Signs of Breast Cancer Could Be Spotted 3-6 Years Before Diagnosis Using AI Screening, Shows Massive Study

Artificial intelligence could detect breast cancer signs up to six years before a clinical diagnosis, according to a study published in the journal Radiology. Researchers from Karolinska University Hospital analyzed over 88,000 mammograms, finding that AI-based computer-assisted detection (AI-CAD) systems identified early warning signs in approximately 20% of cases.

How does AI identify breast cancer earlier than radiologists?

Swedish researchers tested three commercially available AI-CAD radiology systems using data from more than 31,000 patients. The study analyzed 88,963 mammograms performed over a 10-year period between 2008 and 2019 on volunteers aged 40 to 74.

Two radiologists analyzed each mammogram, which were scheduled every two years. Out of the participants, 12,072 individuals, or 38.5%, were eventually diagnosed with cancer by the radiologists.

The AI-CAD systems issued cancer prediction scores that were elevated for those eventually diagnosed and low for those who remained cancer-free. According to senior co-author Professor Fredrik Strand of Karolinska University Hospital, about 20% of cases showed signs visible to AI around six years before diagnosis.

Did You Know? The study’s data set included 88,963 mammograms collected from volunteers across Sweden between 2008 and 2019.

What are the accuracy rates for early AI detection?

The AI systems achieved 90% specificity, meaning they could effectively distinguish between true positive and true negative results. This level of specificity was reached in nearly 20% of participants six years before their recorded diagnosis.

Lab automation at Karolinska University Hospital, Sweden

The detection rate increased as the diagnosis date approached. AI-CAD identified signs in up to 25% of individuals four years before diagnosis and nearly 40% two years before diagnosis.

Professor Strand stated the study confirms AI’s potential to find signs of cancer in mammograms much earlier than radiologists. While AI has previously shown promise in predicting five-year risks, this team focused on signs present up to 10 years in advance.

Expert Insight: Samantha Carter notes that the transition from predicting general risk to identifying specific, visible mammographic signs years in advance represents a significant shift in how AI could be utilized in clinical radiology.

What happens next for breast cancer screenings?

Analyzing AI scores over time could provide insight into how early detectable changes arise. According to Professor Strand, this may potentially allow for earlier intervention.

The findings may add to the growing literature on how AI helps in the earlier detection of breast cancer. Future applications could involve using AI-CAD to flag high-risk individuals more accurately between regular screening intervals.

Frequently Asked Questions

How many years in advance can AI detect breast cancer signs?
AI could provide an early alert for the disease up to six years before a clinical diagnosis.

Who conducted this research?
The study was conducted by Swedish researchers, including Professor Fredrik Strand of Karolinska University Hospital in Stockholm.

What was the specificity rate of the AI systems?
The systems achieved 90% specificity in nearly 20% of participants six years before their diagnosis.

How do you feel about the integration of AI into routine medical screenings?