Primary Surgery for Ovarian Carcinoma

The diagnosis and management of ovarian cancer is the most important challenge facing the gynecologic oncologist today, and it is likely to remain so for the foreseeable future. In 2010, the American Cancer Society estimated that ovarian cancer would be diagnosed in 21,880 women in the United States and that 13,850 would die of their disease.¹ Although it accounts for only 26% of new cases of cancer of the female genitalia, it causes 50% of the deaths caused by these malignancies. These grim figures result from the lack of an effective means of early diagnosis for ovarian cancer and the fact that most patients with advanced disease will eventually die of it.

Surgery is a cornerstone of the management of ovarian cancer and plays a crucial role in diagnosis and treatment. However, major advances have been made in chemotherapy for ovarian cancer during the past 10–15 years, and the optimal treatment of a patient with this disease usually requires surgery and chemotherapy. Ovarian cancer is an important example of a condition that requires the skilled application of multimodal therapy.

This chapter provides an overview of the role of primary surgery in the diagnosis and treatment of ovarian cancer. Discussion is limited to the management of epithelial tumors of the ovary, which constitute approximately 90% of all ovarian malignancies. Discussions of the rarer germ cell and stromal malignancies appear elsewhere on this website.

The surgeon who manages ovarian carcinoma must understand as much as possible about the natural history and patterns of spread of the disease. This understanding is as important as a thorough knowledge of anatomy and good surgical technique. The aggressiveness of treatment and the potential for conservation of reproductive function are dependent on the natural history of the disease, and the search for metastatic disease is based on its patterns of spread.

The ovary develops from the genital ridge in the embryo. The genital ridge is composed of thickened coelomic epithelium, which is thought by some authors to have the potential for developing into ovarian-like tumors. This thickened coelomic epithelium forms the mesothelial (epithelial) covering of the ovary and gives rise to epithelial tumors. The germ cells originate in the primitive streak of the embryo and migrate to the gonad, where they proliferate into a component of the ovarian cortex. The mesenchyma of the medulla is the origin of the ovarian stromal cells. These three cell types are thought to give rise to the three types of malignant ovarian neoplasms: epithelial, germ cell, and stromal.^{2, 3}

Epithelial carcinomas, which constitute approximately 90% of malignant ovarian neoplasms,² spread by direct extension to adjacent organs, via lymphatics, and by dissemination of clonogenic cells into the peritoneal cavity. Blood-borne metastases are infrequent. The incidence of bilateral ovarian involvement is highly dependent on the stage of disease, varying from 6% to 13% in apparent stage IA disease to as high as 70% in stage III disease.^{2, 4} Direct extension to the uterus or fallopian tubes occurs in only 5% of patients thought to have stage IA disease, but involvement is frequent in more advanced stages.⁴

The most significant extraovarian spread of epithelial tumors takes place by direct dissemination of clonogenic cells into the peritoneal cavity. Although this dissemination is usually thought to occur when the tumor penetrates the ovarian capsule, the presence of malignant cells in the peritoneal cavity has been documented even with an intact ovarian capsule.⁵ Once the malignant cells enter the peritoneal cavity, they follow the normal circulation of the peritoneal fluid up the right paracolic gutter to the diaphragm. This clockwise flow of peritoneal fluid is the result of the respiratory motion of the diaphragm, and it accounts for the fact that the right hemidiaphragm is more frequently involved than the left. Exfoliated cells are further distributed by peristaltic movements of the intestine, which can result in implantation on any of the peritoneal surfaces, leading to micrometastases that eventually become clinically evident tumor nodules.

Free malignant cells are cleared into lymphatic channels in the diaphragm (especially the right diaphragm) and enter the pleural space and finally the mediastinal lymphatics.⁶ Clinical studies have found occult involvement of the undersurface of the diaphragm in a significant proportion of patients who appeared to have disease confined to the pelvis.^{7, 8}

The omentum is almost always involved in advanced ovarian cancer. Occult omental involvement is seen in 3–11% of patients thought at exploration to have localized disease.^{7, 8, 9} In many cases, the omentum is the only site of extraovarian spread.

