SimBioSys

SimBioSys, Inc. is a Chicago-based medical software company that builds artificial intelligence tools for breast cancer care from standard breast magnetic resonance imaging (MRI). Its cloud-based TumorSight platform includes two products cleared by the U.S. Food and Drug Administration (FDA): TumorSight Viz, which turns a dynamic contrast-enhanced MRI (DCE-MRI) study into a 3D model of the tumor and surrounding tissue for surgical planning, and TumorSight Risk, which estimates five- and ten-year risk of breast cancer recurrence from that MRI combined with clinical and pathology data.^[1][2][3] TumorSight Risk received 510(k) clearance on 25 September 2026, and SimBioSys described it as the first AI-powered, MRI-derived breast cancer recurrence risk assessment.^[2][4] The company was founded out of the University of Illinois Urbana-Champaign and is a member of the NVIDIA Inception startup program; NVIDIA promoted TumorSight Risk on its main X account on 26 October 2026.^[5][6][24] Its products are an application of AI in healthcare to imaging-based surgical planning and cancer prognosis.

Overview

History

Origins at the University of Illinois

According to the company, Joe Peterson and John Cole were working together in a cancer modeling lab at the University of Illinois when, in 2017, they began discussing whether a newly funded National Cancer Institute grant could become a company. Tushar Pandey joined them, and the three founded SimBioSys.^[8] A 2021 University of Illinois Research Park announcement gives the founding year as 2018 and said the company was then based at the EnterpriseWorks incubator in the Research Park in Champaign.^[5] The name stands for "Simulating Biological Systems," which reflects the company's original approach of combining AI with biophysical simulation of tumors.^[5]

The company's first platform was called TumorScope. SimBioSys described it as software that built a 3D "digital twin" of a patient's tumor from diagnostic data (imaging, pathology, clinical and demographic information) and simulated tumor growth, metabolism, drug delivery and drug sensitivity over time to forecast how the tumor would respond to different treatments.^[6] One output was a "pCR score," the predicted probability that the tumor would be completely eradicated (pathologic complete response) after neoadjuvant therapy.^[6]

Series A and early growth

On 26 October 2021 SimBioSys announced a $15 million Series A to develop and commercialize TumorScope. Genoa Ventures and Northpond Ventures co-led the round, with participation from AV8 Ventures, Heritage Medical Group and Mayo Clinic, as well as existing investors and the founders. The round brought total capital raised to $21 million. Tushar Pandey was CEO, and Andrea Jackson of Northpond and Vikram Chaudhery of Genoa joined the board.^[5]

First FDA clearance and the TumorSight platform

The company's clinical products later carried the TumorSight name. The FDA cleared TumorSight Viz (510(k) K231130) on 26 December 2023, and SimBioSys announced the clearance on 4 January 2024 as its first.^[3][10] Around the same time it announced a strategic collaboration with Mayo Clinic to develop cloud-based tools covering surgical planning, treatment, drug selection and risk stratification for early-stage breast cancer. Under that agreement SimBioSys had access to data from patients in the Mayo Clinic-led BEAUTY trial (NCT02022202), and Mayo Clinic physicians Judy Boughey and Matthew Goetz joined the company's clinical advisory team. The release disclosed that Mayo Clinic has a financial interest in the technology.^[12]

In April 2024 SimBioSys announced a partnership with the venture studio General Inception under which General Inception's oncology drug companies could use PhenoScope, SimBioSys's research platform for biomarker discovery and dose selection.^[11]

Leadership change and commercialization

On 9 January 2025 SimBioSys named Stacey Stevens president and CEO. Stevens had previously been president and CEO of iCAD, a breast imaging AI company, and before that held commercial roles at Philips Healthcare, Agilent Technologies and Hewlett-Packard.^[9] A second TumorSight Viz clearance had come in October 2024. Under Stevens the company received a third clearance in July 2025, signed a non-exclusive distribution agreement with Ricoh 3D for Healthcare in August 2025, and agreed in July 2026 to offer TumorSight Viz through Ferrum Health's Model Hub, an enterprise platform for deploying clinical AI models.^{[13][14][15][16]} In January 2026 it appointed Senthil Periaswamy, a former iCAD and AstraZeneca AI executive, to the new role of chief AI and technology officer.^[17]

