Trial results bring screening for ovarian cancer a step closer,” The Guardian has reported. The newspaper says that results from “the largest screening trial ever” for ovarian cancer could lead to a...
Screening for ovarian cancer is now "a step closer”, The Guardian has reported. The newspaper says results from “the largest screening trial ever” for ovarian cancer could lead to a screening programme to detect early stages of the cancer, which kills 4,500 UK women a year.
The trial involved over 200,000 women aged between 50 and 74, who received either no screening, yearly screening by ultrasound scan, or a blood test with an ultrasound for those at high risk. Screening led to surgery in 1.8% of women in the ultrasound group and 0.2% of women in the combined test/ultrasound group. Of these women receiving surgery 89% were actually found to have non-cancerous abnormalities, with most of these being in the ultrasound-only group. The results therefore demonstrate the dilemma with all screening tests, in that early cancer detection needs to be balanced against potentially treating women unnecessarily.
This highly valuable study has demonstrated the accuracy and potential of screening tests for early detection of ovarian cancer, which might otherwise go undetected until it progresses. This trial is ongoing, and long-term survival rates for these women will be published in the future.
Where did the story come from?
This research was conducted by Dr Usha Menon and colleagues at University College London, Elizabeth Garrett Anderson Institute for Women’s Health, and other UK hospitals and research institutions. The study was funded by Vermillion and Becton Dickinson, and one author received research support from Fujirebio Diagnostics. The study was published in the peer-reviewed medical journal The Lancet Oncology.
What kind of scientific study was this?
This was a randomised controlled trial looking at potential methods to screen for ovarian cancer. Due to the non-specific symptoms of ovarian cancer, most women are diagnosed at an advanced stage of the disease when prognosis is poor.
This study is part of the ongoing United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS) randomised trial to assess the effect of screening upon deaths from ovarian cancer. The findings at this stage of the study are concerned with the diagnostic accuracy of the screening tests and the prevalence of detected ovarian cancer. The full trial (to be complete in 2014) looks at how effectively these screening methods and subsequent treatments reduce mortality.
Women aged 50-74 were invited to participate in the study via 27 Primary Care Trusts across the UK. To be eligible, they had to have gone through (natural or surgical) menopause at least one year previously, or to have been taking HRT for menopausal symptoms for longer than one year.
Women were excluded if they had both ovaries removed, current cancer, any previous history of ovarian cancer or were considered to be at increased risk due to family history. Those who had participated in other ovarian cancer screening trials were also excluded. Women with a past history of cancer were eligible if they had no documented persistent or recurrent disease, and had not received treatment in the past year.
Between April 2001 and October 2005, a total of 202,638 women were recruited to the trial and randomised to receive:
- No screening (101,359 women).
- Annual blood screening for CA125 (a cancer marker) followed by transvaginal ultrasound as a second-line test if risk was indicated by CA125 results (50,640 women).
- Annual transvaginal ultrasound scan alone (50,639 women).
If a transvaginal ultrasound was unacceptable to the patient, abdominal ultrasound was performed instead.
Ultrasound scans were reported as normal (normal size and shape to the ovaries or with small, simple cysts), unsatisfactory (poor view), or abnormal (complex size and shape of one or both ovaries, or cysts of larger size). If an ultrasound showed abnormality, a repeat scan by a more experienced clinician was carried out. Women found to be at high risk through CA125 measurement, or with abnormalities confirmed by their ultrasound, received full clinical assessment, treatment and follow-up as required.
All women are currently flagged through the UK medical record systems, which can give the researchers regular notification of any cancer registrations or deaths among the participants. The current report in this ongoing study includes records up to June 2008. The final screening tests will take place in 2011, and all women will be followed-up until the end of 2014.
What were the results of the study?
Within each screening group, 98.9% women received the combined screening (CA125 blood test with/without ultrasound) that they were randomised to receive, while 95.2% of women randomised to receive ultrasound-only were scanned.
Of the women who underwent combined screening:
- 45,523 (90.9%) were classified as low-risk from blood test results, and they continued to receive annual screening.
- 240 (0.5%) were considered to have increased risk and received ultrasound screening.
- 4,315 (8.6%) women had intermediate risk and were recommended for repeat CA125 blood testing in three months; 169 of these intermediate-risk women also received ultrasound.