The lymphatic vessels of the ovary drain via the ovarian hilus to the subovarian plexus. From the subovarian plexus, the lymphatic drainage follows the course of the ovarian blood supply in the infundibulopelvic ligament to terminate in the high para-aortic lymph nodes. The most frequent site of para-aortic nodal involvement is between the inferior mesenteric artery and the renal vessels. Collateral drainage occurs via the broad ligament to the external iliac and hypogastric lymph nodes and via the round ligament to the inguinal lymph nodes.^{10, 11} The incidence of lymphatic involvement in epithelial ovarian carcinoma appears to be approximately 10–20% in apparent stage I and stage II diseases, and between 60% and 70% in stage III and stage IV diseases.^{9, 12, 13}

In summary, the spread pattern of ovarian carcinoma is characterized by local extension, lymphatic metastases, and intraperitoneal dissemination of tumor cells. These unique routes of spread require that the entire peritoneal cavity be evaluated in any staging procedure for ovarian carcinoma. The potential for intraperitoneal spread must also be taken into account in the treatment plan. Finally, the relatively high incidence of lymphatic spread requires that the retroperitoneum be evaluated, particularly in cases in which the tumor appears clinically to be confined to the ovaries. Nodal sampling must include the high para-aortic lymph nodes that lie between the inferior mesenteric artery and the renal vessels. Both the right and left para-aortic lymph nodes must be evaluated. As initial therapies for ovarian cancer become more effective and patients live longer because of improved control of intraperitoneal disease, distant metastases may be seen more often and may take on greater clinical significance.

Staging systems for human malignancies are designed to provide a standardized assessment of the extent of cancer at the time of diagnosis. The stage of a patient's tumor, assigned at the time of diagnosis, does not change if the disease recurs or progresses. Proper determination of cancer stage is a crucial element in the management of ovarian cancer, because treatment and prognosis are both strongly determined by stage.

In 1964, the General Assembly of the International Federation of Gynecology and Obstetrics (FIGO) approved a staging system for ovarian cancer based on findings at exploratory laparotomy. The staging systems for many other malignancies rely on the results of less invasive diagnostic procedures, such as physical examination and radiographic studies, but diagnostic procedures short of laparotomy are wholly inaccurate in determining the extent of spread of ovarian cancer.

In 1970, the Cancer Committee of FIGO recommended a definitive stage grouping of ovarian cancer based on surgical and pathologic assessment of the extent of disease. This system used four stages, which indicated whether the tumor involved only the ovaries (stage I), other pelvic organs (stage II), abdominal organs (stage III), or distant organs (stage IV). This system incorporated several factors that had recently been recognized to have prognostic significance, including the presence of tumor on the external surface of the ovary, the presence of ascites or peritoneal washings containing malignant cells, and the involvement of retroperitoneal lymph nodes.

In 1985, the Cancer Committee of FIGO¹⁴ revised the surgical staging system for ovarian cancer to reflect not only the results of a thorough exploration of the areas at risk of spread but also a more precise determination of the extent of spread. In this current system, ascites must be cytologically positive to affect the stage. The presence of tumor on the surface of the ovary or the rupture of a localized ovarian tumor, whether spontaneous or during surgery, affects the stage in the same way as positive ascites. Stage III, indicating abdominal spread, has been subdivided into three substages to reflect the size of tumor nodules found in the upper abdomen and the presence of retroperitoneal or inguinal nodal metastases. A pleural effusion must be cytologically positive to permit a diagnosis of stage IV disease. The complete FIGO staging system for ovarian cancer is given in Table 1. Table 2, taken from the FIGO 1991 Annual Report, shows the distribution by stage of more than 8000 patients with epithelial cancers of the ovary reported to FIGO from 95 participating institutions around the world.¹⁵ It can be seen that approximately one quarter of these patients had stage I, and more than half did not have disease diagnosed until stage III or IV was present. The small number of patients listed as “unstaged” includes those who did not undergo exploration because of medical problems or other reasons.

Table 1. FIGO Staging System for Ovarian Cancer¹⁴

Table 2. Distribution by Stage of Ovarian Cancer Patients¹⁵

As outlined in the section on surgical staging, approximately 40% of patients with epithelial cancer have early-stage disease diagnosed.¹⁶ Of these patients, approximately 25% have stage I disease and 15% have stage II disease. In any reported series of ovarian cancer patients, the proportion of women listed as having early-stage disease depends strongly on the adequacy of surgical staging. When full surgical staging is conducted, fewer patients are recorded as having early-stage disease.

A review of the literature reveals a marked variation in reported survival rates for early-stage epithelial ovarian carcinoma. This variability is caused by two factors. First, unless thorough surgical staging is performed, patients with occult advanced disease will be included among early stage patients, resulting in a falsely low survival rate. Second, if patients with stage I and II epithelial ovarian cancer are grouped together, significant variation will occur in the population of patients from one institution to another. Such a discrepancy would result in divergent survival rates that depend on the composition of the various groups within the report.