SimBioSys announced FDA clearance of TumorSight Risk on 15 October 2026.^[4]

FDA clearances

All four clearances listed for SimBioSys in the FDA 510(k) database were found substantially equivalent to a predicate device. As of 28 October 2026:

TumorSight Viz is regulated under 21 CFR 892.2050 (medical image management and processing system).^[3] TumorSight Risk is a Class II device under 21 CFR 864.3755, "pathology software algorithm device analyzing digital images for cancer prognosis."^[2] SimBioSys's website lists other products in development (TumorSight Plan, IQ, Marker and Dose) and states that these are not FDA cleared or approved.^[1]

TumorSight Viz

TumorSight Viz is a cloud-based application that reads breast DCE-MRI studies in DICOM format and displays the results in a web browser. Its features include standard image viewing tools, maximum intensity projections and reformats, image registration and subtraction, kinetic curves, parametric maps, automatic segmentation and 3D volume rendering, and automatic measurements such as tumor volume and the distance from the tumor to the nipple, chest wall and skin.^[3] The FDA-cleared indication is visualization and analysis of breast MRI for patients with biopsy-proven early-stage or locally advanced breast cancer, and the labeling states that patient management decisions should not be based solely on its results.^[3] SimBioSys describes the 3D renderings as a "digital twin" of the tumor shown in the context of automatically segmented skin, vessels, chest, fat, glandular tissue and heart.^[12]

The 510(k) summaries report the following validation results for tumor image segmentation, measured against radiologist ground truth on held-out patients from multiple U.S. sites:

The 2024 clearance added automated DICOM image import and an updated segmentation model; the 2025 clearance (marketed as version 1.3) added another updated segmentation model.^[13][14] In the first clearance, mean absolute error for automated tumor volume and landmark distances was reported to be similar to the variability between radiologists.^[3]

A peer-reviewed validation study of the platform, written mostly by SimBioSys employees with radiologists from the University of Alabama at Birmingham, was published in npj Breast Cancer in November 2024. It evaluated 100 cases whose ground-truth labels were vetted by two breast-specialized radiologists and reported that the automated measurements fell within the range of inter-radiologist variability.^[18]

In October 2025 ABC News and Good Morning America featured a Mercy Cedar Rapids patient whose surgeon, Vincent Reid, used TumorSight Viz to confirm that a lumpectomy was feasible instead of a mastectomy.^[7]

TumorSight Risk

Intended use

TumorSight Risk is a software-only, prescription device that uses age, race, clinical tumor stage, clinical nodal stage and tumor grade together with DCE-MRI data to produce 5- and 10-year risks of breast cancer recurrence and a proprietary prognostic score. It is indicated for adult women with hormone receptor-positive (HR+), HER2-negative, node-negative (N0) or node-positive (N1), Stage I to IIIA breast cancer, to inform a physician's prognostic risk-based decisions together with other clinicopathological factors.^[2] SimBioSys says this subtype represents about 75 percent of breast cancers in the United States.^[4]

Contraindications in the FDA decision summary include breast implants, prior mastectomy, unilateral DCE protocols, MRI acquired after treatment began, patients under 22, cN2 or higher nodal status, and MRI that does not meet American College of Radiology guidelines. The device is specified for GE and Siemens scanners at 1.5 T and 3 T.^[2]

How it works

The FDA decision summary describes a four-step pipeline:^[2]

1. The pre-treatment DCE-MRI is processed by the previously cleared TumorSight Viz device (K251766), which produces a multi-tissue segmentation map.
2. The segmentation map is passed automatically to TumorSight Risk.
3. TumorSight Risk computes intermediate imaging values from the segmentation.
4. These imaging values are combined with the clinical and pathology inputs in a regression-based machine learning model that outputs the TumorSight Risk (TSR) score and a risk category.

