Prognostic factors for patients with stage IV epithelial ovarian cancer receiving intraperitoneal chemotherapy after second-look assessment

Results of Long-Term Follow-up

Authors


Abstract

BACKGROUND.

The aim was to determine the long-term outcome for patients with FIGO stage IV epithelial ovarian carcinoma (EOC) treated with intraperitoneal (IP) chemotherapy after second-look assessment.

METHODS.

By using data from a retrospective cohort of 433 patients who received IP therapy after second-look assessment after primary surgery and initial systemic therapy for EOC between 1984 and 1998 at our institution, all FIGO stage IIIC and IV patients were identified. Standard statistical methods were used.

RESULTS.

Overall, 297 patients met study criteria (246 stage IIIC; 51 stage IV). The median survival for patients with stage IV disease was 34 months compared with 42 months for patients with stage IIIC disease (P = .02). The only significant predictor of overall survival in patients with stage IV disease was the presence of gross residual disease at initiation of IP therapy (P = .027). When comparing stage IV patients with and without pleural effusions to all stage IIIC patients, there was a significant trend toward improved survival in the patients with pleural effusions only compared with other stage IV patients (P = .01).

CONCLUSIONS.

Prolonged overall survival was observed in patients with no gross residual disease at the time of IP chemotherapy initiation. When compared with similarly treated stage IIIC patients, stage IV patients with malignant pleural effusions appear to have a better outcome than those with other sites of metastasis. Future prospective trials should evaluate the use of IP therapy for patients with stage IV EOC by virtue of malignant pleural effusions only who responded to initial systemic therapy. Cancer 2008. ©2008 American Cancer Society.

In the US, epithelial ovarian cancer (EOC) is the eighth most common cancer among women and the fifth leading cause of death, accounting for the most deaths among women with gynecologic malignancies.1 The implementation of comprehensive surgery and platinum plus taxane-based chemotherapy has significantly improved median survival to 57 months over the course of time for patients who are optimally debulked.2 However, as many as 40% to 50% of patients who achieve pathological complete remission after cytoreductive surgery and standard adjuvant chemotherapy will relapse and eventually develop chemotherapy-resistant disease. The majority of the recurrent disease will be confined to the peritoneal cavity.3–5 This has prompted numerous clinical trials to evaluate the role of additional intravenous or intraperitoneal (IP) treatment as consolidation strategies. Studies investigating the use of IP chemotherapy in patients with stage III or IV EOC after second-look assessment have noted prolonged survival in select patients with minimal residual disease at the time of second-look assessment.6–9 Recent large, randomized, controlled phase 3 trials of up-front IP chemotherapy for optimally debulked stage III EOC have shown an improvement in overall survival compared with intravenous chemotherapy alone.10–12

Stage IV EOC, which accounts for only 16% of ovarian cancers, is noted for its poor long-term outcome, with estimated 5-year survival rates of 7% to 17%.13–15 Many authors have reported that optimal cytoreduction is an important determinant of survival in patients with stage IV EOC, advocating optimal debulking of IP disease,13, 15–18 whereas some have reported no difference in survival, arguing that the biology of stage IV disease makes aggressive cytoreductive surgery inappropriate.14, 19 In patients with stage IV EOC, the response rate to intravenous chemotherapy is comparable to that of patients with less-advanced disease.20 Little is known about IP therapy for patients with stage IV disease, and no data are available comparing the different subsets of stage IV diseases with regard to the site of extraperitoneal disease and IP therapy. Many practitioners will not offer up-front IP therapy to patients with optimally debulked stage IV disease because there has been no proven benefit demonstrated in randomized phase 3 trials for this group of patients.

The aim of this study was to evaluate overall survival in patients with stage IV EOC treated with IP chemotherapy after second-look assessment. Within the group of patients with stage IV disease, we also sought to determine how outcome was affected by site of metastasis and considered this in relation to similarly treated stage IIIC patients.

MATERIALS AND METHODS

Patients of this analysis were derived from a previously published retrospective database of 433 patients who received IP chemotherapy after second-look assessment after intravenous primary therapy for ovarian cancer between 1984 and 1998 at Memorial Sloan Kettering Cancer Center. The majority of these patients were treated on phase 2 clinical trials (n = 285), whereas the remainder (n = 126) were treated off-protocol.9, 21–24 In this study we retrospectively reviewed the records of 297 patients with stage IIIC and IV EOC after obtaining Institutional Review Board approval.6 IP therapy was administered as consolidation therapy after negative second-look assessment (n = 59) or for treatment of persistent small volume disease (microscopic disease or residual tumor <1 cm, n = 211) or bulky disease (>1 cm, n = 27) noted at the time of second-look surgical assessment after primary surgery and initial systemic therapy. Patients with nonepithelial histologic subtypes, borderline cancers, earlier stage disease, or those treated with neoadjuvant chemotherapy were not included. The majority of patients (n = 277) had a clinical response to primary therapy (normalization of symptoms, normal clinical examination, and/or normal radiologic findings, and/or normalization of CA 125 <35 U/mL) before undergoing second-look reassessment. The remainder (n = 20) had clinical evidence of persistent or progressive disease.

