Fax: (212) 717-3214
Guidelines and selection criteria for secondary cytoreductive surgery in patients with recurrent, platinum-sensitive epithelial ovarian carcinoma†
Article first published online: 29 MAR 2006
Copyright © 2006 American Cancer Society
Volume 106, Issue 9, pages 1933–1939, 1 May 2006
How to Cite
Chi, D. S., McCaughty, K., Diaz, J. P., Huh, J., Schwabenbauer, S., Hummer, A. J., Venkatraman, E. S., Aghajanian, C., Sonoda, Y., Abu-Rustum, N. R. and Barakat, R. R. (2006), Guidelines and selection criteria for secondary cytoreductive surgery in patients with recurrent, platinum-sensitive epithelial ovarian carcinoma. Cancer, 106: 1933–1939. doi: 10.1002/cncr.21845
Presented at the 36th Annual Meeting of the Society of Gynecologic Oncologists, Miami Beach, Florida, March 19–23, 2005.
- Issue published online: 18 APR 2006
- Article first published online: 29 MAR 2006
- Manuscript Accepted: 6 DEC 2005
- Manuscript Revised: 11 NOV 2005
- Manuscript Received: 6 SEP 2005
- secondary cytoreduction;
- recurrent ovarian cancer;
- ovarian cancer
The benefit of cytoreductive surgery for patients with recurrent epithelial ovarian cancer has not been defined clearly. The objective of this study was to identify prognostic factors for survival in patients who underwent secondary cytoreduction for recurrent, platinum-sensitive epithelial ovarian cancer and to establish generally applicable guidelines and selection criteria.
The authors reviewed all patients who underwent secondary cytoreduction for recurrent epithelial ovarian cancer from 1987 to 2001. Potential prognostic factors were evaluated in univariate and multivariate analyses.
In total, 157 patients underwent secondary cytoreduction, and 153 of those patients were evaluable. After secondary cytoreduction, the median follow-up was 36.9 months (range, 0.2-125.6 months), and the median survival was 41.7 months (95% confidence interval, 36.0-47.2 months). For patients who had a disease-free interval prior to recurrence of between 6 months and 12 months, the median survival was 30 months compared with 39 months for patients who had a disease-free interval between 13 months and 30 months and 51 months for patients who had a disease-free interval >30 months (P = .005). For patients who had a single site of recurrence, the median survival was 60 months compared with 42 months for patients who had multiple sites of recurrence and 28 months for patients who had carcinomatosis (P <.001). The median survival for patients who had residual disease that measured ≤0.5 cm was 56 months compared with 27 months for patients who had residual disease that measured >0.5 cm (P <.001). On multivariate analysis, disease-free interval (P = .004), the number of recurrence sites (P = .01), and residual disease (P <.001) were significant prognostic factors.
In the authors' analysis of secondary cytoreduction for recurrent epithelial ovarian cancer, a significant survival benefit was demonstrated for residual disease that measured ≤ 0.5 cm. The disease-free interval and the number of recurrence sites should be used as selection criteria for offering secondary cytoreduction. Cancer 2006. © 2006 American Cancer Society.
The American Cancer Society estimates that approximately 22,200 women in the United States will be diagnosed with ovarian cancer in 2005, and approximately 16,000 women will die of this disease.1 The benefit of optimal primary cytoreductive surgery followed by platinum-based and paclitaxel-based chemotherapy in the treatment of epithelial ovarian cancer has been well established.2, 3 Using this treatment paradigm, a complete clinical response can be achieved in approximately 80% to 90% of patients with early-stage disease and in 50% of patients with advanced-stage disease.4 Despite these excellent response rates, approximately 50% of patients with advanced-stage disease who are complete responders will develop recurrent disease and, ultimately, will die of their disease.4
Although primary cytoreductive surgery is well accepted as the cornerstone of initial management, the use of cytoreductive surgery in the setting of recurrent disease is defined less clearly. Several studies have demonstrated a benefit of secondary cytoreduction, but a few have not.5 Because of the nonrandomized nature of all of the series published on this topic, patient selection undoubtedly played a significant role in the findings and conclusions of those studies. Moreover, the practicing clinician faces the dilemma of determining which conditions are suitable for offering secondary cytoreduction to a patient with recurrent epithelial ovarian cancer. Should it be offered to all patients with recurrence, or only to those with a disease-free interval >1 year, or >2 years? Does carcinomatosis identified preoperatively preclude offering secondary cytoreduction? The objectives of the current study were to identify prognostic factors for survival in patients who underwent secondary cytoreduction for recurrent platinum-sensitive epithelial ovarian cancer and to establish generally applicable guidelines and selection criteria.
