Face masks and social distancing: Due to the rising prevalence of COVID-19 in our communities, we strongly encourage healthcare staff and visitors to wear a face covering in all of our settings, particularly in clinical areas and those with high footfall. Please exercise a common-sense approach and personal responsibility to help us reduce the impact of COVID-19 on our patients, workforce and services. In addition to wearing a face covering, it is important to continue to maintain social distancing where possible. Thank you for your continued support and co-operation at this time. We continue to regularly review our advice based on prevalence in our communities and our hospitals.
Artificial intelligence could be used to predict the effectiveness of chemotherapy on women with breast cancer.
Genomics, the study of human genes, and AI are two technologies that could potentially transform cancer care.
They are driving precision oncology, allowing personalised treatment focusing on what works for individuals and avoiding what does not.
Main photo above: Oncologist Dr Mark Davies, the trial’s principal investigator
Chemotherapy is the main treatment used for metastatic breast cancer, which is where the cancer has spread throughout the body.
However, the response to chemotherapy varies. For some people it works very well, at least to start with. For others it has little or no effect.
Dr Mark Davies, an oncologist based at Swansea’s Singleton Hospital, said: “Being able to predict who will respond would allow treatment to be targeted to those most likely to benefit.
“Those unlikely to respond could be offered alternative treatments and be spared unnecessary toxicities and side-effects.”
Dr Davies has been awarded a grant of almost £250,000 from Health and Care Research Wales for a two-year study which aims to use machine learning, a form of AI, to predict these outcomes before treatment starts.
Genes are in the DNA of each cell in the human body. They control how the cell works, including how quickly it grows and how often it divides.
When one or more genes change, known as genetic mutation, this can cause cells to multiply uncontrollably and become cancerous.
Dr Davies explained: “The patterns of mutation can vary from one area of cancer to another, and can change over time.
“This can lead to genetically distinct sub-populations of cancer cells, called subclones, to arise. These vary in their sensitivity to chemotherapy.
“We aim to predict which subclones will become predominant, so treatment can be modified to target them.
“Our models may also predict how well a patient will respond, allowing therapy to be individualised.”
Previously, studying the cancerous tumour would have involved taking samples by biopsy – an invasive and potentially harmful procedure.
However, tumours also shed DNA into the bloodstream. This is called circulating tumour DNA (ctDNA).
This ctDNA can now be extracted by taking a blood sample. Advances in technology mean scientists can study around 500 genes, all known to be linked to cancer.
“The problem is that it is very complicated,” said Dr Davies. “You are looking at 500 genes, in multiple patients, all of which change in time.
“This is where machine learning comes in. It’s a mathematical computational technique, where you can essentially look at patterns in data.
“We will be using it to look at the genetic data and work out whether we can predict from the initial profile which will be the resistant cells.”
The study is due to start later this year and will recruit patients with metastatic breast cancer, all being treated with a specific chemotherapy drug, from North, South West and South East Wales.
Blood tests from the volunteer patients will be taken before chemotherapy starts and during its progression.
Dr Davies will be principal investigator. He will work with the All-Wales Medical Genomics Laboratory, with a specialist DNA sequencing company, and with experts from universities in Swansea, Plymouth and Liverpool.
“The study will recruit between 60 and 100 patients. That’s not huge but we will be using machine learning techniques which are optimized for that relatively small sample size,” he added.
“The end result of this will not be an immediate clinically-usable solution. It’s to demonstrate there is predictive information.
“We want evidence that it does work, so we can justify a bigger, more definitive study.”
Dr Davies said it was likely that, in future, all cancer patients in Wales would receive comprehensive genomic profiling of their cancer as part of routine NHS care.
“By applying artificial intelligence approaches to this large-scale genomic and clinical data, we could further refine our models and deploy them into clinical practice,” he added.
“These approaches could be extended to other treatments and cancer types.
“This could improve outcomes, avoid unnecessary toxicities and make effective use of high-cost drugs, leading to better quality and value cancer care within Wales.”
Dr Davies’ research project is one of 23 Health and Care Research Wales funding call awards for 2020-21. These have a combined lifetime value of almost £6.5 million.
The schemes offer varying levels of support, to address different research needs, from supporting talented individuals to become independent researchers to funding high-quality research projects relevant to health and wellbeing needs across Wales.
Michael Bowdery, Head of Programmes, Health and Care Research Wales, said: “As always, the standard of applications was very high this year.
“We are pleased with the range of important topic areas these awards cover, including investigations into the impact of the Covid-19 pandemic in a variety of settings.
“Investing in research and our researchers is vital to our goal, to boost the health and prosperity of people in Wales.”