Canadian researchers have published two new studies using machine learning to improve detection and treatment of cancer.
The first study identified a distinct set of genes that could be used to identify how well patients with cancer are likely to respond to immunotherapy drugs. The second study profiles epigenetic alterations to detect and classify early-stage cancers from a blood test. Both studies were led by Dr Daniel De Carvalho, at the Princess Margaret Cancer Centre in Toronto.
The first study, published in Nature Communications, defined a set of genes that are associated with the failure of immunotherapy cancer treatment.
'The ultimate goal is to find a biomarker that can help the clinician decide if a patient should receive immunotherapy or not,' said Dr De Carvalho.
The study involved one of the largest analyses of extracellular matrix (ECM) genes, using data from more than 8000 cancerous tumour samples. The ECM provides physical scaffolding and biochemical support to cells and is known to be a key factor in how a cancer progresses. Immunotherapy uses the body's own immune system to treat cancer. However, in some cases, the ECM surrounding cancerous cells can stiffen and form a barrier that blocks immune cells from doing their job.
Understanding whether a patient will respond to immunotherapy can help design treatment strategies. The researchers were able to identify a signature in ECM genes linked to blocking immune cells.
The 'liquid biopsy' test, looking at cfDNA is not new, but it has had a low sensitivity for cancers at the earliest stages.
'A major problem in cancer is how to detect it early. It has been a "needle in the haystack" problem of how to find that one-in-a-billion cancer-specific mutation in the blood, especially at earlier stages, where the amount of tumour DNA in the blood is minimal,' said Dr De Carvalho.
The researchers looked for epigenetic changes associated with each of seven cancer types. The researchers were able to find many epigenetic changes that were the same in cfDNA and in the tumour tissue.
As epigenetic changes in bloodborne cfDNA is easier to detect than in early tumours, the authors hope it could form the basis for a more sensitive liquid biopsy test to detect very early stage, even pre-symptomatic cancers.