A ‘LANDMARK’ blood test can detect more than 50 types of cancer before any symptoms show — potentially saving millions of lives, scientists say.
The test, which identifies the tissue in which the cancer originates in 96 per cent of cases, gives a wrong diagnosis less than one per cent of the time, according to a US study.
This compares with about ten per cent for breast cancer tests.
The study’s senior author, president of US Oncology Dr Michael Seiden, said the ability to identify where in the body the cancer is located will ‘help healthcare providers to direct next steps for diagnosis and care’.
The test examines cell-free DNA released by tumours into the blood for a chemical process called methylation.
Abnormal methylation patterns and changes in gene expression can contribute to tumour growth. An algorithm was used to predict the presence and type of cancer among 7,000 volunteers. The researchers found it gave a false positive rate of 0.7 per cent.
In 12 types of cancer — including bowel, stomach and lung — the true positive rate was 67.3 per cent.
These 12 account for about 63 per cent of cancer deaths each year in the US with no way of screening for most of them before symptoms show.
Dr Seiden said the figures met the main requirements ‘for a multi-cancer early detection blood test that could be used for population-level screening’.
The study was published in the journal Annals of Oncology.
Editor Prof Fabrice Andre said: ‘Earlier detection of more than 50 per cent of cancers could save millions of lives every year worldwide and dramatically reduce morbidity induced by aggressive treatments.’
How to save your skin by phone
CANCER researchers have developed Artificial Intelligence that lets you diagnose dangerous skin disorders just by using your smartphone.
The AI tool can spot 134 different skin problems, predict if they are malignant and also suggest treatment.
While it was not as accurate as a dermatologist in tests, it was able to diagnose skin cancer as accurately as medical trainees. And when 47 medics used the software for diagnosing 3,501 images for skin cancer, their success rate jumped to 86.8 per cent, compared with 77.4 per cent without it.
Scientists at Seoul National University in South Korea collected 220,000 images of 174 skin diseases and trained the algorithm to interpret them. Team leader Prof Na Jung-im said he hoped the AI might prompt people ‘to visit specialists for cancerous lesions such as melanoma that might be neglected otherwise.’