ER, PR and HER2 status are required to confirm eligibility but are not inputs to the algorithm.^[2] The model was trained and tuned on 1,133 samples from six U.S. sites and was locked before analytical and clinical validation.^[2]

The report also shows the observed 5- and 10-year recurrence risk for the patient's category in the clinical validation dataset.^[2] SimBioSys's product page says results are returned in one to two hours and positions the tool as a complement to tissue-based gene expression tests that does not require additional tissue.^[19]

Clinical validation for clearance

The FDA reviewed a retrospective study of 2,129 patients from four U.S. sites who were diagnosed between 1 January 2010 and 31 December 2024 and had a pre-treatment MRI. Recurrence at any time occurred in 181 patients (8.5 percent).^[2] SimBioSys identified the institutions as Mayo Clinic, Emory University, Stanford Medicine and Mercy Medical Cedar Rapids.^[4] Recurrence risk by category, estimated with Kaplan-Meier analysis:

The FDA summary treats the High and Low outputs as informative because their confidence intervals lie entirely above and below the overall (pre-test) risk, respectively, while the Intermediate category's risk was similar to the pre-test risk at both 5 and 10 years.^[2] When patients were grouped into finer score bins, 5-year risk rose from 2.2 percent (scores 0-10) to 13.6 percent (scores 58-100) and 10-year risk from 6.2 percent to 34.2 percent.^[2] The summary also reports subgroup results by race, age, nodal stage, treatment and MRI scanner; several subgroups had small numbers and wide confidence intervals (for example, only 7 cN1 patients were classified as low risk).^[2] In a reproducibility test that processed 120 MRIs three times each with added noise (119 processed successfully), overall percent agreement across replicates was 99.4 percent.^[2]

SimBioSys summarized the results as a 3.3-fold greater recurrence risk for high-risk than for low-risk patients and a 22 percent increase in relative recurrence risk for each 10-point increase in score.^[4]

Published abstracts

The validation was reported at the 2026 American Society of Clinical Oncology (ASCO) Annual Meeting in abstract 560 (Journal of Clinical Oncology supplement), whose first author is Matthew Goetz of Mayo Clinic. The abstract described a blinded validation of 2,129 patients from four institutions who had breast MRI before surgery without neoadjuvant therapy, with a median follow-up of 6.24 years. Unlike the cleared three-category device, the abstract used two pre-specified categories (high risk at a score of 47 or more). It reported an unadjusted hazard ratio of 2.71 for high versus low risk, 5-year recurrence-free survival of 95.5 percent (low) versus 89.6 percent (high), and 10-year rates of 88.9 versus 69.6 percent. The score stayed associated with recurrence-free survival after adjustment for pathologic stage, grade, age, race and treatment (p = 0.002).^[20] An OncLive article on the clearance gave slightly different figures (2,138 patients and a hazard ratio of 2.81).^[21]

The model's development was described in an abstract at the 2024 San Antonio Breast Cancer Symposium (Saha et al.). It was trained on 522 women and validated on 496 women from three independent health systems. In that validation cohort, 5-year recurrence-free survival was 93.1 percent for the low-risk group and 78.0 percent for the high-risk group, an unadjusted hazard ratio of 3.8. The authors reported that adding the imaging signature significantly improved prediction over a clinical-only model.^[22]

