Fax: (410) 614-8718
Secondary cytoreductive surgery for localized, recurrent epithelial ovarian cancer
Analysis of prognostic factors and survival outcome
Article first published online: 11 JAN 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 4, pages 685–691, 15 February 2007
How to Cite
Salani, R., Santillan, A., Zahurak, M. L., Giuntoli, R. L., Gardner, G. J., Armstrong, D. K. and Bristow, R. E. (2007), Secondary cytoreductive surgery for localized, recurrent epithelial ovarian cancer. Cancer, 109: 685–691. doi: 10.1002/cncr.22447
- Issue published online: 2 FEB 2007
- Article first published online: 11 JAN 2007
- Manuscript Accepted: 11 NOV 2006
- Manuscript Revised: 7 NOV 2006
- Manuscript Received: 6 SEP 2006
- Pam McDonald Drive for Life Ovarian Cancer Charity Golf Tournament
- ovarian carcinoma;
- secondary cytoreductive surgery;
- localized recurrence;
- diagnosis-to-recurrence interval;
- residual disease
The objective of this study was to evaluate the role of secondary cytoreductive surgery in the outcome of patients who had recurrent epithelial ovarian carcinoma that was limited to ≤5 recurrence sites within the abdomen or pelvis on preoperative imaging studies and attempt to define selection criteria associated with improved survival, with specific attention to the number of lesions suspicious for recurrent disease.
Patients who underwent secondary surgical cytoreduction for recurrent epithelial ovarian cancer between September 1997 and March 2005 were identified retrospectively from tumor registry databases. Study inclusion criteria required a complete clinical response to primary therapy, ≥12 months between initial diagnosis and recurrence, and ≤5 recurrence sites on preoperative imaging studies. Univariate and multivariate logistic regression analyses were used to evaluate the effect of clinicopathologic variables on overall postrecurrence survival.
Fifty-five patients met the study inclusion criteria. The median patient age at recurrence was 57.7 years, and the median diagnosis-to-recurrence interval was 32 months (range, 12–164 months). Complete cytoreduction was achieved in 41 patients (74.5%). On multivariate analysis, the statistically significant and independent predictors of overall survival were a diagnosis-to-recurrence interval ≥18 months (median survival, 49 months vs 3 months; P < .01), the number of radiographic recurrence sites (median survival, 50 months for patients with 1 or 2 sites vs 12 months for patients with 3 to 5 sites; P < .03), and residual disease (median survival, 50 months for patients with no macroscopic residual disease vs 7.2 months for patients with macroscopic residual disease; P < .01). Age, tumor grade, histology, CA-125 level, ascites, and tumor size were not associated significantly with survival.
The current data supported the definition of localized recurrent ovarian cancer as patients with 1 or 2 radiographic recurrence sites. In this select population, a diagnosis-to-recurrence interval ≥18 months and complete secondary surgical cytoreduction, which was achievable in the majority of patients, were associated with a median postrecurrence survival of approximately 50 months. Cancer 2007. © 2007 American Cancer Society.
Ovarian cancer is the leading cause of death from gynecologic malignancy in the United States, accounting for >15,000 estimated deaths in 2006, primarily because >60% of patients with ovarian cancer will experience disease recurrence.1, 2 Primary cytoreductive surgery and combination chemotherapy are the cornerstones of the initial treatment for epithelial ovarian cancer. Current chemotherapy regimens have limited efficacy in platinum-resistant disease; and although, outcomes are improved in patients with platinum-sensitive disease, the results remain insufficient.3–5 Thus, further efforts to improve survival with salvage treatments are warranted. The application of secondary cytoreductive surgery to patients with recurrent disease has been 1 such effort and is associated with extended survival in selected patients.6–9 That survival benefit is predicated on selection criteria, albeit subject to bias; and among the identified selection criteria associated with surgical success and prolonged survival are a disease-free interval >12 months to 18 months, complete surgical resection, and a good performance status.10–26 Recently, attention has been focused on the prognostic significance of a localized recurrence in patients with epithelial ovarian cancer; however, a specific threshold number of disease sites that correlates with clinical outcome has not been defined precisely. Therefore, the objectives of this study were to evaluate the role of secondary cytoreductive surgery in patients with recurrent epithelial ovarian carcinoma limited to ≤5 recurrence sites within the abdomen or pelvis on preoperative imaging studies and attempt to define the selection criteria associated with improved survival, with specific attention to the number of lesions suspicious for recurrent disease.
