'Promising' new test for ovarian cancer developed

Tuesday August 27 2013

Ovarian cancer screening ‘has potential’ reports BBC News.

Many cases of ovarian cancer are only diagnosed at an advanced stage, so accurate early-stage tests for ovarian cancer are urgently needed.

The news is based on research into a two-stage screening test for ovarian cancer among postmenopausal women in the US. Screening was based on measuring a protein called CA125 that is associated with ovarian cancer.

However, a raised CA125 result is not always caused by ovarian cancer, but can be caused by other conditions such as fibroids or endometriosis.

To get round this, the researchers categorised women’s CA125 readings as normal risk, intermediate risk and high risk. Women identified as ‘high risk’ had an ultrasound and were referred to a gynaecologist, who assessed the need for surgery to confirm the cancer.

Over 11 years, 10 women from 4,501 (0.2%) underwent surgery. Of these 10 women, four were found to have high-grade ovarian cancer and two had early stage ovarian tumours. While it is good that these tumours were detected, it does not provide conclusive proof that this is a good screening test. We need larger randomised controlled trials to confirm the findings and see whether screening reduces deaths from ovarian cancer.

Where did the story come from?

The study was carried out by researchers from the University of Texas MD Anderson Cancer Center (sic) in the US. It was funded by grants from the MD Cancer Center at the University of Texas among other foundations and philanthropic support. The study was published in the peer-reviewed medical journal Cancer.

The story was picked up by a variety of UK media sources and most reported the study appropriately. Some of the coverage reported that the research suggests detecting the cancer ‘in time to save lives’. Whether screening for ovarian cancer would save lives is currently unproven, so these statements are incorrect.

What kind of research was this?

This was a prospective cohort study of postmenopausal women in the US. The study aimed to determine how accurate a 2-stage screening strategy was at correctly identifying women who did and didn’t have ovarian cancer. The screening test involved categorising risk depending on levels of a particular protein in the blood, called CA125. This protein is commonly referred to as a ‘tumour marker’, as levels tend to be raised in women with ovarian cancer. However, it is not a specific indicator of cancer as many other conditions can cause raised levels, such as fibroids or endometriosis.

Ovarian cancer is the fifth most common cancer in the UK among women and is the most common among postmenopausal women. It is often diagnosed at an advanced stage of disease and as the symptoms can be ‘non-specific’ and similar to other conditions (such as abdominal pain and bloating), it can be difficult to recognise.

Currently, screening is only available for women who are considered at high risk of developing the disease due to a strong family history or inheritance of a faulty gene. Cervical screening (‘smear tests’) is used to detect cervical cancer only, and cannot detect ovarian cancer.

However, all screening tests include weighing up the risks against the benefits. Risks include “false positive” results which may lead to unnecessary anxiety and further invasive testing – which may involve internal examination, such as vaginal ultrasound, and possibly surgical exploration, such as a laparoscopy.

What did the research involve?

Researchers recruited postmenopausal women living in the US aged between 50 and 74 years. Women were excluded from the study if they previously had ovarian cancer or a family history of breast or ovarian cancer.

Participants underwent yearly blood tests that tested levels of CA125. The researchers were interested to see if there was a rise in the level of CA125 compared to levels from the previous blood test.

Blood tests were analysed using previously researched statistical methods and each woman’s risk of developing ovarian cancer was estimated. Women identified as ‘normal risk’ continued to have annual blood tests. Those with an ‘intermediate risk’ had the blood test repeated at three months. Only women identified as ‘high risk’ had an ultrasound (transvaginal ultrasound or TVA), and also received referral to a gynaecologist. Any decision to have surgery to confirm diagnosis was determined by the gynaecologist.

The researchers then used statistical methods to determine:

  • the proportion of women without ovarian cancer who did not have surgery (to estimate what is known as the specificity of the screening test)
  • the proportion of women who underwent surgery who actually had ovarian cancer (called the positive predictive value of the screening test)

What were the basic results?

The researchers analysed 4,051 women over an 11 year period. The average rate of women classified as normal, intermediate or high risk was:

  • 93.3% were considered low risk
  • 5.8% were considered intermediate risk
  • 0.9% were considered high risk

Over the 11 year period, 83.4% remained in the normal risk category, 13.7% had to repeat a CA125 test in three months, and 2.9% (117 women) were considered high risk. Of the 117 women:

  • 82 had normal ultrasound findings
  • 11 had benign (non-cancerous) ovarian findings
  • 10 had ‘suspicious’ ovarian findings
  • 14 did not have ultrasound testing due to different reasons, including recurrence of previously diagnosed cancer

All ten women with ‘suspicious’ ovarian findings underwent surgery on the basis of ultrasound testing and review by a gynaecologist. Of these women:

  • three had benign (non-cancerous) cysts
  • two had stage 1 ovarian tumours
  • four had early-stage high grade invasive ovarian cancer
  • one had endometrial (womb) cancer

The positive predictive value (PPV) of the two stage screening test was 40% (95% confidence interval [CI] 12.2% to 73.8%) for detecting invasive ovarian cancer (four out of 10 women). PPV is the probability that a test will accurately diagnose a disease when a disease is present. The specificity was 99.9% (95% CI 99.7% to 100%); this means that 99.9% of the women who did not have ovarian cancer, tested negative on both tests. 

How did the researchers interpret the results?

The researchers conclude that although this strategy for ovarian cancer screening in postmenopausal women shows excellent specificity (which in this study was defined as the proportion of women without ovarian cancer who did not have surgery), it is not practice-changing at this time.

They say that more conclusive data is required about the sensitivity of the test (the proportion of people with ovarian cancer correctly identified as having the disease) and the test’s effect on decreasing death from ovarian cancer. They add that results from a large randomised controlled trial in the UK to assess sensitivity and mortality should be available by 2015.

The researchers say that using this two-stage strategy for ovarian cancer screening in the general postmenopausal population should be cost-effective. This is because the majority of women would only need to return on a yearly basis for blood tests, and less than 1% of women would need to proceed to ultrasound testing and referral to a gynaecologist. 


Overall, this study provides positive preliminary findings of a two-stage screening test to detect ovarian cancer in postmenopausal women in the US.

Although this study included 4,015 women, ovarian cancer is relatively rare and only 10 women were identified as in need of surgery. More conclusive evidence is required – ideally from large randomised controlled trials – to see whether the screening test correctly identifies women who have ovarian cancer and also has an effect on decreasing death from ovarian cancer.

Also in the current study, 70% of women considered on CA125 levels to be at ‘high risk’ were found to have normal ovaries on transvaginal ultrasound.

A further 9% were found to have benign ovarian conditions only. Of the 10 who went onto have surgery due to suspicious ultrasound findings, six did not have invasive ovarian cancer (although one had womb cancer). Therefore, it also needs to be ensured that this screening test does not lead to a high level of further unnecessary anxiety and intervention in women with non-cancerous conditions.

The results of a UK study of the test, involving approximately 200,000 postmenopausal women, is likely to be available in 2015.

Analysis by Bazian
Edited by NHS Choices