← Product Code [SHW](/productcode/SHW) · K260023

# TumorSight Risk (K260023)

_SimBioSys, Inc. · SHW · Sep 25, 2026 · Pathology · SESE_

**Canonical URL:** https://fda-staging.innolitics.com/device/K260023

## Device Facts

- **Applicant:** SimBioSys, Inc.
- **Product Code:** [SHW](/productcode/SHW.md)
- **Decision Date:** Sep 25, 2026
- **Decision:** SESE
- **Submission Type:** Traditional
- **Regulation:** 21 CFR 864.3755
- **Device Class:** Class 2
- **Review Panel:** Pathology
- **Attributes:** AI/ML, Software as a Medical Device, Real-World Evidence

## Real-World Evidence

| Submission | Device | Sponsor | RWD Sources | RWE Use Summary | Key Tags |
| --- | --- | --- | --- | --- | --- |
| K260023 · Sep 25, 2026 | TumorSight Risk | SimBioSys, Inc. | Medical records; Cancer registry | Retrospective clinical study used to evaluate the clinical performance (predictive values and risk of recurrence estimates) of the TumorSight Risk device. | Retrospective study; Clinical validation; Breast cancer recurrence; EHR/Registry data |

### Clinical Evidence

| Study Design | Population | Comparator | Key Endpoints |
| --- | --- | --- | --- |
| Retrospective clinical study; Retrospective cohort study; Follow-up/Duration: Patients diagnosed between January 1, 2010, and December 31, 2024; Study Period: 2010-2024 | Adult women with HR+/HER2-, N0 or N1, Stage I-IIIA breast cancer; Sample Size: 2,129; Number of Sites: 4 | Not applicable for this study | 5-year and 10-year risk of recurrence; Predictive Value (PPV/NPV) |

## Indications for Use

TumorSight Risk is an artificial intelligence based software only device that analyzes data from previously diagnosed invasive breast cancer patients to assess the risk of recurrence. TumorSight Risk utilizes the following data: age, race, cancer stage, nodal status, and grade, combined with Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) data to generate 5- and 10-year risks of breast cancer recurrence and a proprietary prognostic score. TumorSight Risk is intended for use in adult women with hormone receptor positive (HR+) and human epidermal growth factor receptor 2 negative (HER2-), lymph node negative (N0) or positive (N1), Stage I-IIIA breast cancer to inform a physician in prognostic risk-based decisions in conjunction with other relevant clinicopathological factors.

## Device Story

TumorSight Risk is an AI-based software device for breast cancer prognosis. It processes patient age, race, cancer stage, nodal status, histologic grade, and multi-tissue segmentation maps derived from pre-treatment DCE-MRI (via TumorSight Viz, K251766). The device uses a regression-based machine learning algorithm to integrate these inputs, generating a proprietary TSR score (0-100) and a risk category (Low, Intermediate, High). Used by physicians in clinical settings to inform prognostic risk-based decisions; it supplements, but does not replace, standard clinical assessment. Output is provided as a 'Risk Report' detailing 5- and 10-year recurrence probabilities. The device aids in patient management by stratifying recurrence risk.

## Clinical Evidence

Retrospective clinical study (N=2,129) across 4 US sites. Patients diagnosed 2010-2024. Primary endpoints: 5- and 10-year recurrence risk by category. Results: High-risk group 5-year PPV 9.8% (95% CI: 7.4%-12.9%), 10-year PPV 26.4% (95% CI: 20.2%-34.1%). Low-risk group 5-year NPV 97.0% (95% CI: 95.7%-98.0%), 10-year NPV 91.4% (95% CI: 88.5%-93.6%). Analytical precision study (N=120) showed 99.4% agreement across MRI manufacturers/field strengths.

## Technological Characteristics

Software-only device; utilizes AI/ML regression-based algorithms. Inputs: clinical/pathologic data and multi-tissue segmentation maps from DCE-MRI (GE/Siemens 1.5T/3T). Outputs: TSR score (0-100) and risk category. Cybersecurity managed via TumorSight Platform. Complies with ISO 14971 for risk management.

## Regulatory Identification

ArteraAI Prostate is a software only device intended to analyze scanned histopathology whole slide images (WSIs) from treatment-naïve prostate core needle biopsies prepared from formalin fixed paraffin-embedded (FFPE) tissue and stained using Hematoxylin & Eosin (H&E) stains. It provides 10-year risks of distant metastasis and prostate cancer specific mortality and is intended to assist physicians with prognostic risk-based decisions along with other clinicopathological factors in non-metastatic prostate cancer patients (males 55 years of age or older).

