Could AI (artificial intelligence) help detect breast cancer early?

content

From how we shop to how we communicate, AI is becoming a staple in our daily lives. Now, studies are exploring whether AI can detect breast cancer faster and more accurately than current methods. This means doctors may be able to diagnose cancer sooner, potentially saving lives.

Style

no-header-background, transparent-header, no-vertical-section-spacing

image

headline

How is breast cancer detected today?

content

Regular breast cancer screening is done with mammography. It uses low-dose X-ray to detect breast cancer before a person has symptoms. However, like all screening tests, mammography is not perfect. It can sometimes miss cancer (a false-negative result). Or a mammogram can be read as cancer but be wrong (a false-positive result).

Some women at higher risk of breast cancer have extra screenings with MRI once a year. MRI is more sensitive and can help detect cancers missed by mammography.

Style

no-vertical-section-spacing

headline

Study 1: Can AI detect cancer in MRI scans earlier?

content

In a 2024 study in Academic Radiology called “Early Detection of Breast Cancer in MRI Using AI,” researchers sought answers to two questions: Can we predict future breast cancer risk from current MRI scans, even those that appear normal? If yes, can this information be seen and evaluated by AI?

Researchers used AI to analyze over 5,000 “normal” breast MRIs from 910 women taken between 2002 and 2014. They wanted to see if AI could accurately predict which women would develop cancer within a year, even though the original MRIs were read as cancer-free by radiologists. Could AI find hidden cancers that doctors missed?

Based on the positive performance of the AI model, scientists concluded that AI could be a powerful tool for radiologists, potentially helping them find 30% more cancers at an earlier stage by highlighting high-risk scans.

headline

Study 2: Can AI predict cancer risk over 1 to 6 years?

content

Risk-based breast cancer screening aims to personalize early detection. But how is risk assessed? The traditional way is with a mathematical model of a person’s risk factors. Some studies now suggest that AI does a better job.

One study focused on predicting future cancer risk. It found that AI isn't just good at finding existing cancers; it can predict future risk too.

headline

Watchouts for AI tools

content

One potential issue is that we don’t fully understand how AI tools make their predictions. We see the overall results, but not the inner workings. That makes it hard to know exactly why they work and if we should completely trust them. Still, the results of these studies offer a basis for future research on AI and may help with early cancer detection.

image

headline

Future directions for AI in breast cancer

content

Currently, the Food and Drug Administration (FDA) has approved at least 20 AI-based products for breast cancer screening and detection. This includes tools for assisting radiologists in identifying suspicious lesions and improving cancer detection rates, especially in dense breasts. Research continues to explore the potential of AI in breast cancer imaging to improve detection and treatment outcomes.

Current state-of-the-art AI tools for breast cancer detection might be repurposed for risk assessment. They may play a vital role in future risk-based breast cancer screening programs, including personalized long-term screening.

Style

no-vertical-section-spacing