Fax: (410) 614-8718
Recurrent micropapillary serous ovarian carcinoma
The role of secondary cytoreductive surgery
Article first published online: 31 JUL 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 4, pages 791–800, 15 August 2002
How to Cite
Bristow, R. E., Gossett, D. R., Shook, D. R., Zahurak, M. L., Tomacruz, R. S., Armstrong, D. K. and Montz, Fredrick. J. (2002), Recurrent micropapillary serous ovarian carcinoma. Cancer, 95: 791–800. doi: 10.1002/cncr.10789
- Issue published online: 31 JUL 2002
- Article first published online: 31 JUL 2002
- Manuscript Accepted: 18 MAR 2002
- Manuscript Revised: 13 MAR 2002
- Manuscript Received: 29 JAN 2002
- Elaine Riccio Ovarian Cancer Research Fund
- micropapillary serous ovarian carcinoma (MPSC);
- secondary cytoreductive surgery;
- disease recurrence
The objectives of the current study were to: 1) characterize the clinical outcome of patients with recurrent micropapillary serous ovarian carcinoma (MPSC) and 2) evaluate the survival impact of secondary cytoreductive surgery and other prognostic variables.
Twenty-six patients with recurrent MPSC were identified retrospectively from hospital and tumor registry databases. Survival curves were generated from the time of tumor recurrence using the Kaplan–Meier method and statistical comparisons were performed using the log-rank test, logistic regression analysis, and the Cox proportional hazards regression model.
The median age of the patients at the time of recurrence was 46 years. The mean progression-free interval was 31.6 months, and 92% of patients had advanced stage disease at the time of the initial diagnosis. Twenty-one patients underwent secondary cytoreductive surgery; tumor debulking was performed in 90.5% of cases and 52.4% of patients required an intestinal resection. Optimal resection (residual disease ≤ 1 cm) was achieved in 15 patients (71.4%). Patients undergoing optimal secondary cytoreduction had a median survival time of 61.2 months from the date of disease recurrence, compared with 25.5 months for those patients in whom suboptimal residual disease remained (P < 0.02) and 29.9 months for nonsurgical patients (P < 0.01). On multivariate analysis, optimal secondary cytoreduction was found to be the only independent predictor of survival. Salvage chemotherapy produced an objective response in 25% of patients with measurable disease. The administration of chemotherapy prior to surgical intervention was associated with a trend toward worse survival and a lower likelihood of optimal secondary cytoreduction.
Optimal secondary cytoreductive surgery is feasible in the majority of patients with recurrent MPSC and is an independent predictor of subsequent survival. Surgical intervention should be considered for those patients with recurrent MPSC. [See editorials on pages 675–6 and 677–80, this issue.] Cancer 2002;95:791–800. © 2002 American Cancer Society.
Micropapillary serous ovarian carcinoma (MPSC) first was described in 1996 and currently is recognized by the World Health Organization as a distinct subgroup of ovarian serous borderline tumors, also referred to as ovarian tumors of low malignant potential.1–3 However, the available literature addressing MPSC to date has focused primarily on the pathologic criteria for diagnosis rather than the clinical management of patients with this disease.1, 2, 4–11 The majority of authors agree that the long-term prognosis is excellent in the absence of surgically confirmed extraovarian disease.1, 5–7 Conversely, advanced stage MPSC reportedly is associated with a high incidence of invasive extraovarian implants and a more aggressive clinical course compared with nonmicropapillary serous borderline tumors (SBTs).1, 2, 4 Although approximately 45–62% of patients with metastatic MPSC ultimately may experience tumor recurrence, effective management strategies for such cases have not yet been defined.1, 2 Therefore, the objectives of the current study were to evaluate the clinical outcome and associated prognostic variables for patients with recurrent MPSC, with particular attention paid to the impact of secondary cytoreductive surgery on survival.
