer (AOC) includes staging and aggressive cytoreductive surgery followed by platinum-based chemotherapy (1). Approximately 70% of patients with advanced stage [International Federation of Gynecology and Obstetrics FIGO III/IV] disease will, however, experience recurrence. The potential benefits of maximal cytoreductive surgery have been shown in multiple studies (2). Hepatic resection has become the cornerstone in the management of patients with different primary and secondary liver tumors. Accordingly, patients with AOC with parenchymal tumor spread into the liver qualify as candidates for liver surgery under the primary intention of tumor debulking.
Only few studies, mostly of case report character or including fewer than 20 patients, have evaluated the benefit of hepatic resection in patients with various gynecological malignant tumors (3-6). Furthermore, no valid predictive factors for optimal surgical outcome after hepatic resection have been identified yet.
The purpose of the systematic analysis of a prospectively assessed database presented here, was to evaluate the surgical and clinical outcome of a large cohort of patients with AOC who underwent hepatic resection and to identify predictive and prognostic factors of survival in the subgroup of ovarian cancer patients with liver metastasis.
Patients and Methods
All patients admitted to the Charité, Virchow Clinic for optimal tumor debulking due to AOC, were included in a prospective database. Patients' data were extracted from this database, from the hospital and office charts, and from interviews with the patients. The collected data included demographics, pathology of the primary ovarian cancer, presentation of the liver tumors, extent and pathology of the liver tumors, details of the surgery, and patient outcome. The detailed tumor pattern was intraoperatively assessed by an independent person, based on the surgical procedures performed and on a postoperative systemetatic interview of the surgeon, as defined by a validated intraoperative documentation tool for ovarian cancer especially developed for this purpose (Intraoperative Mapping of Ovarian Cancer, IMO). Postoperatively, all histological findings were also entered into IMO, with special focus on the description of the tumor dissemination pattern, the maximal tumor burden as well as the postoperative residual tumor mass. All operations were performed by one of three gynecological oncological surgeons. Staging was performed and defined in accordance with the FIGO criteria for ovarian cancer (7).
IMO represents a detailed surgical and histopathological documentation system developed in our clinic in order to obtain a better and more objective description of the ovarian tumor spread within the abdominal cavity and to more precisely define the histopathological features of the malignancy (8-11). Within the Tumor Bank Ovarian Cancer project (www.toc.network.de), tumor tissue, ascites, serum and blood were collected from each patient with malignant tumor. The patient's informed consent was always given prior to surgery, sample collection and documentation.
A systematic personal follow-up was conducted based on the database of the cancer care clinic or contact with the home physician or the patient.
Patients. Overall, 70 women with liver metastases from ovarian cancer were included in the analysis. The total number of patients operated at the same time due to primary malignant or borderline lesions of the ovary of any stage was 675. Seventeen patients suffered from primary ovarian cancer and the remaining 53 patients presented with recurrent disease. The median age of the women was 59 (range=29-76) years. The majority of the women had advanced tumor stages (88.5% FIGO stages III and IV) (7). The clinicopathological characteristics of the patients are shown in Table I.
Preoperative assessment of intra- and extrahepatic tumors. The preoperative diagnostic work-up generally included following diagnostic tools in order to assess extrapelvic and distant metastatic status prior to liver resection: physical and gynecological examination, vaginal and abdominal ultrasound, pre-operative colonoscopy to exclude primary colonic cancer, serum levels, chest X-ray and serial computed-tomographic (CT) scans of the pelvis and abdomen.
Operative procedure. The operative technique was performed with the primary goal of complete or maximal cytoreduction. Anatomical resections of the liver were preferred to achieve tumor-free margins. If the remaining functioning liver tissue or the relationship of the tumor to the vessels did not allow 1-cm margins, smaller resections were accepted. Resectable extrahepatic disease or positive lymph nodes were not considered as a contraindication for the operation. Standard cytoreductive surgery of all extrahepatic tumor manifestations was carried out by an interdisciplinary team. All intraoperative tumor patterns were systematically documented utilizing the previously published IMO system. Unresectability was defined in tumors where a complete resection was not feasible due to advanced liver metastases or extrahepatic disease. On this basis, patients were divided into two groups, resectable and non-resectable.
Pathological assessment of liver metastases. For histology, liver resection specimens were fixed in 10% neutralized formalin. Macroscopic examination documented the number and size of all liver metastases as well as the distance to the surgical resection margins. Representative tissue samples of the tumors and resection margins were obtained for histological examination. De-paraffinized sections were stained using hematoxylin and eosin. The FIGO and TNM stages were determined according to the Union for International Cancer Control (UICC) guidelines and were based on histological confirmation (7). The residual tumor status was categorized as R1 when tumor tissue reached the resection margin directly without any non-neoplastic liver tissue between the tumor and the resection margin. Any non-neoplastic liver tissue between the resection margin and the tumour was categorized as R0. The women were categorized into the following groups: No residual tumour; residual tumour <0.5 cm, residual tumour >0.5 cm and <1 cm; residual tumour >1 cm (R status).
Statistics. All statistical analyses were conducted in an exploratory fashion. Patients' data, main characteristics of the primary tumor, liver metastases, chemotherapy and liver resection were compared between groups. The chi-square test or Fisher's exact test were used for univariate comparisons. Overall survival was calculated using the Kaplan-Meier technique, and data were compared using the log-rank-test. A multivariate analysis using a Cox regression model was used for factors impacting outcome in the univariate analysis (p<0.05).
