Cytoreductive surgery for patients with advanced epithelial ovarian cancer has been practised since the pioneering work of Tom Griffiths in 1975. Further research has demonstrated the prognostic significance of the extent of metastatic disease pre-operatively, and of complete cytoreduction post-operatively. Patients with advanced epithelial ovarian cancer should be referred to high volume cancer units, and managed by multidisciplinary teams. The role of thoracoscopy and resection of intrathoracic disease is presently investigational. In recent years, there has been increasing use of neoadjuvant chemotherapy and interval cytoreductive surgery in patients with poor performance status, which is usually due to large volume ascites and/or large pleural effusions. Neoadjuvant chemotherapy reduces the post-operative morbidity, but if the tumour responds well to the chemotherapy, the inflammatory response makes the surgery more difficult. Post-operative morbidity is generally tolerable, but increases in older patients, and in those having multiple, aggressive surgical procedures, such as bowel resection or diaphragmatic stripping. Primary cytoreductive surgery should be regarded as the gold standard for most patients until a test is developed which would allow the prediction of platinum resistance pre-operatively.
Highlights
- •
Primary cytoreductive surgery is the treatment of choise for most patients with advanced epithelial ovarian cancer.
- •
The goal of surgery should be to cytoreduce the cancer to no macroscopic residual disease.
- •
Significant survival benefit still accrues if the cancer can be cytoreduced to individual nodules ≤1 cm in diameter.
- •
There are usually quality of life benefits if the primary tumour and a large omental cake can be removed.
- •
Neoadjuvant chemotherapy decreases postoperative morbidity in patients who have a poor performance status.
Ovarian cancer is the seventh most common cancer in women worldwide, with 239,000 new cases estimated to have been diagnosed in 2012. About 65,000 cases occurred in Europe. The incidence rate for the United Kingdom is 11.7 per 100,000 women per year.( www.cancerresearchuk.org ).
About 70% of patients with epithelial ovarian cancer present with advanced disease, because of the lack of any satisfactory screening test and the lack of early symptoms. These patients will require a combination of cytoreductive surgery and chemotherapy to give them their best chance of long term survival. Cytoreductive or debulking surgery entails bilateral salpingo-oophorectomy, typically along with hysterectomy, gastrocolic omentectomy, and resection of as much other gross metastatic disease as possible.
With advancements in ovarian cancer surgery and modern chemotherapy, median and overall survivals for patients with even advanced ovarian cancer have improved over the last 15 years. Using data on 40,692 patients from the Surveillance, Epidemiology and End Results (SEER) database taken from 1995 to 2007, Baldwin et al reported that patients with FIGO stages III and IV epithelial ovarian cancer had relative 5-year survivals of 36% and 17% respectively, and relative 10-year survivals of 23% and 8% respectively ( Figure 1 ) . Most of these patients are not disease-free, but some of them are, and it is not possible to predict which patients will apparently be cured of advanced disease with currently available prognostic factors. This will require genetic profiling of all patients in the future.
International differences in ovarian cancer survival are wide, even between high-income countries with similar health systems. The International Cancer Benchmarking Partnership obtained data from population-based cancer registries in Australia, Canada, Denmark, Norway and the United Kingdom, and analysed 20,073 women who were diagnosed between 2004 and 2007 . For patients with FIGO stages III and IV, women in the UK had lower one-year survival than women in the other four countries (61.4% vs 65.8-74.8%). This was particularly so for older women, and the authors suggested that the management of these women should be investigated.
Historical perspective
Although earlier surgeons had recommended “maximal surgical effort” when operating on patients with advanced ovarian cancer , the first to quantify the benefits of aggressive surgery was Griffiths in 1975 . In a retrospective series of 102 patients treated at the Boston Hospital for Women and the Sidney Farber Cancer Institute, Griffiths reported improved survival if what he termed “optimal“ cytoreduction could be performed, meaning no residual disease >1.5 cm in diameter. He confirmed these results in a subsequent prospective study of 26 consecutive patients . Griffith claimed that patients who started with large metastatic disease fared as well as those whose largest metastatic lesions were below 1.5 cm at the outset.
