Ovarian cancer: patterns of care in Victoria during 1993–1995

Descriptive statistics were analysed with the SPSS statistical package.6 All cases were followed up through the VCR and death tapes from the Registrar of Births, Deaths and Marriages and the National Death Index for a minimum of four years. Relative survival proportions were derived by the actuarial method to calculate expected survival and proportional hazards regression to model the relative survival.7 The initial and terminal events from which survival was calculated were diagnosis date and date of death from any cause. No cases were censored; it was assumed that women not known to have died in Australia were still alive. Relative survival proportions were computed separately for each prognostic indicator. This method adjusts the survival proportions for the other causes of mortality associated with age using Victorian life-tables. A multivariate analysis was then performed. Specific relative risks (RR) with 95% confidence intervals were calculated from the regression coefficients of the proportional hazards model for subgroups of women in the study by prognostic factors (eg, stage, age group, tumour grade) and treatment-related factors (eg, chemotherapy, treating specialist, adequate surgery).

Box 3 shows that primary surgery was performed by general gynaecologists in 156 women (35.9%), gynaecological oncologists in 205 (47.2%), general surgeons in 65 (15.0%) and unspecified specialists in 8 (1.8%). Women aged over 50 years or living in rural areas were more likely to be operated on by general surgeons. In contrast, women with early-stage disease (stages I and II) more commonly had their surgery performed by general gynaecologists than gynaecological oncologists or general surgeons (50.7% [79/156] v 29.3% [60/205] v 24.6% [16/65], respectively; P < 0.0001).

When comparing factors related to surgical technique, most women (83.2% [361]) had a vertical incision at surgery.

Only one in three (34.3% [149/434]) surgically treated women had adequate surgery. A quarter with apparent early-stage disease (26.9% [42/156]) had adequate staging surgery, most commonly because of failure to perform pelvic or para-aortic lymphadenectomy; this was performed in only 39.7% (62/156). Gynaecological oncologists were more likely to complete adequate surgery, compared with general gynaecologists or general surgeons. Box 3 shows this comparison for early-stage disease; overall figures were 52% (105/202) for gynaecological oncologists, 21.3% (35/164) for general gynaecologists and 13.3% (8/60) for general surgeons.

In general, women were considered eligible for postoperative chemotherapy if they had stage Ic to IV disease, and in stage Ia and Ib disease if tumours were poorly differentiated, showed clear cell morphology or were densely adherent (Box 1). Of the 373 women considered eligible, 323 (86.6%) received postoperative chemotherapy. The reasons provided for those who did not receive chemotherapy included not relevant (26), contraindicated (6), refused (8) and unspecified (10). Five of 21 who were not considered eligible for chemotherapy actually received it. For 13 of the remaining 40 who received chemotherapy, eligibility for chemotherapy could not be determined because of inadequate information (stage unknown [10] and grade unknown [30]).

Of the 373 patients eligible for chemotherapy, 82.6% (308) received a platinum-based regimen such as carboplatin or cisplatin alone or in combination with cyclophosphamide or paclitaxel. Carboplatin or cisplatin was given alone in 14.2% (53); the remaining women received a combination regimen. Only 4.3% (16) received a platinum–taxane combination. Other agents given first-line included chlorambucil (6), etoposide (3), anthracyclines (4), 5-fluorouracil (1) and gemcitibine (1).

Hormone-receptor (oestrogen and/or progesterone) status was measured in 19.4% (84) and hormone therapy (tamoxifen) prescribed in 7.1% (31) in addition to chemotherapy as part of a multicentre randomised phase III trial. Nine patients (2.1%) received radiotherapy, and two had synchronous rectal and cervical cancers warranting radiotherapy.

Box 4 shows that when analysing early-stage disease separately, a poorer histological grade, surgery performed by a general surgeon and inadequate staging surgery were associated with a worse outcome. Further, in advanced disease, the absence of ascites and use of hormonal therapy were associated with a better outcome, while chemotherapy also improved the outcome but was of marginal significance. The treating specialist, residual disease and tumour grade were not found to be significant predictors for survival.

Numerous prospective studies have evaluated the role of preoperative ultrasound and CA-125 levels over 35 IU/mL8 and have found the sensitivity and specificity of these tests to be comparable (62%–85% and 72%–95%, respectively). However, three studies (including one retrospective study) have combined the results, with or without age, to develop a risk score that improves the negative predictive value of the tests.8-11 Ultrasound in combination with CA-125 assay is recommended for all women with an adnexal mass so that appropriate referral can be made before operation. Only 35.7% of women in our survey had both an ultrasound and a CA-125 assay.

