Intervention Review

You have free access to this content

Surgical cytoreduction for recurrent epithelial ovarian cancer

  1. Thuria Al Rawahi1,
  2. Alberto D Lopes2,
  3. Robert E Bristow3,
  4. Andrew Bryant4,
  5. Ahmed Elattar5,
  6. Supratik Chattopadhyay6,
  7. Khadra Galaal2,*

Editorial Group: Cochrane Gynaecological, Neuro-oncology and Orphan Cancer Group

Published Online: 28 FEB 2013

Assessed as up-to-date: 28 JAN 2013

DOI: 10.1002/14651858.CD008765.pub3


How to Cite

Al Rawahi T, Lopes AD, Bristow RE, Bryant A, Elattar A, Chattopadhyay S, Galaal K. Surgical cytoreduction for recurrent epithelial ovarian cancer. Cochrane Database of Systematic Reviews 2013, Issue 2. Art. No.: CD008765. DOI: 10.1002/14651858.CD008765.pub3.

Author Information

  1. 1

    The Royal Hospital, Department of Obstetrics and Gynaecology, Seeb, Oman

  2. 2

    Princess Alexandra Wing, Royal Cornwall Hospital, Gynaecological Oncology, Truro, Cornwall, UK

  3. 3

    University of California - Irvine, Medical Center, Division of Gynecologic Oncology, Orange, CA, USA

  4. 4

    Newcastle University, Institute of Health & Society, Newcastle upon Tyne, UK

  5. 5

    City Hospital & Birmingham Treatment Centre, Birmingham, West Midlands, UK

  6. 6

    St James's University Hospital, Gynaecological Oncology, Leeds, UK

*Khadra Galaal, Gynaecological Oncology, Princess Alexandra Wing, Royal Cornwall Hospital, Truro, Cornwall, TR1 3LJ, UK. Khadra.Galaal@rcht.cornwall.nhs.uk. khadragalaal@yahoo.co.uk.

Publication History

  1. Publication Status: Edited (no change to conclusions)
  2. Published Online: 28 FEB 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Ayhan 2006

MethodsRetrospective study


Participants64 women, recurrent EOC,

from 1990 to 2001.

Mean age of 51 years.

Inclusion criteria:

Women with recurrent EOC, at least 6 months post primary treatment.

E xclusion criteria:

Women with progressive disease (recurrence within 6 months of initial surgery).

Disseminated intrahepatic or extra-abdominal metastasis.

Palliative surgery rather than a cytoreductive effort.


InterventionsInterventions:

53 women (83%) had optimal SCR ≤ 1 cm of residual disease (28 women without macroscopic disease and 25 women with residual disease ≤ 1 cm)

Comparison:

11 women (17%) had sub-optimal cytoreduction > 1 cm residual


OutcomesPrimary outcome:

MST for optimally cytoreduced: 28 months.

MST for sub-optimally cytoreduced: 18 months.

Secondary outcome:

Multivariable analysis showed 3 factors were associated with a favourable outcome after SCR: optimal cytoreduction during primary surgery, optimal cytoreduction during SCR surgery and endometrioid type of tumour histological type.

Analysis of SCR success:

Multivariate analysis of (age, number of recurrent disease and maximal diameter of the recurrence): only age (≤ 50 years had 57.6% no macroscopic disease vs. > 50 years had 29% no macroscopic disease) and number of recurrent disease (no macroscopic disease in 92% and 31% of women with 1 and ≥ 2 recurrent disease; P value = 0.001) to be significant for maximal SCR.


Notes33 women (52%) had SCR included only resection of the recurrent tumour.

31 women (48%) SCR included additional visceral organ resections (intestinal and bowel resection in 17, splenectomy in 10, partial liver resection in 2 and removal of the bladder in 2).

All women received postoperative adjuvant chemotherapy after the SCR.

56 women (87%) had chemotherapy only.

4 women (6%) had radiation therapy only.

4 women (6%) had chemoradiation.

Median follow-up: 33.7 months.

Outcome of primary surgery was significant for survival after SCR (30 months for optimal vs. 18 months for sub-optimal).