The next contribution to the discussion was from Hacker et al, at the University of California at Los Angeles (UCLA) . They confirmed the importance of small residual disease, but showed that patients who were cytoreduced to individual nodules not greater than 5 mm in diameter had an even better prognosis. However, their data refuted the concept that patients who started with large metastatic disease did as well as those who started with small metastatic disease, and demonstrated for the first time the independent prognostic significance of tumour biology. Within the group having “optimal” cytoreduction, patients who had >1000 cc ascites or metastatic nodules >10 cm diameter had a significantly worse prognosis. The prognostic importance of large metastatic disease is consistent with the Goldie Coldman hypothesis, a mathematical model which assumes that tumours have a spontaneous mutation rate which leads to drug resistance . Larger tumour metastases have, by necessity, gone through more cellular divisions, so would be more likely to harbour chemo resistant clones of cells.
In 1992, Hoskins et al undertook a retrospective review of Gynecologic Oncology Group (GOG) Protocol 52, a randomised trial of two different chemotherapeutic regimes for patients with residual disease of 1 cm or less . There was no difference in the outcome for the 349 evaluable patients. The paper confirmed that a large initial burden of metastatic disease was a poor prognostic factor, and demonstrated, in addition, that the greater the number of residual tumour nodules, (carcinomatosis), the worse the prognosis ( Figure 2 ). A second UCLA paper by Farias-Eisner et al in 1994 confirmed that carcinomatosis was a poor prognostic factor, and in addition, reported for the first time that patients with no residual disease (18 patients) had the best median survival (56.5 months) .
A recent review of 13 papers and 11,999 patients with stages III or IV ovarian cancer determined that the median overall survival for patients with no residual disease was 70 months, compared to 53 months for patients with 1-5 mm, 40 months for patients with 1–10 mm and 30 months for patients with > 10 mm (p < 0.001) . They concluded that although complete cyoreduction should be the aim of cytoreductive surgery, there is still significant benefit to trying to achieve “optimal” residual disease status (i.e., < 1 cm), because such patients have a 10 months longer median overall survival compared to patients with suboptimal residual disease.
Preoperative workup
Most patients with ovarian cancer are investigated with a computed tomographic (CT) scan of the pelvis and abdomen and a chest x-ray, or a CT scan of the chest, pelvis and abdomen. CT has shown accuracy ranging from 60 to 90% in staging the extent of disease, but low sensitivity in the small bowel mesentery and the subdiaphragmatic space .
More recently, 18 F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) has been used to determine the true extent of disease in patients with advanced ovarian cancer . A Finnish study of 30 patients reported supradiaphragmatic lymph node metastases in 20 patients (67%) using FDG PET/CT, compared to only 10 patients (33%) using the conventional CT scan. The location of the positive nodes was parasternal in 14 patients (70%), cardiophrenic in 14 (70%), other mediastinal in 8 (40%), axillary in 6 (30%) and subclavian in 1 (5%). Microscopy of all 4 biopsied lymph nodes (3 axillary and 1 subclavian) confirmed metastatic carcinoma. Patients with supradiaphragmatic nodal metastases had significantly more ascites (p < 0.01) and subdiaphragmatic carcinomatosis (p < 0.03) on preoperative FDG PET/CT, so were a poor prognosis group. The authors felt that further studies were warranted to determine the clinical relevance of their findings, and certainly the cost-effectiveness of a preoperative PET/CT on all patients would need to be studied.
Preoperative workup
Most patients with ovarian cancer are investigated with a computed tomographic (CT) scan of the pelvis and abdomen and a chest x-ray, or a CT scan of the chest, pelvis and abdomen. CT has shown accuracy ranging from 60 to 90% in staging the extent of disease, but low sensitivity in the small bowel mesentery and the subdiaphragmatic space .