For the purpose of study design, the Gynecologic Oncology Group (GOG) has divided patients into two early-disease categories (Table 3). It should be emphasized that this classification is applicable only to patients who have undergone rigorous surgical staging. Low-risk disease group includes patients whose tumors are stage IA and stage IB, grade 1 and grade 2. Patients with grade 3 or stage IC carcinoma are grouped into a high-risk category with all patients with stage II who have no residual disease. Patients with residual disease after primary cytoreductive surgery, even if the disease is confined to the pelvis (stage II), are considered to have advanced disease and are not included in either the low-risk or the high-risk early disease categories.

Table 3. Gynecologic Oncology Group Classification of Early-Stage Epithelial Ovarian Carcinoma

*Stage II patients with residual disease at the conclusion of primary cytoreductive surgery are classified as having advanced disease.

In a study reported by Young and co-workers¹⁷ in 1990, the results of two randomized trials of the management of early-stage ovarian cancer by the GOG seem to justify this classification of patients into low-risk and high-risk categories. In this study, all patients underwent complete surgical staging, and most patients underwent second-look laparotomy at the conclusion of therapy. Low-risk patients (randomized to either observation or treatment with oral melphalan) had 5-year survival rates of approximately 95%, with no difference in the treatment arms. High-risk patients (randomized to either intraperitoneal 32P or oral melphalan) had 5-year survival rates of approximately 85%, with no difference in the treatment arms. Thus, the classification of risk categories seemed valid based on thorough surgical staging and previously published survival data.

Surgical therapy for apparent early ovarian cancer begins with a thorough surgical staging (Table 4) to determine the true extent of disease. Staging may be accomplished via laparotomy with adequate exposure, or via minimally invasive techniques such as laparoscopy. Several retrospective studies have compared different routes for staging and have generally noted comparable adequacy and accuracy between laparoscopy and laparotomy, with surgical outcomes in favor of a minimally invasive approach.^{18, 19, 20, 21}

Ultimately, management depends on the stage of the disease, the histologic grade of the tumor, and the patient's age and reproductive plans. For those who have completed childbearing, any diagnosis of ovarian cancer warrants a total hysterectomy and bilateral salpingo-oophorectomy. The benefit of retaining the contralateral ovary for the purpose of preserving ovarian function is outweighed by the risk that carcinoma may subsequently develop or that an occult bilateral carcinoma may be missed in the retained ovary.

Table 4. Surgical Staging Procedure for Apparent Early Ovarian Cancer

*May be preserved in selected patients.

For the patient who wants to retain reproductive potential, however, conservative surgery may be feasible. If the patient has low-risk disease (histologic grades 1 and 2) confined to one ovary, removal of the involved ovary and ipsilateral fallopian tube is sufficient if a full surgical staging procedure is performed. These patients do not require adjunctive therapy, but close observation is necessary because of the increased risk for subsequent development of ovarian carcinoma in the contralateral ovary. Most authorities recommend removal of the remaining reproductive organs when childbearing is complete, because of the increased risk for carcinoma in the retained ovary.

Patients with high-risk disease confined to one ovary (stage IA, grade 3, or stage IC, confined to one ovary) have the option of conservative surgery followed by chemotherapy. However, only a small number of cases have been managed in this manner. Schilder and colleagues²² conducted a multi-institutional retrospective study of 52 patients with stage IA (42 patients) or IC (10 patients) epithelial ovarian cancer treated with conservative surgery. With median follow-up of 68 months, five patients had recurrence. Overall, 50 patients were alive without evidence of disease, and two died of disease 13 and 97 months after initial treatment, yielding an estimated 5-year survival of 98%. Both patients who died of disease had IA disease diagnosed initially, one with grade 1 histology, and the other with grade 2 histology. Of 24 patients who attempted pregnancy, 17 (71%) conceived and had 26 term deliveries and five spontaneous abortions. Park et al.²³ reported on 59 patients with stage I disease (36 IA, 2 IB, 21 IC). Recurrence was noted in 14% of those with IA disease, and 19% of those with IC disease; 5-year overall survival was 95% and 42%, respectively. Given that 5-year survival for stage I epithelial ovarian cancer is as high as 76−93%,^{24, 25} the reported 42% survival among the IC patients in this study is somewhat alarming. Zanetta and colleagues²⁶ reported on 56 women (32 stage IA, 2 stage IB, and 22 stage IC) undergoing fertility-sparing surgery for epithelial ovarian cancer. At 7 years median follow-up, 27 pregnancies had occurred in 20 patients, with five patients experiencing recurrent disease.

