One of the most important prognostic factors for patients affected by epithelial ovarian carcinoma (OC) is the amount of residual disease at the time of primary cytoreductive surgery.1–3 Advances in surgical training, technique, and perioperative care have allowed surgeons to optimize cytoreductive efforts for patients with advanced stages 3c to 4 ovarian cancer, thus improving the rate of complete cytoreduction and potentially also clinical outcome.4–6

In this context, a more comprehensive surgical approach to disease involving anatomic sites, considered at least some years ago to be unresectable and thus precluding chances of complete cytoreduction,7,8 has been more recently afforded.9–12 Studies investigating technical aspects and feasibility of surgery to remove metastatic disease involving the hepatoceliac lymph nodes (HCLNs) also have been published.13–18 However, the real incidence and clinical relevance of tumor spread to HCLNs still is far from being clarified given the paucity and heterogeneity of studies on this issue.

This retrospective large-scale study aimed to investigate the rate of HCLN involvement, its association with clinicopathologic features, the morbidity of HCLN resection, and the prognostic impact of metastatic HCLN status on advanced ovarian cancer patients undergoing cytoreductive surgery.

Patients and Methods

The clinical and pathologic records of patients with a diagnosis of advanced (stages 3c to 4) epithelial OC referred to the Division of Gynaecologic Oncology of the Catholic University at Rome and Campobasso between January 2010 and September 2016 were searched to select patients undergoing any kind of HCLN resection. The local ethics committee approved the study. All the patients had given informed consent before surgery for their data to be collected and reviewed for scientific purposes.

The inclusion criteria specified age of 18 years or older; International Federation of Gynecology and Obstetrics (FIGO) stages 3c to 4; any histotype; availability of a preoperative workup including computed tomography (CT) scan, Ca125, and ultrasound examination; achievement of residual disease after surgery [either primary debulking surgery (PDS) or interval debulking surgery (IDS)] of 2 cm or less; and Eastern Cooperative Oncology Group (ECOG) performance status of 2 or lower. The exclusion criteria ruled out extraabdominal disease and progressive disease during neoadjuvant treatment.

Intraperitoneal dissemination was always assessed by laparoscopic evaluation, which provided a score [predictive index value (PIV)] used to stratify patients into three groups according to the extent of disease.19 Patients with unresectable disease were triaged to three or four cycles of neoadjuvant chemotherapy (platinum, paclitaxel ± bevacizumab) and submitted to IDS after assessment of their response on the basis of response evaluation criteria in solid tumors (RECIST) criteria.20 Cytoreductive surgery at PDS or IDS was performed with the intent of maximally removing all visible tumor. Debulking surgery was performed and combined with extensive radical procedures if they were likely to result in macroscopically complete cytoreduction.

Indication for HCLN resection was based on documentation of suspicious HCLNs on the preoperative CT scan. Moreover, intraoperative palpation of lymphadenopathy in the HCLN area and peritoneal disease at the omental bursa also were considered as additional indications for HCLN resection, even in the absence of clear radiologic evidence of disease. Peritoneal disease at the omental bursa corresponds to the presence of carcinomatosis on the anterior surface of the caudate lobe, on the surface of pancreas, and on the roof of the supragastric lesser omentum. Surgical complexity was defined according to Aletti et al.21 in all cases.

All the patients had received full mechanical bowel preparation 1 day before surgery. Perioperative antibiotics and thromboembolic prophylaxis with low-molecular-weight heparin were routinely used.

Data about intra- and postoperative complications occurring within 30 days after surgery were retrieved. Surgical morbidity was classified according to the Memorial Sloan Kettering Cancer Center (MSKCC) surgical grading system.22,23 Besides clinical features, data also were collected on histologic subtype and grade, pathologic status of HCLNs, paraaortic lymph nodes (PALNs), and mesenteric lymph nodes (MLNs) at the level of the bowel wall.

Usually, MLNs are resected together with the mesentery resection associated with rectosigmoidectomy in case of disease involvement. Therefore, these lymph nodes are not routinely removed.

The pattern of recurrence presentation was recorded and classified as pure retroperitoneal relapse (i.e., lymph node recurrence only) and mixed (i.e., retroperitoneal plus abdominal and/or extraabdominal recurrence).24

Surgical Technique

The technical aspects of HCLN resection have been reported previously.13,14,16 Briefly, lymph nodes along the hepatic artery, around the celiac trunk, and along the splenic artery were sequentially separated and dissected. Tracking the stomach upward, the proper hepatic artery was identified and exposed from the gastroduodenal artery to the hepatic hilum. Dissection was started by separating the lympho-fatty tissue around the common bile duct and the proper hepatic artery. The hepatic artery was retracted to the right side of the abdomen for dissection of the nodal tissue lying behind it and the nodal tissue surrounding the surface of the portal vein. The dissection was continued by removing the lympho-fatty tissue surrounding the common hepatic artery and the splenic vessels, together the superior border of the pancreas and around the anterior surface of the celiac trunk and the beginning of the left gastric artery.

Statistical Analysis

Fisher’s exact test for proportion or the χ 2 test were used to analyze the distribution of categorical variables between groups. The Mann–Whitney rank-sum test was used to analyze the distribution of nonparametric continuous values.

Progression-free survival (PFS) was calculated from the date of diagnosis to progression of disease or the date the patient was last seen, whereas overall survival (OS) was calculated from the date of diagnosis to the date of death due to disease or the date of the last follow-up visit.

Medians and life tables were computed using the product limit estimate by the Kaplan–Meier method25 and analyzed by the log-rank test.26 Multivariate analysis of the prognostic impact of clinicopathologic parameters was performed by the Cox proportional hazards model.27 Differences were considered statistically significant at a p value lower than 0.05. All statistical calculations were performed by SPSS software, version 17.0 (SPSS, Chicago, IL, USA).

Results

Between January 2010 and September 2016, 566 OC patients with stages 3c to 4 disease were submitted to debulking surgery at the Catholic University of Rome and Campobasso. Of these 566 patients, 85 (15%) underwent HCLN resection. The rate of HCLN resection did not differ over time (data not shown).

The clinicopathologic characteristics of the patients are summarized in Table 1. The vast majority of HCLN removal was performed in the context of PDS (n = 78, 91.8%), and most of the patients (81.2%) had serous adenocarcinomas. At time of cytoreduction, laparoscopy documented abdominal tumor load as low in 26 patients (30.6%), intermediate in 29 patients (34.1%), and high in 30 patients (35.3%).

Table 1 Patient characteristics
Full size table

Rectosigmoid resection was performed for 58 patients (68.2%), whereas splenectomy, distal pancreatectomy, or both were performed for 34 patients (40%). Paraaortic lymphadenectomy was performed for 80% of the patients and mesenteric lymphadenectomy for 41.2% of the patients. Overall, a surgical complexity score of 2 or 3 was recorded for 48.2% of the patients. Absence of visible tumor at the end of surgery was documented for 73 patients (85.9%), and residual tumor smaller than 1 cm was achieved for 11 patients (12.9%). Eight patients were classified as FIGO stage 4b for parenchymal liver and spleen metastasis.

Regarding HCLN resection, the median number of HCLNs removed was 6 (range 1–18). Histopathologic evaluation was able to identify HCLN metastasis in 45 (52.9%) of the 85 cases (Table 2). Of 66 cases considered to be negative for HCLN involvement at preoperative imaging, only 32 were found to be negative (negative predictive value, 48.5%). Conversely, of 19 cases judged to harbor HCLN metastasis at imaging, 13 were shown to be positive at histologic evaluation (positive predictive value, 68.4%). Three of the eight patients submitted to IDS showed HCLN involvement. The HCLN involvement was strictly correlated with metastatic PALNs (p = 0.033), metastatic MLNs (p = 0.011), and peritoneal disease shown at the omental bursa level (p = 0.0008).

Table 2 Association of clinicopathologic features with pathologic hepatoceliac lymph node (HCLN) status
Full size table

Perioperative Details and Postoperative Measures

Table S1 shows that in the whole series, the median surgery time was 390 min (range, 180–630 min). The median hospital stay was 10 days (range 3–37 days), and the median time to the start of chemotherapy was 33 days (range 19–40 days). The peri- and postoperative measures did not differ according to pathologic HCLN status.

No intraoperative complications related to HCLN surgery were documented. Table S2 summarizes the details about postoperative morbidity (any grade) occurring within 30 days after surgery. Overall, 58 complications occurred, 30 (51.7%) of which were grade 3 or higher. The most frequent complications concerned the hematologic/vascular system (n = 12, 40%), followed by pulmonary complications (n = 9, 30%). Three postsurgical deaths occurred due to sepsis (n = 1) and heart failure (n = 2). No difference in the rate of surgical morbidity according to pathologic status of HCLN was observed (data not shown).

Clinical Outcome

As of December 2016, the median follow-up period was 36 months (range, 6–54 months). Recurrence of disease was observed in 35 (41.2%) of the 85 cases. Relapse of disease most frequently occurred for patients with metastatic HCLN involvement (65.7%) compared with patients who had no HCLN involvement (34.3%) (p = 0.048) (Table S3). The median PFS values were 16 months (95% confidence interval [CI], 12–19 months) for the patients with metastatic HCLNs and 22 months (95% CI, 12–19 months) for the patients with no HCLN involvement (p = 0.035) (Fig. 1a). The pathologic status of HCLNs was not significantly related to the pattern of progression of disease.

Fig. 1
figure 1

a Median progression-free survival of patients with negative (blue line) and positive (yellow line) hepatoceliac lymph nodes. b Overall survival of patients with negative (blue line) and positive (yellow line) hepatoceliac lymph nodes

Full size image

Death due to disease was documented for 22 (25.9%) of the patients. The median OS in the whole series was 43 months. The patients with HCLN involvement did not differ from those with no HCLN involvement (Fig. 1b).

Multivariate analysis showed that stage 4 disease and metastatic HCLN status retained an independent, unfavorable prognostic role for PFS (Table 3).

Table 3 Uni- and multivariate analyses for progression-free survival
Full size table

Discussion

To our knowledge, this is the largest study investigating the rate of HCLN involvement, its association with clinicopathologic features, and the prognostic impact for advanced OC submitted to cytoreductive surgery. Overall, HCLN resection was performed for 15% of advanced OC patients surgically managed with a complete eradication intent. This figure closely matches results obtained in other series of patients enrolled during a similar time frame.13–15,17,18 (Table 4).

Table 4 Main features of studies investigating the clinical role of HCLN resection in ovarian cancer patients
Full size table

The rate of metastatic HCLN involvement was 52.9%, thus appearing to be comparable with the results obtained by Martinez et al.,14,17 but much different from those reported by other authors.15,18 These discrepancies could be related to several factors. First, it must be acknowledged that the few available studies are considerably heterogeneous in terms of study design, sample size, and clinical setting (e.g., including primary patients, recurrent patients, and patients triaged to neoadjuvant chemotherapy) (Table 4).

Second, the criteria used in the decision to triage patients to HCLN resection must be taken into account. Indeed, in the prospective series by Raspagliesi et al.,15 the indication for omental bursa and HCLN resection was based on intraoperative findings and a CT scan suspicious for HCLN involvement. However, the CT scan failed to identify HCLN or porta hepatitis involvement in almost all cases. On the other hand, in the study by Tozzi et al.,18 the CT scan was able to detect correctly all cases with metastatic HCLNs. In this context, integration of data coming from CT or positron emission tomography (PET)/CT approaches with intraoperative palpation or laparoscopic visualization of porta hepatis could represent the best effective strategy for improving diagnostic accuracy of HCLN involvement, as sustained by Tozzi et al.18

Finally, it must be considered that to date, resection of HCLN stations has been confined to suspicious bulky lymphadenopathy (i.e., diameter, ≥1 cm), thus leading clinicians to underestimate the true prevalence of disease in this area.

