Credit: Anna Widdup
The inability to reliably diagnose ovarian cancer early has long been thought of as a major reason for the disease’s high mortality. But a 20-year trial of ovarian cancer screening, which published its final results in May, was unable to show that detecting that cancer sooner resulted in fewer deaths (U. Menon et al. Lancet 397, 2182–2193; 2021). Usha Menon, a gynaecological cancer researcher at University College London who led the UK Collaborative Trial of Ovarian Cancer Screening, spoke to Nature about the impact of this surprising finding.Why was early detection of ovarian cancer expected to reduce mortality?Treatments for many cancers generally work better in the earlier stages, when the tumour is small. Whether you’re having surgery or some other treatment, the earlier you strike the better, and that means being able to detect lower tumour volumes.How was your trial designed?The idea began back in the late 1990s, when the mood in the country was in favour of a national screening programme for ovarian cancer. The trial itself began in 2001, when we started recruiting 200,000 women. Of these, 100,000 were controls, with no screening; the other half were screened with blood testing for a protein called CA125. The evidence at the time suggested that change in CA125 over time would be a good biomarker for increasing early detection. Part of Nature Outlook: Ovarian cancer You did catch some cancers earlier. Why did that not result in fewer deaths?We achieved early diagnosis, but not as early as we needed. Early-stage cancer detection was increased by about 10%. Some people think detecting more cancers even earlier might make a difference, but I’m not sure. I think we might have to make detections a great deal earlier than we currently can.What was it like to get these results after 20 years of work?It was difficult for us, because we didn’t know this was coming. We were blinded from the results, and only came to know that we had not improved mortality last December, so we haven’t had a huge amount of time to process it all yet.We worked so long on this, and there’s been a huge investment, not only in funding for the research, but also from the thousands of women involved, and the doctors who looked after them. We had hoped that we would make some sort of a dent, even if it was not a huge one, but we were not able to do that. I feel bad for the whole ovarian cancer community, but we did make progress in a number of areas.What positives can we take from your trial?One major outcome is that we now have a valuable bioresource. Throughout the trial, all the women donated blood when they had a blood test, which means that we now have a unique set of annual serum samples for individual women. The women’s records were connected — with their permission — to the cancer and death registration records in the United Kingdom, and, for all of these 200,000 women, we know who got cancer, which type of cancer and at what point. So there’s an opportunity now to assess whether the cancers would have been picked up a year or two earlier — or even 5–10 years earlier — had we been using other biomarkers than CA125.Have trials such as yours been successful for other cancers?Absolutely. It didn’t work in our trial, but all cancers are different. In ovarian cancer, we’re only beginning to understand the natural history, such as whether the majority of the aggressive cancers develop in the ovaries or fallopian tubes. But with other cancers we know more. The widespread use of cervical cancer screening programmes in the United States and Europe, for example, has decreased mortality rates from that disease. More from Nature Outlooks Breast cancer screening is more controversial — there is some debate about whether current breast screening reduces mortality. Barring cervical cancer, all of these screening programmes have been initiated through a randomized controlled trial, like the one we did.Where do you think research around ovarian cancer detection is going next?At the moment, everybody is focusing on new biomarkers for early detection. One form of testing, which is being developed to detect multiple cancers, is based on circulating tumour DNA (ctDNA). The test tries to pick up tumour DNA, which is different from normal DNA. So it should be much more specific than using CA125.However, the tumours are extremely small when we need to catch them, so whether there will be a detectable amount of ctDNA in the blood at that point is unclear. There are also other biomarkers that are being studied for multi-cancer detection, such as microRNAs.Could that work take another 20 years?I hope not. It will demand further trials, but they will need to be big trials that necessarily take a long time — possibly more than a decade. Deaths have to be the main outcome to monitor, so that we can show we can reduce mortality, but that takes time to become clear. Also, we don’t know how much of a detection shift we need to make to save lives. Will 20% earlier detection help? Do we need 30%? These are the questions for the next generation of trials. Meanwhile, the multi-cancer detection efforts I mentioned have already started. These trials are currently focused on reducing the number of women diagnosed with advanced disease. This interview has been edited for length and clarity.This article is part of Nature Outlook: Ovarian cancer, an editorially independent supplement produced with the financial support of third parties. About this content.
https://www.nature.com/articles/d41586-021-03721-x