MATERIALS AND METHODS
Approval to conduct this study was obtained from the Johns Hopkins Medical Institutions (JHMI) Institutional Review Board. All patients who underwent secondary cytoreduction for recurrent epithelial ovarian cancer at the JHMI between September 1997 and March 2005 were identified retrospectively from tumor registry databases. Typical selection criteria for secondary cytoreductive surgery during the study interval included a disease-free interval ≥6 months, a Gynecologic Oncology Group performance status ≤2, and surgically resectable disease. Eligibility criteria for study inclusion required the following: 1) a complete clinical response to primary therapy, 2) ≥12 months between initial diagnosis and recurrence, 3) ≤5 recurrence sites limited to the abdominal cavity on preoperative imaging studies, and 4) attempted secondary cytoreduction.
All women had recurrent ovarian carcinoma diagnosed or confirmed by radiographic imaging studies, which consisted of a computed tomography (CT) scan, positron emission tomography (PET) scan, magnetic resonance image (MRI), or a combination. Patients were excluded if they underwent an interval debulking or a second-look procedure with findings of macroscopically positive disease. All surgical specimens were assessed by gynecologic pathologists at the JHMI.
The major statistical endpoint of this study was the association of clinical factors with postrecurrence survival outcome. These factors included the amount of residual disease, the diagnosis-to-recurrence interval, the number of sites of recurrent disease on preoperative imaging studies, the number of sites of recurrent disease identified at surgery, age at recurrence, tumor histology, serum CA-125 level at the time of recurrence, the greatest tumor dimension, second-look findings (if applicable), tumor grade, and the presence of ascites. The diagnosis-to-recurrence interval was chosen as a measure because some patients received consolidation treatment beyond the prescribed initial course of chemotherapy, which would have resulted in unequal starting points if the chosen measures were disease-free interval, progression-free interval, or treatment-free interval.
For statistical analysis, factors that were associated with overall postrecurrence survival were based on cross tabulations and logistic regression modeling. Cross tabulations were analyzed by using chi-square or Fisher exact tests where appropriate. Multivariate logistic regression models27 were used to determine the effects of multiple factors on the probability of ≥3 recurrence sites. Event time distributions for this endpoint were estimated with the method of Kaplan and Meier28 and were compared by using the log-rank statistic29 or a proportional hazards regression model.30 The simultaneous effect of ≥2 factors was studied by using multivariate Cox proportional hazards models. All statistical computations were performed using the SAS31 or EGRET (Statistics and Epidemiologic Research Corp.)32 software packages. All confidence intervals are at the 95% level), and all P values are 2-sided.
Patient Characteristics and Preoperative Imaging Studies
Fifty-five patients with recurrent epithelial ovarian carcinoma met the study inclusion criteria. Clinicopathologic patient characteristics are summarized in Table 1. After the initial surgery, 53 patients received adjuvant chemotherapy. A platinum- and taxol-based regimen was given to 48 of 53 patients (90.6%), a platinum agent and cyclophosphamide were given to 4 patients (7.5%), cisplatin alone was given to 1 patient (1.9%), and 2 patients did not receive chemotherapy. Twenty patients (36.4%) underwent either a second-look laparoscopy or a laparotomy procedure. Thirteen of those 20 procedures (65.0%) resulted in negative findings. The remaining 7 procedures were macroscopically negative, and the findings were limited to positive washings or microscopic disease on pathology evaluation. Patients who received chemotherapy after a second-look procedure had an interval of ≥6 months (range, 6–52 months) from the time they completed chemotherapy to time they were diagnosed with a recurrence.