MATERIALS AND METHODS
Approval to conduct this study was obtained from the Johns Hopkins Medical Institutions Clinical Research Committee and Joint Committee on Clinical Investigation. All patients undergoing clinical management and/or review of pathologic material for a diagnosis of MPSC at the Johns Hopkins Medical Institutions (JHMI) were identified from a search of the JHMI Tumor Registry, the Kelly Gynecologic Oncology Service clinical database, and the Department of Pathology database. Patients with primary nonmicropapillary SBTs of the ovary and those with high-grade invasive serous carcinomas were not reviewed. Mandatory inclusion criteria specified that a primary ovarian tumor diagnosis of MPSC had been rendered by an attending gynecologic pathologist from the Department of Pathology according to previously defined criteria.1, 2, 11 Briefly, the cardinal pathologic feature of MPSC is the formation of micropapillae in a hierarchical branching pattern with little or no fibrovascular support (Fig. 1). Detached epithelial clusters are common, and solid epithelial nests also may be present. Cytologically, the cells are small and rounded with a high nuclear:cytoplasmic ratio, but may be deceptively bland with infrequent mitoses.1 For the purposes of the current study, micropapillary architecture had to constitute at least 5 mm in greatest dimension of a serous tumor, irrespective of the total tumor size, to qualify as MPSC. Tumors with lesser degrees of micropapillary architecture were excluded from further analysis. Inclusion criteria also required clinical, radiographic, or histologic confirmation of disease recurrence after primary treatment.
Individual subject data were collected retrospectively from inpatient and ambulatory medical records, referring physicians, and, when necessary, direct patient contact. Surgical and pathology reports from both primary and secondary surgeries were reviewed in all cases and the following information was abstracted: the pathologic diagnoses of primary and recurrent tumor, the International Federation of Gynecology and Obstetrics (FIGO) stage of disease, the primary cytoreductive surgical outcome, and whether invasive extraovarian implants were present at the time of initial surgery.12 Implants present at the time of initial surgery were classified as invasive or noninvasive according to previously published criteria.11 Specifically, only extraovarian tumor implants exhibiting invasion of underlying normal tissue or micropapillary architecture were classified as invasive. The type of tumor implants present at the time of initial surgery was used for statistical analyses because not all patients in the current series underwent secondary surgical exploration and hence could not have their recurrent tumor implants classified. Data collection also included the secondary cytoreductive surgical procedures performed, the presence of clinically apparent ascites at the time of disease recurrence, and the sites of recurrent metastatic tumor spread. Recurrent metastatic tumor sites were categorized according to the following anatomic regions based on surgical or radiographic (in nonsurgical cases) findings: pelvis, abdominal peritoneum, residual omentum/gastrocolic ligament/lesser sac, bowel mesentery/serosa, diaphragm/liver surface, extraabdominal/liver parenchyma/spleen, retroperitoneal lymph nodes, and porta hepatis/gallbladder fossa. Postoperative residual disease after secondary cytoreductive surgery was classified as either optimal (residual tumor nodules ≤ 1.0 cm in greatest dimension) or suboptimal (residual tumor nodules > 1.0 cm in greatest dimension).
Follow-up information included the interval from diagnosis to clinical or radiographic progression of disease (progression-free interval [PFI]), the date of last follow-up or death, and disease status at the time of last follow-up. Information regarding the administration of salvage chemotherapy after disease recurrence also was recorded. For patients with measurable disease, the response to salvage chemotherapy was categorized using standard criteria according to the impression of the treating clinician. A complete response required the disappearance of all clinically measurable disease. A partial response was defined as a decrease of at least 50% in the size of all measurable lesions for at least 60 days, whereas progressive disease was considered to be an increase in size of at least 50% for any measurable lesion or the development of any new lesion. Stable disease was defined as a change in measurable disease that was insufficient to meet criteria for a partial response or progressive disease over a period of at least 60 days.
The primary outcome variable was overall survival (OS), which was calculated from the date of disease recurrence to the date of death or last follow-up. The date of disease recurrence was used as the initial reference for OS to assure equivalent starting points from which to compare all patients irrespective of whether they underwent secondary surgery initially, received salvage chemotherapy prior to secondary surgery, or had their disease managed nonsurgically. Demographic, pathologic, and surgical variables were evaluated for their influence on this outcome measure. Patient age was evaluated both as a continuous variable and as a function of the median cohort age at the time of disease recurrence. For survival analyses, event time distributions were estimated using the method of Kaplan and Meier, and differences in survival rates were compared using the log-rank test or Cox proportional hazards regression model.13–15 Covariates with potential prognostic value in univariate analyses were entered into a multivariate Cox proportional hazards model and nonsignificant factors were removed in a stepwise fashion. Hazard ratios were expressed relative to a baseline reference category for each covariate, and all P values reported are two-sided. An additional statistical endpoint was to identify predictors of optimal secondary cytoreduction. Factors associated with this surgical outcome were selected based on cross-tabulations and logistic regression modeling.16 All computations were performed using the Statistical Analysis System or EGRET.17, 18
Examination of all database records identified 78 patients who potentially were eligible for study. Of these, 13 patients were excluded secondary to insufficient pathologic criteria for MPSC and 14 patients had incomplete medical information. An analysis of the primary management and clinical outcome of the remaining 51 patients is the subject of a separate report. Twenty-six of these 51 patients had a documented recurrence of disease by clinical examination or by radiographic imaging studies, and it is this group that forms the basis for the current study. Twenty of these patients (76.9%) had received clinical care (other than pathologic consultation) at JHMI at some point during their disease course. Two patients (7.7%) with recurrent MPSC had FIGO Stage I disease whereas 24 patients (92.3%) had FIGO Stage III or Stage IV disease. The median patient age was 46 years (range, 24–67 years) and the median gravidity and parity were 2 (range, 0–6) and 1 (range, 0–4), respectively. In all 26 cases of recurrent disease, institutional review of pathologic material had confirmed a primary histologic diagnosis of MPSC in an ovarian tumor. In addition, 23 patients (88.5%) had invasive extraovarian tumor implants present at the time of initial diagnosis. After primary surgery, 22 of 24 patients with FIGO Stage III/IV disease received adjuvant chemotherapy, which was platinum-based in 21 cases. Thirteen patients received a platinum agent combined with paclitaxel and 8 patients received a platinum agent combined with cyclophosphamide. The two patients with Stage I disease were not treated with chemotherapy after the initial diagnosis.
For all 26 patients, the mean interval from diagnosis to disease progression (PFI) was 31.6 months (median, 15.7 months; range, 4.4–195.4 months). Fifteen patients (57.7%) had a PFI ≥ 12 months whereas 11 patients (42.3%) had a PFI < 12 months. All 21 patients received platinum-based chemotherapy had a PFI ≥ 6 months. The only patient with a PFI < 6 months had FIGO Stage IIIC disease with noninvasive peritoneal tumor implants. This patient presented with recurrent disease 4.4 months after undergoing a suboptimal primary surgical effort. Clinically evident ascites was present at the time of recurrence in 9 patients (34.6%). The anatomic regional distribution of recurrent tumor based on surgical or radiographic (in nonsurgical cases) findings is shown in Table 1. The mean number of anatomic regions involved with disease was 2.8 (median 3 regions; range, 1–5 regions).
|Anatomic region||% with disease (n)|
|Abdominal peritoneum||80.8% (21)|
|Residual omentum/gastrocolic ligament/lesser sac||61.5% (16)|
|Bowel mesentery/serosa||57.7% (15)|
|Diaphragm/liver surface||42.3% (11)|
|Extraabdominal/liver parenchyma/spleen||26.9% (7)|
|Retroperitoneal lymph nodes||23.1% (6)|
|Porta hepatis/gallbladder fossa||7.7% (2)|
Twenty-one patients with recurrent MPSC underwent surgical exploration for the purpose of cytoreduction. In 15 cases the surgery was performed immediately after the diagnosis of recurrent disease, whereas 6 patients received salvage chemotherapy prior to secondary surgery. A total of five patients were treated with chemotherapy alone. Eight secondary surgical procedures (38.1%) were performed at JHMI. A gynecologic oncologist was the primary surgeon in all cases, and in two instances a general surgeon also participated in the surgery. Of the 21 patients undergoing secondary cytoreductive surgery, 15 (71.4%) were left with residual disease ≤ 1 cm in greatest dimension and 6 patients (28.6%) had residual disease measuring > 1 cm. In many cases, extensive surgical procedures were required, the details of which are shown in Table 2. Tumor cytoreduction or debulking was attempted in the majority of patients (90.5%). Eleven patients (52.4%) required some form of intestinal resection; 7 patients underwent either a small or large bowel resection whereas 4 patients underwent both small and large bowel resections. The mean estimated blood loss was 603 mL (median, 400 ml; range, 50–2000 ml). Four patients (19.0%) required intraoperative or postoperative blood product replacement of 2–4 units of packed red blood cells (mean, 3 units). There were no reported postoperative deaths; however, significant postoperative morbidity (thromboembolic event, bacteremia, and aspiration pneumonia) was documented in 5 patients (23.8%). Four patients (19.0%) experienced minor postoperative complications (incisional cellulitis and urinary tract infection). There were no instances of intestinal anastomosis breakdown among the 11 patients undergoing bowel resection.