In most other cases of high-risk, early-stage disease, preservation of reproductive functioning is impossible because the uterus or both ovaries are involved. Because of the rapid advancement in reproductive techniques, however, partial conservation of the reproductive organs should be considered in any young patient who desires further childbearing. A patient with an intact uterus who has had her ovaries removed can still carry a pregnancy from a donor's egg and her partner's sperm by in vitro fertilization. A woman with a single ovary can produce her own genetic offspring by using her ova and her partner's sperm through in vitro fertilization and the use of a surrogate mother. All of these options must be explained to the patient preoperatively if any possibility of early-stage ovarian cancer exists.

Although most cases of ovarian cancer are considered to be sporadic (i.e., with no discernible familial tendency), it is now recognized that approximately 10% of ovarian cancers occur in association with germline mutations in one of several known cancer predisposition genes.²⁷ In 1994, the BRCA1 gene was identified and cloned.²⁸ In that same year, BRCA2 was localized to chromosome 13q.²⁹  BRCA1 and 2 are involved in homologous recombination in DNA damage repair. The frequency of BRCA1 mutations in the general population is approximately 1 in 833; in Jewish women of Eastern European descent, it is approximately 1 in 100.³⁰

Women carrying a germline mutation of BRCA1 have a significantly increased risk of breast and ovarian cancer compared with the general population. Brose and associates³¹ examined 483 BRCA1 mutation carriers from 147 families and noted the risk for ovarian cancer to be 41% and the risk for female breast cancer to be 73% by age 70. The lifetime probability for ovarian cancer associated with BRCA2 mutations ranges from 15 to 40%. Hereditary ovarian cancer not related to mutations in the BRCA genes may be related to the hereditary nonpolyposis colorectal cancer (HNPCC) genes.²⁷

Risk-reducing bilateral salpingo-oophorectomy should be considered for women with documented mutations in cancer susceptibility genes. Rebbeck and associates³² compared 259 women with BRCA1 or BRCA2 mutations undergoing prophylactic salpingo-oophorectomy with 292 matched controls who did not have the procedure and noted a marked reduction in the risk of primary peritoneal cancer. After excluding six patients with early ovarian cancer diagnosed at the time of prophylactic surgery, only two women (0.8%) had subsequent primary peritoneal carcinoma in the group undergoing oophorectomy compared with 58 women (19.9%) having subsequent ovarian cancer in the control group.

A later meta-analysis demonstrated an 80% risk reduction after bilateral salpingo-oophorectomy in BRCA mutation carriers.³³ A decision analysis by Kurain et al. suggests age 40 as the most effective age for prophylactic surgery.³⁴ Given that up to 12% of women undergoing risk-reducing surgery have occult disease detected, care must be taken during the procedure to examine all abdominal and pelvic peritoneal surfaces, and to biopsy any suspicious lesions. Given the risk of fallopian tube carcinoma, complete resection of all visible fallopian tube should be performed.

Despite decades of research directed at the problem of early diagnosis of ovarian cancer, most cases are still not detected before metastasis. Although the development of serum tumor markers³⁵ and sensitive imaging techniques such as vaginal color-flow sonography³⁶ have focused attention on the problem of early detection, neither method has been proved effective as a screening tool for the general population or even for a high-risk population.

As a result of the lack of effective methods for early diagnosis and the fact that ovarian cancer produces no specific symptoms in its early stages, patients often present with obvious ascites and palpable tumor masses in the pelvis and upper abdomen. On exploration, it is often not possible to resect all gross tumor, and such disease certainly cannot be cured by surgery alone. The surgeon confronted with such a situation must decide what level of surgical aggressiveness is appropriate. For many human tumors, aggressive surgery can be justified only if all known tumor can be removed, thus making the operation at least potentially curative. For epithelial cancer of the ovary, however, there is considerable theoretic and clinical evidence that cytoreduction, or debulking, of large tumor masses may be beneficial to the patient even if all gross tumor cannot be removed.

