FDA Clears AI Tool to Score Breast Cancer Recurrence

TumorSight RiskWhat to know

30-second read

For Oncology · Radiology · Pathology

Who

Women with hormone receptor-positive, HER2-negative, stage I–IIIA breast cancer being worked up for adjuvant treatment decisions.

What it detects

A 0-100 recurrence-risk score and a high/intermediate/low category for 5- and 10-year recurrence-free survival, read from an MRI already on file.

Evidence

Validated on 2,129 patients across 4 centers (Mayo, Emory, Stanford, Mercy); 99.4% run-to-run agreement. No peer-reviewed publication yet.

Where it fits

Alongside — not in place of — tissue-based assays like Oncotype DX and MammaPrint, at the same point: before the chemotherapy conversation.

Key consideration

No trial yet ties the score to chemotherapy benefit the way TAILORx validated Oncotype DX’s intermediate range; this measures risk, not proven benefit.

Availability

Cleared 25 September 2026 (510(k) K260023), announced 15 October; no launch date, price or payer coverage decision announced yet.

Practice impactConsider alongside tissue-based assays

A genuinely new, biopsy-free option for a decision point oncologists already own. Order it as additional information this year, not as a replacement for a validated chemo-benefit test, until its own outcome data matures.

Read the full physician analysis

Full analysis · 9 min readFDA 510(k) K260023 · SimBioSys validation study · TAILORx (Sparano et al., NEJM 2018)

FDA cleared SimBioSys’ TumorSight Risk on 25 September 2026, an AI software device that scores a breast-cancer patient’s five- and ten-year recurrence risk from an MRI, pathology report and clinical factors already sitting in her chart. Regulatory News covered the clearance itself, the company and the broader recurrence-testing market the day it was announced. This is the clinical layer underneath that: where this score actually sits next to the tests oncologists already order, and the one piece of evidence it doesn’t have yet.

The decision this is built to inform

For a patient with hormone receptor-positive, HER2-negative, early-stage breast cancer, the question after surgery is whether adjuvant chemotherapy adds enough benefit on top of endocrine therapy to justify its toxicity. Oncologists have answered that question for two decades with gene-expression assays run on tumor tissue — Oncotype DX (Exact Sciences) and MammaPrint (Agendia) chief among them — sent to a central lab and returned days to weeks later. TumorSight Risk proposes a different route to a related number: instead of reading gene expression in a tissue sample, it reads the shape and texture of the tumor on an MRI the patient already had, plus pathology and clinical factors already in the chart. No new procedure, no tissue shipment, no dedicated scanner.

What 2,129 patients actually proved

The clearance rests on a blinded, multi-institutional validation study of 2,129 patients drawn from Mayo Clinic, Emory University, Stanford Medicine and Mercy Medical, alongside a precision study showing 99.4% agreement across replicate runs on the same image and a 30-clinician human-factors study. FDA listed the device under product code SHW, a Class II designation, and named ArteraAI Breast — another imaging-and-AI prognostic tool, not a tissue-based assay — as the predicate device for the 510(k). What that validation study demonstrates is that the score is reproducible and correlates with recurrence-free survival outcomes in the population studied. What it does not demonstrate, because it wasn’t designed to, is whether a patient’s score predicts how much she benefits from adding chemotherapy. Those are different questions, and the gap between them is the one a prescriber needs to hold onto before ordering this instead of, rather than alongside, an established assay.

The evidence Oncotype DX has that this doesn’t, yet

Oncotype DX’s clinical authority doesn’t rest only on correlating its Recurrence Score with outcomes — it rests on TAILORx, a prospective randomized trial of 10,273 women that assigned patients with intermediate scores (11–25) to endocrine therapy alone or endocrine therapy plus chemotherapy and found equivalent outcomes at roughly nine years of follow-up. That trial is why a mid-range Oncotype score can tell a patient, with randomized evidence behind it, that chemotherapy is very unlikely to help her. TumorSight Risk has nothing equivalent yet: its 2,129-patient study establishes that the score tracks recurrence risk, not that a specific score threshold identifies patients who can safely skip chemotherapy without losing benefit. SimBioSys calls TumorSight Risk the world’s first AI-powered, MRI-derived prognostic tool for breast-cancer recurrence — a claim about itself that Regulatory News has not independently verified against every competing imaging-based tool, and one that describes a novel approach, not yet a validated chemo-benefit test.