|Characteristic||No. of patients (%)|
|Primary surgery outcome|
|Optimal (<1 cm)||44 (80.0)|
|Suboptimal (≥1 cm)||11 (20.0)|
|Median age [range], y||57.7 [36.5–84.8]|
|Median diagnosis-to-recurrence interval [range], mo||32 [12.9–164.9]|
|Median CA-125 level [range], U/mL||56 [7–81,604]|
|Median greatest dimension of disease [range], cm||4 [1–30]|
|Largest disease site|
|Retroperitoneal lymph nodes||9|
|No. of recurrence sites by imaging study|
|1–2 Recurrence sites||42 (76.4)|
|3–5 Recurrence sites||13 (23.6)|
|No. of recurrence sites by surgical evaluation|
|1–2 Recurrence sites||35 (63.6)|
|3–5 Recurrence sites||20 (36.4)|
|Secondary cytoreduction outcome|
|Residual ≤1 cm||8 (14.5)|
|Residual >1 cm||6 (10.9)|
Among patients who underwent secondary cytoreductive surgery, the median age was 57.7 years (range, 36.5–84.8 years), and the median diagnosis-to-recurrence interval was 32 months (range, 12.9–164.9 months). Prior to secondary cytoreductive surgery, serum CA-125 levels were available for 48 of 55 patients (87.3%); and, in 28 patients (58.3%), the CA-125 level was >35 U/mL. At the time of secondary cytoreductive surgery, the median CA-125 level was 56 U/mL (range, 7–81,604 U/mL).
The median time between preoperative imaging and secondary cytoreductive surgery was 27 days (range, 3–120 days). Patients were subdivided into 2 groups based on the number of recurrence sites on radiographic imaging. There were 42 patients (76.4%) with 1 or 2 recurrence sites and 13 patients (23.6%) with 3 to 5 recurrence sites. Preoperative imaging studies correctly predicted the number of recurrence site(s) in 34 of 55 patients (61.8%), and those predictions were confirmed by the operative and pathology reports. Of 21 patients who were diagnosed incorrectly, 16 patients (76.2%) had imaging studies that reported fewer recurrence sites than were encountered at the time of surgery, and imaging studies from 5 patients (23.8%) overstated the number of lesions compared with surgically documented findings (Table 2).
|Imaging study||Diagnosed correctly||Diagnosed incorrectly|
Secondary Surgery and Postoperative Follow-up
After secondary cytoreductive surgery, complete surgical cytoreduction (no macroscopic residual disease) was achieved in 41 patients (74.5%). Of the remaining 14 patients who had evidence of macroscopic disease, 8 of 55 patients (14.5%) had optimal but not complete cytoreduction (macroscopic disease with a maximal dimension <1 cm), and 6 of 55 patients (10.9%) had suboptimal cytoreduction (>1 cm of residual disease). The median operative blood loss was 200 mL (range, from <50 mL to 900 mL), and the median length of hospital stay was 5 days (range, 2–23 days). Postoperative complications occurred in 14 patients (25.5%) and included febrile morbidity in 5 patients, prolonged ileus in 5 patients, respiratory complications in 2 patients, wound dehiscence in 1 patient, and pelvic hematoma in 1 patient. There was 1 perioperative death (1.8%) caused by septic shock.
Forty-eight patients received adjuvant chemotherapy after secondary cytoreductive surgery, and the specific treatment regimens were determined by the treating gynecologic oncologist or medical oncologist. Salvage chemotherapy consisted of the following regimens: platinum plus paclitaxel in 16 patients, platinum plus gemcitabine in 8 patients, platinum only in 8 patients, and gemcitabine only in 4 patients. Paclitaxel and topotecan were administered to 3 patients each, 2 patients received platinum plus liposomal doxorubicin, and 1 patient received cyclophosphamide, platinum plus cyclophosphamide, or etoposide. Seven patients did not receive chemotherapy either because the patient refused or because the patient was unable to tolerate additional chemotherapy.