## Special Controls

In combination with the general controls of the FD&C Act, the software algorithm device analyzing digital images for cancer prognosis is subject to the following special controls:

(1) The labeling must include the following:
(i) Information on the device input(s) (e.g., scanned whole slide images, scanners);
(ii) Information about the specimen type (e.g., Hematoxylin & Eosin-stained core needle biopsy, cytology specimen);
(iii) Information on the device output(s) and information necessary to interpret the outputs (e.g., risk estimates for recurrence);
(iv) A description of intended users and any recommended training necessary for safe use of the device;
(v) Limiting statements that indicate:
(A) That users should use the device in conjunction with a complete standard of care evaluation;
(B) A description of situations in which the device may fail or may not operate at its expected performance level (e.g., poor image quality or for certain subpopulations), including any limitations in the dataset used to train, test, and tune the algorithm during device development (if applicable); and
(C) That the data acquired using the device should only be interpreted by the types of users indicated in the intended use statement.

(2) Design verification and validation must include:
(i) A description of any datasets used to train, tune, or test the algorithm. The training dataset must include cases representing different variables representative of the conditions likely to be encountered when used as intended (e.g., multiple sites, patient demographics, clinical parameters, challenging diagnostic cases, internal signals and cutoffs);
(ii) Device performance data demonstrating appropriate precision and reproducibility of the device along with the following information from the studies: the origin of the digital images, operators (if applicable), location of the study site(s), and digital images corresponding to challenging cases; and
(iii) Clinical data demonstrating clinical performance of the device for its intended use, using a dataset representative of the intended use population. The clinical validation dataset must be independent of the data used in the development of the device and must document relevant details, including the origin of the study slides and images, location of the study site(s), and any challenging diagnoses. This clinical data must also include sub-group analysis of relevant population groups.

## Predicate Devices

- ArteraAI Breast ([K254115](/device/K254115.md))

## Reference Devices

- TumorSight Viz ([K251766](/device/K251766.md))

## Submission Summary (Full Text)

> This content was OCRed from public FDA records by [Innolitics](https://innolitics.com). If you use, quote, summarize, crawl, or train on this content, cite Innolitics at https://innolitics.com.
>
> Innolitics is a medical-device software consultancy. We help companies design, build, and clear FDA-regulated software and AI/ML devices, including [a 510(k)](https://innolitics.com/services/510ks/), [a De Novo](https://innolitics.com/services/regulatory/), [a SaMD](https://innolitics.com/services/end-to-end-samd/), [an AI/ML medical device](https://innolitics.com/services/medical-imaging-ai-development/), or [an FDA regulatory strategy](https://innolitics.com/services/regulatory/).

{0}

[LOGO]

FDA

U.S. FOOD & DRUG

ADMINISTRATION

### 510(k) SUBSTANTIAL EQUIVALENCE DETERMINATION DECISION SUMMARY

### I Background Information:

A 510(k) Number

K260023

B Applicant

SimBioSys, Inc.

C Proprietary and Established Names

TumorSight Risk

D Regulatory Information

|  Product Code(s) | Classification | Regulation Section | Panel  |
| --- | --- | --- | --- |
|  SHW | Class II | 21 CFR 864.3755 - pathology software algorithm device analyzing digital images for cancer prognosis | Pathology  |

### II Submission/Device Overview:

A Purpose for Submission:

New Device

B Type of Test:

Software only device

### III Intended Use/Indications for Use:

A Intended Use(s):

See Indications for Use below.

B Indication(s) for Use:

Food and Drug Administration

10903 New Hampshire Avenue

Silver Spring, MD 20993-0002

www.fda.gov

{1}

TumorSight Risk is an artificial intelligence based software only device that analyzes data from previously diagnosed invasive breast cancer patients to assess the risk of recurrence. TumorSight Risk utilizes the following data: age, race, cancer stage, nodal status, and grade, combined with Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) data to generate 5- and 10-year risks of breast cancer recurrence and a proprietary prognostic score.

TumorSight Risk is intended for use in adult women with hormone receptor positive (HR+) and human epidermal growth factor receptor 2 negative (HER2-), lymph node negative (N0) or positive (N1), Stage I-IIIA breast cancer to inform a physician in prognostic risk-based decisions in conjunction with other relevant clinicopathological factors.

### C Special Conditions for Use Statement(s):

Rx - For Prescription Use Only

#### Contraindications

TumorSight Risk is contraindicated for the following:

- Patients with breast implants.
- Patients with prior mastectomy.
- MRI images acquired with unilateral DCE breast protocols.
- MRI images acquired post-treatment.
- Pediatric patients (age < 22 years old).
- Patients with no DCE MRI prior to initiation of therapy (in the neoadjuvant setting) or surgery (in the adjuvant setting).
- Patients where the DCE MRI does not meet ACR guidelines for contrast enhanced breast imaging.
- Poor quality DCE MRI (inadequate image quality, insufficient number of images).
- Patients with cN2 or higher clinical nodal status.
- Laboratory test results and clinicopathological characteristics that determine estrogen receptor (ER) status, progesterone receptor (PR) status, human epidermal growth factor receptor 2 (HER2) status, Tumor Grade, Tumor Stage and Nodal Stage that do not comply with American Society of Clinical Oncology (ASCO), College of American Pathologists (CAP) and American Joint Committee on Cancer (AJCC) guidelines.