MATERIALS AND METHODS
After obtaining Institutional Review Board approval, the Virginia K. Pierce Gynecology Service Data Base at our institution was used to identify all patients with recurrent epithelial ovarian cancer who underwent secondary cytoreductive surgery from January 1987 to December 2001. Patients who were included in the study were those with documented epithelial ovarian carcinoma who had undergone primary surgery and received platinum-based chemotherapy followed by a complete clinical remission of at least 6 months. In patients who underwent second-look surgery and received further systemic or intraperitoneal, platinum-based chemotherapy, the time of clinical remission was measured from the completion of the therapy after the second-look surgery. The pathology slides from each patient's first surgery were reviewed prior to the secondary cytoreduction. The secondary cytoreduction, which was defined as a surgical exploration with the intent to resect all visible disease, had to be performed at our institution. Preoperative computed tomography scans were obtained on all patients and were reviewed by the surgical team. Patients who were deemed to have recurrent disease that was amenable to resection then underwent an attempted secondary cytoreduction. Patients who had ovarian carcinoma of low malignant potential and those who underwent surgery for the correction of malignant bowel obstruction were excluded.
Records for all patients were reviewed, and pertinent demographic, clinical, operative, and follow-up information was abstracted. Patients who had >100 cc of peritoneal fluid at the time of laparotomy were considered to be positive for ascites. The fluid was not always confirmed histologically as malignant; because, in many patients, the ascites was noted in the operative record but was not sent for evaluation, because these procedures were not done for staging purposes. Patients with less fluid were considered negative. Carcinomatosis was defined as the presence of ≥20 tumor nodules noted at the time of surgery. Residual disease was documented as the greatest dimension diameter (cm) of the largest residual tumor nodule as determined by the operating physician at the end of the operation.
Survival was calculated in months from the secondary cytoreduction to death or the date of last follow-up visit for patients who remained alive. Survival curves were estimated by using the Kaplan–Meier method.6 Univariate analysis of potential prognostic factors was performed with the log-rank test for categorical factors7 and with the Cox proportional hazards model for continuous factors.8 Stepwise model-selection methods, using a cut-off P value of <.05, were used to select factors that were included in the multivariate Cox proportional hazards model. Disease-free interval was included as a categorical variable in the multivariate model in which the categories were determined by using a combination of methodologies—the optimal cut-point method for prognostic variables and applied smoothing techniques to look at survival as a function of disease-free interval.9, 10
During the study period, 157 patients were identified who underwent secondary cytoreductive surgery for recurrent epithelial ovarian cancer at our institution. Four patients did not have any follow-up after surgery, however, leaving 153 evaluable patients for the current analysis. The primary surgeon in all 153 patients was an attending gynecologic oncologist with an assistant at the fellowship level in either gynecologic oncology or surgical oncology. The median age of the study group at the time of secondary cytoreduction was 56.5 years (range, 27.6-79.6 years). Table 1 lists patient and tumor characteristics prior to the secondary cytoreduction, and Table 2 summarizes the findings, procedures performed, and outcomes after the surgery.
|Characteristic||No. of Patients (%)|
|Stage IA-IIC||50 (33)|
|Stage IIIA-IV||100 (65)|
|Papillary serous||70 (46)|
|Well differentiated||12 (8)|
|Moderately differentiated||36 (24)|
|Poorly differentiated||91 (60)|
|Carcinomatosis at primary cytoreduction|
|Residual after primary cytoreduction|
|≤2 cm||75 (49)|
|>2 cm||23 (15)|
|No disease||43 (28)|
|Microscopic disease||13 (9)|
|Macroscopic disease||19 (12)|
|Unknown/not done||78 (51)|
|Initial method of detection of recurrence|
|Computed tomography scan||48 (31)|
|Physical examination/symptoms||49 (32)|
|Characteristic||No. of Patients (%)|
|Site of largest tumor at secondary cytoreduction|
|No. of sites|
|One (no carcinomatosis)||41 (27)|
|Multiple (no carcinomatosis)||68 (44)|
|Procedures performed (some patients underwent >1 procedure)|
|Large-bowel resection||48 (31)|
|Small-bowel resection||32 (21)|
|Biopsy only||20 (13)|
|Partial bladder/ureter resection||3 (2)|
|Hepatic resection||2 (1)|
|Presacral mass resection||2 (1)|
|Residual after secondary cytoreduction|
|No macroscopic disease||62 (41)|
|0.1–0.5 cm||17 (11)|
|0.6–1.0 cm||21 (14)|
|1.1–2.0 cm||11 (7)|
|> 2 cm||41 (27)|
A complete macroscopic resection of all visible peritoneal disease was achieved in 41% of patients; and, overall, 52% of patients had residual disease that measured ≤0.5 cm greatest dimension. There were no perioperative mortalities. There were 6 intraoperative complications (4%), all of which involved bowel injuries that required repair or bowel resection for correction. There also were 6 postoperative complications (4%), which included the following: infectious process requiring antibiotics in 3 patients, venous thromboembolism in 2 patients, and bleeding secondary to gastritis in 1 patient.