DFI significantly affected the OS duration (If DFI < 12 months, the MST was 18 months and if DFI > 12 months the MST up to 39 months).

Chemotherapy before SCR has no effect of survival outcome.

Authors carried out a multivariate analysis and found that optimal cytoreduction during primary cytoreduction (P value = 0.003), endometrioid-type tumour histology (P value = 0.005) and SCR (HR 0.47, 95% CI 0.22 to 0.99; P value = 0.04) were all significant factors, which indicates that women who had no visible disease after SCR surgery were at less risk of death than those with visible disease. In the meta-analysis, the reference group was reversed so that all forest plots were consistent and a HR > 1 indicated that women with residual disease > 0 cm had increased risk of death compared to those with microscopic disease


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot done as it was retrospective study. Women who had SCR surgery were recruited after an extensive discussion by the multidisciplinary team.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques that were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

Chi 2006

MethodsRetrospective study


Participants153 women with recurrent EOC from January 1987 to December 2001.

Median age of 56.5 years.

Inclusion criteria:

Women diagnosed with EOC who had undergone primary surgery and received platinum-based chemotherapy.

Had completed clinical remission of at least 6 months.

For women who underwent second-look surgery and received further chemotherapy, the time of the clinical remission was measured from completion of the therapy after the second-look surgery.

Exclusion criteria:

Women with low malignant potential.

Women who underwent surgery for correction of malignant bowel obstruction.


InterventionsInterventions:

62 women (41%) had no macroscopic disease.

17 women (11%) had residual disease of 0.1-0.5 cm.

21 women (14%) had residual disease of 0.6-1.0 cm.

52% of women had residual disease that measure ≤ 0.5 cm.

Comparison:

11 women (7%) had residual disease of 1.1-2.0 cm.

41 women (27%) had residual disease of > 2 cm.

1 woman (1%) had residual disease status unknown.


OutcomesPrimary outcome:

Median survival of women who had optimal cytoreduction (≤ 0.5 cm)

was 56.2 months.

MST of women who had sub-optimal cytoreduction (≥ 0.5 cm) was 26.7 months.

105 women (69%) died of disease.

23 women (15%) remained alive with no evidence of disease.

25 women (16%) remained alive with disease.

Secondary outcome:

No perioperative mortalities.

6 (4%) intraoperative complications, all which involved bowel injuries.

6 (4%) postoperative complications.

3 women had infectious process requiring antibiotics.

2 women had venous thromboembolism.

1 woman had bleeding secondary to gastritis.

On multivariate analysis:

DFI,

the number of sites of recurrence and

residual disease after SCR retained prognostic significance


NotesMedian OS was 41.7 months.

All women had a CT scan preoperatively.

129 women (84%) received platinum- based chemotherapy after their SCR.

21 women (14%) received non-platinum- based chemotherapy.

3 women (2%) received unknown treatment.

No statistically significant difference between survival of women who were left no macroscopic residual disease that measured from 0.1 cm to 0.5 cm.

Median follow-up was 36.9 months.

Median DFI was 17 months.

MST was 60 months for single site of recurrence.

MST was 42 months for multiple sites of recurrence.

MST was 28 months for women who had carcinomatosis.

The only continuous factor that had prognostic significance in the univariate analysis was DFI.

In general, the median survival improved significantly with longer DFIs, fewer sites of recurrence and SCR to residual disease that measured ≤ 0.5 cm.

The suggestion from this study was that the objective of SCR should be to achieve residual disease that measures ≤ 0.5 cm.

The following selection criteria are suggested:

For women with only 1 of site of recurrence

DFI is ≥ 6 months

Offer SCR surgery for women with multiple recurrence sites but no carcinomatosis who have a DFI > 12 months.

For women with carcinomatosis who have a DFI > 30 months (SCR may be beneficial).

SCR not recommend for women who have a DFI of 6-12 months with evidence of carcinomatosis.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskThis was a retrospective non-randomised study.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques that were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasHigh riskLack of specific information on outcomes of comparable women with recurrent disease who were managed without SCR.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

Eisenkop 2000

MethodsProspective study


Participants106 women with recurrent epithelial ovarian carcinoma from 1990 to 1998.