More recently, 18 F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) has been used to determine the true extent of disease in patients with advanced ovarian cancer . A Finnish study of 30 patients reported supradiaphragmatic lymph node metastases in 20 patients (67%) using FDG PET/CT, compared to only 10 patients (33%) using the conventional CT scan. The location of the positive nodes was parasternal in 14 patients (70%), cardiophrenic in 14 (70%), other mediastinal in 8 (40%), axillary in 6 (30%) and subclavian in 1 (5%). Microscopy of all 4 biopsied lymph nodes (3 axillary and 1 subclavian) confirmed metastatic carcinoma. Patients with supradiaphragmatic nodal metastases had significantly more ascites (p < 0.01) and subdiaphragmatic carcinomatosis (p < 0.03) on preoperative FDG PET/CT, so were a poor prognosis group. The authors felt that further studies were warranted to determine the clinical relevance of their findings, and certainly the cost-effectiveness of a preoperative PET/CT on all patients would need to be studied.
Complete cytoreduction
The first person to set out to try to resect all gross disease in patients with advanced ovarian cancer was Eisenkop . Between 1990 and 1996, his team operated on 163 consecutive patients with stage IIIC or IV disease with the intension of excising or ablating all visible disease. No patients were excluded on the basis of age, performance status or extent of disease. Complete cytoreduction was achieved in 139 patients (85.3%), 22 (13.5%) had cytoreduction to ≤ 1 cm, and 2 (1.2%) had bulky residual disease. Significant variables determining the cytoreductive outcome were performance status, stage of disease (IIIC vs IV), and the number of mesenteric and serosal implants.
In order to achieve complete cytoreduction, 85 patients (52.1%) had a modified posterior exenteration, 32 (19.6%) had an extra-pelvic bowel resection, 66 (40.5%) had diaphragmatic stripping and/or resection, 145 (89%) had peritoneal implant ablation, 152 (93.2%) had a retroperitoneal lymph node resection, and 31 (19%) had such miscellaneous operations as splenectomy, liver resection, and distal pancreatectomy ( Figure 3 ). The mean operating time was 254 minutes (range 75-435) and the mean blood loss was 1190 ml (range 100-6000). Three patients (1.8%) died within 30 days of surgery.
Median follow-up for the group was 27 months. The median survival for patients with no residual disease was 62.1 months, compared to 20 months for those with any residual disease ( p = 0.001). Significant factors for poor survival in multivariate analysis were age >61 years ( p = 0.003), FIGO stage IV disease ( p = 0.04), ascites >1000 ml ( p = 0.01), any residual disease ( p = 0.02), and mucinous or clear cell histology ( p = 0.03).
The prognostic importance of complete cytoreduction was confirmed by Du Bois et al in a retrospective review of 3126 patients who had been entered onto three randomized European chemotherapy trials (AGO-OVAR 3, 5 and 7) . No residual disease was present in 1046 patients (33.5%) Group A), 1-10 mm in 975 (31.2%) (Group B) and >10 mm in 1105 (33.3%) (Group C). Median survival was 99.1 months for patients in Group A, compared to 36.2 months for Group B and 29.6 months for Group C ( p = < 0.0001). Multivariate analysis revealed that other independent factors for overall survival were age, performance status, grade, FIGO stage and mucinous histology. They concluded that complete cytoreduction improved the prognosis in any substage, but could not completely overrule the prognostic impact of preoperative tumour burden (as indicated by the FIGO stage).
Until better chemotherapy or targeted therapy becomes available, better surgery is the only variable that is amenable to improvement. Harter et al showed the benefits of introducing a specific ovarian cancer quality management program into their hospital in 2011. The main aspects of the program were the formation of dedicated surgical teams, and the restriction of ovarian cancer surgery to those teams . The rate of cytoreduction to no macroscopic residual disease increased from 33% for the period 1997-2000 (n = 57) to 62% for the period 2004-2008 (N = 259). The utilization of extended surgical procedures increased over time, and the median overall survival increased from 26 months in 1997-2000 to 45 months in 2004-2008 (p<0.003).
