Surveillance of women at high or potentially high risk of ovarian cancer

A family history of breast and ovarian cancer is the most significant risk factor for ovarian cancer but does not necessarily imply an inherited genetic cause. Of women in Australia with non-mucinous invasive epithelial ovarian cancer, 14% were found to have a heritable (germ-line) mutation in the BRCA1 or BRCA2 genes.21 However, over one-third of women in Australia with invasive epithelial ovarian cancer in whom a heritable BRCA1/2 mutation is identified have no known family history of breast or ovarian cancer.21-23 For women with a BRCA1 mutation, the risk of developing ovarian cancer by the age of 80 is 44%, and for women with a BRCA2 mutation the corresponding risk by age 80 is 17%.24

In some ethnic populations such as the Ashkenazi Jewish population, prevalence of heritable BRCA1/2 mutations in women with ovarian cancer is high due to common founder mutations. The frequency of BRCA1/2 mutations for the Ashkenazi Jewish population has been estimated at approximately 2.5% in Australia compared to less than 1% in the general population.25,26 Heritable BRCA1/2 mutations are found in approximately 30% of Ashkenazi Jewish women with ovarian cancer.26,27

A small proportion of hereditary ovarian cancers are due to mutations in a number of DNA mismatch repair genes associated with Lynch syndrome (e.g. MLH1, MSH2, MSH6).28 Heritable mutations in other genes are also associated with ovarian cancer, though the ovarian cancer risk is uncertain.

Definition of high or potentially high risk

The category of high or potentially high risk of ovarian cancer covers less than 1% of the female population. Lifetime risk of ovarian cancer varies but may be up to 50%.24 Individual risk may be higher or lower if genetic test results are known.

Women are at high risk of ovarian cancer if they:

  • have had genetic testing and been found to have a high-risk ovarian cancer-related gene mutation (e.g. in BRCA1/ 2, or one of the Lynch syndrome-associated genes such as MLH1, MSH2, MSH6).

Women are at potentially high risk if they:

In a family that has had genetic testing

  • are an untested member of a family in which the presence of a high-risk ovarian cancer gene mutation has been established.

In a family where no genetic testing has been done

  • have one first-degree relative diagnosed with invasive epithelial ovarian cancer aged less than 60 years; 
  • have more than one first-degree or second-degree relative with invasive epithelial ovarian cancer diagnosed at any age (on the same side of the family); 
  • have a strong family history of both breast and epithelial ovarian cancer; or
  • have a family history of breast and/or epithelial ovarian cancer with Ashkenazi Jewish heritage.

For women at potentially high risk who have not undergone genetic testing, consider referral to a family cancer clinic.

Clinical trials of surveillance in women at high risk or potentially high risk

Two large cohort studies assessed surveillance of women at high or potentially high risk of ovarian cancer using a ROCA triage strategy, and one cohort study used threshold testing of CA125 for surveillance.

Studies to date have been focused on sensitivity, specificity and stage shift. No randomised controlled trials of surveillance among high-risk or potentially high-risk women have been conducted. Due to ethical and practical limitations in randomising this population group between surveillance and control groups, it is unlikely that a randomised controlled trial will be conducted in the future.

The Cancer Genetics Network and Gynecologic Oncology Group (CGN/GOG) study combined the results from two screening trials that used the same ROCA triage testing strategy; CGN (United States; 2001-2011) and GOG-0199 (United States and Australia; 2003-2006). Women over 30 years of age were eligible if: (1) they or a first or second degree relative had a BRCA1/2 mutation, or (2) they or a first or second degree relative had been diagnosed with ovarian and/or breast cancers (at least two cancer diagnoses). A total of 3,818 women were recruited for ROCA evaluation of CA125 every 3 months (with triage to additional CA125 and/or TVUS where indicated) plus annual TVUS.29 Nine ovarian cancers were screen-detected but the rate of false-positive surgeries was high, occurring at a rate of 20.7 per screen-detected cancer.

CA125 thresholds were used for surveillance in the United Kingdom Familial Ovarian Cancer Screening Study Phase I (UKFOCSS – Phase I), using different thresholds for pre-menopausal and post-menopausal women. From 2002 to 2008, 3,563 women aged at least 35 years and with a lifetime risk of ovarian cancer of 10% or more were recruited for annual testing with both CA125 and TVUS. Nine women were diagnosed with ovarian cancer after a positive test and three women were diagnosed within 12 months of a negative test.30 There was evidence of a shift to earlier-stage cancers in women diagnosed within a year of screening compared to those diagnosed in the post-screening phase. Although the false-positive surgery rate was within what may be considered acceptable limits, occurring at a rate of 2.4 per screen-detected cancer, sensitivity was poor, which prompted Phase II of the trial, incorporating the ROCA triage test.

From 2007 to 2012, the UKFOCSS – Phase II study enrolled 4,531 women aged at least 35 years with a lifetime risk of ovarian cancer of 10% or more. Testing took place using ROCA evaluation of CA125 every 4 months (triaging to additional CA125 and/or TVUS where indicated), plus TVUS every year.31 Nineteen cancers were diagnosed within 1 year of prior screening; 13 diagnoses were screen-detected and six occult cancers were identified at RRSO. The rate of false-positive surgeries was high at 11.5 per screen-detected cancer (13 cancers and 149 false-positive surgeries).

Phase II of the UKFOCSS study reported a stage shift, finding a significantly higher proportion of low volume disease, defined as stage IIIa or less (corresponding to less macroscopic peritoneal metastasis outside the pelvis) in women diagnosed within one year of screening compared with women diagnosed more than one year after the last screen (p<0.001). However, half (6/12) of the low-volume diagnoses made within one year of the last screen were occult cancers discovered during bilateral salpingo-oophorectomy unrelated to screening. The study also reported that the proportion of all cases with zero residual disease after surgery (an important prognostic factor in ovarian cancer) was higher in women diagnosed within one year of screening compared with women from the same cohort in whom cancer was diagnosed more than one year after screening was completed, but this finding was not significant.31

Overall, the evaluation of surveillance tests in these uncontrolled trials demonstrates that an acceptable approach to surveillance has not yet been established. The current tests have unacceptably high rates of false positives and/or insufficient sensitivity to detect ovarian cancer.

There is currently insufficient evidence to support ovarian cancer screening and it is not recommended in women at either population or at high or potentially high-risk of ovarian cancer, outside of a clinical trial setting. This recommendation is in line with international guidelines.32-34

The utility of ovarian cancer surveillance in high-risk women will be further explored in a clinical trial underway in the UK known as the Avoiding Late Diagnosis in Ovarian Cancer (ALDO) which aims to detect ovarian cancer among asymptomatic women aged 35 years or over with a BRCA1 or BRCA2 mutation, who have chosen to delay risk-reducing surgery to remove their ovaries and fallopian tubes (RRSO).35

RCT of population screening re-analysed for women who may have elevated risk

The PLCO population screening RCT was not designed to identify women at high risk of ovarian cancer but, using data that was collected at enrolment, it was possible to identify a subgroup of over 22,000 women who had either a personal history of breast cancer or at least one first degree relative with breast cancer or ovarian cancer.36 These criteria are less stringent than those defining women at high risk of ovarian cancer and would include some women with no increased risk. When analysed as a group, however, it could be expected that the overall risk is higher than population risk.

Annual screening using CA125 (single threshold) and TVUS was associated with a significant reduction in advanced bulky disease at diagnosis. The clinical significance of this is not clear but, for the important outcome of mortality, no statistically significant benefit was found in this analysis.