New test predicts lung cancer survival better than current standards

Researchers at the Francis Crick Institute, the UCL Cancer Institute and UCLH have shown that a test called ORACLE can predict lung cancer survival at the point of diagnosis better than currently used clinical risk factors. This could help doctors make more informed treatment decisions for people with stage 1 lung cancer, potentially reducing the risk of the cancer returning or spreading.

In research published today in Nature Cancer, the team tested ORACLE in 158 people with lung cancer as part of the Cancer Research UK-funded TRACERx study. They found that ORACLE could better predict patient survival than currently used clinical standards like tumour stage.

ORACLE was developed in 2019 to overcome the lack of biological markers in lung cancer, which can indicate to doctors who might be at a greater risk of their cancer coming back or spreading to another part of the body.

This is particularly important for people with stage 1 lung cancer, who are normally given surgery without chemotherapy. For a quarter of stage 1 patients, their cancer returns, suggesting they may have benefitted from more frequent monitoring or chemotherapy.

When doctors take a sample from a tumour, they typically only capture less than 1% of the tumour, and the genetics can vary massively from region to region within the same tumour. ORACLE overcomes this by looking at genes expressed at a high or low level in every part in the tumour.

The new findings show that ORACLE could predict which patients with stage 1 lung cancer had a lower chance of survival, and might benefit from chemotherapy as well as surgery. Currently used clinical standards weren't able to give this information for stage 1 patients.

The researchers also found that high ORACLE risk scores were linked to regions of the tumour that were more likely to spread to another part of the body.

Finally, by looking at 359 current and potential lung cancer drugs, they found that a high ORACLE risk score predicted better response to some types of chemotherapy, particularly platinum drugs like cisplatin

This is because tumour regions with high ORACLE scores are associated with unstable DNA (called 'chromosomal instability'), which is particularly targeted by platinum drugs. The same lab has recently found that changes in a key gene called FAT1 drive chromosomal instability, which is also one of the genetic variations ORACLE looks for.

The next steps for the researchers are to compare people with high ORACLE scores receiving standard care and those receiving more surveillance or chemotherapy to determine if the test improves survival, even for people diagnosed at the earliest stage.

Dhruva Biswas, Translation Fellow at the Crick, Postdoctoral Fellow at the UCL Cancer Institute, Associate Research Scientist at Yale School of Medicine, and co-first author, said: "ORACLE can now predict survival rates in patients diagnosed at the earliest stage. If validated in larger cohorts of patients with lung cancer, doctors could one day use ORACLE to help make informed treatment decisions, bringing lessons from cancer evolution into the clinic."

We wanted to build on the previous work developing ORACLE and show that it can predict survival at the point of a lung cancer diagnosis. We've also shown that it can predict who would benefit from certain types of chemotherapy drugs or if someone's cancer is likely to spread, giving a holistic measure of how a patient's cancer might progress and respond."

Yun-Hsin Liu, Research Assistant at the UCL Cancer Institute, and co-first author

Charles Swanton, Deputy Clinical Director and Head of the Cancer Evolution and Genome Instability Laboratory at the Crick, medical oncologist at University College London Hospitals, Chair in Personalised Cancer Medicine at the UCL Cancer Institute, Chief Investigator for TRACERx, and co-senior author of the study, said: "Lung cancer is the leading cause of cancer-related death throughout the world, so it's clear we need better markers to accurately classify tumours and predict who is at high risk. We're now working with the Translation team at the Crick and industry partners to progress ORACLE into a test which could hopefully be used in the clinic as soon as possible."

Paul Mercer, Head of Industry Partnerships in the Crick Translation team, said: "This is an important step forward, translating our understanding of the infinite complexities of lung cancer mutation into a diagnostic tool, prioritising patients for the most effective therapies. We look forward to working with partners to take this work forward and maximise patient benefit from ORACLE."

Dani Edmunds, Science Engagement Manager at Cancer Research UK, said: "In the last 50 years, cancer survival has doubled in the UK. However, progress has not been equal across all types of cancer. Although survival for lung cancer has improved since the 1970s, it's still one of the most challenging cancers to treat.

"New tests to predict lung cancer's behaviour could help doctors tailor treatment strategies to each person's condition, giving the best chance of a successful outcome.

"This research reflects Cancer Research UK's commitment to tackle this hard-to-treat cancer. While ORACLE still needs testing in larger-scale trials, these initial results show it could take us a step closer to more personalised approaches to treating lung cancer, so more people live longer, better lives." 

The study is supported by the National Institute for Health and Care Research UCLH Biomedical Research Centre.