Median age of 60.5 years.

Inclusion criteria:

Completion of primary surgery and chemotherapy with a clinical, radiographic and serological DFI of at least 6 months after primary adjuvant chemotherapy.

Absence of unresectable extra-abdominal or hepatic metastases.

Patient willingness to be treated with chemotherapy or radiation therapy after surgery.

Absence of medical contraindications to an extensive surgical procedure.

Gynecologic Oncology Group (GOG) performance status < 4.


InterventionsIntervention:

87 women (82%) had complete cytoreduction, no visible residual disease.

3 women (3%) had residual disease < 5 mm.

Comparison:

16 women (15%) had residual disease > 5 cm.


OutcomesPrimary outcome:

Complete cytoreduction MST 44.4 months.

Optimal/sub-optimal cytoreduction (any residual disease); MST 19.3 months (P value = 0.007).

Secondary outcome:

2 women (2%) died (1 of multiple organ failure and 1 of sepsis).

34 women (32%) had postoperative complications.

11 women (10%) had wound infection.

27 women (26%) had prolonged ileus.

10 women (9%) had septicaemia.

3 women (3%) had enterocutaneous fistula or pneumonia.

2 women (2%) had systematic candidiasis, partial fascial separation or mechanical small bowel obstruction managed surgically.

1 woman (1%) had a superior vena cava syndrome, cholecystitis, RDS, DVT, pseudomembranous colitis and a vesicovaginal fistula.

Multivariate analysis: use of salvage chemotherapy before surgery, the GOG performance status and size of the largest site of recurrent disease were independent predictor of survival.

Preoperative GOG performance status: survival: (0 (100%), 1 (91%), 2 (82%), 3 (47%); P value = 0.001).

This study confirmed that the completeness of surgical resection independently determines the prognosis and is proven to improve survival.


NotesOverall MST from date of SCR was 34.4 months.

64 women (60.5%) had secondary surgery before salvage chemotherapy.

42 women (39.5%) had salvage chemotherapy before secondary surgery.

Among the women with bulky unresected disease:

6 women (38%) had open and closed procedure

7 women (44%) had palliative procedure such as gastrostomy or colostomy.

After recovery from surgery, women received salvage therapy based on initial treatment.

34 women (32%) received IV platin-paclitaxel therapy.

24 women (23%) were treated with other platinum-based systemic combination therapy.

7 women (6.5%) were treated with either paclitaxel or platin-based intraperitoneal chemotherapy.

4 women (4%) received whole abdomen radiation therapy.

5 women (5%) did not receive any therapy.

Median survival for women who were not treated with salvage chemotherapy before SCR was 47% and 40% of those women survived > 5 years after recurrence.

MST for women treated with salvage chemotherapy before secondary surgery was 15.8 months and 15% of those women survived > 5 years after recurrence.

Majority of women had advanced disease at primary diagnosis and bulky, multifocal disease at time of recurrence (87 women (82%) had multifocal disease sites).

Survival was influenced by:

DFI after primary treatment (6-12 months (median 56.8 months) vs. 13-36 months (median 44.4 months) vs. > 36 months (median 56.8 months); P value = 0.007)

The use of salvage chemotherapy before secondary surgery (chemotherapy given (median 24.9 months) vs. chemotherapy not given (median 48.4 months); P value = 0.005)

Largest size of recurrent tumour (< 10 cm (median 37.3 months) vs. > 10 cm (median 35.6 months); P value = 0.04).

The probability of complete cytoreduction was influenced by:

The largest size of recurrent tumour (< 10 cm (90%) vs. > 10 cm (67%); P value = 0.001). Women with metastases > 10 cm in largest dimension were rendered visibly disease free 67% of the time.

Use of salvage chemotherapy before secondary surgery (chemotherapy given (64%) vs. chemotherapy not given (94%); P value = 0.001).