In total, 409 (0.8%) of women in the total combined screening group received ultrasound, after which 167 (0.3%) women were referred for clinical assessment and 81 proceeded to surgery. There were also women who proceeded to clinical assessment and treatment without any repeat assessment. This gave a total of 97 women (0.2% of this group) who received surgery.
Of the women who received ultrasound screening alone:
- 42,451 (88.0%) women had normal scans and were returned to annual screening.
- 2,774 (5.8%) women showed abnormalities and were given a further ultrasound by an experienced clinician.
- 3,005 (6.2%) women had unsatisfactory initial scans, requiring another scan by a clinician of the same experience. Of these women, 110 were then given another scan by a more experienced clinician.
Overall, 5,779 (12.0%) women in this group required a repeat test and 2,785 (5.8%) women were referred for a scan by an experienced clinician. Of these women, 1,894 (3.9%) were referred for clinical assessment, and 775 women proceeded to surgery. Additionally, 70 women had clinical assessment and surgery following an initial abnormal screen without additional scans. In total, 845 (1.8%) women in the ultrasound group went on to have surgery.
Overall, of the 98,308 women receiving screening in either group, 942 (0.95%) had surgery. Significantly, more women in the ultrasound-only group received surgery compared to those in the combined group (8.7 women in the ultrasound group for every one woman from the combined group).
Of all women who received surgery, 834 (47 of the combined screen group; 787 of the ultrasound group) were found to have benign (non-cancerous) growths or normal ovaries, and of these 24 (2.9%) had a major complication as a result of surgery.
Cancers of the ovaries or fallopian tubes were detected in 87 women undergoing surgery (42 in the combined screen group and 45 in the ultrasound group). There was a significant difference in the specificity of the two tests. The ultrasound-only screen had lower specificity than the combined screen, i.e. women who did not have ovarian cancer were more likely to have an ultrasound scan incorrectly detecting cancer, leading to further unnecessary assessment and surgery, etc.
The specificities (the proportion of people without cancer who have a negative test) were significantly different (99.8% in the combined screening group against 98.2% for ultrasound alone).
The sensitivity (the proportion of people with cancer who have a positive test) of combined screening and ultrasound alone was similar (89.5% compared to 75%); a difference that was not statistically significant. This means that for a woman with ovarian cancer, both types of screening test were equally likely to correctly indicate that she had cancer.
What interpretations did the researchers draw from these results?
The researchers conclude that the sensitivity of both screening tests (either ultrasound alone or CA125 testing with/without ultrasound) for detecting ovarian cancer is similar. However, the nature of the combined test means that fewer women will receive unnecessary repeat testing and surgery because it has higher specificity. This is due to the high prevalence of non-cancerous abnormalities, which are more often identified as potential cancers by ultrasound.
These initial results suggest that ovarian cancer screening is feasible. Results of the complete trial are currently being awaited to see whether screening has any impact upon deaths from ovarian cancer.
What does the NHS Knowledge Service make of this study?
This is a very large and high-quality trial following the effects of two potential methods of ovarian cancer screening. This early analysis is an ongoing trial that has demonstrated that among about 50,000 women screened with each method, 845 women (1.8%) in the ultrasound group and 97 women (0.2%) in the combined screen group went on to receive surgery.
Eighty-nine percent (834) of the women receiving surgery were actually found to have non-cancerous abnormalities, with most of these being in the ultrasound-only group. This left 42 in the combined screen group, and 45 in the ultrasound group with cancer detected by the single phase of screening.
The reduced specificity of ultrasound, as the researchers say, is due to the high prevalence of benign abnormalities among women, which are more often detected by ultrasound. Therefore, these results demonstrate the difficult dilemma with all screening tests – balancing the benefits of early cancer detection (sensitivity) against the drawback of investigating and treating women unnecessarily (specificity).
This study is highly valuable in demonstrating the potential screening tests for this cancer, which (due to few or non-specific symptoms) are often only detected at an advanced stage when prognosis is poor. The report suggests that large-scale screening is a possibility, and that both screening methods are equally able to identify people who have ovarian cancer (at the expense of also detecting women with non-cancerous abnormalities).
The researchers have not yet looked at the number of cancers detected in the control group, nor the results on any reduction in ovarian cancer deaths as a result of screening, which are still to come. They will also look at the psychological effects and costs involved. The researchers say, "What we need to show everybody is not only that this screening programme can pick up the cancer early, but also that we are saving lives."