How can I get screened?

A systematic review of
Screening for Ovarian Cancer

Reprinted from Winter 1998 Newsletter

Since November 1997, a concerted effort has been made to involve consumers in the Health Technology Assessment programme (HTA), from identifying and prioritising research topics to commissioning research and reporting the findings. Following this, they have recently commissioned research on screening for ovarian cancer and a report was sent to us for our information. I now have pleasure in conveying some of the findings to you in the following article.

Background

Ovarian cancer is the seventh most common cancer in women, with a very poor survival rate. This statistic has led to an interest in screening methods that might result in early diagnosis and reduce both mortality and morbidity.

Current screening methods include ultrasound scanning and the CA125 blood test. When used for screening the blood test is followed by an ultrasound scan.

Women with persistently abnormal findings are referred for diagnostic abdominal surgery for removal of ovarian tissue.

Objectives

The objectives of this review are to:

• evaluate the performance of current screening tests for ovarian cancer
• assess the adverse effects of screening, including morbidity associated with surgical intervention and psychological morbidity associated with false-positive diagnosis
• report on the stage of development of newer methods of screening
• investigate the potential cost- effectiveness of screening in different risk groups

Results

The effectiveness of screening
Although there are three trials in progress at present none has yet been completed. In the absence of evidence of effectiveness, it would be premature to establish any kind of screening programme.

Screening test performance
The evidence suggests that both CA125-based screening and ultrasound screening can detect a higher proportion of ovarian cancers at stage I than currently observed in the UK. About 50% are diagnosed at stage I in CA125 based screening studies and about 75% in ultrasound screening studies.

We need to be cautious in interpreting these results because they are based on a small number of cancers detected in diverse studies carried out mainly on self-selected women.

From the limited data available, annual screening with ultrasound appears to have a detection rate close to 100% and CA125 has a rate of about 80%.

About 0.5-1% of women will suffer a significant complication due to surgery and most of those who do not have ovarian cancer will have a benign gynaecological condition. There is a risk that detection of benign and borderline tumours may become a target of ovarian screening, even though they would not have been associated with any morbidity during a patient’s lifetime.

Intervals for ultrasound scanning of between one and thre years are under investigation in the trials. The CA125 screening has been carried out annually. The effect of different screening intervals on the detection rate and false-positive rate has not been formally investigated.

About 3-12% of screened women are recalled for further testing and assessment, resulting in potential distress and anxiety to otherwise healthy women.

Potential impact of screening
The low positive predictive value of ovarian screening is due mainly to the relatively low prevalence of ovarian cancer, which limits the potential cost-effectiveness of general population screening.

Evidence suggests that ultrasound screening is more sensitive than CA125 screening, but that the latter may result in fewer false positives and hence a higher positive predictive value. However, a less sensitive test must be repeated more frequently to achieve the same overall detection rate, which may reduce the apparent advantages of CA125-based screening.

The most efficient screening method and interval is unknown but studies suggest that annual CA125 screening may provide lower overall benefits but be more cost effective at detecting early-stage cancers than annual ultrasound screening. It has been suggested that the addition of colour doppler imaging (CID) to ultrasound screening may reduce the false-positive rate, but reported results are mixed.

Screening a higher risk population
A family history of ovarian cancer is one of the strongest risk factors for developing the disease and, as mentioned in a previous newsletter, there are many UK centres offering screening to such women. Until these trials have been completed there is no evidence as to whether, or by how much, screening women at higher risk reduces mortality.

For some women with an extensive family history of ovarian cancer the increased risk is associated with an inherited genetic mutation. Carriers of some specific mutations may have a lifetime risk of developing ovarian cancer as high as 50-60%. The identification of some of these mutations raises the possibility of testing individuals in these families to determine whether they are carriers, potentially enabling more accurate assessment of risk.

Conclusions

Implications for policy
Further evidence is required before a decision can be made about the potential benefits, harms and costs of screening for ovarian cancer. While awaiting the results of the current trials, demand for screening is likely to increase, and a strong national lead will be required.

The relatively low prevalence of ovarian cancer means that the positive predictive value of screening tests is low. Since the consequence of a false-positive result is a surgical procedure, consideration of the overall impact of screening is important. The low prevalence also limits the potential cost effectiveness of population screening.

Screening women who are at risk because of a strong family history may be more cost effective but this has not been established. No trials are planned in this group, but a screening study has been established. This will provide some evaluation using intermediate outcomes of screening but may also increase demand for screening services.

Implications for research
In a few years, trials should provide an estimate of the impact of screening on mortality.

New or modified screening tests should be compared with those being evaluated in current trials. Test developments which require further evaluation include the marginal impact of adding CDI to ultrasound screening; the use of CA 125 levels in multivariate algorithms to determine thresholds for ultrasound and surgical intervention and the marginal value of adding CA125 measurement to ultrasound screening. The screening modalities will require continuous re-evaluation in line with technical developments.

Research efforts should be directed towards evaluating both the clinical and cost effectiveness of screening strategies for patients at high risk. This includes investigation of any differences in the natural history, performance of screening tests compared with the strategies used in trials, investigation of age-specific risks of developing ovarian cancer and psychological impact and value of risk assessment.

Research is also needed into the impact of genetic testing on health outcomes and the level of demand for such services.