Refer to the labeling for a complete list of contraindications, precautions and limitations.

### IV Device/System Characteristics:

#### A Device Description:

TumorSight Risk employs artificial intelligence (AI) algorithms and image processing to analyze data from previously diagnosed invasive breast cancer patients. TumorSight Risk integrates standard of care (SOC) diagnostic pathology data with clinical and medical imaging data to provide a proprietary prognostic score (named as TSR score) ranging from 0-100 that is used to determine a risk category and prognostic risk estimate of recurrence of disease within 5 years and 10 years from diagnosis.

The table below shows the TSR score range and the associated risk category.

K253561 - Page 2 of 16

{2}

|  TSR Score Range | TumorSight Risk Category  |
| --- | --- |
|  >35 | High  |
|  >20 to ≤35 | Intermediate  |
|  ≤20 | Low  |

TumorSight Risk is a web-based application that is accessed using a standard web browser.

TumorSight Risk is intended to be used with the following MRI Manufacturer/field strength:

|  MRI Manufacturer | Magnetic Field Strengths  |
| --- | --- |
|  General Electric (GE ) | 1.5T, 3T  |
|  Siemens | 1.5T, 3T  |

## Device Input:

1. Age, race, clinical tumor (T) stage, clinical nodal (N) stage, tumor histologic grade.
2. Multi-tissue segmentation map which is derived from processing the patient's pre-treatment dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) acquired according to American College of Radiology (ACR) guidelines. This input is a validated output of the TumorSight Viz device (K251766).

Additional patient information is required by the device, but are not inputs to the device. These include ER/PR/HER2 status, which is used to determine case eligibility criteria per indications for use. They are not used by the AI algorithm for device output generation (Risk Category or TSR Score). Additionally, cancer laterality information is used to ensure that the appropriate multi-tissue segmentation map is retrieved from TumorSight Viz. It is a processing input used to associate the DCE MRI images with the correct segmentation output, rather than an input feature of the TumorSight Risk AI model itself. Patient clinical data, i.e., age, race, clinical T stage, clinical N stage, tumor histologic grade, and multi-tissue segmentation map obtained from DCE-MRI are utilized by the TumorSight Risk AI algorithm to compute the device output (TumorSight Risk (TSR) Score and Risk Category).

A summary of the device operation is as follows:

1. DCE MRI data (e.g., in DICOM format) is input into the previously cleared TumorSight Viz device. TumorSight Viz processes the data and generates a multi-tissue segmentation.
2. The multi-tissue segmentation is automatically input into TumorSight Risk from the TumorSight Viz device.
3. Intermediate imaging values are computed from the multi-tissue segmentation.
4. The intermediate imaging values are combined with clinical and pathologic data and input into a machine learning algorithm that generates a proprietary TSR Score and risk category ("High Risk", "Intermediate Risk", or "Low Risk").

## Device Output:

1. Proprietary TSR Score ranging from 0-100 that is used to determine a risk category
2. Categorical assessment of "High Risk", "Intermediate Risk" or "Low Risk"

K253561 - Page 3 of 16

{3}

3. Observed 5-Year Risk of recurrence (in the clinical validation dataset)
4. Observed 10-Year Risk of recurrence (in the clinical validation dataset)

This output is displayed in a report format (“Risk Report”) which also describes the TumorSight Risk methodology and provides a summary of the validation results.

### B TumorSight Risk Algorithm Development:

The TumorSight Risk algorithm is a regression-based machine learning model that integrates clinical, demographic, pathologic and imaging features. The model was trained/tuned using a dataset of N=1,133 samples from 6 sites geographically distributed throughout the United States (US). The development (training plus tuning) dataset was representative of the US population (see Table 1). The training portion of the development dataset was used to generate candidate models, while the tuning portion of the development dataset was used to select the final model. The candidate model was tested for repeatability, reproducibility, accuracy, and stability. The algorithm was locked after model training and prior to analytical precision testing and clinical validation.