It is virtually always possible for an experienced surgeon to remove the pelvic disease, but this will sometimes require rectosigmoid resection, which can be achieved with low morbidity in the absence of any prior chemotherapy or radiation therapy. Peiretti et al retrospectively reviewed the charts of 238 patients who underwent rectosigmoid resection for advanced ovarian, fallopian tube or peritoneal cancers between 1998 and 2008 at the European Institute of Oncology, Milan and The Johns Hopkins Medical Institutions in Baltimore . Complete cytoreduction was achieved in 41% of cases. Stapled reanastomoses were performed in 98% of cases, and a protective stoma was used in only 2.9% of cases. Morbidly was low, with only 7 patients (3%) developing an anastomotic leak, and 9 patients (3.7%) a pelvic abscess.
Two or more bowel resections may be justified if it will allow the patient to be optimally cytoreduced. Salani et al identified 34 patients who underwent ≥ 2 bowel resections at The Johns Hopkins Medical Institutions in Baltimore, from 1997 to 2006. Although estimated blood loss was higher and length of postoperative hospitalization longer in these patients, the only significant factor with respect to overall survival was the amount of residual disease .
Aggressiveness of the surgeon
A 2006 study from the Mayo Clinic retrospectively examined 194 consecutive patients with stage IIIC ovarian cancer who underwent primary surgical exploration between 1994 and 1998, in order to correlate the rate of optimal cytoreduction (defined as residual disease less than 1 cm) with clinico-pathological variables and the aggressiveness of the surgeon . They reported that the only independent variables were the American Society of Anesthesiology (ASA) score, the presence of carcinomatosis, and the aggressiveness of the surgeon. Factors which were not significant included the age of the patient, the CA125 titre, the presence of ascites, and the presence of diaphragmatic or mesenteric involvement.
A subsequent analysis of the same group of patients demonstrated that disease-specific survival was markedly improved for patients with carcinomatosis if they were operated on by surgeons who most frequently used radical procedures compared to those least likely to use radical procedures (5-year disease-specific survival rates 44% versus 17% respectively; p < 0.001) .
Role of lymphadenectomy
Several retrospective studies have suggested that lymphadenectomy may improve survival for patients with advanced ovarian cancer .
Du Bois et al examined the impact of systematic pelvic and paraaortic lymphadenectomy on patients entered onto the same three European randomized chemotherapy trials that were used to validate the prognostic role of complete cytoreduction . There were 1924 patients eligible for analysis, which represented 57.3% of the originally randomized cohort. The rate of pelvic and paraaortic lymphadenectomy ranged from 68% in patients with clinically suspicious nodes and no gross residual disease to 16.3% in patients without suspicious nodes and small residual disease. In patients with no residual disease, lymphadenectomy increased median survival time from 84 to 103 months, and the 5-year survival rate from 59.2% to 67.4% ( p = 0.0166). For patients with small residual disease, lymphadenectomy made no significant difference. The authors concluded that a randomized, prospective trial was required for patients with complete cytoreduction.
The only randomized, prospective trial to investigate the role of lymphadenectomy in patients with advanced ovarian cancer was a multicentre study reported by Benedetti-Panici in 2005 . Patients who had residual intraperitoneal nodules ≤ 1 cm were randomized to either systematic pelvic and paraortic lymphadenectomy (n = 216) versus resection of bulky nodes only (n = 211) ( Figure 4 ). Both arms were well matched for clinical variables, and to be eligible for the study, patients in the lymphadenectomy arm had to have a minimum of 15 nodes removed from the pelvis and 10 nodes from the paraortic area. With a median follow-up of 68.4 months, there was a 6-month benefit in terms of progression-free interval ( p = 0.02), but no benefit in terms of overall survival ( Figure 5 ). Systematic lymphadectomy increased the median operating time by 90 minutes, the transfusion rate by 12%, and increased the incidence of lymphocysts and lower limb lymphoedema.