The theoretic evidence in support of debulking comes largely from concepts of the kinetics of tumor cell growth and the mechanisms of tumor cell destruction by cytotoxic chemotherapy. For diseases in which chemotherapy is essentially ineffective, such as certain gastrointestinal malignancies, the concepts of cytoreduction do not hold. In ovarian cancer, however, the cancer cells are usually sensitive to chemotherapy, and the removal of large tumor masses may improve the response to subsequent chemotherapy. The blood supply of these masses may be relatively poor, especially centrally. Although poor blood supply can ultimately cause central necrosis of the tumor mass, it can also lead to the development of adjacent areas of viable tumor that cannot be reached by adequate concentrations of chemotherapeutic agents. Additionally, cells in masses with a poor blood supply tend to have a low growth fraction and a higher proportion of cells in the nonproliferating (Go) phase of the cell cycle, making them relatively insensitive to the effects of chemotherapy.³⁷ The removal of large tumor masses may actually decrease the number of cycles of chemotherapy required to eliminate all remaining tumor cells.

Skipper and associates³⁸ demonstrated the phenomenon of fractional cell kill, which applies to tumors that grow in an exponential fashion. In this phenomenon, a given dose of a cytotoxic agent will kill a constant fraction of the tumor, rather than a constant number of tumor cells, regardless of the initial cell population. Although solid tumors such as ovarian cancer do not grow in a true exponential fashion, it is still likely that surgical reduction of tumor burden may reduce the number of cycles of chemotherapy that are needed. The likelihood of acquired chemoresistance caused by exposure of the tumor cells to the cytotoxic agent will therefore be decreased, as will the risk for chemotherapy-related toxicity. According to the mathematical model of Goldie and Coldman,³⁹ in addition to increasing the response to chemotherapy, effective cytoreduction may improve the patient's comfort by improving or restoring intestinal function and decreasing the formation of ascites. The patient will be better able to maintain adequate nutrition and reduce the adverse metabolic consequences of the tumor, which in turn can increase her ability to withstand intensive combination chemotherapy. Blythe and Wahl⁴⁰ reported improvement in the quality of life of patients who had undergone successful cytoreductive surgery. These patients were more likely to be able to continue their normal activities, including employment, compared with the patients who did not undergo aggressive cytoreduction. Nearly 80% of the patients who underwent optimal cytoreduction were able to tolerate a regular diet, compared with only 40% of patients who had suboptimal surgery. Morton⁴¹ proposed that debulking might be beneficial in terms of decreasing the immunosuppression caused by the tumor.

Clinical studies that support debulking now span more than three decades and include patients treated with radiotherapy, nonplatinum-based chemotherapy, and platinum-based combination regimens, suggesting that all of these treatments are more effective in patients with small tumor burdens. In 1968, Munnell,⁴² who developed the concept of “maximal surgical effort” for ovarian cancer, reported that patients who underwent a “definitive operation” had an improved survival rate compared with patients who had “partial removal” or “biopsy only.” In a 1969 report, Elclos and Quinlan⁴³ reported that 25% of patients with stage II ovarian cancer whose disease was surgically reduced to “nonpalpable” survived, compared with 9% of those left with “palpable” disease.

The measures of cytoreduction and residual disease used in these early studies were obviously quite crude. Griffiths,⁴⁴ in a seminal report published in 1975, was the first to quantify accurately the amount of residual disease after the primary operation for ovarian cancer and to attempt to correlate this with outcome after chemotherapy. Reporting on 102 patients who received alkylating agent chemotherapy for stage II or stage III ovarian cancer, he demonstrated with the use of a linear regression model that survival duration was significantly related to the size of the residual tumor at the completion of primary debulking. In his report, patients with no residual tumor had a median survival of 39 months, compared with 12.7 months for patients with residual tumor of greater than 1.5 cm in maximum diameter. Griffiths also made the important observation that aggressive resection of bulk tumor without removal of all tumor masses greater than 1.5 cm in diameter did not improve survival.

Subsequent reports of clinical trials of both nonplatinum-based combination regimens⁴⁵ and platinum-based regimens⁴⁶ have supported the concept of debulking. Reports from many institutions have evaluated the effect of surgical cytoreduction in ovarian cancer. Patients undergoing optimal cytoreduction are generally shown to have an increased chance of negative surgical reexploration after chemotherapy,⁴⁷ an increase in median survival,⁴⁸ and a decrease in the rate of recurrence after a negative second-look operation.⁴⁹Table 5 shows a compilation of 10 reports,^{50, 51, 52, 53, 54, 55, 56, 57, 58, 59} comprising a total of 1012 patients, that demonstrates the effect of the amount of residual disease after primary cytoreduction on the frequency of negative surgical reexploration after chemotherapy (second-look laparotomy). Table 6, based on data from nine reports,^{60, 61, 62, 63, 64, 65, 66, 67, 68} shows the effect of the amount of residual disease on median survival in patients with advanced ovarian cancer treated with chemotherapy.