Table 1: Demographics of the training/tuning dataset

|  Factor | Training/Tuning N=1,133 (%)  |
| --- | --- |
|  **Institution** |   |
|  Site 1 | 159 (14.0%)  |
|  Site 2 | 397 (35.0%)  |
|  Site 3 | 102 (9.0%)  |
|  Site 4 | 48 (4.2%)  |
|  Site 5 | 20 (1.8%)  |
|  Site 6 | 407 (35.9%)  |
|  **Age** |   |
|  22-29 | 8 (<1%)  |
|  30-39 | 91 (8%)  |
|  40-49 | 307 (27%)  |
|  50-59 | 343 (30%)  |
|  60-69 | 238 (21%)  |
|  70-79 | 124 (11%)  |
|  80+ | 20 (2%)  |
|  Unknown | 2 (<1%)  |
|  **Race** |   |
|  White | 844 (75%)  |
|  Black or African American | 204 (18%)  |
|  Asian | 21 (2%)  |
|  Native Hawaiian or Other Pacific Islander^{†} | 0 (0%)  |
|  American Indian or Alaska Native^{†} | 5 (<1%)  |
|  Other^{†} | 50 (4%)  |
|  Unknown^{†} | 9 (<1%)  |
|  **Cancer Grade** |   |

K253561 - Page 4 of 16

{4}

|  1 | 242 (21%)  |
| --- | --- |
|  2 | 606 (54%)  |
|  3 | 255 (22%)  |
|  Unknown | 30 (3%)  |
|  **Clinical Tumor Stage** |   |
|  T1 | 537 (47%)  |
|  T2 | 404 (36%)  |
|  T3 | 138 (12%)  |
|  T4 | 1 (<1%)  |
|  Unknown | 53 (5%)  |
|  **Clinical Nodal Stage** |   |
|  N0 | 752 (66%)  |
|  N1 | 277 (24%)  |
|  N2 | 0 (0%)  |
|  N3 | 0 (0%)  |
|  Unknown | 104 (9%)  |
|  **Breast Laterality** |   |
|  Left | 577 (50.9%)  |
|  Right | 556 (49.1%)  |
|  **Recurrence (Any Time)** |   |
|  True | 144 (12.7%)  |
|  False | 989 (87.3%)  |

### V Substantial Equivalence Information:

**A Predicate Device Name(s)**

ArteraAI Breast

**B Predicate Device Number(s)**

K254115

**C Comparison with Predicate(s):**

|  **Device & Predicate Device(s):** | K260023 | K254115  |
| --- | --- | --- |
|  Device Trade Name | TumorSight Risk | ArteraAI Breast  |
|  **General Device Characteristic Similarities**  |   |   |

K253561 - Page 5 of 16

{5}

|  Intended Use/Indications For Use | TumorSight Risk is an artificial intelligence based software only device that analyzes data from previously diagnosed invasive breast cancer patients to assess the risk of recurrence. TumorSight Risk utilizes the following data: age, race, cancer stage, nodal status, and grade, combined with Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) data to generate 5- and 10-year risks of breast cancer recurrence and a proprietary prognostic score. TumorSight Risk is intended for use in adult women with hormone receptor positive (HR+) and human epidermal growth factor receptor 2 negative (HER2-), lymph node negative (N0) or positive (N1), Stage I-IIIA breast cancer to inform a physician in prognostic risk-based decisions in conjunction with other relevant clinicopathological factors. | ArteraAI Breast is a software only device intended to analyze scanned histopathology whole slide images (WSIs) from treatment-naive breast resection specimens prepared from formalin fixed paraffin-embedded (FFPE) tissue and stained using Hematoxylin & Eosin (H&E) stains. Additional inputs to ArteraAI Breast include the following physician-provided clinical variables: • Age • Tumor size • Nodal status ArteraAI Breast provides 5- and 10-year risks of distant metastasis and ArteraAI risk score for adult patients with HR+/HER2-, N0 or N1, early-stage invasive breast cancer without clinically or pathologically defined metastases after surgical tumor resection and who are candidates for standard of care adjuvant therapy. ArteraAI Breast is intended to assist physicians with prognostic risk-based decisions along with other clinicopathological factors. ArteraAI Breast is intended to utilize WSIs acquired from FDA-cleared interoperable scanners and file formats that have been validated for use with this device.  |
| --- | --- | --- |
|  Physical Characteristics | Artificial intelligence/machine learning (AI/ML) based software only Device | Same  |
|  General Device Characteristic Differences  |   |   |
|  Data Source (Input) | DCE-MRI and Clinicopathological data | Scanned histopathology whole slide images (WSIs) from treatment-naive breast resection specimens prepared from formalin fixed paraffin embedded (FFPE) tissue and stained using Hematoxylin & Eosin (H&E) stains  |
|  Output | • Proprietary TSR Score ranging from 0-100 that is used to determine a risk category • Categorical assessment of "High Risk", "Intermediate Risk" or "Low Risk" • Observed 5-Year Risk of recurrence (in the clinical validation dataset) • Observed 10-Year Risk of recurrence (in the clinical validation dataset) | • An ArteraAI Risk Score • An ArteraAI categorical risk group (Low or High), based on the ArteraAI Risk Score • Observed 5-Year Distant Metastasis (DM) Risk (in the clinical validation dataset) • Observed 10-Year DM Risk (in the clinical validation dataset)  |

### VI Standards/Guidance Documents Referenced:

1. Guidance for Industry and Food and Drug Administration Staff: Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions (June 2025).
2. Guidance for Food and Drug Administration Staff: Content of Premarket Submissions for Device Software Functions (June 2023).