One hundred twenty-nine of 153 patients (84%) received platinum-based chemotherapy after their secondary cytoreduction. Twenty-one patients (14%) had nonplatinum-based treatment, and the treatment was unknown in 3 patients (2%). Overall, 105 of 153 patients (69%) died of disease, 23 patients (15%) remained alive with no evidence of disease, and 25 patients (16%) remained alive with disease. After secondary cytoreduction, the median follow-up for the entire cohort was 36.9 months (range, 0.16-125.6 months). The median disease-free interval was 17 months (range, 2.3-251 months), and the median overall survival was 41.7 months (95% confidence interval, 36.0-47.2 months) (Fig. 1).
On univariate analysis, it was observed that patients who had no macroscopic disease and patients who had residual disease that measured from 0.1 cm to 0.5 cm had significantly improved survival compared with patients who had larger residual disease (Fig. 2). There was no statistically significant difference between the survival of patients who were left with no macroscopic residual disease and the survival of patients who had residual disease that measured from 0.1 cm to 0.5 cm. There also was no significant difference in the survival of patients who had residual disease that measured from 0.6 cm to 1.0 cm, from 1.1 cm to 2.0 cm, and >2 cm. Therefore, the cut-off point of 0.5 cm was used in all subsequent statistical analyses to differentiate “optimal” and “suboptimal” secondary cytoreduction (Fig. 3). By using the cut-off size of 0.5 cm to differentiate between “optimal” and “suboptimal” secondary cytoreduction, the median survival of patients who achieved optimal cytoreduction was 56.2 months compared with 26.7 months for patients who had suboptimal results.
Fourteen categorical and continuous clinical and pathologic variables were analyzed for their prognostic significance (Table 3). Categorical variables that were significant on univariate analysis were the site of the largest tumor, the number of sites of disease found at exploration, the presence of ascites, and residual disease after secondary cytoreduction. The only continuous factor that had prognostic significance in the univariate analysis was disease-free interval.
|Variable||Total No.||% Alive||Median Survival (95% CI), Mo||HR (95% CI)||P|
|Stage IA–IIC||50||36||47.3 (38.7–57.4)||1.00||.18|
|Stage IIIA–IV||100||29||39.6 (30.1–45.9)||1.39 (0.62–2.09)|
|Papillary serous||70||37||40.7 (36.0–46.5)||1.00||.99|
|Endometrioid||31||26||38.7 (27.2–61.2)||0.98 (0.59–1.64)|
|Adenocarcinoma||18||22||45.5 (31.0–56.2)||1.02 (0.56–1.87)|
|Other||34||29||44.1 (28.0–57.4)||0.98 (0.60–1.61)|
|Well differentiated||12||58||57.4 (37.2–NR)||1.00||.14|
|Moderately differentiated||36||22||36.9 (26.3–51.4)||1.40 (0.74–2.63)|
|Poorly differentiated||91||32||45.5 (36.0–48.6)||1.25 (0.71–2.20)|
|Carcinomatosis at first debulking|
|Present||19||21||41.7 (26.5–43.9)||1.09 (0.63–1.88)|
|Residual after first debulking|
|≤2 cm||75||37||45.9 (39.5–50.3)||1.00||.16|
|>2 cm||23||13||36.9 (21.4–43.9)||1.50 (0.92–2.44)|
|No disease||43||30||42.2 (38.7–51.4)||1.00||.46|
|Microscopic disease||13||54||26.7 (20.9–NR)||0.99 (0.43–2.28)|
|Macroscopic disease||19||16||37.2 (24.3–53.5)||1.38 (0.81–2.37)|
|Method of detection|
|Computed tomography scan||48||35||40.5 (30.1–51.7)||0.99 (0.62–1.59)|
|Physical examination/symptoms||49||29||38.7 (27.5–48.9)||1.10 (0.69–1.74)|
|Site of largest tumor|
|Abdomen||56||29||30.1 (21.4–41.8)||1.51 (1.00–2.27)|
|Retroperitoneal||14||36||56.2 (40.7–80.9)||0.71 (0.35–1.44)|
|Yes||29||14||28.0 (23.7–33.7)||2.25 (1.42–3.57)|
|No. of sites|
|Multiple, no carcinomatosis||68||32||41.7 (33.7–51.4)||1.85 (1.10–3.11)|
|Carcinomatosis||44||18||27.5 (20.8–36.0)||3.81 (2.17–6.69)|
|Residual after second debulking|
|≤0.5 cm||79||44||56.2 (48.2–66.6)||1.00||<.001|
|>0.5 cm||73||16||26.7 (21.9–31.0)||3.12 (2.08–4.67)|
|CA-125, log||1.14 (0.95–1.36)||.16|
|Disease-free interval, log||0.73 (0.55–0.97)||.03|