The median follow-up of surviving patients was 30 months from the time of surgery, and the median survival from the time of secondary cytoreductive surgery for the entire study group was 48 months (Fig. 1). In total, 10 different clinical and pathologic variables were analyzed for prognostic significance on univariate and multivariate analysis (Table 3). Only 3 factors were associated independently and significantly with postrecurrence and overall survival and retained prognostic significance on multivariate analysis: a diagnosis-to-recurrence interval ≥18 months (49 months vs 3 months; P = .0013) (Fig. 2), complete cytoreduction after secondary cytoreductive surgery (50 months vs 7.2 months; P = .0001) (Fig. 3), and 1 or 2 recurrence sites on radiographic imaging studies (50 months vs 12 months; P = .026) (Fig. 4). Patients who had only 1 or 2 surgically documented sites of recurrent disease also had a statistically significant survival advantage (median survival, 52 months) compared with patients who had ≥3 surgically documented disease sites (median survival, 12 months; P < .001). The number of surgically documented disease sites was associated significantly with postrecurrence survival on univariate analysis but dropped to borderline significance on multivariate analysis (P = .06). A median preoperative serum CA-125 level ≥56 U/mL was of borderline significance, and these patients had more than twice the risk of overall death compared with patients who had CA-125 levels below the median (hazard ratio, 2.2; 95% confidence interval, 0.97–4.83; P = .06).
|Greatest dimension of recurrence||.82|
|No. of recurrence sites|
|Diagnosis-to-recurrence interval (≥18 mos)||.001|
|No. of recurrence sites|
During the past 20 years, secondary cytoreductive surgery has emerged as a treatment option for a select subgroup of patients with recurrent epithelial ovarian cancer. Despite a growing body of retrospective literature illustrating an inverse relation between residual disease and postrecurrence survival, the clinical applicability of secondary cytoreductive surgery for recurrent ovarian cancer remains a controversial topic.9–13, 15, 19–26, 33 Proponents point to the theoretical and clinical benefits of cytoreductive surgery in the primary setting as a basis for similar benefits in patients with platinum-sensitive, albeit recurrent, ovarian cancer. Critics of secondary cytoreductive surgery counter that the same favorable prognostic factors that define ideal candidates for secondary cytoreductive surgery also characterize those patients who will experience extended postrecurrence survival with salvage chemotherapy alone.34, 35 These prognostic factors include the amount of residual disease after the procedure and the length of the disease-free interval.9, 10, 12, 14–26, 33 A number of studies have shown that patients who are left with no or minimal residual disease at the time of secondary cytoreductive surgery have a median survival that ranges from 38 months to 61 months compared with 4.5 months to 27 months for patients who undergo suboptimal cytoreduction.9–13, 15, 16, 18–26, 33 A longer disease-free interval (from >12 months to 36 months) typically has been associated with an improved survival outcome.10, 12, 14, 15, 17, 20–26 Although these factors have been associated consistently with postrecurrence survival, defining the impact of the extent of disease has been more problematic. The objectives of the current study, therefore, were to attempt to define a localized recurrence of ovarian carcinoma based on the number of recurrent lesions detected and to evaluate the prognostic impact on overall survival of multiple factors in patients undergoing secondary cytoreductive surgery.
