FDA clears AI tool to estimate breast cancer recurrence risk before surgery

The FDA has granted 510(k) clearance to SimBioSys’s TumorSight Risk, an AI tool designed to estimate breast cancer recurrence risk before surgery. Announced by the company on October 15, 2026, the clearance covers a specific group: adults with hormone receptor–positive, HER2-negative stage I to IIIA breast cancer.

The software uses routine breast MRI, pathology and clinical information to estimate the likelihood of remaining free of recurrence at five and 10 years. Its output is intended to sit alongside other clinical and pathology findings, rather than serve as a stand-alone assessment, according to OncLive’s October 19 coverage of the clearance.

A forecast from data already in the workflow

TumorSight Risk builds on the FDA-cleared TumorSight Viz platform. Its spatial AI modeling considers tumor characteristics, the surrounding tissue and clinical factors. A cloud-based service returns a score from zero to 100, accompanied by a low-, intermediate- or high-risk category. The score summarizes the recurrence assessment rather than simply identifying a tumor in an image.

The system requires no additional procedure, tissue shipment or specialized hardware, OncLive reports. Results are designed to arrive before surgery within standard workflows. That timing is central to the proposed use: a recurrence assessment could become available before definitive surgery, rather than waiting for surgical pathology, physician Matthew Kurian wrote in commentary reproduced by OncoDaily.

The intended population is important. This is not a recurrence tool for every breast cancer subtype. Kurian, a physician at St. Elizabeth Healthcare, emphasized that its current use does not extend to triple-negative or HER2-positive disease. The clearance described by OncLive also specifies adults with stage I through IIIA disease.

Testing the forecast against later recurrence

The validation described by OncLive was a blinded, prospectively designed retrospective study presented at the 2026 American Society of Clinical Oncology meeting. Its primary prespecified analysis examined recurrence at five years after diagnosis. A secondary analysis assessed the score and risk category alongside other variables, testing whether the association remained after accounting for established factors.

95.5%Low-risk group

Reported five-year recurrence-free survival in the retrospective validation’s low-risk group.

89.6%High-risk group

Reported five-year recurrence-free survival in the retrospective validation’s high-risk group.

At 10 years, OncLive reported recurrence-free survival of 88.9% in the low-risk group and 69.6% in the high-risk group. The corresponding 95% confidence intervals were 86.5%–90.9% and 59.1%–77.9%, showing greater uncertainty around the high-risk estimate. The score remained associated with recurrence-free survival after adjustment for standard prognostic factors.

Recurrence risk is not chemotherapy benefit

Kurian drew a distinction between prognosis—estimating what may happen—and predicting whether chemotherapy will help. In his commentary, published by OncoDaily on October 18, he cautioned that clearance permits marketing for the intended use; it does not establish improved patient outcomes. Clinical utility and comparisons with genomic assays remain unproven, he wrote.

Jasmin Hundal, identified by OncoDaily as a former chair of a hematology-oncology research fellowship working group at Cleveland Clinic, framed the next step as proving usefulness in care. Her questions concern decisions, not just forecasts: whether these tools can reduce unnecessary therapy, improve outcomes and widen access to personalized risk assessment. She described combining imaging, pathology, genomics and clinical data as a possible future direction, not an established result of this clearance.