|Procedure||% of cases (n)|
|Tumor cytoreduction (debulking)||90.5% (19)|
|Excision of residual omentum/gastrocolic ligament||71.4% (15)|
|Small bowel resection||38.1% (8)|
|Large bowel resection||33.3% (7)|
|Lymph node excision||33.3% (7)|
|Resection of diaphragm/abdominal wall||19.0% (4)|
|Partial vaginectomy||19.0% (4)|
|Liver resection||14.3% (3)|
|Distal pancreatectomy||4.8% (1)|
|Partial gastrectomy||4.8% (1)|
|Hysterectomy/unilateral salpingo-oophorectomy||4.8% (1)|
Pathology specimens from secondary surgery were reviewed at JHMI in 15 cases and all were interpreted as recurrent MPSC. Pathology findings for the remaining six surgical cases that were not reviewed at JHMI were described as either recurrent MPSC (five cases) or well differentiated serous carcinoma with psammoma bodies (one case) by extramural pathologists. No patient was found to have moderately or poorly differentiated serous carcinoma at the time of tumor recurrence.
For patients undergoing surgical intervention, multiple factors were evaluated for their strength of association with an optimal secondary cytoreductive surgical outcome (Table 3). After controlling for all measured variables, age ≥ 46 years (odds ratio [OR], 0.06, 95% confidence interval [95% CI], 0.01–0.68; P = 0.02) was found to be the only statistically significant independent predictor of surgical outcome, with advancing age associated with a lower likelihood of optimal residual disease. The administration of chemotherapy prior to surgical intervention was associated with a trend toward a lower likelihood of optimal secondary cytoreduction (OR, 0.10, 95% CI, 0.01–1.07; P = 0.06), but this difference did not reach statistical significance. The PFI, the presence of clinical ascites at the time of disease recurrence, and anatomic regional distribution of recurrent metastatic tumor sites were not found to be associated significantly with secondary cytoreductive surgical outcome. There were no patients without invasive tumor implants at the time of initial surgery who did not undergo optimal secondary cytoreduction; consequently, this variable could not be evaluated by logistic regression analysis. Using the Fisher exact test, the presence of invasive implants at the time of initial surgery was not found to be associated significantly with surgical outcome (two-sided P value = 0.24).
|Variable||OR||95% CI||Significance (P value)|
|Age ≥ 46 yrs||0.11||0.02–0.72||0.02|
|Primary surgery ≤ 1 cm||3.33||0.59–18.89||0.17|
|PFI ≥ 12 mos||2.40||0.48–11.89||0.28|
|Anatomic regions ≥ 3||1.67||0.34–8.26||0.53|
|Chemotherapy prior to surgery||0.21||0.04–1.11||0.07|
|Multivariate logistic regression model|
|Variable||OR||95% CI||Significance (P value)|
|Age ≥ 46 yrs||0.06||0.01–0.68||0.02|
|Chemotherapy prior to surgery||0.10||0.01–1.07||0.06|
Four patients were treated with secondary surgery alone whereas 22 patients received some form of systemic chemotherapy prior to surgery, after surgery, or as the sole treatment modality after disease recurrence. The following agents were used either alone or in various combinations: cisplatin, carboplatin, paclitaxel, etoposide, topotecan, cyclophosphamide, ifosfamide, and pyrazoloacridine. One patient was treated with an investigational angiogenesis inhibitor. Fourteen patients received platinum-based salvage chemotherapy, with the most common regimen being a platinum agent combined with paclitaxel (n = 7). The next most frequently used agent was topotecan (five patients). Because of the wide variety of agents administered, comparisons of response rates between individual agents were impractical.