Log rank analysis revealed the DFI, use of salvage chemotherapy before secondary surgery, the largest size of recurrence disease and cytoreductive outcome to influence the probability of survival.

Patient age, GOG performance status, tumour grade, histology, presence or absence of ascites, location of largest recurrence tumour, sub-speciality training of physician involved at primary surgery, the number or specific types of procedures performed at secondary surgery and the presence or absence of symptoms, physical findings, or preoperative radiographic findings of ascites, retroperitoneal lymph nodes, or other intra-abdominal masses did not influence the probability of survival.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskA prospective study, no randomisation.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques that were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

Harter 2006

MethodsRetrospective study (DESKTOP OVAR TRIAL)


Participants267 women with recurrent EOC.

Median age 60 years

From January 2000 to December 2003 from 25 institutions

Exclusion criteria:

Women with non-EOC.

Women with borderline tumours.

Women who had operations with palliative purposes or within primary therapy (second-look or interval operations).


InterventionsIntervention:

133 women (50%) had complete cytoreduction.

69 women (26%) had optimal cytoreduction (residual tumour of 1-10 mm).

Comparison:

22 women (8%) had sub-optimal cytoreduction (residual tumour of 11-20 mm).

43 women (16%) had sub-optimal cytoreduction (residual tumour of > 20 mm).


OutcomesMST of complete cytoreduction (without residual tumour) was 45.2 months.

MST with residual tumour, irrespective of its size was 19.7 months (HR 4.33, 95% CI; P value < 0.0001).

The size of residual tumour did not impact survival in women not completely debulked.

MST of women with residual tumour of 1-10 mm and > 10 mm was 19.6 and 19.7 months, respectively (P value = 0.502).

Multivariate analysis:

Significant factors for survival following SCR.

Complete cytoreduction (residual tumour at surgery for recurrence 0 vs. > 0 mm; P value < 0.001).

Absence of ascites.

Application of a platinum-containing chemotherapy.


Notes168 women (63%) had DFI was 12 months from primary surgery.

92% had a good performance status (ECOG).

69% had advanced disease at initial diagnosis.

Some women had salvage chemotherapy before the surgery.

All women received platinum-based first-line chemotherapy.

73% had recurrent disease localised beyond the pelvis.

Post SCR treatment with platinum-based chemotherapy in 47%.

43% had received other chemotherapy.

No postoperative chemotherapy was documented for 10.5%.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskA retrospective study, no randomisation.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques that were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

Oksefjell 2009

MethodsRetrospective study


Participants789 women treated for the first recurrence EOC from 1985 to 2000

217 women with EOC who had any surgical procedure following primary debulking and chemotherapy.

Inclusion criteria:

Surgery for cytoreduction or for bowel obstruction.

Non-responders to chemotherapy during primary treatment.

Tumour relatively localised in pelvis or upper abdomen.

Age, good performance status and TFI > 6 months.

Exclusion criteria:

Women with borderline tumours.

571 women who got only chemotherapy at first relapse.


InterventionsIntervention:

68 women (35%) without macroscopic tumour.

Complete cytoreduction was achieved in 49% of women operated with SCR intentions.

Comparison:

33 women (17%) having tumour nodules ≤ 2 cm residual tumours.

21 women operated for bowel obstruction were excluded from statistical analysis biopsy the residual tumour after SCR was not registered.


OutcomesPrimary outcome:

MST was 4.5 years in women who were SCR to no residuum compare to

0.7 years in women left with residual disease > 2 cm and

2.3 years when ≤ 2 cm macroscopic disease was left.

Multivariate analysis:

Residual tumour after SCR, TFI and age as independent prognostic factors for survival.

Localised tumour remained as prognostic factor in binary logistic regression.

There is a clear survival benefit for women who had undergone a secondary complete cytoreduction followed by chemotherapy compared with chemotherapy alone at the time of first recurrence.


NotesTreatment of the first relapse was registered in 3 groups:

  • no surgery;
  • surgery for local disease (1 or 2 lesions in abdomen or pelvis and refers to what is found during the operation);
  • surgery for disseminated disease.