Context and limitations

Recurrence risk in HR+/HER2- early breast cancer has mainly been estimated from clinicopathologic features and tissue-based genomic assays. SimBioSys argues that genomic tests add cost, logistics and turnaround time and that some groups, such as premenopausal and node-positive women, have had fewer validated options.^[4] Goetz said in the company's release that MRI-derived assessments "may provide prognostic information beyond standard clinicopathologic factors."^[4] The clearance rests on retrospective data, the device is limited to patients who had a qualifying pre-treatment breast MRI on GE or Siemens equipment, and SimBioSys's release disclosed that Mayo Clinic, where the abstract's first author works, has a financial interest with SimBioSys.^[2][4] The predicate, ArteraAI Breast, is a tissue-based tool that analyzes H&E-stained whole-slide pathology images rather than MRI.^[2]

Relationship with NVIDIA

SimBioSys is a member of NVIDIA Inception, NVIDIA's program for startups. In a 2022 company blog post, Joseph Peterson and Dorys Lopez wrote that TumorScope was built on NVIDIA GPUs and that, with help from NVIDIA engineers, SimBioSys used the NVIDIA Clara and MONAI frameworks to cut training time for new medical imaging networks from about a month to under a day. They also wrote that moving CPU-based simulation code to CUDA reduced a single patient-specific response simulation from over 10 hours to about 15 minutes.^[6] SimBioSys said in 2025 that NVIDIA computing platforms and NVIDIA MONAI power the AI behind TumorSight Viz.^[23] ABC News reported that the system uses NVIDIA GPUs to process MRI data in minutes, and NVIDIA healthcare vice president Kimberly Powell appeared in the segment.^[7]

On 26 October 2026 NVIDIA's X account described TumorSight Risk as "Powered by NVIDIA technology" and said it "analyzes an existing breast MRI alongside pathology and clinical information to estimate the risk of cancer returning at five and ten years."^[24]

Leadership

As of the company's most recent announcements:

See also

• AI in Healthcare
• Digital twin
• Image segmentation
• Nvidia

References

1. ^1 ^2SimBioSys. "SimBioSys: Redefine precision medicine" (home page and product list). simbiosys.com (accessed 28 October 2026).
2. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9 ^10 ^11 ^12 ^13 ^14 ^15 ^16 ^17 ^18 ^19 ^20U.S. Food and Drug Administration. "510(k) Premarket Notification K260023: TumorSight Risk" and "510(k) Substantial Equivalence Determination Decision Summary K260023." accessdata.fda.gov/...pmn.cfm ; accessdata.fda.gov/...K260023.pdf
3. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8 ^9U.S. Food and Drug Administration. "510(k) Premarket Notification K231130: TumorSight Viz" (clearance letter and 510(k) summary). accessdata.fda.gov/...pmn.cfm ; accessdata.fda.gov/...K231130.pdf
4. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8SimBioSys (15 October 2026). "SimBioSys Secures FDA Clearance for TumorSight Risk, Introducing the World's First AI-Powered, MRI-Derived Breast Cancer Recurrence Risk Assessment." simbiosys.com/...fda-clearance-for-tumorsight-risk
5. ^1 ^2 ^3 ^4 ^5 ^6 ^7 ^8University of Illinois Research Park (26 October 2021). "EnterpriseWorks Startup SimBioSys Raises $15 Million Series A to Develop the Future of Precision Cancer Care." researchpark.illinois.edu/...iosys-raises-series-a
6. ^1 ^2 ^3 ^4Peterson, J. R.; Lopez, D. (20 September 2022). "Enabling personalized cancer care through AI and computer simulations with NVIDIA." SimBioSys. simbiosys.com/...-computer-simulations-with-nvidia
7. ^1 ^2Good Morning America / ABC News (23 October 2025). "40-year-old mom's cancer journey made easier thanks to AI." goodmorningamerica.com/...made-easier-ai-126682821
8. ^1 ^2SimBioSys. "Our Company." simbiosys.com/about-us (accessed 28 October 2026).
9. ^1 ^2 ^3 ^4SimBioSys (9 January 2025). "SimBioSys Appoints Stacey Stevens as President and CEO." simbiosys.com/...acey-stevens-as-president-and-ceo
10. ^1 ^2 ^3SimBioSys (4 January 2024). "SimBioSys receives milestone FDA clearance for its foundational digital precision medicine technology, TumorSight." simbiosys.com/...-future-cutting-edge-applications
11. ^1 ^2SimBioSys (3 April 2024). "Precision Oncology Innovator SimBioSys and General Inception Announce Strategic Partnership to Accelerate Cancer Drug Development." simbiosys.com/...ccelerate-cancer-drug-development
12. ^1 ^2SimBioSys (2 January 2024). "SimBioSys and Mayo Clinic Enter Strategic Collaboration to Advance AI-powered Digital Precision Medicine Platform for Cancer Patients." simbiosys.com/...cine-platform-for-cancer-patients
13. ^1 ^2 ^3 ^4U.S. Food and Drug Administration. "510(k) Premarket Notification K243189: TumorSight Viz" (510(k) summary). accessdata.fda.gov/...pmn.cfm ; accessdata.fda.gov/...K243189.pdf
14. ^1 ^2 ^3 ^4 ^5SimBioSys (9 July 2025). "SimBioSys Secures Third FDA Clearance for TumorSight Viz, Advancing AI-Powered Precision Surgery in Breast Cancer Care," with FDA 510(k) K251766 summary. simbiosys.com/...ion-surgery-in-breast-cancer-care ; accessdata.fda.gov/...K251766.pdf
15. ^SimBioSys (27 August 2025). "SimBioSys and Ricoh 3D for Healthcare announce strategic alliance expanding access to TumorSight Viz for breast cancer surgery." simbiosys.com/...ght-viz-for-breast-cancer-surgery
16. ^SimBioSys (23 July 2026). "SimBioSys and Ferrum Health Partner to Expand Access to AI-Powered Surgical Planning Through Ferrum's AI Governance Suite." simbiosys.com/...rough-ferrums-ai-governance-suite
17. ^1 ^2SimBioSys (27 January 2026). "SimBioSys Appoints Senthil Periaswamy as Chief AI & Technology Officer." simbiosys.com/...my-as-chief-ai-technology-officer
18. ^Weitz, M.; Pfeiffer, J. R.; Patel, S.; et al. (2024). "Performance of an AI-powered visualization software platform for precision surgery in breast cancer patients." *npj Breast Cancer* 10, 98. doi.org/...s41523-024-00696-6
19. ^SimBioSys. "TumorSight Risk." simbiosys.com/tumorsight-risk (accessed 28 October 2026).
20. ^Goetz, M. P.; Xu, Q.; Reid, V.; et al. (2026). "Prognostic value of an MRI-derived recurrence assay in women with hormone receptor-positive, HER2-negative breast cancer: A retrospective, multi-center analysis." *Journal of Clinical Oncology* 44 (16 suppl), abstract 560. doi.org/...JCO.2026.44.16_suppl.560
21. ^OncLive Staff (19 October 2026). "FDA Clears TumorSight Risk, an MRI-Derived AI Recurrence Tool for HR+, HER2- Breast Cancer." OncLive. onclive.com/...ence-tool-for-hr-her2-breast-cancer
22. ^Saha, P.; Howard, F. M.; Stringer-Reasor, E. M.; et al. (2025). "Leveraging AI to Predict Recurrence-Free Survival in Breast Cancer Patients through Image-Based assessment of Tumor Characteristics." San Antonio Breast Cancer Symposium 2024, abstract PS11-06. *Clinical Cancer Research* 31 (12 Suppl). doi.org/...1557-3265.SABCS24-PS11-06
23. ^SimBioSys (27 October 2025). "SimBioSys Showcased on ABC News Digital and Good Morning America Highlighting the Power of AI-Driven 3D Visualization in Breast Cancer Surgery." simbiosys.com/...lization-in-breast-cancer-surgery
24. ^1 ^2NVIDIA (@nvidia) (26 October 2026). Post on X about SimBioSys TumorSight Risk. x.com/...2114827978748383342