7th Sep 2018
7th Sep 2018
Curated by Alejandro Noval
In a major achievement, scientists have created an algorithm that can spot the similarities of tumour evolution in different cancer patients. This will help to predict the progress of the disease and find the most efficient personalised treatment.
In 10 seconds? Researchers blew a hole in cancer’s armour – they designed a tool called REVOLVER that can help healthcare workers to tell in advance how tumours will evolve, improving outcomes for individual patients. (Read the science)
So, how does predicting tumour evolution help? It is important! A lot of cancers are now treatable, but tumours change and can become drug resistant. Knowing what to expect allows doctors choose medications that help the patient survive or live longer. (Find out more about the tumour mutations)
And now we'll know what to expect? Indeed – thanks to the new tool, REVOLVER, a.k.a Repeated Evolution of Cancer. It uses a machine learning algorithm to spot patterns and thus help make more accurate predictions. It is revolutionary in the sense that it allows to simultaneously analyse different tumour regions in different patients, making common mutation patterns stand out more. (Read the research paper)
Why mutations in different patients matter? Cancer develops and survives through mutations. Clinicians track these through “family trees” describing the changes, but up to now only individual patients’ “trees” could be created and compared. However, mutations even within the same cancer seem so random, that one patient’s “tree” can't be used to fully predict the progress of another patient’s tumour. (More on the genetic sequencing of clonal evolution in cancer)
And now we will be able to predict? Yes, with REVOLVER, revealing common patterns in the genetic evolution of cancer cells. It proved its worth by successfully identifying cell mutations that drive lung, breast, kidney and colorectal cancer by making their evolutionary steps stand out. The team used almost 800 various cancer samples from more than 170 patients to validate the method with a successful outcome. (See how AI predicts breast cancer risk)
So, how will individual patients benefit? By allowing doctors to better classify patients into subgroups, which in practice means being able to fine-tune their treatment based on advance knowledge of potentional changes in their tumours. And, as I mentioned, the tool - to be introduced in the future - will aid doctors to know in advance if a patient would become resistant to a certain type of chemotherapy. This will allow to switch to another, more efficient drug. So, we can say, REVOLVER fires an important shot in the fight against cancer!
Tumour evolution – how does it happen?
The evolution of the tumour is completely dynamic but is not unidirectional.
Often, when a mutation occurs, it triggers subsequent mutations that contribute to make the cell more aggressive.
When a tumour develops, the mutated cells keep dividing, and each of them can acquire different sets of mutations, that will pass to the next generation.
Hence within a tumour there are actually many different subpopulations of cancer cells, each with their own mutations.
However, mutations are not always bad news – certain tumours with a higher number of mutations may become more visible to the immune system, which then can attack them.
(Psst, Alejandro distilled 24 research papers to save you 963.4 min)