NCCN Updates Breast Cancer Guidelines to Include AI-Based Risk Assessment

The National Comprehensive Cancer Network (NCCN) has formally updated its 2026 Clinical Practice Guidelines in Oncology for Breast Cancer Screening and Diagnosis to include image-based artificial intelligence (AI) risk assessment as a primary tool for identifying individuals at increased risk for the disease.¹ The update marks a significant shift in the standard of care, integrating imaging-based AI models to predict 5-year future breast cancer risk directly from routine screening mammograms.

Establishing a New Threshold for Personalized Care

The 2026 guidelines introduce a specific 5-year breast cancer risk threshold of ≥1.7% determined by AI-based analysis as a criterion for identifying those at increased risk. This assessment utilizes pixel-level data from standard bilateral 2D mammograms to detect subtle tissue patterns associated with future malignancy, many of which are imperceptible to the human eye.

Clairity Breast is the first FDA-approved model to predict 5-year breast cancer risk using AI-based mammography and the only model currently available for commercial use. Unlike traditional models such as the Gail or Tyrer-Cuzick models, which rely on patient questionnaires, family history, and genetic markers, the imaging-based AI model generates a prognostic score using only the mammographic image. This capability addresses a long-standing gap in oncology: the identification of "sporadic" cancers in those who lack high-risk genetic mutations or significant family histories, a group that constitutes roughly 85% of all breast cancer diagnoses.

Clinical Implementation and Supplemental Imaging

The NCCN’s move links AI-derived risk assessment directly to clinical intervention. For those identified as having a 5-year risk ≥1.7%, the guidelines now recommend:

• Supplemental Imaging: Consideration of MRI or ultrasound as an adjunct to annual mammography.
• Risk-Reduction Strategies: Discussion of chemoprevention or lifestyle modifications tailored to individual risk levels.
• Earlier Screening: Expanding the identification of increased-risk individuals starting at age 35, rather than the traditional age 40 baseline.
• Dynamic Reassessment: Recognizing that breast tissue and risk profiles are not static, the guidelines call for periodic risk re-evaluation over time.

"Advances in AI now allow us to extract critical information from a mammogram about a woman's future risk in a clinically meaningful way," said Connie Lehman, MD, PhD, professor of radiology at Harvard Medical School and CEO of Clairity Inc, in a news release. "This provides a foundation for more precise, individualized care that is accessible at the point of care."

Validation and Trial Data

The inclusion of AI-based models in the guidelines is supported by large-scale validation studies. Findings presented at the 2025 Annual Meeting of the Radiological Society of North America (RSNA) demonstrated that the imaging-based AI model provides significantly stronger risk stratification than radiologist-reported breast density. In a multicenter analysis of 245,344 bilateral mammograms, those categorized as high risk by the AI model had a breast cancer incidence of 5.9% over 5 years, compared with 1.3% in the average-risk group—a more than 4-fold difference. By contrast, breast density alone showed only a modest separation (3.2% for dense vs. 2.7% for nondense).²

Ongoing prospective validation of these AI models is being conducted through clinical trials, including the MIRAI-MRI study (NCT05968157), which quantifies the clinical benefit of using AI-based guidelines to trigger supplemental MRI screening.³

REFERENCES

1. Clairity Announces Update to NCCN Guidelines to Include AI Mammogram-Based Risk Tool to Identify Future Breast Cancer Risk in Women Missed by Traditional Methods. News release. Clairity Inc. April 14, 2026. Accessed April 15, 2026. https://tinyurl.com/mr2mf3ny

2. Lehman CD, Kuhl C, et al. Image-only AI model compared with breast density for identifying women at 5-year increased risk of breast cancer: Multicenter U.S.–European analysis. Presented at: RSNA Annual Meeting; December 2025; Chicago, IL.

3. MIRAI-MRI: Comparing Screening MRI for Patients at High Risk for Breast Cancer Identified by Mirai and Tyrer-Cuzick. ClinicalTrials.gov. Updated January 7, 2026. Accessed April 15, 2026. https://clinicaltrials.gov/study/NCT05968157