K253561 - Page 6 of 16

{6}

3. IEC 62304 Edition 1.1 2015-06 CONSOLIDATED VERSION, Medical Device Software – Software Life Cycle Processes (recognition Number 13-79).
4. ISO 14971 Third Edition 2019-12, Medical devices – Applications of risk management to
5. medical devices (recognition Number 5-125).
6. CLSI document EP05-A3: Evaluation of Precision of Quantitative Measurement Procedures; Approved Guideline—Third Edition; 2014
7. CLSI document EP12-Ed3: Evaluation of Qualitative, Binary Output Examination Performance – Third Edition; 2023
8. CLSI document EP49-Ed1: Framework for Developing Evidence of Clinical Validity of Medical Laboratory Test Methods – First Edition; 2026

## VII Performance Characteristics (if/when applicable):

### A Analytical Performance:

#### 1. Precision/Reproducibility from MRI

A study using simulated noise added to MRIs was conducted to assess the precision/reproducibility of the TumorSight Risk algorithm. One hundred twenty (120) patient images were selected which were evenly split across GE/1.5T, GE/3T, Siemens/1.5T, Siemens/3T MRI machines. These images were processed through the system three times with 3% added Rician noise. One hundred nineteen (119) samples were successfully processed and 0.83% (1/120] samples were not successfully processed. The reproducibility study included a minimum of 4 samples near the cutoff C1[20] between low and intermediate and 4 samples near the cutoff C2 [35] between intermediate and high (borderline cases). The sample characteristics are provided in Table 2 below.

Table 2: Characteristics of Samples Used in Reproducibility Study

|  Characteristic | GE/1.5T N=30 | GE/3T N=30 | Siemens/1.5T N=30 | Siemens/3T N=30  |
| --- | --- | --- | --- | --- |
|  # Successfully Processed | 30(100%) | 30(100%) | 30(100%) | 29(97%)  |
|  Age (mean ± std) | 57 ± 12 | 51 ± 13 | 53 ± 14 | 60 ± 12  |
|  Age < 50 | 11 (37%) | 17 (57%) | 10 (33%) | 8 (27%)  |
|  Age ≥ 50 | 19 (63%) | 13 (43%) | 20 (67%) | 22 (73%)  |
|  Race |  |  |  |   |
|  White | 25 (83%) | 16 (53%) | 12 (%) | 29 (%)  |
|  Black | 4 (13%) | 1 (3%) | 18 (%) | 0 (0%)  |
|  Other† | 1 (3%) | 13 (43%) | 0 (0%) | 1 (3%)  |
|  Cancer Grade |  |  |  |   |
|  1 | 7 (23%) | 6 (20%) | 6 (20%) | 6 (20%)  |
|  2 | 21 (70%) | 20 (67%) | 18 (60%) | 16 (53%)  |
|  3 | 2 (7%) | 4 (13%) | 6 (20%) | 8 (27%)  |
|  Clinical Tumor Stage |  |  |  |   |
|  T1 | 10 (33%) | 10 (33%) | 10 (33%) | 10 (33%)  |
|  T2 | 10 (33%) | 10 (33%) | 10 (33%) | 11 (36%)  |

K253561 - Page 7 of 16

{7}

|  Characteristic | GE/1.5T N=30 | GE/3T N=30 | Siemens/1.5T N=30 | Siemens/3T N=30  |
| --- | --- | --- | --- | --- |
|  T3 | 10 (33%) | 10 (33%) | 10 (33%) | 9 (30%)  |
|  **Clinical Nodal Stage** |  |  |  |   |
|  N0 | 15 (50%) | 15 (50%) | 15 (50%) | 25 (83%)  |
|  N1 | 15 (50%) | 15 (50%) | 15 (50%) | 5 (17%)  |
|  **Laterality** |  |  |  |   |
|  Left | 15 (50%) | 14 (47%) | 16 (53%) | 18 (60%)  |
|  Right | 15 (50%) | 16 (53%) | 14 (47%) | 12 (40%)  |
|  **TumorSight Risk Category** |  |  |  |   |
|  High Risk | 17 (57%) | 19 (63%) | 19 (63%) | 12 (41%)  |
|  Intermediate Risk | 4 (13%) | 7 (23%) | 5 (17%) | 6 (21%)  |
|  Low Risk | 9 (30%) | 4 (13%) | 6 (20%) | 11 (38%)  |
|  **Borderline Group** |  |  |  |   |
|  Borderline Case | 9 (30%) | 7 (23%) | 10 (33%) | 10 (34%)  |
|  Non-Borderline Case | 21 (70%) | 23 (77%) | 20 (67%) | 19 (66%)  |