Data were extracted for patient age, stage, histologic subtype, grade, surgical debulking status (optimal vs suboptimal), presence of paclitaxel in the primary regimen, presence of platinum in the IP regimen, size of residual disease at initiation of IP therapy, and overall survival. All patients included in the study had undergone a second-look surgical assessment of disease, which defined the size of residual disease at initiation of IP therapy. Stage IIIC and IV EOC was defined according to the criteria outlined by the Oncology Committee of the International Federation of Gynecology and Obstetrics (FIGO).25 Follow-up data were available for 286 patients, of whom 51 had documented stage IV disease and 235 had stage IIIC disease.

Second-look assessments were performed both via laparotomy and laparoscopy. Secondary cytoreductive surgery was performed in 10 patients with bulky disease at the time of second-look assessment. IP treatment was administered through a subcutaneous titanium implanted port with a fully fenestrated catheter (Bard Access Systems, Salt Lake City, Utah, part #0603006 or similar), which was generally placed at the time of second-look assessment. Chemotherapy agents used alone or in combination included cisplatin, carboplatin, etoposide, 5-flourouracil, floxuridine, cytarabine, mitoxantrone, paclitaxel, and gemcitabine. Overall survival was defined as the interval from second-look surgery until death or last follow-up. Statistical analyses included tests for associations within or between potential prognostic factors and survival. Survival probabilities were estimated by using the method of Kaplan and Meier. Prognostic factors for survival were evaluated by the log-rank test. Multivariate Cox regression models were used to test for independent associations between survival and clinical-pathologic covariates.

RESULTS

Patient Demographics

The clinical characteristics of the study population are listed in Table 1. The mean age of the 297 patients in the study population was 53 years (range, 26–76 years). Fifty-one patients (17%) were diagnosed with stage IV EOC. Most cancers were high-grade tumors (73%) of serous histology (58%). The distribution of stage IV-defining metastatic sites includes cytologically positive pleural effusions (n = 33), parenchymal liver (n = 9), abdominal wall (n = 3), vagina (n = 2), vulva (n = 1), and mediastinal (n = 1) and supraclavicular lymph nodes (n = 2). Of all 297 patients, optimal debulking at initial surgery with residual tumor no more than 1 cm in greatest dimension was performed in 37%. Optimal surgical debulking was achieved in 39% of patients with stage IIIC EOC and 28% in the stage IV group (P = .15). The size of residual disease at initiation of IP therapy was as follows: none, 67 (23%); microscopic, 89 (30%); less than 1 cm, 114 (38%); and greater than 1 cm, 27 (9%). There was no statistically significant difference in the size of residual disease at initiation of IP therapy between patients with stage IIIC or stage IV disease. The median follow-up of the surviving patients was 9.1 years (range, 0.57-19.6 years).

Table 1. Clinicopathologic Characteristics*
VariableStage IVStage IIIC
  • NOS indicates not otherwise specified; IP, intraperitoneal.

  • *

    Numbers do not sum to the total because of unavailable values.

No. of patients51246
Age, y  
 Mean, [median], range51.8 [51] 34–7653.3 [54] 26–75
Tumor grade  
 10 (0%)15 (7%)
 28 (17%)53 (23%)
 339 (83%)163 (70%)
Histology  
 Serous30 (59%)143 (58%)
 Endometrioid8 (16%)32 (13%)
 Adenocarcinoma, NOS4 (8%)27 (11%)
 Mixed6 (12%)21 (9%)
 Clear cell2 (4%)12 (5%)
 Other1 (2%)10 (4%)
Primary debulking  
 Optimal (≤1 cm)13 (28%)81 (39%)
 Suboptimal (>1 cm)34 (72%)127 (61%)
Residual at initiation of IP therapy  
 None7 (14%)60 (24%)
 Microscopic19 (37%)70 (29%)
 <1 cm21 (41%)93 (38%)
 >1 cm4 (8%)23 (9%)
Chemotherapy  
 Received initial paclitaxel15 (29%)86 (36%)
 Received IP platinum50 (98%)204 (83%)