Results
In 41 out of 70 (58.6%) women with hepatic metastases, a liver resection was performed. In the remaining 29 patients, the disease was unresectable and radical surgery was restricted. Operative procedures of both groups are presented in detail in Table II. Additional procedures performed were: hysterectomy 14.3%, bilateral salpingo-oophorectomy 24.3%, omentectomy 32.9%, partial colic resection 51.4%, partial small bowel resection 32.9%, partial gastric resection 5.7%, and partial pancreatic resection 4.3%. There were no significant differences in the type of procedures between the two groups. Only the number of partial small bowel resections was greater in the group of women without liver resection (20% vs. 52%, p=0.01).
Operative morbidity. The overall 3-month mortality rate was 14.6% in the liver resection group and 41% in the group of women without liver resection (p=0.025) and therefore this difference was statistically significant. The overall numbers of complications were similarly distributed between the groups. Overall numbers and types of complications are listed in Table III.
Long-term survival. The median follow-up of the 70 women was 7 months (range=0–145 months). The median survival was significantly increased in women who underwent liver resection. Median survival in patients with R0 resection was 42 (95% confidence interval (CI)=17-66 months), versus 4 months (95% CI: not computed) for R1, 6 (95% CI=0-11 months) for R2 and 5 (95% CI=0-9 months) for those without liver resection. (Figure 1, p<0.001).
Furthermore, the presence of pre-operative ascites (hazard ratio (HR)=4.1, 95% CI=1.7-9.9) and bilobular liver metastases (HR=4.7, 95% CI=1.7-13.6) were also significantly associated with a poorer overall outcome in the multivariate Cox regression (p=0.002 and 0.004, respectively) (Figure 2).
Predictors of resectability of hepatic metastases. Of all the patients included in this analysis there were no women without liver metastases. All patients who did not undergo liver resection displayed hepatic spread of ovarian cancer. In the group of women who underwent liver resection, the post-operative maximal tumor diameter was <0.5 cm in 21/41 (51%) of the cases. We included the following variables to predict optimal (i.e. <0.5 cm residual tumor) tumor resection in our patients collective: pre-operative ascites, increased CA-125 levels, response to platinum-based chemotherapy, recurrent disease, age, ascites and bilobular liver metastases. Out of the investigated factors, only the presence of ascites (odds ratio, OR=0.10; 95% CI=0.02-0.55) was negatively-associated with an optimal tumor debulking with residual tumor <0.5 cm.
Discussion
In contrast to primary ovarian cancer, the role of cytoreductive surgery for recurrent ovarian cancer has not been clearly established and randomized trials are lacking. A large number of retrospective analyses as well as reviews indicate a favorable outcome for those patients where optimal tumor resection is accomplished (12-15). However, in literature there are no data of larger trials in terms of outcome of patients with metastasized disease. In our study, we report on the first large analysis of patients with AOC with distant metastasized disease limited to the liver, and compare these women to a second group in which hepatic tumor debulking was not achievable. Although the trial was not randomized and the groups were not comparable in terms of tumor load, we clearly showed that women with advanced tumor stages and liver metastases showed a satisfactory survival when optimal cytoreductive surgery with less than 0.5 cm residual tumor was feasible. This favorable prognosis compared to women with unresectable liver disease was underlined by the fact that none of the women without liver resection survived longer than three years after surgery. We were unable to show, however, any beneficial effect of hepatic resection in patients with residual tumor greater than 0.5 cm, indicating that only optimal debulking results in improved survival. This has been shown for patients with AOC without liver resection in a large number of trials (16, 17). We were only able to identify the presence of ascites and bilateral liver involvement to be significantly related to impaired outcome, whereas the presence of ascites was also related to a decreased feasibility of optimal tumor debulking. The median survival, in our analysis, for women with R0 resection was 42 months.
Reported mean survival times of the International Collaborative Ovarian Neoplasm 4/Arbeitsgemeinschaft and the Gynecologic Cancer Intergroup studies showed survival times of more than 18 and 29 months in the respective superior arms. This is very similar to that for our patients with metastasized disease and complete debulking, showing again that a complete tumor debulking may lead to improved long-term outcome (16, 17). More importantly, none of the patients without liver resection in our series had long-term survival. Yoon et al. cautioned about performing hepatic resections in patients with gynecological malignancies and other sites of suspected metastases (18). However, this was not confirmed by recent studies published by Bristow et al., who performed hepatic resections and tumor debulking at other sites of tumor dissemination (16). We therefore believe that hepatic resection should be attempted at the time of surgical intervention due to AOC without increasing morbidity. The operative morbidity should not represent a relevant factor in the decision making of whether a woman with ovarian cancer metastasized to the liver should undergo optimal cytoreductive surgery in terms of hepatic resection. The extent of radicality in women with AOC metastasized to the liver should be based on the feasibility of achieving a residual tumor less than 0.5 cm. To identify the cohort of patients who would be optimal candidates for liver resection, the definition of pre-operative predictive factors in terms of outcome and resectability are warranted. In our study, only ascites and metastases in both liver lobes were associated to impaired outcome; neither tumor marker CA-125, nor tumor dissemination pattern, number of tumor sites, age or initial tumor stage appeared to have any significant impact on outcome or resectability. This should be, however, confirmed by further prospective multicenter trials.
To conclude, extensive tumor debulking treatment, including liver resection, appears to be feasible with acceptable morbidity in patients with AOC, and is associated with a significant improvement in overall outcome.
Footnotes
-
Disclosures
No funding sources or any conflicts of interest to disclose.
- Received July 3, 2012.
- Revision received August 14, 2012.
- Accepted August 15, 2012.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved