Cytoreductive Surgery in Advanced Epithelial Cancer of the Ovary: The Impact of Aortic and Pelvic Lymphadenectomy
Nick M. Spirtos, M.D., Gary M. Gross, M.D., James L. Freddo, M.D., and Samuel C. Ballon, M.D.
Beginning in July 1998, a planned program was undertaken to assess the role of aortic and pelvic lymphadenectomy in patients with advanced epithelial cancer of the ovary (Stages IIIa-IVa) undergoing cytoreductive surgery. Our intent was to perform a complete aortic and pelvic lymphadenectomy in all patients in whom we could surgically remove all intra- or retroperitoneal disease measuring 1 cm or greater. Accordingly, 56/77 patients (73%) underwent complete aortic and pelvic lymphadenectomy. The remaining 21/77 patients (27%) did not, either because the lymphadenectomy would not have impacted on the patient's cytoreductive status or because intraoperative conditions precluded it. Positive lymph nodes were found in 36/56 patients (64%). Of these, 23/36 (64%) were macroscopically positive, and if left in situ would have affected the patient's cytoreductive status. Thirteen of 36 (36%) were positive microscopically. Reassessment laparotomy was performed in 44/56 (79%) of the patients having had a lymphadenectomy and is correlated to disease status. Median follow-up is 30 months (range 2-64 months). Survival analysis reveals: 10/20 patients (50%) with negative lymph nodes; 6/13 patients (46%) with microscopically positive lymph nodes; 10/23 patients (43%) with macroscopically positive lymph nodes; 10/23 patients (43%) with macroscopically positive, but surgically removed lymph nodes; and only 2/21 patients (10%) with residual disease measuring at least 1 cm in diameter are alive without evidence of disease. These preliminary findings suggest that the removal of macroscopically negative lymph nodes offers little benefit to the patient with advanced epithelial cancer and minimal residual (less than 1 cm) disease. However, the concept of cytoreductive surgery, whether it be intra- or retroperitoneal, appears to be validated by the fact that the patients undergoing removal of macroscopically positive lymph nodes have approximately the same chance of survival as those with microscopically positive and/or negative lymph nodes.
The value of optimal cytoreduction in patients with Stage III and IV ovarian cancer has been demonstrated in numerous studies. Removal of all tumors greater than 1 cm in diameter improves progression-free interval survival and establishes the therapeutic value of a maximal surgical effort in patients with advanced ovarian cancer [1-3]. Resection of portions of the gastrointestinal and genitourinary tracts is accepted to achieve optimal tumor removal [4,5]. Diaphragmatic resection and splenectomy have also been performed in pursuit of this goal [6,7]. Yet, despite evidence that the pelvic and para-aortic lymph nodes are frequent sites of metastases in patients with advanced ovarian cancer, the value of lymphadenectomy in this maximal surgical effort has not been well defined [8,9]. This prospective clinical trial assesses the role of pelvic and para-aortic lymphadenectomy in patients with advanced epithelial ovarian cancer.
MATERIALS AND METHODS
From July 1, 1988 to June 30, 1991, 77 patients with Stage III or IV
epithelial cancer of the ovary underwent exploratory laparotomy. Complete
pelvic and para-aortic lymphadenectomy was attempted whenever all intra- or
retroperitoneal disease measuring 1 cm or greater could be resected. All
patients underwent similar operations through a midline incision extending
from the symphysis pubis to the xiphoid. Initial intraperitoneal and
retroperitoneal exploration identified patients with metastases to the
hepatic parenchyma or the porta hepatitis, involved lyph nodes (greater
than 1 cm) above the celiac trunk, and/or unresectable tumors of the bowel
mesentery. These patients were excluded from this study. A systematic
approach minimized the extent of operation prior to determining
resectability. The entire greater omentum was resected, as was the lesser
omentum if involved by tumor. The lesser sac and its contents were then
explored. The phrenicocolic ligament was transected and the transverse and
descending colon mobilized medially to expose the left renal vessels.
Mobilizing the duodenum anteriorly exposed the vena cava at the level of
the ovarian vein insertion. The lymph node status was evaluated and if
nodal metastases measuring greater than 1 cm were identified above the
level of the celiac trunk, the operation was redirected toward relief of
symptoms and confirmation of the primary site of disease. Prior to lymph
node dissection, diaphragmatic disease was resected using an argon beam
coagulator after mobilizing the liver by transecting the falciform,
triangular, and coronary ligaments as necessary. If metastatic disease
penetrated the muscle of the diaphragm so as to involve the pleura, further
cytoreductive efforts were abandoned. At this point, remaining
intraaperitoneal disease was resected prior to the lymphadnectomy.
The mean age of the 56 patients undergoing complete lymphadenectomy and
optimal cytoreduction was 60.2 (range 32-76) years. The mean age of the 21
patients in whom optimal cytoreduction and complete lymph node dissection
was not achieved was 76.4 (range 47-85) years. The stage, histology, and
grade of the cancer are listed in Table 1, categorized according to the
three pathologic findings for resectable disease and also for unresectable
disease. Gross inspection of the table suggests that there is no
difference in pathologic category for resectable vs. unresectable disease,
regardless of lymph node status, but because of the large number of
categories, and high risk of Type-II statistical error, separate analysis
was made by forming the separate groupings: Stage < 3c, Stage * 3c,
papillary cell type, endometrioid cell type, adenocarcinoma, Grade < 3,
Grade = 3, lymph node positive, lymph node negative, resectable, and
unresectable tumors. Using these categories, grade was not different
between resectable and unresectable tumor (P = 0.20), but Grade 3
was more common for lymph-node-positive than negative patients (83 vs 55%,
P = 0.025). Stage * 3c was more likely for
lymph-node-positive patients than lymph-node-negative (86 vs 60%, P
= 0.03), and foir unresectable than resectable patients (100 vs 77%,
P = 0.01). Both papillary serous cell type and adenocarcinoma were
more likely than endometrioid cell type for resectable tumors compared to
unresectable tumors (P = 0.04 for both comparisons), suggesting
endometrioid cell type as a risk factor for unresectability in our series.
Two patients had Stage IIIC disease based only on the presence of positive
retroperitoneal lymph nodes which were macroscopically involved. Of the 77
patients, 21 (27%) did not undergo optimal cytoreductive surgery or
complete lymph node dissection. Seventeen of the 21 (81%) were not
candidates for lymph node dissection because unresectable intraperitoneal
disease precluded optimal cytoreduction. Two of 21 (10%) were found to
have lymph node metastases (greater than 1 cm) above the level of the renal
vessels and 2 (10%) were found to have liver metastases measuring greater
than 1 cm in diameter.
Fifty-six of the 77 patients (73%) had optimal intraperitoneal cytoreduction and complete pelvic and aortic lymph node dissection. Twenty of these 56 patients (36%) had histologically negative lymph nodes and 36 (64%) had positive lymph nodes. Of the 36 patients with positive lymph nodes, 23 (63%) were macroscopically positive (greater than 1 cm in diameter) and 13 (36%) were microscopically positive. Aortic lymph node metastases were present in 18 of the 23 patients (78%) with macroscopically positive lymph nodes and in 9 of 13 patients (69%) with microscopically positive lymph nodes (Table 2). The average number of lymph nodes counted was 43.3 (range 24-74). The number of lymph nodes resected (counted) was similar in those whose lymph node status was negative (41.6 +/- 13.8), microscopically positive (44.6 +/- 13.1), or macroscopically positive (46.9 +/- 17.1), with P > 0.32 for all three possible comparisons among these groups.
Twenty-three of 77 patients (30%) with Stage III and IV ovarian cancer required pelvic and/or para-aortic lymphadenectomy in order to achieve an optimal cyto-reductive status as the lymph nodes involved measured greater than 1 cm in diameter. Eighteen of these 23 patients (78%) required resection of metastatic lymph nodes between the level of the inferior mesenteric artery and renal vessels. Ten of these 23 patients (43%) required en bloc resection of the sigmoid colon and reproductive organs to achieve optimal intraperitoneal cytoreduction before retroperitoneal tumor resection was indicated. In addition, 8 of 23 (35%) required resection of diaphragm metastases; 1 of 23 (4%) required partial gastrectomy; and 2 of 23 (8%) required small bowel resection to achieve equal status. In only 2 of our patients were the lymph nodes the only site of metastatic disease. Both had macroscopically positive para-aortic lymph nodes near the insertion of the ovarian vein into the left renal vein. Intraoperitoneal procedures of similar magnitude were required in the 13 patients with microscopic nodal metastses and in the 20 patients with negative lymph nodes (Table3) .
The mean operative times for patients with negative lymph nodes, microscopically positive, and macroscopically positive lymph nodes were 2.9, 3.75, and 4.25 hr, respectively. In addition to removal of the uterus, tubes, ovaries, and omentum, the procedures required to achieve an optimal cytoreductive status in each of these groups are listed in Table 3. In the 21 patients in whom optimal cytoreduction was not achievable the mean operative time was 2.75 hr. Four (19%) of these patients underwent en bloc resection of the omentum and transverse colon in order to access the peritoneal cavity, assess tumor resectability, and relieve obstructive symptoms. Two (10%) of these patients required resection of the small bowel or sigmoid colon resection with diverting colostomy for relief of obstructive symptoms.
The average hospital stays for patients with negative lymph nodes ("neg"), microscopically positive ("micpos"), and macroscopically positive ("macpos") lymph nodes were 9.2 +/- 4.0 (range 5-18) days, 11.5 +/- 3.6 (range 7-17) days, and 12.6 +/- 6.6 (range 7-35) days, respectively. Comparison of number of hospital days among these groups was not statistically significant, even without excluding the one outlier of 35 days in the last group (neg-micpos, P = 0.096; neg-macpos, P = 0.052; micpos-macpos, P = 0.595; excluding one outlier: neg-micpos, P 0.096; neg-macpos, P = 0.081; micpos-macpos, P = 0.972). The average hospital stay for patients with suboptimal residual disease was 14.1 (range 6-31) days. Morbidity and mortality were similar in both the lymph-node positive and -negative groups (Table 4). Mortality was zero in all resected groups. Major morbidity (as defined in Table 4) occurred in 3/20 (15%) of lymph-node-negative, 2/13 (15%) of lymph-node-microscopically positive, and 3/23 (13%) of lymph-node-macroscopically positive patients (P = ns among these three groups). By comparison, the unresectable and suboptimally resected patients had major morbidity in 3/21 (14%) and mortality in 1/21 (4.8%).
Fifty-five of 56 patients (98%) with optimal cytoreduction receieved platinum-based multi-agent chemotherapy. One patient died secondary to myocardial infarction 45 days after exploration and 35 days after discharge. Three patients with macroscopically positive lymph nodes, one patient with microscopically positive lymph nodes, and one patient with negative lymph nodes progressed within the first four cycles of treatment for an overall progression rate of 10%. One patient died secondary to renal failure after the second cycle of platinum-based chemotherapy. Therefore, 50 of these 56 patients (89%) completed the planned course of chemotherapy. Seveteen of 21 patients (80%) with suboptimal cytoreduction received similar platinum-based chemotherapeutic regimens. One patient received melphalan; one Adriamycin and and cyclophosphamide; and two died prior to the initiation of chemotherapy. Five of these 21 (23%) progressed within the first three cycles of chemotherapy and 1 patient died of sepsis after her first course of chemotherapy. Therefore, the reason that a greater percentage of optimal cyto-reduction patients received platinum-based chemotherapy than suboptimal patients is that 2 suboptimal patients died prior to chemotherapy.
Forty-four of 56 patients (79%) with optimal cytoreduction and lymph node dissection underwent reassessment laparotomy as previously described. (10) Only 6 patients considered for laparotomy did not undergo reassessment. The disease status of these patients is presented in Table 5. In the 30 of 36 patients with either macroscopically or microscopically positive lymph nodes undergoing reassessment laparotomy, 17 (57%) were negative and 13 (43%) were positive. The lymph nodes were the sole site of persistent tumor in only 1 patient and in that case the aortic lymph nodes were the only site of metastases identified at the time of initial operation. In the patients with negative intraperitonealfindings, a retroperitoneal dissection at or above the level of the ovarian vessels was undertaken to document the presence of persistent retroperitoneal disease. Fourteen of 20 patients with negative lymph nodes underwent reexploration: 8 were negative (57%) and 6 were positive (43%). At the time of the second operation, no retroperitoneal lymph nodes were identified as the sole site of persistent disease in this group of patients.
Patients with positive second operations were treated with various salvage regimens. Table 6 summarizes the approaches and results of these regimens and categorizes the patients by lymph node status. Fifteen of 17 patients (88%) were treated as noted in Table 6 as 2 patients died following reassessment laparotomy. One died of bilateral cerebral infarcts as a result of prolonged hypotension and another died of sepsis. Long-term survival information for those patients undergoing cytoreductive surgery and lymphadenectomy is summarized according to the status at reassessment laparotomy in Table 7. Ten of 23 patients (43%) with macroscopically involved but resected lymph nodes are alive, as are 6 of 13 patients (46%) with microscopically positive nodes and 10 of 20 patients (50%) with negative lymph nodes. There was no difference in the survival when patients with either microscopically or macroscopically positive aortic lymph nodes were analyzed serparately. Eight of 18 patients (44%) with macroscopically positive aortic lymph nodes are alive without evidence of disease as are 3 of 9 patients (33%) with microscopically positive aortic lymph nodes. (P = 0.45) Median follow-ups for these groups of patients are 40.4 (range 2-64), 34.3 (range 9-63), and 31.8 (range 2-58) months, respectively. One patient with macroscopically positive lymph nodes and 1 with negative lymph nodes are alive with disease and for the purpose of this paper are not considered long-term survivors. Of 18 patients who had resection of macroscopically positive aortic lymph nodes, 8 are alive at long-term follow-up. This compares with 9 patients who had resection of microscopically negative aortic lymph nodes, 3 of whom are alive at long-term follow-up. thus, the presence of macroscopically positive aortic nodes did not adversely affect long-term survival if the nodes were resected (P = 0.45, Fisher's test). Recurrences in patients with negative reassessment operations occurred in 4 of 12 patients (33%) with macroscopically positive lymph nodes; one of 6 patients (17%) with microscopically positive lymph nodes; and 5 of 13 patients (38%) with negative lymph nodes. The sites of recurrence are presented in Table 8. Two of 21 patients (10%) who were left with residual disease greater than 1 cm in diameter are alive without evidence of disease. One of the 2 patients underwent reassessment laparotomy and has remained without evidence of disease for 39 months. The other patient has not been reexplored and is alive 43 months after her first operation. Seventeen other patients were treated with chemotherapy with a median survival of 12.6 (range 1-30) months. Sixteen of these patients are dead of disease and 1 patient is alive with disease at 30 months. These results can be compared with the 21 patients with unresectable disease at the time of exploration. Nineteen of 21 (90%) died before 24 months, and 2 of 21 (10%) are alive at 5 years follow-up.
Tumor recurred in 4 of 25 (16%) patients with a negative reassessment laparotomy. Two patients with macroscopically positive lymph nodes and negative reassessment developed recurrent tumor, 1 with cerebral metastases 3 years after initial operation and 1 with pericardial metastases 12 months after her initial laparotomy. The other 2 patients with tumor recurrence did not have a second operation. One recurrence was intraperitoneal and the other involved the retropancreatic lymph nodes. No patient with microscopically positive lymph nodes and a negative second operation has suffered a recurrence, but there are only 6 patients in this group. One patient with microscopically positive lymph nodes developed a recurrence, but did not undergo reexploration. Recurrent tumor has been diagnosed in 2 of 8 patients with negative lymph nodes and a negative reassessment laparotomy. One patient recurred in the vagina and another patient in a pararenal lymph node. In the second case, this site was most likely microscopically positive at the time of her first and second operations and went undetected. The other 3 patients developed intraperitoneal recurrence and did not have a reassessment laparotomy.
The high prevalence rates for both pelvic and paraaortic lymph node
metastases in patients with advanced ovarian cancer is reconfirmed in this
report [8,9]. The relationship between para-aortic lymph node metastases
and intraperitoneal disease had been analyzed  and the role of pelvic
and para-aortic lymph node sampling as part of operative staging is well
established. Equally well established is the importance of intraperitoneal
cytoreduction in patients with advanced ovarian cancer [1,2]. The value of
retoperitoneal cytoreduction is less well understood. Retrospective
studies of Allen et al.  and Kigawa et al.  suggest
that lymphadenectomy benefits patients with advanced ovarian cancer as
indicated by improved survival. However, neither author defines those
patients with ovarian cancer most likely to benefit from retroperitoneal
lymphadenectomy. Burghardt et al.  found a correlation between
the number of positive lymph nodes present and their inability to perform
optimal cytoreduction, but they do not indicate how many patients required
lymphadenectomy to achieve optimal cytoreductive status.
In this series, if optimal intraoperitoneal cytoreduction was achieved, a complete pelvic and para-aortic lymph node dissection was performed in order to determine its impact on both cytoreductive status and survival. In those patients in whom optimal cytoreductive status could not be achieved, a maximal surgical effort was not undertaken, and the long-term survival was very low.
The morbidity and mortality associated with this approach are detailed in Table 4 and compare favorably with those in other reports in which cytoreductive efforts independent of lymphadenectomy have not been universally performed . However, it should be noted that reassessment laparotomy is made difficult in patients having undergone complete lymphadenectomy, as adhesions to retroperitoneal structures increase the difficulty of dissection. Loops of small bowel frequently adhere to the iliac vessels and the vena cava. En bloc resection of the sigmoid colon and reproductive organs compounds this problem. Retroperitoneal fibrosis often makes reevaluation of this area impossible. In addition, after the diaphragmatic peritoneum is stripped, the liver becomes adherent and precludes subsequent evaluation. Some of the technical difficulties discussed above were encountered in our patients, but we did not attempt to quantitate adhesions, nor are we sure how this could be done retrospectively. Morbidity and mortality for those patients found to have unresectable disease are acceptably low, in our opinion, and this therfore lends support to offering this cytoreductive procedure to all suitable candidates, thus allowing resection for 73% at the cost of incomplete resection in 27%.
Our results suggest that the resection of macroscopically involved lymph nodes is as important a part of cytoreductive surgery as is the removal of intraperitoneal tumor, regardless of location of lymph nodes. It is more difficult to determine the benefit of resecting microscopically positive lymph nodes. Some authors suggest that nodal metastases are unaffected by chemotherapy . However, these claims are difficult to assess in patients with unresected lymph nodes since it is not clear whether persistent microscopic retroperitoneal disease had responded or remained stable. Lymphogrophy studies indicate that it is unlikely that all microscopic metastases are removed by performing lymphadenectomy in patients with involved lymph nodes [14,15]. Therefore, resection of microscopically involved lymph nodes appears to have less impact on cytoreductive status and survival. To ensure that macroscopically involved lymph nodes are not left unresected, thorough retroperitoneal exploration of the lymphatic drainage of the ovaries to the level of the renal vessels is mandatory in all patients undergoing operation with curative intent.
Based on our data, tumor recurrence is not likely during a 30-month follow-up after negative ressessment laparotomy (incidence of 16% for the present series). It is difficult to compare recurrence rates in our patients following negative reassessment laparotomy to those of others, but it is noteworthy that the lymph nodes were rarely a site of isolated persistent disease compared to other published reports . There has been much speculation regarding the biology of ovarian cancer that has extensively metastasized to both intraperitoneal and retroperitoneal sites. Hoskins et al. concluded that cytoreduction of large-volume to small-volume disease did not raise survival rates to equal that of patients with initial small-volume disease. It was suggested that this might be because of the disease being present longer or simply being more aggressive biologically . Our findings suggest that there is little biologic difference between bulky disease which can be resected optimally and that which presents with multiple small metastases. Only 7 of 57 patients (13%) did not require cytoreduction beyond removal of the reproductive organs and omentum. Additionally, the poor outcome of those patients with residual disease measuring more than 1 cm in diameter again suggests the importance of a maximal surgical effort for patients with advanced ovarian cancer. To help better understand this issue, we are currently attempting to correlate the results of flow cytometry with patterns and extent of metastases in patients with advanced ovarian cancer.
These preliminary findings appear to validate retroperitoneal cytoreduction in that patients undergoing removal of macroscopically positive lymph nodes have approximately the same survival as those with microscopically positive and/or negative lymph nodes.
1. Delgado, G., Oram, D.H., and Petrelli, E.S. Stage III ovarian cancer:
The role of maximal surgical cytoreduction. Gynecol. Oncol.
18, 293-298 (1984).