How AI Could Speed Up A Cancer Diagnosis
The development and growing use of artificial intelligence (AI) has prompted a wide range of reactions, from negative notions of millions of human jobs being lost and even the machines turning against humanity in the manner of the Skynet system in the Terminator films, through to extremely positive hopes that a whole range of problems could be solved or at least aided.
A new development that definitely falls in the second category is the use of AI in clinical diagnosis, since if this can be done more accurately and sooner it can make it easier to treat various conditions with early intervention. This could never be more so than with cancer.
Such a benefit looks closer than ever after a study of an algorithm-based AI system developed by experts at the Royal Marsden NHS foundation trust, the Institute of Cancer Research, London, and Imperial College London.
Published in the Lancet Journal eBioMedicine, the research found the system was better than current methods at identifying if unusual grows spotted in CT scans were cancerous. Studying 500 CT scans on large lung nodules, the system proved more adept than the human eye at identifying features that marked out the growths as being cancer.
An assessment tool called Area Under The Curve (AUC) was used to test the accuracy of the method, with an AUC of 1 denoting perfection and 0.5 random guesses. The AI system scored 0.87.
If this indication of effectiveness is backed up by further research and peer reviews, it could mean a Sheffield cancer diagnosis can be arrived at more easily and provide the best chance of an early intervention taking place, helping to save or at least significantly extend the life of the affected patient.
Commenting on the discovery, Dr Benjamin Hunter, a research fellow at Imperial College and an oncology registrar at the Royal Marsden, said: “In the future, we hope it will improve early detection and potentially make cancer treatment more successful by highlighting high-risk patients and fast-tracking them to earlier intervention.”
He added: “Next, we plan to test the technology on patients with large lung nodules in clinic to see if it can accurately predict their risk of lung cancer.”
The importance of early detection in the case of lung cancer cannot be underestimated, not just because it accounts for 21 per cent of cancer cases in the UK, but also because more than half of cases are discovered at stage three or four, when little or nothing can be done to save the patient’s life.
Fortunately, recent research by the Office for National Statistics has shown that the level of smoking, the main cause of lung cancer, was at its lowest level on record in 2021. This means there should be far fewer cases of the disease arising to begin with.
However, having proved effective in the early detection of lung cancer, AI tools may also offer similar benefits when it comes to the early detection of other forms of cancer, raising survival rates across the board.