Role of thoracoscopy
The first report of thoracoscopy in the management of patients with advanced ovarian cancer was from Eisenkop in 2002 . Between 1998 and 2000, 30 patients underwent thoracoscopy at the time of primary cytoreductive surgery. Four procedures were performed through the right chest wall by a thoracic surgeon, and the remaining 26 patients underwent trans-diaphragmatic thoracoscopy by a gynaecologic oncologist at the time of laparotomy. Among 24 patients with stage IV disease, 11 (45.8%) had no macroscopic pleural disease, 10 (41.7%) underwent pleural implant ablation with the Argon Beam Coagulator and /or excision, as well as nodal excision, and 3 (12.5%) had an effort to achieve complete intra-abdominal cytoreduction abbreviated after unresectable intra-thoracic disease was found. In a crude comparison with historical controls, Eisenkop claimed significantly better survival for the group having thoracoscopy.
Subsequently, other groups have used video-assisted thorascopic surgery (VATS) to investigate the extent and resectability of intrathoracic disease prior to abdominal exploration .
Diaz et al reviewed the experience at Memorial Sloan-Kettering Hospital of all patients with untreated advanced ovarian cancer and moderate to large pleural effusions who underwent VATS between 2001 and 2008. There were 42 evaluable patients, and the primary surgeon in all cases was a thoracic surgical oncologist. VATS revealed macroscopic disease in 29 patients (69%), with nodules > 1 cm in 18 cases (43%). Of these 18 patients, 6 (33%) had intrathoracic cytoreduction. Patients who had optimal intrathoracic disease (≤1 cm) de novo, or those who could be cytoreduced to optimal status in the thorax, went on to have abdominal cytoreduction, while patients who had disease >1 cm in the thorax were given neoadjuvant chemotherapy .
A subsequent German study evaluated the use of VATS to improve the accuracy of FIGO staging and to assess operability more reliably than using imaging alone . Their median operating time was 46.5 minutes, there was negligible morbidity, and for them, VATS altered the therapeutic management in 6 of 17 patients (35%).
Cytoreductive surgery in patients with stage IV disease
The data on complete surgical resection of all metastatic disease is based on patients with stage III ovarian cancer. The data regarding patients with stage IV disease is less clear, because most studies are heterogeneous and retrospective , and the exact extent of any residual disease in the thorax is not usually known.
Several studies have nevertheless shown a significant survival advantage when optimal residual disease status could be achieved in the pelvis and abdomen. Aletti et al reviewed 109 patients with stage IV ovarian cancer treated at the Mayo Clinic, of whom 58 had positive pleural cytology only . In the latter group, patients having optimal intra-abdominal cytoreduction (≤1 cm) had a median survival of 3.1 years, compared to 1.3 years for patients having suboptimal cytoreduction.
The GOG reported a retrospective analysis of 360 patients with stage IV ovarian cancer who were entered onto four prospective randomized phase III trials, and received primary cytoreductive surgery followed by intravenous platinum-based chemotherapy plus paclitaxel . It was a heterogeneous group of patients, with the majority (172; 47.8%) having a malignant pleural effusion. Median overall survival for patients with no macroscopic residual disease was 64 months, which compared favourably with the 72 months for similar patients with stage III disease.
A recent study evaluated the prognostic impact of residual disease after intra-abdominal cytoreduction in 326 consecutive patients with stage IV ovarian cancer treated at Essen, Germany, from 2000 to 2014 . FIGO stage IV was due to a pleural effusion/pulmonary metastasis in 134 cases (41.1%), metastases to the abdominal wall in 133 (40.8%), extra-regional lymph nodes in 63 ((19.3%), liver in 45 (13.8%) and spleen in 22 (6.7%). Primary cytoreductive surgery was performed in 286 patients (87.7%). Median survivals for patients with no residual disease, 1-10 mm and >10 mm were 50, 25 and 16 months respectively (p = 0.001).
Abdominal wall and splenic metastases can usually be resected at the time of primary cytoreduction. Splenic metastases are usually related to haematogenous dissemination and are picked up on CT scan, but they may occasionally be related to direct infiltration from a large omental “cake” ( Figure 6 ).