Treatment Results

The median survival for patients with stage IIIC disease was 42 months compared with 34 months for patients with stage IV disease (P = .02, Fig. 1). Of the 51 patients with stage IV disease, 48 died of disease, 1 patient died of metachronous pancreatic carcinoma, and 2 patients are alive without evidence of disease. A univariate analysis for overall survival for all patients with stage IV disease was performed on the following parameters: age, grade, histology, size of residual disease at initiation of IP chemotherapy, the presence of paclitaxel in the primary regimen, the presence of platinum in the IP regimen, and stage IV-defining metastatic disease site (Table 2). Of all variables, the only significant predictor of survival in patients with stage IV disease treated with IP chemotherapy was the presence of gross residual disease at initiation of IP therapy. Median survival for those stage IV patients receiving IP therapy as consolidation with no gross residual disease was 41 months, and 18 months for those with residual disease at the time of second-look assessment (P = .027; Fig. 2).

Figure 1.

Overall survival according to stage of disease.

Figure 2.

Overall survival based on size of residual disease at the initiation of intraperitoneal therapy.

Table 2. Univariate Analysis of Survival for Patients With Stage IV Epithelial Ovarian Cancer
VariableRisk ratio95% CIP
  1. IP indicates intraperitoneal; CI, confidence interval.

Age, y >51 vs ≤51, median1.060.602–1.876.835
Pleural effusion vs other metastases0.790.437–1.425.432
Histologic grade, 3 vs other0.730.337–1.573.419
Histology, serous vs other0.800.438–1.460.466
Residual disease at initiation of IP therapy, gross vs other1.921.066–3.445.027
Primary paclitaxel given1.000.560–1.929.903
IP platinum given1.260.127–9.217.821
Primary optimal cytoreduction1.620.786–3.332.191

When comparing patients with pleural effusions to patients with other sites of stage IV-defining metastatic disease, the median overall survival was 38 months compared with 25 months, respectively. Stage IV patients with pleural effusions at initial diagnosis were less likely to have gross residual disease at the time of second-look assessment than those with other metastatic sites (P = .014, Table 3). Optimal debulking was achieved in 41% of patients with pleural effusions compared with only 5% in patients with other metastatic sites (P = .008). The median overall survival was 42 months for stage IIIC patients, 38 months for stage IV patients with pleural effusions only, and 25 months for patients with other stage IV-defining disease (P = .036, log-rank; P = .01, trend test; Fig. 3). The subgroup of stage IV patients with no gross residual disease at the time of second-look assessment had a median overall survival of 41 months, which compares favorably with that observed in stage IIIC patients with or without gross residual disease (34 months vs 59 months; P = .001, trend test), respectively. This implies that some stage IV patients fare better than certain stage IIIC patients. On multivariate analysis, stage and size of residual disease at the initiation of IP therapy were both independent predictors of survival (hazard ratio [HR] = 0.67 and 0.56, respectively).

Figure 3.

Overall survival showing differences based on site of stage IV-defining disease.

Table 3. Characteristics of Stage IV Patients
Site of metastasisPleural effusion n = 33, 65%Other* n = 18, 35%P
  • IP indicates intraperitoneal; SLL, second look; NS, not significant.

  • *

    9 liver parenchyma, 3 abdominal wall, 2 each supraclavicular lymph node and vagina, one each mediastinal lymph node and vulva.

  • 4 missing values.

Median survival, mo3825NS
Optimal debulking12 (41%)1 (6%).008
Residual disease at initiation of IP therapy12 (36%)13 (72%).014
Long-term survivors7 (21%)3 (17%)NS

Ten stage IV patients (20%) survived more then 5 years (median age, 49 years; range, 42-57 years, Table 4). The stage IV-defining metastatic sites in the long-term survivor group were as follows: pleural effusion, 7 patients; liver parenchyma, 2; and vagina, 1. Six patients did not have gross residual disease at the time of second-look surgical assessment. In the remaining 4 cases the residual disease was less than 1 cm. Eight patients had poorly differentiated tumors. Seven patients died of their disease, 1 patient died of secondary pancreatic carcinoma after 172 months' follow-up, and 2 patients are still alive with no evidence of disease after 15 years of follow-up. Of the patients who died of disease in the long-term survivor group, all patients except 1 had near-death findings showing extensive intraabdominal carcinomatosis with hepatic involvement; the remaining patient presented with brain metastases.