Summary precision statistics for: i) the variability of SD standard deviation (SD) (mean values of SD, the median of SD); ii) the variability of percent coefficient of variation (%CV) ((mean values of %CV, the median of %CV)), and iii) the percent agreement are shown in Table 3 below. The average SD is less than 0.25 risk units, and the average %CV is less than 1.1% across all manufacturer/field strength combinations. Overall percent agreement was 99.4% (only 2 out of 119 patients exhibited less than perfect agreement).

Table 3: Summary Precision Results for the Processed Cases (n=119) for the TSR Score

|  Parameter | Overall N=119 | GE/1.5T N=30 | GE/3T N=30 | Siemens/1.5T N=30 | Siemens/3T N=29  |
| --- | --- | --- | --- | --- | --- |
|  SD (mean, median) | 0.23 (0.12) | 0.25 (0.09) | 0.21 (0.11) | 0.21 (0.17) | 0.25 (0.12)  |
|  %CV (mean, median) | 0.61 (0.31) | 0.48 (0.25) | 0.44 (0.26) | 0.51 (0.36) | 1.06 (0.33)  |
|  %Agreement (N *3 replicates) | 99.4% (355/357) | 98.9% (89/90) | 100% (90/90) | 98.9% (89/90) | 100% (87/87)  |
|  %Disagreement (N*3 replicates) | 0.6% (2/357) | 1.1% (1/90) | 0% (0/90) | 1.1% (1/90) | 0% (0/87)  |

Precision profiles, including average SD vs. mean score of TSR, and %CV vs. mean score of TSR for the overall dataset are shown in the following Figure 1.

K253561 - Page 8 of 16

{8}

![img-0.jpeg](img-0.jpeg)

Figure 1: Precision profiles for all N=119 subjects included in the study for the TSR Score.

### B Clinical Performance:

The clinical performance of TumorSight Risk device was evaluated in a retrospective clinical performance study which included a total of 2,129 patients across 4 sites in the US. Enrollment included identification of patients that were diagnosed between January 1, 2010, and December 31, 2024, and had a pre-treatment MRI available. The clinical, pathologic and recurrence data elements were extracted from the medical record and/or cancer registry. The cases were selected based on the inclusion and exclusion criteria described below.

#### Inclusion criteria:

- Subject has previously granted the institution consent for their data to be used for additional research purposes
- Adult ( \( \geq \) 22 years) female subjects with biopsy proven early-stage or locally advanced invasive breast cancer
• Any race/ethnicity
• Clinical node negative (cN0) and node positive (cN1) patients are eligible
• Hormone receptor positive (HR+)
• Human epidermal growth factor receptor 2 negative (HER2-)
• Pre-treatment DCE MRI available
• Patient received treatment and was followed for first recurrence
• GE or Siemens MRI manufacturer
• 1.5T or 3T field magnetic strength
- Imaging conforms to ACR guidelines for contrast enhanced breast imaging

#### Exclusion criteria:

• HR- or HER2+ disease
- No DCE MRI prior to initiation of therapy (in the neoadjuvant setting) or surgery (in the adjuvant setting)
• Distant metastases at the time of diagnosis
- Bilateral disease

K253561 - Page 9 of 16

{9}

• T4 disease
• cN2 or higher clinical nodal status
• Patients meeting the following criteria:

○ Breast implants
○ Patients with prior mastectomy
○ Unilateral DCE breast protocols
○ Post-treatment DCE breast MRI
○ Pediatric patients (age < 22 years old)

• Poor quality images (inadequate image quality, insufficient number of images)

• DCE MRI does not meet ACR guidelines for contrast enhanced breast imaging

The summary of patient characteristics for the clinical validation dataset is presented in Table 4 below.