A localized recurrence of ovarian cancer often is cited as a favorable prognostic factor, both in terms of the likelihood of successful secondary cytoreduction and in terms of postrecurrence survival. However, the literature has been limited to comparison of a solitary site with multiple sites of recurrence.9, 11–13, 21, 22, 24, 33
Munkarah et al. evaluated the role of cytoreductive surgery for solitary versus multiple intra-abdominal recurrence sites of ovarian carcinoma.33 Although those investigators did not evaluate survival on the basis of recurrent disease sites, they were able to achieve optimal cytoreduction in a greater proportion of patients with isolated recurrence sites, which resulted in a trend toward improved survival.33 Gronlund et al. also noted that a solitary recurrence was associated significantly with the ability to achieve complete tumor resection. Furthermore, in that study, the patients who achieved complete cytoreduction experienced an improved overall survival.11 Zang et al. not only noted the ability to perform optimal cytoreduction in patients with a solitary lesion but also observed a significant 5-year survival advantage for patients who had 1 recurrent disease site (49.8%) compared with patients who had multiple recurrent disease sites (5.4%).22 Gadduci et al. reported a median survival of 40 months for patients who had an isolated, solitary recurrence versus 19 months for patients who had multiple recurrence sites.12 Recently, Chi et al. demonstrated that patients who had a single site of recurrence had a median survival of 60 months compared to 42 months for patients who had multiple sites of recurrence and 28 months for patients who had carcinomatosis.24 In addition, other studies, including an evaluation of isolated lymph node recurrences, demonstrated the notable survival benefit of secondary cytoreductive surgery for patients who have a single recurrence site.9, 13, 21, 25, 36
In the current study, we evaluated patients who had ≤5 recurrence sites on preoperative imaging studies. The median survival of 48 months from the time of secondary cytoreductive surgery suggests that patients with ≤5 lesions on imaging studies, in general, comprise a group with a good prognosis. Furthermore, we were able to demonstrate a median survival advantage >3 years for patients who were limited to 1 or 2 recurrence sites compared with patients who had 3 to 5 recurrence sites based on both imaging studies and surgical exploration. However, the number of lesions identified by the preoperative imaging studies allows for the selection of patients for secondary cytoreductive surgery, making it more relevant than the number of sites identified at the time of surgery. The data from the current study also confirm the findings of previous reports of a significantly improved survival for patients who undergo complete cytoreduction (50 months vs 7.2 months) and for patients who experience a diagnosis-to-recurrence interval ≥18 months (49 months vs 3 months). Table 4 provides a brief summary of studies that evaluated similar prognostic factors in the setting of secondary cytoreduction for recurrent ovarian cancer.
|DFI, months||Residual disease, cm||No. of sites|
|Chi et al., 200624||<12 (30)||<0.5 (56)||1 (60)|
|13–30 (39)||>0.5 (27)||≥2 (42)|
|Eisenkop et al., 200021||<12 (25)||No residual(44)|
|13–36 (44)||Residual (19)|
|Gadducci et al., 200012||<17.5 (15)||No residual (37)||1 (40)|
|>17.5 (25)||Residual (19)||≥2 (19)|
|Segna et al., 199315||<12 (9)||<1 (27)|
|>12 (23)||>1 (9)|
|Current study||<18 (3)†||No residual (50)||1–2 (50)|
|≥18 (49)†||Residual (7)||3–5 (12)|
Because this study was retrospective, it was limited by interpretation of the data and potential selection bias from the individual selection of patients for secondary cytoreductive surgery and for study inclusion. There also was variability among the preoperative imaging studies, which, along with the pathology specimens, were not rereviewed for this study. In addition, the surgeries were performed by several surgeons, which may represent the possibility of variations in surgical objectives as well as technical capabilities. Because of the variety of salvage chemotherapy regimens, we were unable to incorporate the potential impact of this factor on overall survival.
Despite these limitations, to our knowledge, the current study is the first to quantify a localized recurrence of epithelial ovarian cancer as 1 or 2 radiographic lesions based on superior survival outcome compared with patients who have more extensive disease. These data support the idea t hat, when considering secondary cytoreductive surgery for an individual patient, preoperative images showing 1 or 2 target lesions can be incorporated into the decision analysis along with the already accepted selection criteria of prolonged disease-free interval and the likelihood of complete surgical resection. Ideally, both selection criteria as well as the proposed clinical benefits of secondary cytoreductive surgery should be evaluated by a large, multiinstitutional or cooperative group study with prospectively collected data and uniform postoperative treatment strategies.
Supported by the Pam McDonald Drive for Life Ovarian Cancer Charity Golf Tournament.
- 1American Cancer Society. Cancer facts and figures, 2006. Available from URL: http://www.cancer.org/downloads/STT/CAFF2006PWSecured.pdf [accessed March 14, 2006].
- 27The Analysis of Binary Data. London: Methuen; 1970..
- 30Regression models and life-tables (with discussion). J R Stat Soc. 1972; 34: 187–220..
- 31SAS Institute Inc. SAS User's Guide: Statistics, Version 5. Cary, NC: SAS Institute Inc.; 1985.
- 32Statistics and Epidemiology Research Corporation. EGRET Users' Manual. Seattle, Wash; Statistics and Epidemiology Research Corporation; 1988.