Twenty of the 22 patients receiving chemotherapy for disease recurrence had adequate follow-up information with which to evaluate the response to treatment. Eight patients were left with no measurable disease and had a median time to subsequent disease progression of 24.6 months. Three of these patients remained without clinical evidence of disease at follow-up times of 24 months, 26 months, and 36 months, respectively. Of the 12 patients with measurable disease, 2 achieved a complete clinical response and 1 patient achieved a partial clinical response, for an overall response rate to salvage chemotherapy of 25% (3 of 12 patients). Three additional patients achieved stable disease while receiving chemotherapy, whereas 6 patients (50%) experienced disease progression.
The median follow-up time from the date of diagnosis of recurrent disease for surviving patients was 29.5 months (mean, 33.7 months; range, 12.1–58.6 months). At the time of last follow-up, 12 patients (46.1%) were dead of disease, 10 patients (38.5%) were alive with disease, and 4 patients (15.4%) were without clinically detectable tumor. The median overall postrecurrence survival for all patients was 56.2 months (Fig. 2).
Univariate survival estimates were obtained for demographic, surgicopathologic, and therapeutic covariates (Table 4). Age at recurrence, PFI, and the extent of anatomic regional tumor involvement were not found to be associated significantly with postrecurrence survival. Without controlling for other factors, the administration of chemotherapy prior to surgical intervention (hazard ratio [HR], 5.17, 95%CI, 1.34–19.90; P = 0.02) and optimal secondary cytoreduction (HR, 0.05, 95%CI. 0.01–0.43; P = 0.01) were found to be associated significantly with survival, whereas an optimal primary surgical outcome was of borderline significance (HR, 0.32, 95%CI, 0.09–1.06; P = 0.06). Kaplan–Meier survival event time distributions demonstrated that optimal secondary cytoreductive surgery was associated with a median OS time from the date of recurrence of 61.2 months. This was statistically significantly longer than both patients undergoing suboptimal secondary cytoreduction (median survival of 25.5 months; P = 0.01) and those patients treated with chemotherapy alone (median survival of 29.9 months; P = 0.01) (Fig. 3). There was no statistically significant difference with regard to survival between patients undergoing suboptimal secondary cytoreduction and those treated with chemotherapy alone (log-rank P = 0.60); consequently, the survival estimates for these groups were combined for statistical comparison by multivariate regression analysis. In the final multivariate model, optimal secondary cytoreductive surgery was found to be the only significant and independent predictor of OS after disease recurrence, whereas the administration of chemotherapy prior to surgical intervention was found to be of borderline significance (Table 3). After controlling for these variables, there was no demonstrable association between residual disease after primary surgery and post-recurrence survival.
|Variable||HR||95% CI||Significance (P value)|
|Age ≥ 46 yrs||2.19||0.63–7.62||0.22|
|Primary surgery ≤ 1 cm||0.32||0.09–1.06||0.06|
|PFI ≥ 12 mos||0.57||0.17–1.93||0.37|
|Anatomic regions ≥ 3||1.12||0.32–3.94||0.86|
|Chemotherapy prior to surgery||5.17||1.34–19.90||0.02|
|Secondary surgery ≤ 1 cm||0.05||0.01–0.43||0.01|
|Cox proportional hazards regression model|
|Variable||HR||95% CI||Significance (P value)|
|Chemotherapy prior to surgery||3.89||0.90–16.87||0.07|
|Secondary surgery ≤ 1 cm||0.06||0.01–0.53||0.01|
The recognition of MPSC as a distinct subgroup of SBTs stems from the observation that although early-stage disease appears to have a very favorable long-term outcome, advanced stage MPSC predicts a high risk of recurrence and tumor-related death compared with nonmicropapillary SBTs.1, 2, 4 For patients with FIGO Stages II–IV MPSC, the risk of recurrent disease ranges from 45–62%.1, 2 Furthermore, the disease-related mortality rate may be as high as 40–50% in the presence of extrapelvic metastases.1, 2 These data, in addition to molecular evidence distinguishing MPSC from both nonmicropapillary SBTs and invasive ovarian carcinomas, have generated controversy regarding whether MPSC should be classified separately as a type of low-grade invasive ovarian carcinoma or should remain a subgroup of the borderline category.4, 5, 8, 10 Two comprehensive review studies examining advanced stage ovarian tumors of low malignant potential with invasive