Table 4: Summary of Clinical Validation Patient Cohort Characteristics

|  Variables | Validation N=2129 (%)  |
| --- | --- |
|  **Institution** |   |
|  Site A | 712 (33.4%)  |
|  Site B | 718 (33.7%)  |
|  Site C | 447 (21.0%)  |
|  Site D | 252 (11.8%)  |
|  **Age*** |   |
|  22-29 | 16 (<1%)  |
|  30-39 | 128 (6%)  |
|  40-49 | 505 (24%)  |
|  50-59 | 580 (27%)  |
|  60-69 | 567 (27%)  |
|  70-79 | 300 (14%)  |
|  80+ | 33 (2%)  |
|  **Race**** |   |
|  White | 1532 (72%)  |
|  Black or African American | 338 (16%)  |
|  Asian^{†} | 182 (9%)  |
|  Native Hawaiian or Other Pacific Islander^{†} | 65 (3%)  |
|  American Indian or Alaska Native^{†} | 6 (<1%)  |
|  Other^{†} | 6 (<1%)  |
|  Unknown | 0 (0%)  |
|  **Cancer Grade*** |   |
|  1 | 736 (35%)  |
|  2 | 1084 (50%)  |
|  3 | 309 (15%)  |
|  Unknown | 0 (0%)  |
|  **Clinical Tumor Stage***** |   |
|  T1 | 1299 (61%)  |
|  T2 | 681 (32%)  |
|  T3 | 149 (7%)  |

K253561 - Page 10 of 16

{10}

|  Variables | Validation N=2129 (%)  |
| --- | --- |
|  T4 | 0 (0%)  |
|  Unknown | 0 (0%)  |
|  **Clinical Nodal Stage***** |   |
|  N0 | 1811 (85%)  |
|  N1 | 318 (15%)  |
|  N2 | 0 (0%)  |
|  N3 | 0 (0%)  |
|  Unknown | 0 (0%)  |
|  **Recurrence (Any Time)** |   |
|  True | 181 (8.5%)  |
|  False | 1948 (91.5%)  |

Within the TumorSight Risk device and algorithm, races other than White and Black are categorized as “Other”. The measures of the clinical performance for the device with three outputs (High, Intermediate, Low) are i) Predictive Value (PV) for each output and ii) percentage of the results for each output. Predictive Values are evaluated using Kaplan-Meier analyses, from the “High”, “Intermediate” and “Low” populations, respectively.

### Summary of Clinical Validation Results:

#### 1. Measures of Clinical Performance

Clinical performance of the TumorSight Risk device is described by Table 5 below. Predictive Value (PV) for output “X” is the probability that a Target Condition (Recurrence) occurs within a specific time frame (5 years and 10 years) given the device output for the patients is “X” (this PV for output “X” is called also “risk of recurrence”)

Table 5: Recurrence Rate by TumorSight Risk Category for 5 years and 10 years

|  Risk Category by TumorSight Risk device | Total number | Percentage of patients with the Risk Category | Number of recurrences for 5 years | Risk of Recurrence within 5 years (95% CI) | Number of recurrences for 10 years | Risk of Recurrence within 10 years (95% CI)  |
| --- | --- | --- | --- | --- | --- | --- |
|  **High** | 534 | 25.1% | 46 | 9.8% (7.4%, 12.9%) | 70 | 26.4% (20.2%, 34.1%)  |
|  **Intermediate** | 559 | 26.3% | 28 | 5.7% (4.0%, 8.2%) | 43 | 13.4% (9.6%, 18.5%)  |
|  **Low** | 1036 | 48.7% | 27 | 3.0% (2.0%, 4.3%) | 52 | 8.6% (6.4%, 11.5%)  |
|  **Total** | 2129 | 100.0% | 101 | 5.4% (4.5%, 6.5%) | 165 | 13.5% (11.4%, 15.9%)  |

- The risks of recurrence for TumorSight Risk category “High” patients were 9.8% (95%CI: (7.4%, 12.9%)) for 5 years and 26.4% (95%CI: (20.2%, 34.1%)) for 10 years. The TumorSight Risk output “High” is statistically and clinically informative because i) the 5-year risk is higher than the pre-test risk of 5.4% (the lower bound of the 95%CI, 7.4%, is higher than the pre-test risk of 5.4% (7.4% > 5.4%), and ii) the 10-year risk is higher than

K253561 - Page 11 of 16

{11}

the pre-test risk of 13.5% (the lower bound of the 95%CI, 20.2%, is higher than the pre-test risk of 13.5% (20.2% > 13.5%).

- The risks of recurrence for TumorSight Risk category “Low” patients were 3.0% (95%CI: (2.0%, 4.3%)) for 5 years and 8.6% (95%CI: (6.4%, 11.5%)) for 10 years. The TumorSight Risk output “Low” is statistically and clinically informative because i) the 5-year risk is lower than the pre-test risk of 5.4% (the upper bound of the 95%CI, 4.3%, is lower than the pre-test risk of 5.4% (4.3% < 5.4%) and ii) the 10-year risk is lower than the pre-test risk of 13.5% (the upper bound of the 95%CI, 11.5%, is lower than the pre-test risk of 13.5% (11.5% < 13.5%).
- The risk of recurrence for TumorSight Risk category Intermediate is similar to the pre-test risk for both 5 years (5.7% vs 5.4%) and 10 years (13.4% vs 13.5%. Percents of patients with “intermediate” results in the clinical performance study was 26.3% (559/2129).