To determine useful clinical guidelines, we attempted to analyze disease-free interval as a categorical variable. Figure 4 is a graph that demonstrates all 153 study patients in which circles represent patients who remained alive, and crosses represent patients who died. The median survival was measured from the date of secondary cytoreduction. We applied smoothing techniques10 to construct a smooth curve of the median survival as a function of disease-free interval. The graph demonstrates that there was little change in survival for patients with a disease-free interval of 6 months to 12 months. There was a significant change in survival for patients as the disease-free interval increased from 13 months to 30 months. Then, after 30 months, the survival benefit of prolonged disease-free interval plateaued. Using the optimal cut-point method, a disease free-interval cut-point value of 30 months was associated with the most significant Miller– Siegmund adjusted P value of .009.9 In other words, because longer disease-free intervals generally were associated with a better prognosis, this cut-point method determined that a 30-month disease-free interval was the specific disease-free interval that most significantly divided the study cohort into better versus worse prognostic groups. This combination of these 2 methods helped determine the categories of disease-free interval that were used in our multivariate analysis.
On multivariate analysis, disease-free interval (P = .004), the number of sites of recurrence (P = .01), and residual disease after secondary cytoreduction (P<.001) retained prognostic significance. The median survival was 30 months for a disease-free interval of 6 months to 12 months, 39 months for a disease-free interval of 13 months to 30 months, and 51 months for a disease-free interval >30 months (P = .005). The median survival was 60 months for patients who had a single site of recurrence, 42 months for patients who had multiple sites but not carcinomatosis, and 28 months for patients who had carcinomatosis (P <.001). The median survival was 56 months for patients who had residual disease that measured ≤0.5 cm after secondary debulking and 27 months for patients who had residual disease that measured >0.5 cm. In general, the median survival improved significantly with longer disease-free intervals, fewer sites of recurrence, and secondary cytoreduction to residual disease that measured ≤0.5 cm.
Recurrent epithelial ovarian carcinoma continues to be a therapeutic dilemma for physicians; to date, there is no consensus for optimal treatment strategies. Various therapeutic options have been investigated, including retreatment with platinum-based chemotherapy with or without a taxane for platinum-sensitive tumors.4 Patients who have disease that is identified as platinum-resistant have been treated with topotecan, liposomal doxorubicin, etoposide, and gemcitabine with overall response rates that vary from 20% to 40%.4 Numerous reports in the literature have focused on the survival advantage of secondary cytoreductive surgery for patients with recurrent ovarian cancer; however, there remains a great deal of controversy with regard to which patients would most benefit from this procedure and with regard to the size of residual disease associated with the greatest survival benefit.
It is believed that secondary cytoreduction has a survival benefit in select platinum-sensitive patients. The first study to describe the value of secondary cytoreduction for recurrent ovarian cancer was published by Berek and colleagues more than 20 years ago.11 Those authors demonstrated that patients who underwent optimal debulking (defined as residual disease ≤1.5 cm) at the time of secondary cytoreduction had a median survival of 20 months compared with 5 months for patients who were debulked suboptimally. They also reported that duration from therapy, symptoms, ascites, and initial tumor mass were important prognostic factors for survival. Morris et al. subsequently studied 30 patients with recurrent ovarian cancer and, using a cut-off size of 2 cm for optimal debulking, found no survival benefit for secondary cytoreduction.12 In follow-up study of 25 patients, Munkarah et al. from the same institution also found no statistically significant benefit for secondary cytoreduction.13 However, in all other published series, a survival benefit has been demonstrated for patients with recurrent ovarian cancer who undergo optimal secondary cytoreduction. Table 4 summarizes the series on secondary cytoreduction for recurrent ovarian cancer.11–24 In most of the series listed, residual disease and disease-free interval prior to secondary cytoreduction were identified as independent prognostic factors.
