AI tool boosts early cancer detection by identifying hidden patterns
What's the story
An artificial intelligence tool, "C the Signs," has successfully increased cancer detection rates by 8% among general practitioners in England. The tool scans patient records to identify hidden patterns that could suggest a higher risk of cancer. It examines a patient's past medical history, test results, prescriptions, treatments, and personal characteristics such as postcode, age, and family history. Currently used in about 1,400 practices across England, "C the Signs" covers approximately 15% of the population.
Successful trial
'C the Signs' proves effective in initial testing
"C the Signs" was first tested in 35 practices in the east of England in May 2021, serving a population of 420,000 patients. The Journal of Clinical Oncology published results showing that cancer detection rates at GP practices using the AI tool rose from 58.7% to 66.0% by March 31, 2022. Meanwhile, practices not utilizing "C the Signs" maintained a similar rate as before.
Diagnostic aid
AI tool aids in early and faster cancer diagnosis
Bea Bakshi, a GP who co-created "C the Signs" with Miles Payling, stated that their system has detected over 50 different types of cancers. She emphasized that the tool not only facilitates an earlier diagnosis but also a faster one. The system prompts GPs to inquire about new symptoms from patients and recommends tests or clinical pathways based on detected patterns.
Precision diagnosis
The AI tool shows high accuracy in cancer detection
In a study assessing 118,677 patients, "C the Signs" successfully identified 7,056 out of 7,295 diagnosed with cancer. When the tool concluded that it was unlikely a patient had cancer, only 239 out of 8,453 went on to have a confirmed cancer diagnosis within six months. Peter Holloway, co-author on the study, highlighted a case where "C the Signs" recommended a faecal test for a patient with common symptoms, leading to an early diagnosis and successful treatment of colorectal cancer.