implants provide a useful benchmark for comparison. Summarizing the available literature, both studies calculated a recurrence risk of approximately 45% and a tumor-related death rate of 34–37% with long-term follow-up.4, 19 These outcomes are similar to those reported for advanced stage MPSC and consistent with the high incidence of invasive implants (70–91%) reported among patients with extraovarian disease.1, 2, 11 Therefore, from a practical standpoint, micropapillary tumor morphology can be viewed as a marker of invasive disease when extraovarian tumor implants are present. In all likelihood, it is the presence of this feature that accounts for the adverse outcome associated with both advanced stage MPSC and selected nonmicropapillary ovarian tumors of low malignant potential. The recent description of MPSC, combined with the limited experience of any single institution, has restricted the amount of clinically useful information regarding the management of patients with recurrent disease. As the unique histopathologic features of MPSC become more widely appreciated, clinicians will likely be challenged to treat such patients with increasing frequency. With this in mind, the objectives of the current study were to evaluate variables affecting the clinical outcome of patients with recurrent MPSC and to characterize the survival impact of secondary cytoreductive surgery.
Cytoreductive surgery is well established as an integral component of the initial management of advanced invasive ovarian carcinoma. In contrast, the indications for and efficacy of surgical cytoreduction for recurrent disease have been more controversial. Although some authors have questioned the advisability of repeat debulking surgery for recurrent ovarian carcinoma, there currently is a growing body of literature demonstrating that, in appropriately selected patients, successful secondary cytoreduction is associated with a significant survival advantage.20–30 Recent reports indicate that patients left with optimal residual disease (variably defined) after secondary surgery have a median postrecurrence survival time that is 18–25 months longer than that for patients left with bulky residual tumor.25–29 In the majority of studies published to date, residual disease after secondary cytoreduction has been shown to be an independent predictor of survival outcome. The survival of patients with recurrent invasive ovarian carcinoma after secondary surgery also is linked closely to the effectiveness of subsequent chemotherapy. A PFI of at least 6–12 months from the completion of primary treatment usually is used to identify patients as having chemotherapy-responsive or “platinum-sensitive” disease. Indeed, the PFI has been shown to be an important predictor of not only survival but secondary cytoreductive surgical outcome as well.25, 28, 29 Other factors associated with survival outcome after secondary surgery include performance status and whether salvage chemotherapy is administered prior to surgery.28 Several studies have found patient age, the primary surgical outcome, and the size of the largest recurrence to be significant factors, whereas other studies have not.25–27, 29 Weighing the balance of available literature, it appears that patients with recurrent invasive ovarian carcinoma most likely to benefit from secondary cytoreductive surgery will be highly functional, have platinum-sensitive disease with a PFI of at least 6–12 months, and undergo surgery prior to receiving salvage chemotherapy.
Similar principles of cytoreductive surgery have been advocated for patients with advanced stage ovarian tumors of low malignant potential, although the body of supporting literature is less extensive. In the setting of primary treatment, the amount of residual tumor after initial surgery has emerged consistently as a significant predictor of both progression-free survival and OS.19, 31–33 Contrary to invasive ovarian carcinoma, the effectiveness of adjuvant chemotherapy for advanced stage borderline tumors has yet to be established. Although several retrospective reports have described pathologic responses to chemotherapy based on second-look laparotomy findings, no study published to date has been able to demonstrate a survival advantage for patients with advanced stage borderline tumors who were treated with adjuvant chemotherapy.31, 34–41 Therefore, in the absence of efficacious pharmacologic treatment, cytoreductive surgery takes on potentially greater therapeutic importance for patients with metastatic ovarian borderline tumors.