|Study||Y||No. of Patients||Survival Benefit?|
|Berek et al.11||1983||32||If ≤1.5 cm residual disease|
|Morris et al.12||1989||30||None|
|Janicke et al.14||1992||30||If microscopic residual disease|
|Segna et al.15||1993||100||If ≤2.0 cm residual disease|
|Vaccarello et al.16||1993||57||If ≤0.5 cm residual disease|
|Lichtenneger et al.17||1998||81||If ≤2.0 cm residual disease|
|Eisenkop et al.18||2000||106||If microscopic residual disease|
|Gadducci et al.19||2000||30||If microscopic residual disease|
|Munkarah et al.13||2001||25||Trend if ≤2.0 cm (P = .08)|
|Scarabelli et al.20||2001||149||If ≤1.0 cm residual disease|
|Tay et al.21||2002||46||If microscopic residual disease|
|Zang et al.22||2004||117||If ≤1.0 cm residual disease|
|Gronlund et al.23||2004||38||If microscopic residual disease|
|Onda et al.24||2005||44||If microscopic residual disease|
|Current series||2005||153||If ≤0.5 cm residual disease|
However, for the practicing clinician, the decision regarding whether to offer a patient who has recurrent ovarian cancer a secondary cytoreductive procedure may not be clear from the available studies. Should secondary cytoreduction be offered to all patients who have platinum-sensitive, recurrent disease? How should the disease-free interval and other potential variables be factored into the decision?
To our knowledge, this is the largest single-institution study to date that has evaluated the survival benefit associated with secondary cytoreduction for recurrent ovarian cancer. The findings on multivariate analysis of disease-free interval, number of sites of recurrence, and residual disease not only confirm the findings of other studies: They also make clinical sense. Surely, a patient who has an isolated recurrence and a disease-free interval of 30 months most likely is going to live longer after secondary cytoreduction than a patient who develops carcinomatosis 7 months after completing her primary treatment.
Based on the findings in this study, we suggest that the objective of secondary cytoreduction should be to achieve residual disease that measures ≤0.5 cm. In patients who are able medically to tolerate a major surgical procedure, the following selection criteria are suggested: For patients with only 1 of site of recurrence, as long as their disease-free interval is ≥6 months, offer secondary cytoreduction; for patients with multiple recurrence sites but no carcinomatosis who have a disease-free interval >12 months, offer secondary cytoreduction; and, for patients with carcinomatosis who have a disease-free interval >30 months, secondary cytoreduction also may be beneficial (Table 5). We do not recommend offering secondary cytoreduction to patients who have a disease-free interval of from 6 months to 12 months with evidence of carcinomatosis. For patients who have multiple sites of recurrence and a disease-free interval of from 6 months to 12 months or who have carcinomatosis with a disease-free interval of from 13 months to 30 months, secondary cytoreduction may be considered, and the decision may be individualized based on various factors, such as the exact disease-free interval (closer to 6 months than to 30 months, or vice versa), patient age, performance status, overall general medical condition, and the patient's wishes.
|DFI||Single Site||Multiple Sites: No Carcinomatosis||Carcinomatosis|
|6–12 Mo||Offer SC||Consider SC||No SC|
|12–30 Mo||Offer SC||Offer SC||Consider SC|
|>30 Mo||Offer SC||Offer SC||Offer SC|
We acknowledge that the current study suffered from selection bias, a modest number of patients, and most of the problems associated with retrospective analyses, such as a lack of specific information on the outcomes of comparable patients with recurrent disease who were managed without secondary cytoreduction. However, while we await prospective, randomized data on this issue, we hope that these guidelines and selection criteria prove practical and useful for the practicing gynecologic and medical oncologist and aid investigators in the design of future clinical trials.
- 4Epithelial ovarian cancer. In HoskinsWJ, PerezCA, YoungRC, BarakatR, MarkmanM, RandallM, editors. Principles and Practice of Gynecologic Oncology. 4th ed. Philadelphia, PA: Lippincott-Raven Publishers, 2005: 895–987., , , .
- 8Regression models and life tables. R Stat Soc B. 1972; 34: 187–220..
- 10Applied Smoothing Techniques for Data Analysis: The Kernel Approach with S-Plus illustrations. Oxford: Oxford University Press, 1997., .
- 15Secondary cytoreduction for ovarian cancer following cisplatin therapy. J Clin Oncol. 1993; 57: 61–65., , , et al.