Table 4. Long-term Stage IV Survivors: Characteristics, Treatment, Survival
Pt. no.Year of diagnosisAge, yHistology, gradeResidual disease at initiation of IP therapyInitial chemotherapy regimenIP-Chemotherapy; no. of cyclesSurvival mo, aliveSites & time to failure/comments
  • CDDP indicates cisplatin; CBDCA, carboplatin; VP16, etoposide; IP, intraperitoneal; NED, no evidence of disease; DFS, disease-free survival.

  • *

    Vagina.

  • Liver parenchyma.

Pleural Effusion       
1198754Serous, 3MicroscopicCDDP/CytoxanCDDP; (4)104, (−)35 mo after primary diagnosis upper abdominal metastases, peritoneal carcinomatosis
2198848Serous, 3<1 cmCDDP/CytoxanCDDP, (4)67, (−)DSF 50 mo (rising CA 125), followed by portal vein and hepatic vein occlusion, pericardial effusion
3198857Endometrioid, n/aNo diseaseCDDP/CytoxanCDDP; (4)172, (−)Advanced pancreatic carcinoma diagnosed 164 mo after diagnosis of ovarian carcinoma
4199048Adenocarcinoma, 3No diseaseCDDP/CytoxanCDDP/VP16; (3)184, (+)NED
5199157Adenocarcinoma, 3No diseaseCDDP/CytoxanCDDP/VP16; (1) (catheter failure, IP chemotherapy discontinued)180, (+)NED; renal failure after 165 mo, requiring dialysis, awaiting kidney transplant
6199348Serous, 3MicroscopicCDDP/PaclitaxelCDDP/Mitomycin; (5)97, (−)First recurrence after 59 mo (retroperitoneal), findings on last follow-up: multiple liver and visceral metastases, small bowel obstruction
7199351Serous, 2<1 cmCDDP/PaclitaxelCDDP; (4)114, (−)DSF 41 mo (lymph node), followed by extensive thoracic metastases, carcinomatosis with portal obstruction
Other       
1*198747Serous, 3MicroscopicCDDP/CytoxanCDDP; (6)117, (−)Brain metastasis 96 mo after primary diagnosis
2199057Serous, 3<1 cmCBDCA/CytoxanCBDCA/VP16; (6)69, (−)Inguinal metastases after 41 mo disease free interval
3199642Endometrioid, 3<1 cmCDDP/PaclitaxelCDDP; (5)99, (−)First recurrence after 26 mo; hepatic and pulmonary metastases, carcinomatosis, retroperitoneal adenopathy

DISCUSSION

The prognosis of stage IV epithelial ovarian cancer is generally poor, and the natural history is often one of persistent disease or progression after primary treatment. These circumstances warrant the consideration of alternative therapeutic strategies. Studies investigating the use of IP chemotherapy in patients with stage III and IV EOC after second-look assessment have noted prolonged survival in select patients with minimal residual disease at the time of second-look assessment.6–8 Recent large, randomized, controlled phase 3 trials of up-front IP chemotherapy for optimally debulked stage III EOC have shown an improvement in overall survival and progression-free survival compared with intravenous chemotherapy.10–12 Up-front IP therapy in optimally debulked patients with stage IV EOC is controversial, however, because its anticipated added benefit derived from enhanced drug exposure is thought not applicable to extraperitoneal sites.

This study evaluated patients with stage IV disease treated with IP therapy after second-look assessment. Technical problems with the fully fenestrated peritoneal catheter were relatively uncommon and have been reported elsewhere.6 With the exception of 2 patients, all patients with stage IV disease died despite primary surgery and platinum-based chemotherapy followed by IP therapy after second-look assessment.

The overall survival of patients with stage IV disease in this series is significantly shorter than in similarly treated patients with stage IIIC disease. In 2002 Barakat et al.6 reported that stage was not a predictor of long-term survival in patients receiving IP therapy after second-look assessment. The only significant predictors of survival were grade and size of residual disease at initiation of IP therapy. In that study, 433 patients of all stages who received IP therapy after second-look assessment were included, and no discrimination had been made between stage IIIA, IIIB, and IIIC disease. The median patient follow-up was 2.7 years (range, 0.06–13.2 years). In the present study the median length of follow-up was 3.4 years, and all but 2 patients died in the stage IV group. The longer follow-up may have contributed to the statistically significant difference in survival when comparing stages IIIC and IV. These results support the FIGO staging system and other data stating that the survival of patients with stage IV disease remains worse than stage IIIC disease.14, 26 Previous reports, such as the study by Eitan et al,26 have also shown that survival of stage IV patients with optimally debulked EOC based on pleural effusions only is significantly worse than the survival of patients with optimally debulked stage IIIC disease. One explanation suggested for the survival difference in that study is the presence of undetected bulky intrathoracic disease.

