AI Matches Radiologists in Breast Cancer Detection
A new study in The Lancet Oncology has revealed that artificial intelligence (AI) can spot breast cancer in mammograms as accurately as two seasoned radiologists. This breakthrough could change how breast cancer is detected and diagnosed across the UK and beyond.
Massive Trial with 80,000 Swedish Women
Researchers tested over 80,000 women from Sweden, average age 54. Half underwent the usual screening by two radiologists. The other half had AI-assisted mammograms, checked by one or two radiologists afterwards. The results were eye-opening: AI spotted 244 cancer cases, while standard checks caught 203.
AI Cuts Radiologist Workload by Nearly Half
Crucially, AI did not increase false alarms — both groups had a false positive rate of just 1.5%. Even better, AI slashed radiologists’ screening workload by 44%, reducing 36,886 readings. This means quicker results and less strain on healthcare staff.
Experts Cautiously Optimistic
“The results are very promising,” said lead author Dr Kristina Lang from Lund University, Sweden. “AI could eliminate the need for double reading of most mammograms, easing the burden on radiologists and cutting patient waiting times. But more trials are needed before full rollout.”
With radiologist shortages impacting the NHS, AI could be the game-changer needed to spot cancer earlier and save more lives. The NHS is already exploring how to use AI to speed up and improve breast screening nationwide.
