Incidence and survival
In 2019, it is estimated that 1,510 new cases of ovarian cancer will be diagnosed in Australia and there will be 1,046 deaths from ovarian cancer.3 The age-standardised incidence rate is estimated to be 9.8 cases per 100,000 females in 2019.3 In Australia, ovarian cancer is the tenth most common cancer in women and the sixth most common cause of death from cancer in women.
In 2011-2015 Australian women with ovarian cancer had a 45.7% chance of surviving for 5 years compared to their counterparts in the general Australian population.3 Ovarian cancer is often diagnosed at an advanced stage when survival outcomes are poor. For women diagnosed with advanced disease (Stage III and Stage IV), the 5-year survival rates are reported to be less than 30%, whereas for patients diagnosed with Stage I disease, the 5-year survival is reported to be around 90%.4 Therefore, in order to improve the mortality rate for ovarian cancer, detection in the early stages of the disease is required.
Screening principles and ovarian cancer
General principles have been developed as criteria for screening programs by the World Health Organization.5 Based on these principles, the Australian Department of Health has developed an Australian Population Based Screening Framework,6 which highlights the need for a strong evidence base in making the decision to introduce a screening program, and the requirement that the screening program should offer more benefit than harm to the target population. The Framework also provides criteria for the condition (including that it is an important health problem and that it has a recognisable latent or early symptomatic stage) and criteria for the test (including high sensitivity and specificity, adequate validation, safety, and a relatively high positive and negative predictive value).
Ovarian cancer satisfies some of the criteria for screening, but some criteria represent a significant challenge (see http://www.health.gov.au/internet/screening/publishing.nsf/Content/population-based-screening-framework for further information on population based screening framework). Although an opportunity exists to alter the natural history of ovarian cancer if it can be detected earlier, our understanding of early disease progression from latent to declared disease is incomplete and continuing to develop. Recent evidence suggests that many ovarian cancers originally thought to have arisen in the ovary may have had precursor lesions in the fallopian tubes.7
Although various subtypes of ovarian cancer are recognised on the basis of histologic morphology, a new classification scheme now divides ovarian cancers into two broad types based on tumour characteristics.8 Type I are low-grade, relatively non-aggressive carcinomas, often arising from recognisable precursor lesions (e.g. borderline ovarian tumours). Type I tumours include low-grade serous, mucinous, endometrioid, clear cell, and transitional cell carcinomas. While clear cell carcinomas are categorised as a type I tumour, they may actually belong to an intermediary category due to their mutations and behaviour. Type II tumours are high-grade, genetically unstable, aggressive carcinomas, which have a tendency to metastasise from small, or even microscopic, primary lesions. Precursor lesions for type II tumours have not been described clearly and tumours may develop de novo from the epithelium of the fallopian tube and/or the ovarian surface epithelium.8 Type II tumours include high-grade serous carcinomas, undifferentiated carcinomas, and carcinosarcomas. Type II tumours constitute the majority of ovarian cancers and, due to their rapid progress from microscopic to widespread disease, are more difficult to detect at earlier stages and have the poorest prognosis. It is now apparent that in addition to identifying Type I tumours, a successful screening test would need to identify Type II tumours in what appears to be a brief window of opportunity prior to their early dissemination. These insights were not available at the inception of the recent large population screening trials for ovarian cancer.
In terms of designing screening trials, the relatively low incidence of ovarian cancer presents a significant challenge; particularly large clinical trials are needed to detect a sufficient number of cases for meaningful analysis of the screening test. Also, the need for surgical removal of ovaries to make a final diagnosis places a stringent requirement on the positive predictive value of any screening test for ovarian cancer, as the consequences of a false positive test result are serious.
Current tests
To be suitable for screening or surveillance, any test for ovarian cancer should be highly sensitive (i.e. most women with ovarian cancer are identified by the test) and have a high positive predictive value (i.e. most test-positive women have ovarian cancer). The latter is critical where diagnosis is associated with surgical removal of ovaries. Additional criteria are a low false negative rate (i.e. not many women with ovarian cancer are missed by the test) and the ability to detect cancers at an earlier stage to allow earlier treatment. Currently, there are several tests available that have been studied alone and in combination for use in screening and/or surveillance:
- CA125, a high molecular weight glycoprotein, is the most thoroughly assessed serum biomarker for ovarian cancer. The utility of CA125 for screening, however, is limited by its poor sensitivity in early stage disease, with CA125 levels elevated in only 50-60% of women with Stage I disease, whereas levels are elevated in 75-90% of women with advanced disease.9 The specificity of CA125 is also limited, due in part to elevation of the marker in other conditions, including other cancers, benign diseases and physiological conditions,9,10 which reduces the positive predictive value of the test.
- Transvaginal ultrasound (TVUS) utilises morphology and ovarian volume to detect changes that may signify developing malignancy. Although highly sensitive, TVUS has limitations in distinguishing between benign and malignant masses due to the complexity of ovarian morphology, which can result in unnecessary surgery (low positive predictive value).7,11
- Combining CA125 and TVUS tests can increase the sensitivity and positive predictive value of screening or surveillance compared to either test alone. Testing can be done in sequence, by limiting a second-line test (e.g. TVUS) to women with abnormal results in the first-line test (e.g. CA125). For example, the Risk of Ovarian Cancer Algorithm (ROCA) is a proprietary test that incorporates changes in CA125 over time, among other factors, to inform secondary testing intervals and triage to other tests such as TVUS. This approach has been used in large screening and surveillance trials in the UK, the USA and Australia, as described below.