Scientists predict time before cancer becomes drug resistant – study

Scientists can use genetic “weather forecasting” to predict how long it will take for bowel cancer to evolve resistance to a drug before a patient has even started treatment, a new study suggests.

Using similar principles to those used in meteorology to predict which mutations might arise, and how they would undergo selection, researchers made long-range forecasts of how cancers might evolve drug resistance.

Knowing when drugs are likely to stop working could give clinicians more time to prepare for the next step in a person’s treatment, and to consider offering them alternative therapies or to enrol them in clinical trials.

A team at The Institute of Cancer Research, London (ICR) and The Royal Marsden NHS Foundation Trust analysed 42 tumour samples from patients with advanced bowel cancer treated with a drug called cetuximab.

This is an anti-EGFR antibody, which blocks growth signals to cancer cells.

The researchers used genetic profiling to measure how new mutations were generated in bowel cancers prior to treatment with cetuximab.

These so-called mutational signatures are distinct footprints left by the different processes that alter the DNA sequence of cancer cells.

They can drive the generation of further mutations and be more or less active in tumours.

Researchers were able to develop forecasting models to make precise, long-term predictions about how the cancer would evolve.

This was done by understanding the complex interplay between active mutational signatures before treatment, the DNA alterations they drive, and resistance mutations that occur in response to this treatment.

While several cancer drugs are highly effective for bowel cancer, their effects are often short-lived, with resistance developing within months.

This is because of the cancer’s ability to adapt to new environments through Darwinian evolution.

A specific mutational signature detected before treatment, called SBS17b, indicated that tumours would quickly develop resistance to cetuximab.

Much like long-term weather forecasting which simulates fundamental atmospheric processes to generate accurate predictions, the study modelled how mutational signatures, the fundamental processes that fuel evolution – influence the development of resistance.

They showed that the SBS17b signature leads to specific mutations which cause drug resistance.

A specific group of mutations were particularly common in cancer cells that developed resistance to cetuximab, and they only evolved in tumours that already had an active SBS17b signature before treatment was started.

This suggests SBS17b could be used to predict that someone will develop resistance quickly, researchers say.

Dr Marco Gerlinger is team leader in translational oncogenomics at the ICR, and consultant medical oncologist at The Royal Marsden NHS Foundation Trust.

He said: “Over the last few years, we’ve developed ways of tracking early signs of drug resistance by taking blood samples while patients are receiving treatment, and assessing which resistance mutations are starting to evolve.

“Our new study goes several steps further, by instead making long-range forecasts about the development of drug resistance in bowel cancer before patients have even started on a drug.

“Our approach is similar to weather forecasting as it analyses the most basic processes that enable cancer evolution in order to make longer-term predictions.

“This new technique will allow us to predict before a patient starts treatment how long it will take for a cancer to evolve resistance, allowing us to stay a step ahead of their cancer, and shape their treatment accordingly.”

Professor Paul Workman, chief executive of the ICR, said: “Cancer evolution is the biggest challenge we face in cancer medicine.

“This new study assesses mutational signatures like a meteorologist measures long-term weather patterns and could offer a completely new way to predict how and when bowel cancers will evolve to evade treatment.

“Being able to make accurate long-range forecasts could be hugely valuable in planning a patient’s treatment, and ensuring we keep patients alive and healthy for much longer.”

The study is published today in the journal Nature Ecology and Evolution and was funded by the European Research Council under the European Union’s Horizon 2020 research and innovation programme.