Table 5. Effect of Amount of Residual Disease After Primary Cytoreduction on Findings at Second-Look Laparotomy

*As defined by each author.

Table 6. Survival of Patients With Advanced Ovarian Cancer According to the Amount of Residual Tumor After Primary Cytoreduction

*As defined by each author.

To evaluate the effect of residual disease on survival, Hoskins and colleagues⁶⁹ reviewed data on 458 patients treated in accordance with a GOG protocol that evaluated standard versus high-dose chemotherapy with cisplatin and cyclophosphamide in suboptimal (residual disease greater than 1 cm) stage III and stage IV epithelial ovarian cancer. They found that patients with residual disease of 1–2 cm had better survival than patients with disease greater than 2 cm (Fig. 1). They found that cytoreduction that did not achieve residual disease of less than 2 cm had little effect on survival. When they combined the results of this analysis with an analysis of a similar GOG protocol (cisplatin and cyclophosphamide versus a cisplatin, cyclophosphamide, and doxorubicin combination) that enrolled patients with optimal (residual disease less than 1 cm) stage III epithelial ovarian cancer, they found that survival at 48 months was 58% for microscopic disease, 34% for residual gross disease up to 2 cm in diameter, and 16% for patients with residual disease greater than 2 cm (Fig. 2).

The finding of Hoskins and co-workers that disease that could not be cytoreduced to less than 2 cm in diameter was associated with uniformly poor survival and should be viewed in relation to a study of interval cytoreductive surgery by the European Organization for Research and Treatment of Cancer reported in 1995 by Van der Burg and associates.⁷⁰

These authors randomized patients with suboptimal disease that could not be cytoreduced to six cycles of cisplatin and cyclophosphamide versus three cycles of the same chemotherapy followed by interval cytoreductive surgery. The surgical patients then received an additional three cycles of cisplatin and cyclophosphamide. They found a median survival of 26 months for patients undergoing interval cytoreduction versus 20 months for those who did not (p = 0.012) (Fig. 3).

Patients who present with suspected advanced ovarian cancer should undergo the routine preoperative tests before any major operation, including a complete blood count, biochemical profile, chest radiograph, and electrocardiogram. In addition, special radiographic studies such as computerized tomography, ultrasonography, or magnetic resonance imaging may be helpful in confirming the presence of a mass and identifying ascites or other evidence of metastatic disease. Intestinal imaging studies or upper and lower endoscopy is performed if there is clinical suspicion of a gastrointestinal primary tumor. Most gynecologic oncologists use a regimen of mechanical and antibiotic bowel preparation to cleanse the intestine in case intestinal resection is necessary or the bowel is entered inadvertently. Many gynecologic oncologists also use a regimen of low-dose subcutaneous heparin or intermittent calf compression devices to prevent deep vein thrombosis and perioperative intravenous antibiotics to decrease infections.

Before surgery, the patient and her family should be counseled thoroughly about the risk of malignancy and the surgical procedures that might be necessary. If there is a suspicion of rectal involvement, the possibility of low rectal anastomosis or colostomy should be mentioned. The patient should also understand that although every attempt will be made to remove as much tumor as possible, advanced ovarian cancer cannot be cured by surgery alone, and chemotherapy will be necessary.

The abdomen should be opened through a vertical incision extending from the pubic symphysis to well above the umbilicus to allow access to the upper abdominal structures. Before undertaking any pelvic operation, a complete exploration of the upper abdomen should be performed, with particular attention to the hemidiaphragms and omentum. The stomach and the small and large intestines should be evaluated, as should the retroperitoneal structures, including the lymph nodes, pancreas, and kidneys. An estimate of the resectability of upper abdominal disease may be important in determining the appropriate procedure for dealing with the pelvic tumor. In evaluating the pelvis, the surgeon must consider whether the primary tumor in the ovaries can be completely removed, the extent to which hysterectomy would contribute to debulking, and the resectability of any areas of pelvic spread. An experienced gynecologic oncologist should be able to assess the extent of possible tumor removal and to decide which operative procedures would be most appropriate for cytoreduction.