To our knowledge only three studies to date have addressed the potential benefit of secondary cytoreductive surgery for recurrent ovarian borderline tumors.31, 33, 42 In 1986, Bostwick et al. reported 17 patients with recurrent borderline ovarian tumors of serous or seromucinous histology with a mean PFI of 57 months.31 All patients underwent secondary surgery and complete excision of all macroscopically visible recurrent disease was possible in 15 cases. The only disease-related deaths reported occurred in two patients in whom macroscopically visible residual recurrent tumor remained after incomplete surgical excision. These authors were unable to identify any survival benefit when adjuvant therapy was administered after secondary surgery.31 In 1992, Leake et al. described their experience with 34 patients undergoing secondary cytoreductive surgery for recurrent SBTs diagnosed after a median PFI of 26 months.33 Although the specific details of residual disease were not provided, the long-term disease-related mortality rate was low (38%). It is interesting to note that only five patients in this series received chemotherapy after secondary surgery.33 Recently, Crispens et al. reported 53 patients with recurrent or progressive SBTs who were accrued over a 43-year period at the M. D. Anderson Cancer Center.42 In this series, the median PFI was 5.6 years and 36% of patients with recurrent tumor ultimately died of disease. Thirty-five patients underwent secondary cytoreductive surgery and optimal residual disease (≤ 2 cm) was achieved in 48.6% of cases. The long-term disease-related mortality was 12% for optimally resected patients, compared with 60% for those patients in whom suboptimal residual disease was present. In the final analysis, optimal residual disease was found to be a statistically significant predictor of postrecurrence survival. Nonsurgical therapy was administered in 45 cases, with 42 patients receiving chemotherapy. The overall clinical response rate to nonsurgical treatment was only 26%.42 Taken together, these studies suggest that favorable long-term survival can be achieved with the surgical management of recurrent borderline tumors and cast further doubt on the comparative benefit of adjuvant chemotherapy.
These data provide a contextual backdrop for interpreting the results of the current study. The mean PFI of 31.6 months indicates that patients with recurrent MPSC typically will present at a time remote from the time of initial diagnosis and that close clinical surveillance after primary treatment is required. The high proportion of patients with advanced stage disease is similar to other series of recurrent ovarian borderline tumors.31, 33 Gynecologic oncologists performed secondary surgery in all cases and achieved a 71.4% rate of optimal secondary cytoreduction. In centers with extensive experience in tumor reductive techniques, optimal secondary cytoreduction of recurrent invasive ovarian carcinoma has been reported in 61–87% of cases, with surgical radicality and associated morbidity comparable to that reported in the current series.22, 25, 26, 28–30, 43 Survival data were calculated from the date of recurrence to ensure equivalent starting points for the comparison of patients undergoing secondary surgery initially, those treated with chemotherapy prior to surgery, and patients treated with chemotherapy alone. After a median postrecurrence follow-up of 29.5 months, 46.1% of patients were dead of disease whereas 53.9% were alive with or without detectable tumor. Multiple variables were analyzed for their effect on the clinical outcome of patients with recurrent MPSC. In the final multivariate model, the only independent predictor of postrecurrence survival was optimal residual disease after secondary cytoreductive surgery (HR, 0.06, 95%CI, 0.01–0.53; P = 0.01). Optimal secondary cytoreduction was associated with a median postrecurrence survival of 61.2 months, compared with 25.5 months and 29.9 months, respectively. for patients undergoing suboptimal surgery and nonsurgical patients. Patient age, residual disease after primary surgery, PFI, and the anatomic extent of recurrent disease were not found to be predictive of subsequent survival. The lack of statistical significance may reflect the more indolent disease course associated with MPSC, such that these other factors may be relatively less important determinants of survival after the diagnosis of recurrence. Alternatively, the relatively small number of patients in our database may have limited our ability to detect statistically significant effects.