Ascites and performance status were not always registered.

Types of chemotherapy:

Either as chemotherapy alone or as post-SCR therapy.

SCR was chosen for 217 (27%) of 789 women with recurrence EOC.

At relapse significantly more women with TFI > 24 months had SCR.

Significantly more women > 70 years had chemotherapy.

84 of 217 women had localised disease and MST was 3.4 years.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskA retrospective study, no randomisation.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques which were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

Salani 2007

MethodsRetrospective study


Participants55 women with recurrent EOC from September 1997 to March 2005.

Median age at recurrent was 57.7 years.

Inclusion criteria:

Complete a clinical response to the primary therapy.

≥ 12 months between the initial diagnosis and recurrence.

Performance status ≤ 2.

Attempted SCR.

≤ 5 recurrence sites within the abdomen or pelvis on preoperative imaging studies.

Exclusion criteria:

Women who underwent an interval debulking or second-look procedure with findings of macroscopically positive disease.


InterventionsIntervention:

Complete cytoreduction was achieved in 41 women (74.5%).

8 women (14.5%) had optimal cytoreduction (macroscopic disease with a maximal dimension < 1 cm).

Comparison:

6 women (11%) had sub-optimal cytoreduction (> 1 cm of residual disease).


OutcomesPrimary outcome:

MST for women who had complete cytoreduction after SCR was 50 months (41 women).

MST for women who had sub-optimal cytoreduction (macroscopic residual disease) was 7.2 months (14 women) (P value = 0.0001).

Secondary outcome:

Postoperative complications occurred in 14 women (25.5%):

5 women had febrile morbidity.

5 women had ileus.

2 women had respiratory complications.

1 woman had wound dehiscence.

1 woman had pelvic haematoma.

1 (2%) woman had perioperative death (septic shock).

Multivariate analysis: the statistical significant and independent predictors of OS were:

Diagnosis to recurrence interval ≥ 18 months (MST was 49 months vs. 3 months; P value < 0.01).

Number of radiological recurrence sites (MST of 50 months for women with 1 or 2 sites vs. 12 months for women with 3 to 5 sites; P value < 0.03).

Residual disease (MST was 50 months for women with no macroscopic residual disease vs. 7.2 months for women with macroscopic residual disease; P value < 0.01).

Age, tumour, grade, histology, Ca125 level, ascites and tumour size were not associated significantly with survival.


NotesMedian follow-up was 30 months.

Median diagnosis to recurrent interval was 32 months.

A time of SCR the median Ca125 level was 56 U/mL.

MST from time of SCR was 48 months suggesting that women with ≤ 5 lesions on imaging studies, comprise a group with a good prognosis

Median survival advantage > 3 years for women who were limited to 1 or 2 recurrence sites compared with women who had 3 to 5 recurrence sites based on both imaging studies and surgical exploration.

The major end point of this study was the association of clinical factors with post recurrence survival outcome.

Preoperative imaging studies correctly predicted the number of recurrence site in 34 of 55 women (62%) and those predications were confirmed by the operative and pathology reports.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskA retrospective study, no randomisation.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques which were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

Scarabelli 2001

MethodsProspective study


Participants149 women with recurrent EOC from June 1993 to June 1999

Inclusion criteria:

age ≤ 75 years.

Karnofsky performance status ≥ 60.

RFI > 6 months from primary treatment.

Radiographic or physical findings suggestive recurrence.

Absence of unresectable extra-abdominal or hepatic metastases.

Patient willingness to be treated with chemotherapy after recovery of surgery.

Absence of medical contraindications to an extensive surgical procedure.

Exclusion criteria:

SCR during second-look laparotomy.

Interval cytoreduction.

SCR in women with radiographic or physical findings suggestive of progressive disease during primary or any salvage chemotherapy.

Palliative surgery in women with intestinal obstruction.

Presence of other malignancies except for basal cell carcinoma.


InterventionsIntervention:

53 women (35.5%) had complete cytoreduction.

51 women (34.5%) had optimal cytoreduction (residual disease < 1 cm in diameter).