The importance of surgical stage as an independent prognostic factor emphasises the value of adequate surgery where disease is apparently confined either to the ovary or pelvis. One study has shown the importance of a vertical incision at surgery in a retrospective study of 291 women with ovarian cancer, as women were more likely to be incompletely evaluated intraoperatively if a transverse incision was performed.12 Eighty per cent of patients in our survey had a vertical incision.

Internationally accepted standard practice for adequate staging surgery has been described in the report of the National Institutes of Health (NIH) Consensus Conference,5 which calls for peritoneal washings and sampling of pelvic and para-aortic nodes, as well as omentectomy and biopsy of suspicious areas. In our study, only one in four women with apparent early-stage disease received adequate surgery, most commonly because pelvic and para-aortic lymph node samples were not taken. Although controversy exists about the therapeutic role of pelvic lymphadenectomy, most studies have shown that lymph nodes can give positive results when disease is clinically confined to the ovary,13 indicating use of adjuvant therapy with cytotoxic agents. Although no randomised prospective studies have examined survival benefit, one large retrospective, population-based study suggested the outcome for early-stage disease was not affected by surgical staging.14 By contrast, our study did show staging to be associated with improved survival when tumours were adequately staged.

Gynaecological oncologists were more likely than general gynaecologists and general surgeons to complete adequate staging surgery. One study also showed that gynaecological oncologists complied best with accepted optimal practice,10 with 97%, 52% and 35% of women having adequate intraoperative evaluation when operated by gynaecological oncologists, general gynaecologists and general surgeons, respectively, in United States university hospitals. However, "adequate surgery" was not defined and 17% of the study population had borderline tumours. A retrospective study of 533 cases of ovarian cancer diagnosed at the West of Scotland Cancer Surveillance Unit not only demonstrated that "adequate surgery" was more likely to be completed by gynaecologists than general surgeons, but also that survival was better.15 This finding is supported by our study in women with early-stage disease, but such a survival difference may be explained by "stage migration", where failure to complete "adequate surgery" may result in underestimation of some patients' tumour stage, thereby giving the appearance that they had worse results. A US cohort study showed that 31% of early-stage disease was upstaged to stage III if re-operated after referral to one of four Ovarian Cancer Study Group participating institutions.16

In late-stage disease the NIH Consensus Conference recommended aggressive efforts at maximal cytoreduction.5 The aim of cytoreduction is to reduce tumour bulk, because small residual tumours theoretically have better perfusion and a higher growth rate, improving diffusion and efficacy of chemotherapeutic agents17 in addition to providing symptomatic relief from large, compressive masses. Many retrospective studies, including a meta-analysis of 58 reports, have shown residual disease with tumours less than 2 cm to be associated with a survival benefit.18 Further, a meta-analysis of four randomised trials of platinum-based chemotherapy showed residual tumour size to be a major determinant of survival.19 We were unable to confirm optimal debulking surgery to be a significant predictor of survival in women with advanced disease, but this may be because of the large proportion of cases in our study (30%) in which residual disease was not documented. In our study, gynaecological oncologists were more likely to perform optimal debulking surgery. However, the proportion of tumours optimally resected was relatively low, even when surgery was performed by gynaecological oncologists. A retrospective series of 263 cases not only showed that gynaecological oncologists were more likely to perform optimal cytoreductive surgery (defined by residual disease less than 1 cm) in late-stage disease, but also that there was an associated survival advantage in their cohort.20 We were unable to confirm this improvement in survival, again possibly because of the 30% of cases for which residual disease status was unknown.

Chemotherapy is recommended for women with high-risk early-stage and advanced-stage disease after surgery, provided there is no contraindication.5 Current standard practice is to administer combination chemotherapy consisting of platinum (cisplatin or carboplatin) and a taxane (docetaxel or paclitaxel).5,21-23 The chemotherapeutic regimens administered in this cohort of women reflected the standard practice of the study time period, which was to use platinum, most commonly alone or in combination with cyclophosphamide. Taxanes became available only in the latter part of the study.

We found an improvement in outcome in women with advanced disease who received tamoxifen. Tamoxifen has been used for relapsed ovarian cancer, as there is some evidence from observational studies that it may produce a response in a modest proportion of women with relapsed ovarian cancer.24 The results of a local randomised trial are awaited.