The only other factor that approached statistical significance in the final multivariate model of OS was the administration of chemotherapy prior to secondary surgery, which was associated with an increased risk of tumor-related death (HR, 3.89, 95% CI, 0.90–16.87; P = 0.07). A similar effect has been reported for invasive ovarian carcinoma, but this finding may carry greater relevance for patients with MPSC in light of the reportedly low objective response rate to salvage chemotherapy.28 The overall response rate of 25% observed in the current study is consistent with the findings of Crispens et al. in their series of recurrent SBTs.42 However, it is unclear whether such responses translate into an associated survival benefit because too few patients in the current study were treated with surgery alone to allow for a meaningful comparison. Analysis of surgical outcome also revealed a trend toward a lower likelihood of achieving optimal residual disease when secondary surgery was preceded by initial treatment with salvage chemotherapy (OR, 0.10, 95%CI, 0.01–1.07; P = 0.057). Theoretically, the administration of chemotherapy prior to secondary surgery may select for those patients with chemotherapy-resistant disease, thereby resulting in apparently worse OS. Conversely, initial treatment with salvage chemotherapy may provide additional time for tumor growth, subsequently decreasing the likelihood of achieving optimal secondary cytoreduction with the attendant adverse effect on long-term outcome.
To our knowledge, the current series represents the largest published report of patients with recurrent MPSC to date and is the first to address the issue of secondary cytoreductive surgery. Nevertheless, the recent description of the distinguishing morphologic features of MPSC coupled with the small number of well documented cases place necessary limitations on the methodology available to study this disease. Although vigorous attempts were made to evaluate all eligible patients, the retrospective study design introduces the potential for the selection bias inherent to all investigations of this nature. We also must acknowledge that the observed survival advantage associated with optimal secondary cytoreductive surgery may have been influenced by unmeasured factors or criteria used to select patients for surgical intervention, rather than the actual treatment provided. A second limitation of the current study stems from the consultative nature of many of the cases and the possibility of referral bias. Although these data are not population-based, the majority of patients did receive some part of their clinical care at JHMI such that our analysis may be seen as representative of a university-based academic practice. Third, a re-review of pathologic material was not undertaken specifically for the expressed purposes of the current study because in the majority of cases the original slides were returned to the referring institutions. In an attempt to minimize interobserver variability, study inclusion criteria required that an attending gynecologic pathologist at JHMI had confirmed each primary diagnosis of MPSC according to standardized institutional criteria. A fourth limitation of the current study is that not all secondary surgeries were performed at the study institution. Although our review of clinical information was comprehensive, we were unable to collect data concerning several parameters (e.g., performance status) that may have affected survival outcome. Finally, and perhaps most important, due to the limited number of cases and wide variety of treatment regimens we were unable to clarify the precise role of chemotherapy for patients with recurrent MPSC.
Despite these limitations, the current data indicate that recurrent MPSC is associated with an unusually protracted disease course, which more closely approximates that of recurrent borderline ovarian tumors with invasive implants than frankly invasive ovarian carcinomas. In the current series, the strongest predictor of post-recurrence survival was optimal residual disease after secondary cytoreductive surgery. Until additional data become available that address the effectiveness of chemotherapy against these tumors, the decision to administer such treatment should be individualized and based on clinical judgement. However, the low objective response rate to chemotherapy observed in the current study makes surgical intervention the preferred treatment for patients with recurrent MPSC. Establishing a centralized database with uniform pathologic review criteria would help to clarify the clinical behavior and appropriate management of MPSC further.
The authors thank the following individuals for providing clinical information: Fouad Abbas, M.D.; Joseph Ashwal, M.D.; Charles Boice, M.D.; Joseph Buscema, M.D.; David L. Child, M.D.; Michael B. Dillon, M.D.; Wesley J. Harris, M.D.; Kenneth D. Hatch, M.D.; B. Frederick Helmkamp, M.D.; Christine Holschneider, M.D.; Sarah Hosford, M.D.; Peter Johnson, M.D.; Isidro Martinez, M.D.; Steven Noskow, M.D.; Michael Purtel, M.D.; and Joan L. Walker, M.D.
- 3Histological typing of ovarian tumors. World Health Organization international histological classification of tumours. 2nd edition. New York: Springer, 1999..
- 12Announcements. FIGO stages: 1988 revision. Gynecol Oncol. 1989; 35: 125–127.
- 15Regression models and life tables [with discussion]. J R Stat Soc. 1972; 34: 187–220..
- 16The analysis of binary data. London: Methuen, 1970..
- 17SAS Institute Inc. SAS user's guide: statistics. Version 5. Cary, NC: SAS Institute Inc., 1985.
- 18Statistics and Epidemiology Research Corporation. EGRET user's manual. Seattle:, 1988.