Three routes to the same adjuvant-chemotherapy decision

	TumorSight Risk	Oncotype DX	MammaPrint
Input	Existing breast MRI, pathology, clinical data	Tumor tissue, sent to a central lab	Tumor tissue, sent to a central lab
New procedure required	None	None beyond the original biopsy	None beyond the original biopsy
Output	0-100 score; high/intermediate/low, 5- & 10-yr	Recurrence Score, 0-100	Low/high risk, binary
Chemo-benefit evidence	Recurrence-risk correlation only; no chemo-benefit trial yet	TAILORx: randomized, 10,273 patients, intermediate range validated	MINDACT: randomized, ~6,700 patients, low-risk range validated
U.S. clearance	510(k), September 2026	2004 (LDT; components FDA-cleared since)	510(k), 2007

From each product's own FDA record and published trial literature; not a clinical recommendation. MINDACT and TAILORx particulars from their published NEJM reports.

What changes at the point of ordering

• Faster, cheaper information is now available at the point a recurrence-risk conversation starts — no added turnaround time for tissue shipment, no added biopsy cost.
• It is not yet a chemo-benefit test. Treat a TumorSight Risk score as additional prognostic information, not as grounds on its own to withhold chemotherapy the way a validated intermediate Oncotype score can support.
• Coverage is undetermined. SimBioSys says it is working with Medicare, Medicaid and private insurers toward reimbursement; no coverage decision has been announced, so ordering it today may mean an uncovered cost to the patient or the practice.
• The underlying study isn’t published yet. A journal reviewer hasn’t had the chance to stress-test the 2,129-patient validation cohort the way peer review would; that is due diligence a prescriber would otherwise rely on a journal to have already done.
• It doesn’t cover triple-negative or HER2-positive disease — the clearance is specific to hormone receptor-positive, HER2-negative, stage I–IIIA breast cancer.

What radiologists and pathologists should expect to field

Because TumorSight Risk is built on imaging already ordered for staging or surgical planning, it puts radiologists in the position of having generated an input to a prognostic score without having been the one to order the score itself — a workflow question practices will need to settle locally: who requests TumorSight Risk, from what existing study, and when in the diagnostic sequence. Pathologists supply the companion clinical and histologic inputs (grade, nodal status, receptor status) the same way they already do for tissue-based assays, so the data-entry burden is familiar even though the specimen requirement isn’t. Neither specialty needs new equipment or training to participate; the open question is process, not capability.

SimBioSys’s fourth clearance, and what’s still ahead

TumorSight Risk is SimBioSys’ fourth FDA 510(k) clearance, following three clearances of TumorSight Viz, a 3D breast-tumor surgical-visualization platform, since December 2023. The company has not announced a launch timeline, a price, or which health systems will be first to offer TumorSight Risk clinically, and the 2,129-patient validation study has not yet appeared in a peer-reviewed journal as far as Regulatory News can confirm. For a practice deciding whether to offer it, those three gaps — publication, pricing and coverage — are the practical blockers between a cleared device and a device actually available to order.

Frequently asked questions

Does TumorSight Risk replace Oncotype DX or MammaPrint?

Not based on what's cleared so far. TumorSight Risk reports a recurrence-risk score from imaging; it has not been shown, in a published trial, to predict which patients actually benefit from adding chemotherapy the way Oncotype DX's intermediate-score range was validated in TAILORx. Until that evidence exists, it's a new option to weigh, not a validated substitute for a chemo-benefit decision.

Does a patient need a new procedure to get a TumorSight Risk score?

No. SimBioSys says the score is computed from a breast MRI, pathology report and clinical factors already collected as part of routine diagnostic workup — no additional biopsy, tissue shipment or specialized scanner.

Is TumorSight Risk covered by insurance?

Not yet determined. SimBioSys has said it is working with Medicare, Medicaid and private insurers toward reimbursement, but no coverage decision had been announced as of this story's publication.