Our series showed no statistically significant difference in the distribution of clinicopathologic features or pathologic HCLN status, except for omental bursa involvement, metastatic PALNs as previously reported,14–17 and MLNs. Indeed, despite the limits of the relatively low number of patients undergoing para-aortic and mesenteric lymphadenectomy, the patients with metastatic PALNs and MLNs showed metastatic involvement of HCLN in 62.8 and 70% of cases, respectively.

The documentation that HCLN involvement increases with the number of metastatic PALNs14 and that relapse at HCLN stations is higher in cases with metastatic MLNs11 should encourage systematic intraoperative exploration of the HCLN area even in the absence of any suspicious lesion at preoperative imaging if metastatic PALNs and MLNs are documented.

Concerning clinical outcome, we showed that the patients bearing metastatic HCLNs experienced worse PFS than the patients with uninvolved ones, thus corroborating preliminary observations that had failed to achieve statistical significance only because of the relatively shorter follow-up period and the smaller series17,18 (Table 4). Moreover, we first reported a detailed evaluation of the multivariate analysis, which identified metastatic HCLN status as having an independent, unfavorable prognostic role for PFS. Although this finding aligns with those reported in other malignancies,28,29 no evidence indicates that HCLN resection might confer a survival advantage for advanced OC patients. Indeed, studies addressing this issue are expected to be hardly feasible considering the frequent association of metastatic HCLNs with diffuse abdominal and even extraabdominal disease.14,15,17 Therefore, it could be argued that in the absence of any evidence demonstrating the efficacy of macroscopic HCLN resection in prolonging patient survival, this additional supra-radical procedure could lead to overtreatment and related morbidity.

In this context, it must be acknowledged that very recent data have documented that systematic lymphadenectomy does not provide any survival benefit compared with lymphadenectomy for patients with stages 2b to 4 ovarian cancer submitted to primary cytoreduction.

Conversely, systematic lymphadenectomy has been shown to carry higher rates of serious postoperative complications, hospital re-admittance, and deaths within 2 months after surgery.30

During the wait for the ongoing final evaluation results from the Scorpion trial31 and the future findings from the Trust trial (NCT02828618, http://clinicaltrials.gov), achievement of complete cytoreduction in advanced OC remains the standard (http://www.esmo.org/Guidelines and http://www.nccn.org/professionals/physician_gls/pdf/ovarian.pdf), thus highlighting the relevance of taking into consideration resection of bulky HCLN disease among all surgical procedures required to achieve this goal.

Probably, resection of bulky HCLNs should never be omitted when palliation is required given the risk for obstruction of biliary ducts and/or bowel, as well as celiac plexus infiltration. Therefore, the decision to proceed to maximal cytoreductive effort should be based on a careful evaluation of the balance between the chances of complete cytoreduction and the number and types of all planned surgical procedures rather than on overestimation of morbidity related to surgical maneuvers at the HCLN site. Indeed, possible complications related to HCLN resection could include vascular and bile duct injuries. In our case series, these intraoperative complications did not occur. As described in literature, the rate of overall G3-4 postoperative complications ranged between 10.8 and 31.7% of patients (Table 4). Moreover, the rate of morbidity appears to be acceptable due to the relatively low postoperative rate of lymph-vascular complications (11.8%) and considering the ability to manage these complications by percutaneous drainage of infected and noninfected collections.

Therefore, a gradual but progressive increase in adoption of HCLN resections within maximal cytoreductive effort for advanced OC is expected in the future, thanks to the closer interaction of gynecologic oncologists with hepatobiliary surgeons who are more familiar with the HCLN area and vascular variants.32

In conclusion, our data suggest that HCLN resection of clinically suspicious HCLNs should be included among the surgical procedures to be adopted for achievement of complete cytoreduction. Moreover, systematic evaluation of the HCLN area could be helpful when metastatic PALNs and MLNs are documented. The edge approached over time by the criteria predicting little or no resectability in ovarian cancer has moved further and further.33–35 Perhaps the time is right to include systematic HCLN evaluation and eventual resection in upfront cytoreductive surgery as well as in other steps of the relentless trajectory of this neoplasia. Future prospective studies are needed to assess the role of HCLN resection in advanced ovarian cancer.