Comparison:

45 women (30%) had sub-optimal cytoreduction of > 1 cm intra-abdominal residual disease.


OutcomesPrimary outcome:

The OS according to residual tumour after secondary surgery in 69 women with DFI of 7-12 months: the estimated 2-year survival rates were:

56% for no macroscopic disease.

9% for < 1 cm residual disease.

6% for > 1 cm residual disease; P value < 0.001.

The OS according to residual tumour after secondary surgery in 59 women with DFI of 13-24 months:

2-year survival rates were:

91% for no macroscopic disease.

49% for < 1 cm residual disease.

17% for > 1 cm residual disease.

The estimated 5-year survival was 53% for women with no macroscopic disease.

None of the women with < 1 cm and > 1 cm outlived the 4-year estimate.

The estimated 2-year survival rates for women with DFI > 24 months were:

14% for no macroscopic disease

53% for < 1 cm residual disease

14% for > 1 cm residual disease

None of these women outlived the 4-year estimate

Secondary outcome:

Median blood units transfused was 3 units

Perioperative complications in 39 women (26%):

10 women (7%) had severe haemorrhages (estimated blood loss > 1000 mL).

7 women (5%) had ureteral lesions.

10 women (7%) had vesical lesions.

12 women (8%) had bowel lesions.

Postoperative complications in 24 women (16%).

2 women (1%) had enterocutaneous.

5 women (3%) had enterovaginal fistula.

2 women (1%) had vesical fistula.

5 women (3%) had ureteral fistula.

7 women (5%) had septicaemia.

3 women (2%) had pulmonary embolism

5 women (3%) died (2 of multiple organ failure, 2 of sepsis and 1 of pulmonary embolism) within 30 days of surgery.

Multivariate analysis RFI, prior chemotherapy combination and residual tumour after secondary surgery were associated independently with OS.

Residual tumour after secondary surgery was by far the most strongly predictive factor for patient's survival.


NotesMedian follow-up was 27 months.

Multivariate analysis

Unexpectedly, RFI > 24 months was not correlated significantly with OS. The explanation for this difference was that 17 women (81%) with RFI > 24 months were heavily treated with chemotherapy before surgery.

The number of chemotherapy combinations from primary surgery significantly influenced survival for women with RFI 13-24 months and > 24 months.

The median estimated survival related to 1 or > 1 previous chemotherapy treatment was 16 vs. 12 months in women with RFI of 7-12 months, 45 vs. 10 months in women with RFI 13-24 months and 35 vs. 13 months in women with RFI > 24 months.

In conclusion, the current study demonstrates a significant survival advantage for women who undergo complete SCR surgery, but despite improved survival, the long-term prognosis remained poor

Women with platinum-resistant disease have an extremely unfavourable prognosis


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskA prospective study, no randomisation.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques which were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

TIAN 2010

MethodsRetrospective study


Participants123 women with recurrent epithelial ovarian tumour from 2002 to 2006.

Median age 51 years.

Inclusion criteria:

PFI > 6 months.

ECOG performance status ≤ 2

Exclusion criteria:

Women with low malignant potential tumours.

Non-EOCs.


InterventionsIntervention:

97 women (79%) optimal cytoreduction ≤ 1 cm.

51 women (41.5%) had complete cytoreduction (R0 cytoreduction) (complete).

46 women (37.5%) had optimal cytoreduction (R1 cytoreduction) (residual disease 0.1-1 cm).

Comparison:

26 women (21%) had sub-optimal cytoreduction (R2) (residual disease > 1 cm).


OutcomesPrimary outcomes:

Mean survival of women who had (R0) complete cytoreduction was 63.2 months.

Mean survival of women who had (R1) optimal cytoreduction was 31.1 months.

Mean survival of women who had (R2) sub-optimal cytoreduction 15.6 months.

The estimated 5-year survival had reached 54% for women with relapsed ovarian cancer who underwent cytoreduction of R0.

When optimal cytoreduction was defined as residual disease ≤ 1 cm, 5-year survival reached 38% in 79% of women with optimal surgical outcomes.