In this study we looked at several prognostic factors for stage IV disease and found that the size of residual disease at initiation of IP therapy is the only determinant of long-term survival. Median survival for those stage IV patients receiving IP therapy as consolidation with no gross residual disease was 41 months, and 18 months for those with residual disease at the time of second-look assessment. Thus, patients who failed to respond to primary intravenous therapy by evidence of bulky disease at the time of second-look reassessment did not benefit from further IP chemotherapy.

From the data it appears that the site of stage IV-defining disease may also play a role in outcome after IP therapy. When comparing survival for patients with pleural effusions to patients with other sites of metastasis, the difference did not reach statistical significance. However, when both groups of stage IV patients (with or without pleural effusions) are compared with similarly treated stage IIIC patients, there was a significant trend for improved survival in those patients with pleural effusions only. Interestingly, the median survival of stage IV patients with pleural effusions was more similar to those with stage IIIC disease than stage IV patients with visceral or other stage IV-defining metastasis. The trend test suggests that the site of metastasis may be an important determinant of survival. These findings are supported by the outcome of the long-term stage IV survivors in this study. Near-death findings in long-term stage IV survivors defined by pleural effusions who subsequently died of the disease all had extensive hepatobiliary disease, which may support the use of regional therapy in this ‘favorable’ group of stage IV patients. In other words, these patients died of intraabdominal disease, the target of IP therapy, and not distant spread.

In this analysis the median survival of 34 months for all patients with stage IV disease was longer than that shown in other series. Bonnefoi et al.14 reported a median survival of 13.4 months and a 5-year survival rate of 7.6%. Naik et al.13 reported a median survival of 11 months and a 5-year survival of 9%. Several factors could have accounted for this relatively favorable outcome. The majority of these patients (n = 277, 93%) were in clinical remission after primary surgery and initial intravenous chemotherapy. Also, our study population is relatively young (median age, 51 years) when compared with the other series. Recent studies have investigated the relationship between age at diagnosis of ovarian cancer and prognosis. A recent retrospective analysis of a large group of patients with stage III EOC who received primary cytoreduction and intravenous platinum and paclitaxel chemotherapy during participation in a Gynecologic Oncology Group (GOG)27 trial confirmed that age was an independent predictor of prognosis. Whereas this outcome in our study may be explained by the inclusion of only a highly selected, mostly favorable group of stage IV patients (ie, those in clinical remission, young and healthy enough to undergo second-look reassessment and IP chemotherapy), it nonetheless raises the possibility that more favorable survival is an achievable goal in a subpopulation of patients with stage IV disease. Another important difference between this series and others is that the majority of patients (64%) had malignant pleural effusions as their stage IV-defining site of disease.

In our series, patients with stage IV who were treated with primary cytoreduction and intravenous chemotherapy and who were subsequently treated with IP therapy had a better median overall survival if no gross residual disease was found at second-look laparotomy. Drugs delivered by the IP route penetrate only a depth of a few millimeters beneath the tumor surface.28–30 Thus, patients with small-volume residual disease benefit most from this approach. Although randomized trials evaluating the role of consolidation strategies have not demonstrated a statistically significant improvement of overall survival, they have suggested that these strategies are most effective for patients who have responded to primary therapy.31–35 Undoubtedly, there are methodological issues that may not permit generalization of these findings to contemporary patients. The heterogeneity of treatment strategies throughout the study period makes it difficult to account for differences in treatment regimens used during and after IP consolidation therapy. These findings are also limited by a long study period. During this time period, both the standard chemotherapy agents and surgical techniques have improved. Cyclophosphamide and cisplatin are no longer the standard agents used for treating primary EOC. Surgical technique, instrumentation, and approach have all improved as evidenced by the current optimal cytoreduction rate at this institution, which is near 80%, compared with only 37% in the reported series.36

In conclusion, the survival of patients diagnosed with stage IV EOC continues to be short, and even in this ‘favorable’ group selected for IP therapy after achieving a clinical remission, only 2 patients survived. Whereas new approaches are undoubtedly needed, identification of a group of patients with stage IV disease who may benefit from IP consolidation is warranted. Optimal cytoreduction rates may prove helpful in enhancing survival rates even for stage IV patients, and certain subsets may benefit from implementation of strategies heretofore thought applicable only to stage III patients. Because stage IV patients all receive intravenous therapy, consolidation IP therapy may offer an opportunity to extend the benefits of the IP route to a population with a historically poor outcome.

Ancillary