2023-06-11 10:05:00
It’s an ongoing study at the Aberdeen Royal Infirmary, one of Scotland’s leading teaching hospitals, that might show promise: During the pilot project, the use of artificial intelligence (AI) by radiologists identified early-stage breast cancer in a patient who will now undergo surgery.
The use of an AI would not be intended to replace human medical expertise in screening, but to support it: in Scotland, of the 5,000 mammograms examined on average by radiologists each year, between 250 and 300 patients are recalled, and 30 to 40 of them require further investigation. “With such a number, it is possible to miss cancers”says Gerald Lip, the clinical director of the North East Scotland Breast Cancer Screening Programme.
The AI model, developed by Kheiron Medical Technologies, might thus help practitioners identify areas of concern from the results of mammograms or biopsies performed on patients. “As part of a screening, you have to know how to spot things when they are small, before they become big”, says Dr. Lip. This is where automating the analysis of results can be effective, so as not to miss certain symptoms, or even to detect them earlier.
Win time
Currently, with the National Screening Counsel prohibiting the automatic deployment of AI in screenings, it is only used by the service’s radiologists as an additional check at the end of the exam.
Ultimately, however, AI might make it possible to compensate for the chronic lack of medical personnel, to which the Royal College of Radiologists has already alerted, by speeding up screening and double-checking procedures: currently, two radiologists are needed to read and report mammogram results, yet patient wait times for vital exams and cancer treatments are too long.
“The objective of this test is to see what is the best way to work with AI, whether it is replacing a radiologist, reading part of normal mammograms or improving cancer detection, by using it as a safety net”Dr. Lip told the BBC.
1686481374
#artificial #intelligence #detect #breast #cancer