During follow-up:

63 women (51%) died of disease progression.

32 women (26%) were alive with disease.

28 women (23%) were still alive without evidence of recurrence.

Women with R2 were at highest death risk at 12 months after SCR.

Women with R1 were at highest death risk at 2 peaks at 12 and 38 months after SCR.

Women with R0 were at highest death risk at 24 months after SCR.

Secondary outcome:

No perioperative mortality.

5 women (4%) developed major perioperative complications:

1 woman had pelvic haematoma

1 woman had uretero-vaginal fistula

1 woman had vesicovaginal fistula

1 woman had bacterial infection of cutaneous wound.

1 woman had mycotic infection of gastrointestinal tract.

Residual disease after SCR alone was the independent prognostic factor determined by multivariate analysis. It suggests that R0 is the strongest determinant of survival and should be the first consideration in patient selection for SCR.


NotesMST after SCR was 31.7 months.

Median follow-up duration: 26.1 months.

PFI was 16.1 months.

48 women (39%) developed the recurrence within 12 months.

39 women (32%) between 13 and 24 months.

36 women (29%) had more than 24 months.

Obvious ascites (> 500 mL) was found in 13 women (10.5%).

Solitary recurrence in 49 women (40%).

Multiple recurrence in 75 women (60%).

5 years' observation study.

MST with solitary lesion was 44.1 months vs. 28.1 months for multiple sites of recurrence; P value = 0.03.

MST of women with lesions > 10 cm was 14.6 months vs. 38.3 months for lesion < 10 cm; P value = 0.03.

Optimal cytoreduction obtained for 43 women (88%) with solitary lesion and for 55 women (73%) with multiple lesions.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskNot done as this was a retrospective study.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques that were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasUnclear riskInsufficient information to assess whether an important risk of bias exists.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

Zang 2000

MethodsRetrospective study


Participants60 women with recurrent EOC from January 1986 to December 1997.

Median age of 49 years.

Inclusion criteria:

The interval of the clinical remission had to be at least of 6 months from primary treatment (surgery and platinum-based chemotherapy) for stage III and IV EOC.

Secondary surgery done with therapeutic intent.

Exclusion criteria:

Interval cytoreductive surgery.

Second-look laparotomy.

Women who have stable or progressive disease while receiving first-line chemotherapy after primary cytoreductive surgery.


InterventionsIntervention:

23 women (38%) had optimal cytoreduction (residual disease ≤ 1 cm).

Comparison:

37 women (62%) had sub-optimal cytoreduction (residual disease > 1 cm).


OutcomesPrimary outcome:

MST was 19 months for women who had optimal cytoreduction.

MST was 8 months for women with sub-optimal cytoreduction.

Secondary outcome:

No postoperative deaths.

Postoperative complications in 5%:

wound infection, lung infection and ileus in 1 woman each.

Logistic stepwise regression revealed that:

recurrent ascites (P value = 0.007) and residual disease after SCR (P value = 0.01) were important determinant of SCR outcome.


NotesAscites presented in 15 women (25%) at disease recurrence.

Second-line chemotherapy was administered in 11 women before SCR.

Multivariate analysis showed:

3 major prognostic factors affected survival after SCR surgery: residual disease, ascites at recurrence and PFI.

Residual disease after initial operation was not an independent factor of survival but it affected the secondary surgical attempt.

Preoperative second-line chemotherapy was not a variable influencing second debulking surgery (optimal cytoreduction was achieved in 7 of 11 women (64%) who received chemotherapy preoperatively).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)High riskThis was a retrospective non-randomised study.

Allocation concealment (selection bias)High riskConcealment of allocation irrelevant to this study.

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNot reported.

Incomplete outcome data (attrition bias)
All outcomes
Low riskWomen were analysed for OS using appropriate statistical techniques that were used to account for any censoring.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement.

Other biasHigh riskNo details of women who received chemotherapy pre- and post-SCR.

Representative sample?Low riskAll women had recurrent ovarian cancer that had been cytoreduced via SCR surgery.