In some cases, an omental tumor can be removed adequately by an infracolic omentectomy with division of the omental attachments to the transverse colon. If the tumor involves the gastrocolic omentum, the dissection will have to be carried upward to detach the omentum from the greater curvature of the stomach. Care must be taken to avoid injury to the mesentery of the transverse colon, to which the omentum is often adherent, and to the gastroepiploic vessels along the greater curvature of the stomach wall. Omental tumors often extend toward the spleen. An appropriately large incision usually allows resection of the tumor without the need for a splenectomy; however, a splenectomy is occasionally indicated. If the omental tumor infiltrates into the wall of the transverse colon, resection of the colon may be necessary to allow optimal debulking. Bulky tumor nodules on the diaphragm are often the limiting factor in achieving optimal cytoreduction, although to some extent they can be debulked sharply or with the use of the ultrasonic aspirator⁷¹ or argon beam coagulator.⁷² If resection of small areas of residual disease on the diaphragm will render the patient's disease optimal, small areas of the diaphragm can be resected with primary repair. It is usually necessary to mobilize the liver to perform this procedure. Although a chest tube may be required after resection of a portion of the diaphragm, one can often insert a catheter into the chest before tying the last stitch; with the end of the tube under water, the anesthesiologist can expel the air in the chest with a deep inspiration. Isolated nodules on the small intestine can be resected.

In the pelvis, bulky tumor often distorts or eliminates normal intraperitoneal tissue planes and anatomic landmarks. The retroperitoneal tissue planes, however, are usually well preserved and should be used as primary landmarks during the pelvic operation. The retroperitoneal space can usually be entered via incision of the peritoneum lateral to the infundibulopelvic ligament or the colon. After development of the space and direct identification of the ureter, the infundibulopelvic ligament can be ligated and transected. Continuing the retroperitoneal dissection, the uterine artery can also be ligated. After this is repeated contralaterally, mobilization of the ovarian tumors can proceed with less blood loss. In the case of a large tumor adherent to the pelvic sidewall, removal may require dissection of the ureter from the mass. If the tumor involves the lateral aspect of the uterus, one can perform the type of dissection used in a modified radical hysterectomy, unroofing the ureter as it passes through the parametrial tunnel and ligating the uterine vessels at this point. Extensive involvement of the rectosigmoid may require en bloc resection along with hysterectomy, as is performed in a posterior exenteration. Intestinal continuity can usually be reestablished by low anastomosis using intestinal stapling instruments. A colostomy is rarely indicated. Keeping in mind the observation first made by Griffiths that unless “optimal” cytoreduction can be accomplished the outcome is not improved, the surgeon should be reasonably certain that all bulky disease can be resected before undertaking an aggressive procedure.

Management of the pelvic and aortic lymph nodes in advanced ovarian cancer is not well defined. Although most patients with advanced ovarian cancer have involvement of pelvic or paraaortic nodes,^{73, 74} most gynecologic oncologists agree that there is no need for routine diagnostic node sampling in patients with obvious intraperitoneal dissemination, because the information gained will have no bearing on treatment or prognosis. Whether an extensive lymphadenectomy has any effect on survival is unknown. One group has reported a retrospective series in which patients undergoing lymphadenectomy had an improved survival rate compared with historical controls.⁷⁵ These data await confirmation. If the resection of bulky nodal disease allows optimal cytoreduction, it should be performed. It is our practice to remove enlarged lymph nodes in patients able to undergo optimal resection of intra-abdominal tumor.

The closure of the abdominal incision in advanced ovarian cancer patients deserves special attention. These patients are often elderly, obese, and nutritionally depleted, and they may have concomitant medical problems, such as diabetes, that interfere with wound healing and increase the risk of dehiscence. This closure may be performed with interrupted sutures or with a continuous suture using nonabsorbable or delayed absorbable material.^{76, 77} Most studies show improved results with continuous mass closure techniques.

A number of centers have reported their experiences with the primary debulking of advanced ovarian cancer. Table 7 shows data from 11 reports,^{64, 67, 78, 79, 80, 81, 82, 83, 84, 85} published since 1979, on a total of 2282 patients. The great variation among these individual studies in the ability to achieve optimal cytoreduction (range 17–87%) must in part be explained by patient selection, varying degrees of surgical skill and aggressiveness, and different definitions of optimal cytoreduction. In general, skilled and experienced gynecologic oncologists should be able to accomplish optimal debulking in at least one-half of their patients with advanced ovarian cancer.

