FDA Clears TumorSight Risk, an MRI-Derived AI Recurrence Tool for HR+, HER2– Breast Cancer

The FDA has granted 510(k) clearance to TumorSight Risk, an artificial intelligence (AI)–based software that estimates 5- and 10-year recurrence-free survival (RFS) from routine breast MRI, pathology, and clinical data in adult patients with hormone receptor–positive, HER2-negative stage I to IIIA breast cancer.¹

TumorSight Risk is the first AI-powered, MRI-derived breast cancer recurrence risk assessment, and the output is intended for use alongside other clinicopathologic factors.

In a retrospective multicenter validation analysis of 2138 patients, the 5-year RFS rate was 95.5% (95% CI, 94.6%-96.3%) in the low-risk category (n = 1815) vs 89.6% (95% CI, 86.4%-92.6%) in the high-risk category (n = 323; HR) for recurrence, 2.81; 95% CI, 2.03-3.88; P < 5 × 10⁻¹⁰).² At 10 years, the RFS rates were 88.9% (95% CI, 86.5%-90.9%) vs 69.6% (95% CI, 59.1%-77.9%), respectively.

“MRI-derived assessments may provide prognostic information beyond standard clinicopathologic factors,” Matthew P. Goetz, MD, stated in a news release.¹

Goetz is the Erivan K. Haub Family Professor of Cancer Research and director of the Mayo Breast Cancer SPORE at Mayo Clinic in Rochester, Minnesota.¹

How does TumorSight Risk generate a risk estimate?

The software is built on the FDA-cleared TumorSight Viz platform and applies spatial AI modeling to tumor characteristics, context for surrounding tissue, and clinical factors. It requires no additional procedure, tissue shipment, or specialized hardware, and a cloud-based platform returns a score from 0 to 100 with a high, intermediate, or low risk category. The results are designed to be available before surgery within standard workflows.

How was the validation cohort assembled?

Findings from a validation cohort of 2129 patients and described a blinded, prospectively designed retrospective study presented at the 2026 ASCO Annual Meeting.¹ The primary prespecified analysis was the recurrence rate at 5 years after diagnosis, and a prespecified secondary analysis was a multivariable assessment of the risk score and category.²

Most patients were 40 to 69 years of age (78%). In total, 72% of patients were White, 16% of patients were Black or African American, and 9% of patients were Asian. Tumors were grade 1 in 35% of patients, grade 2 in 51% of patients, and grade 3 in 14% of patients, and clinical T stage was cT1 in 61% of patients, cT2 in 32% of patients, and cT3 in 7% of patients. Clinical nodal stage was cN0 in 85% of patients and cN1 in 15% of patients. A total of 183 of 2138 patients (8.5%) had a recurrence at any time, and 1955 patients (91.5%) were disease free.

What did the multivariable and time-dependent analyses show?

In the multivariable analysis, the HR for risk category was after adjustment for clinical parameters was 2.18 (95% CI, 1.35-3.52). In exploratory analyses by time interval, the HR for the high-risk vs low-risk category was 2.8 (95% CI, 2.0-3.9; P < .005) from 0 to 5 years (low risk, n = 1814; high risk, n = 323) and 3.4 (95% CI, 2.0-5.6; P < .005) from 5 to 10 years (low risk, n = 1242; high risk, n = 167). The investigators concluded that the risk score and category provide prognostic information independent of therapy given and pathologic data.

The news release stated that patients in the high-risk group had a 3.3-fold greater recurrence risk than those in the low-risk group, and that each 10-point increase in score reflected a 22% increase in relative recurrence risk.¹ The estimated 5-year recurrence rates ranged from 2.2% (scores 0-10) to 13.6% (scores 58-100), and the estimated 10-year recurrence rates ranged from 6.2% to 34.2% across the same groups. The score remained associated with RFS after adjustment for standard prognostic factors (P = .002).

References

1. SimBioSys secures FDA clearance for TumorSight Risk, introducing the world's first AI-powered, MRI-derived breast cancer recurrence risk assessment. News release. SimBioSys. October 15, 2026. Accessed October 19, 2026. https://simbiosys.com/2026/10/15/simbiosys-secures-fda-clearance-for-tumorsight-risk/
2. Goetz MP, Xu Q, Reid V, et al. Prognostic value of an MRI-derived recurrence assay in women with hormone receptor-positive, HER2-negative breast cancer: a retrospective, multi-center analysis. J Clin Oncol. 2026;44(suppl 16):560. doi:10.1200/JCO.2026.44.16_suppl.560