Comparability of groups?Low riskMultivariate analysis was used to adjust for important prognostic factors in Cox model for OS.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Benedetti 200647 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Berek 198332 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Bristow 200322 women in the study (the review excluded case-control studies and case series of fewer than 30 women). The aim of the study was to evaluate the utility of combined PET/CT for identifying ovarian cancer tumour ≥ 1 cm in women with clinically occult recurrent disease by conventional CT imaging.

Bristow 2009Retrospective study with no comparison group.

The study concentrated on the clinical outcome of women undergoing rectosigmoid colectomy as a component of secondary cytoreductive surgery for recurrent ovarian cancer.

Eisenkop 199536 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Gadducci 200036 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Goto 2011Included women with disease-free interval less than 6 months.

Gronlund 200536 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Gungor 200544 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Helm 200718 women in the study (the review excluded case-control studies and case series of fewer than 30 women). 

Karam 2007A retrospective study of tertiary cytoreduction.

Landoni 199838 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Matsumoto 200646 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Morris 198930 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Munkarah 200125 women in the study (the review excluded case-control studies and case series of fewer than 50 women). 

Park 200614 women with recurrent epithelial ovarian carcinoma in the study (the review excluded case-control studies and case series of fewer than 50 women). The study was mainly to assess the safety, efficacy and impact on survival of low anterior resection and primary anastomosis at the time of en bloc resection for primary and recurrent epithelial ovarian carcinoma.

Tay 200246 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

Tebes 2007Used percentage rather than actual participant numbers in analysis of the data.

Vaccarello 199523 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

van der Vange 20005 women involved (the review excluded case-control studies and case series of fewer than 50 women). 

Zang 2000aThis study combined the 6 months of recurrent epithelial ovarian cancer group and the interval cytoreductive surgery group.

Zang 2003This retrospective study did not report the results of secondary surgery by residual disease. Instead the authors included the effect of giving the women 'redebulking surgery' or secondary chemotherapy in a multivariate Cox regression analysis for overall survival.

Zang 2004Included women with disease-free interval less than 6 months.

Zanon 200430 women in the study (the review excluded case-control studies and case series of fewer than 50 women).

 
Comparison 1. Residual disease (RD) > 0 cm versus microscopic disease

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Overall survival2331Hazard Ratio (Random, 95% CI)2.62 [1.67, 4.11]

 
Comparison 2. Residual disease (RD) ≤ 1 cm versus microscopic disease

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Overall survival2Hazard Ratio (Random, 95% CI)2.91 [1.78, 4.77]

 
Comparison 3. Residual disease (RD) ≤ 2 cm versus microscopic disease

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Overall survival1Hazard Ratio (Random, 95% CI)Subtotals only

 
Comparison 4. Residual disease (RD) >1 cm versus microscopic disease

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Overall survival2Hazard Ratio (Random, 95% CI)6.80 [3.95, 11.71]

 
Comparison 5. Residual disease (RD) > 2 cm versus microscopic disease

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Overall survival1Hazard Ratio (Random, 95% CI)Subtotals only

 
Comparison 6. Residual disease (RD) > 1 cm versus RD ≤ 1 cm

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Overall survival1Hazard Ratio (Random, 95% CI)Subtotals only

 
Comparison 7. Overall survival

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Overall survival5Hazard Ratio (Random, 95% CI)3.59 [2.45, 5.24]

    1.1 Residual disease (RD) of any size vs. microscopic disease
2Hazard Ratio (Random, 95% CI)2.62 [1.67, 4.11]

    1.2 RD ≤ 1 cm vs. microscopic disease
2Hazard Ratio (Random, 95% CI)2.91 [1.45, 5.84]

    1.3 RD ≤ 2 cm vs. microscopic disease
1Hazard Ratio (Random, 95% CI)2.18 [1.08, 4.42]

    1.4 RD > 1 cm vs. microscopic disease
2Hazard Ratio (Random, 95% CI)6.80 [3.15, 14.66]

    1.5 RD > 2 cm vs. microscopic disease
1Hazard Ratio (Random, 95% CI)7.69 [3.72, 15.88]