Table 7. Optimal Cytoreduction at Primary Surgery in Stage III And Stage IV Ovarian Cancer

Although some authors have questioned the value of debulking surgery in patients with stage IV ovarian cancer, reports from several institutions, including Memorial Sloan-Kettering and the University of Pennsylvania,^{87, 88, 89, 90} indicate a significant survival benefit to optimal cytoreduction in stage IV patients. Data from these studies are summarized in Table 8. These authors concluded that attempted cytoreduction was indicated in all patients with advanced ovarian cancer, even in patients with stage IV disease.

Table 8. Effect of Debulking on Survival in Stage IV Ovarian Cancer

The morbidity of aggressive cytoreductive surgery appears to be within acceptable limits, based on publications from major centers. Heintz and co-workers⁸² reported that with increasing experience, improved cardiovascular monitoring, and the increased use of parenteral hyperalimentation, the risk of serious perioperative morbidity decreased to 10%, the most commonly encountered problem in their series being postoperative pneumonia.

Although patients with small-volume residual tumor at the initiation of primary chemotherapy have an improved prognosis, these patients may have had smaller, less aggressive tumors to begin with. Their improved prognosis may relate more to the biology of the tumor than to the extent of the surgery. Several authors have attempted to address this issue by retrospective studies that examine the prognostic significance of the extent of disease before and after cytoreduction. In a report by Hacker and co-workers,⁹¹ patients with large-volume metastatic disease (greater than 10 cm) were analyzed according to the success of cytoreduction. Of the patients with bulky disease, those for whom tumor was reduced to optimal status had a longer median survival than those for whom it was not. However, patients who started with smaller volume disease and achieved optimal status had an even longer survival. In this series, published in 1983, Hacker and associates used only a single alkylating agent as the chemotherapy after surgery. In a 1988 series, Heintz and colleagues⁹² reported on the effect of preoperative tumor size in patients treated with cisplatin-based combination chemotherapy. They observed that optimal tumor resection in patients with large-volume metastatic disease improved median survival by approximately 12 months. Once again, the survival of patients with optimal cytoreduction (smallest residual disease) who started with large-volume disease was not as good as that for their counterparts who started with smaller volume disease.

In an attempt to evaluate the relationship between cytoreductive surgery and tumor biology, Hoskins and colleagues⁹³ reviewed 349 patients with stage III optimal (less than 1 cm) epithelial ovarian cancer. They divided patients into those found to have disease less than 1 cm in diameter on exploration and those who required cytoreductive surgery to achieve disease less than 1 cm. They theorized that if only surgery was important, survival would be equal between the two groups. They observed that patients found to have small-volume abdominal disease had a better survival than those who achieved small-volume residual disease via debulking (Fig. 4) (p = 0.0001). When evaluated by multivariate analysis, a variety of factors in addition to cytoreduction were found to be significant. The authors concluded that although cytoreduction was important, the biology of the tumor also influenced outcome. These studies suggest that although the finding of large-volume metastatic disease confers a poorer prognosis, and although the biology of the tumor is important, optimal tumor resection in these patients can improve median survival.

No prospective study randomizing ovarian cancer patients to aggressive cytoreduction versus less aggressive surgery has ever been performed, nor is such a study likely to be completed in the future. The GOG attempted such a trial with its protocol 80, but the trial came to a halt because of slow patient accrual; this failed attempt also was a reflection of the strong belief among gynecologic oncologists in the benefits of debulking.

The use of neoadjuvant chemotherapy is emerging as an attractive option in management of certain patients with either bulky unresectable disease, or those with medical morbidity. Patients initially undergo chemotherapy and then interval surgical cytoreduction. Retrospective analyses have demonstrated equivalent overall survivals when comparing patients treated with primary chemotherapy followed by interval debulking with conventionally treated patients.^{94, 95} In 2010, Vergote et al.⁹⁶ reported results of the first randomized control trial of 670 patients assigned to either neoadjuvant chemotherapy followed by debulking or traditional debulking surgery followed by adjuvant platinum-based chemotherapy. As expected, postoperative morbidity and mortality were higher on the primary surgery arm.  Overall and progression-free survival were statistically similar between the two groups. Notably, complete cytoreduction of all visible disease was the strongest independent factor influencing survival, whether performed at initial surgery or at interval surgery: optimal debulking was achieved in 80.6% and 41.6% of patients, respectively.

Few procedures offer as great a challenge to the skill and judgment of the surgeon as the cytoreduction of advanced ovarian cancer. Because optimal cytoreduction is such an important part of treatment, we believe that patients with suspected advanced ovarian cancer should be referred to a center where trained and experienced gynecologic oncologists are available.