Ovarian cancer has a high mortality risk because it is so often diagnosed at a very late stage. In a new study, our team has shown that detection rates can be significantly improved by screening for a specific set of proteins in the bloodstream. This could mean detection of ovarian cancer up to two years before current screenings allow.

Cancer tests walk a harsh line between missing cancer and misdiagnosing healthy people. If you make your test too strict, you will fail to detect traces of real cancers that are present. If it is too lenient you will falsely detect cancer where it doesn’t exist.

While it might seem obvious we should tip the scales in favour of catching every cancer, the burden this places on the health system can be unsustainable. Not to mention the stress and potentially dangerous treatment it can mean for healthy people.

Our group is working to improve this balancing act in the diagnosis of ovarian cancer, the statistics for which speak for themselves. In 2016, 4,227 women died from ovarian cancer. The overall five year survival for stage III & IV ovarian cancer (late stage) is only 22%, making this the most lethal female reproductive cancer. When ovarian cancer is detected early, the patient’s prospects are much better, with approximately 90% of women diagnosed at stage I (early stage) surviving five years or more.

Red herrings in current cancer detection

Unfortunately, ovarian cancers are often not caught early enough. Almost six in ten ovarian cancer cases are diagnosed at a late stage in England and Northern Ireland today, resulting in a high death rate.

Late detection of ovarian cancers is mostly due to the “non-specific” nature of its symptoms. Currently most blood tests measure the levels of a protein CA125. But using this as a marker for cancer in the blood is not reliable as it can also be elevated in pregnancy, during a woman’s period, and other non fatal conditions (such as endometriosis). Additionally, not all cancer patients show this marker. Studies have shown it is only elevated in around 50% of early stage cancers.

For this reason, our research team has been working with an international group of experts from the universities of New South Wales, Milan and Manchester, to develop a combination of proteins we can use to identify cancer earlier than is currently possible.

Our previous work lead to the identification of four possible markers we could look for in the blood to maximise the power of our tests (protein CA125, phosphatidylcholine-sterol acyltransferase, vitamin K-dependent protein Z, and C-reactive protein).

A time capsule of blood samples

Armed with our panel of blood markers, the next stage was to test for effectiveness in a large group of women. Part of the success of our study came from the incredible dataset we had access to thanks to the United Kingdom Collaborative Trial for Ovarian Cancer Screening (UKCTOCS). From the 200,000 women registered in this database, we were able to select 80 women (49 ovarian cancer cases, and 31 healthy people). Blood samples from these women had been taken every year for seven years before the ovarian cancer patients were diagnosed with the disease.

This provided us with a very powerful tool. With this time capsule of samples we were able to observe how levels of our four target proteins changed over time between patients and healthy individuals. In essence, it allowed us to see directly if the proteins differed between patients and the healthy people over time. Because the samples were taken over such a long period, we could track back over the seven years to see when the first time the proteins would allow us to detect the disease.

We then developed a computerised tool that assessed changes in protein levels to determine the “predicted risk” of ovarian cancer for each person (cancer patients and healthy people). The tool was asked to grade the risk of ovarian cancer for each person at each point along the seven year timeline (rating them severe, elevated, intermediate, or normal).

This initial work demonstrated that the screening tool has the potential to diagnose ovarian cancer one to two years earlier than current diagnosis. Our results also suggest it could identify 60% of the most aggressive ovarian cancers around one year before currently possible.

The future - digital biological maps

The results of this study are encouraging but this research is still at an early stage. We are now setting up a study to verify our results in a large group, which will provide us with around 1,000 blood samples. By using a technique called SWATH mass spectrometry, we can create a digital map of all the proteins in a person’s system at that time.

This means we do not have to rely on physical blood samples, which can get used up or expire. If future research suggests a new protein is important, we can come back in the future to look at these samples again and can share them easily with other research groups. We hope this study will provide the data we need to advocate for an ovarian cancer screening programme.

By optimising these methods we also hope we can move towards a stage where ovarian cancer is diagnosed in stages I and II in most women, when treatment can really make a difference.