2. Figure 2 below shows the risk of recurrence curves of the “High”, “Intermediate” and “Low” categories patients along with all subjects in the study population through 10 years:

![img-1.jpeg](img-1.jpeg)

|  Low (n=1,035)  |   |   |   |   |   |   |   |   |   |   |   |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
|  At risk | 1035 | 993 | 949 | 888 | 812 | 735 | 601 | 487 | 374 | 299 | 218  |
|  Censored | 0 | 41 | 75 | 132 | 200 | 273 | 401 | 509 | 616 | 689 | 765  |
|  Events | 0 | 1 | 11 | 15 | 23 | 27 | 33 | 39 | 45 | 47 | 52  |
|  Int (n=559)  |   |   |   |   |   |   |   |   |   |   |   |
|  At risk | 559 | 540 | 514 | 482 | 433 | 374 | 294 | 239 | 179 | 142 | 96  |
|  Censored | 0 | 14 | 35 | 60 | 106 | 157 | 235 | 288 | 342 | 377 | 420  |
|  Events | 0 | 5 | 10 | 17 | 20 | 28 | 30 | 32 | 38 | 40 | 43  |
|  High (n=535)  |   |   |   |   |   |   |   |   |   |   |   |
|  At risk | 535 | 510 | 484 | 433 | 373 | 298 | 225 | 170 | 123 | 79 | 51  |
|  Censored | 0 | 16 | 31 | 70 | 125 | 191 | 259 | 310 | 354 | 390 | 414  |
|  Events | 0 | 9 | 20 | 32 | 37 | 46 | 51 | 55 | 58 | 66 | 70  |

Figure 2: Kaplan-Meier curves for 5- and 10-year risk of recurrence for TumorSight Risk categories (High, Intermediate (Int), and Low), and the whole study population (N=2,129). The vertical lines delineate the 5-year and 10-year analysis endpoint.

### 3. Subgroup Analyses

Subgroup analysis of recurrence rates observed in the clinical validation study at 5- and 10-years post diagnosis were performed based on race, age, nodal stage, treatment, MRI Manufacturer/ Magnetic Field Strength, as shown in tables 6 - 10.

#### a. Subgroup Analysis by Race

Clinical performance of the TumorSight Risk device analyzed by race: (White, African American and Other) is described below:

K253561 - Page 12 of 16

{12}

Table 6: Subgroup Analysis of Recurrence Risks at 5- and 10-years Post Diagnosis by Race

|  Race | Risk Category | Recurrence Rate at 5 years (95% CI) | Recurrence Rate at 10 years (95% CI) | N (% of subjects)  |
| --- | --- | --- | --- | --- |
|  White | High | 11.1% (7.6%, 16.0%) | 26.5% (18.5%, 37.1%) | 268 (17.5%)  |
|   |  Intermediate | 5.4% (3.4%, 8.4%) | 11.3% (7.4%, 17.1%) | 381 (24.9%)  |
|   |  Low | 3.1% (2.1%, 4.6%) | 9.3% (6.3%, 12.5%) | 883 (57.6%)  |
|   |  Overall | 5.0% (4.0%, 6.3%) | 12.2% (10.0%, 14.9%) | 1532 (100.0%)  |
|  African American | High | 4.8% (2.2%, 10.3%) | 13.8% (7.3%, 24.5%) | 134 (39.6%)  |
|   |  Intermediate | 7.8% (3.8%, 15.7%) | 22.9% (12.7%, 39.3%) | 102 (30.2%)  |
|   |  Low | 2.2% (0.6%, 8.6%) | 3.9% (1.2%, 12.0%) | 102 (30.2%)  |
|   |  Overall | 5.0% (3.0%, 8.2%) | 15.1% (9.6%, 23.3%) | 338 (100.0%)  |
|  Other | High | 12.8% (7.9%, 20.5%) | 42.0% (24.4%, 65.2%) | 132 (51.0%)  |
|   |  Intermediate | 4.3% (1.4%, 12.9%) | 6.8% (2.5%, 17.9%) | 76 (29.3%)  |
|   |  Low | 2.4% (0.4%, 16.1%) | 2.4% (0.3%, 16.1%) | 51 (19.7%)  |
|   |  Overall | 8.2% (5.3%, 12.7%) | 21.7% (13.6%, 33.6%) | 259 (100.0%)  |

# b. Subgroup Analysis by Age

Clinical performance of the TumorSight Risk device analyzed by age: (<50 years and ≥50) is described below:

Table 7: Subgroup Analysis of Recurrence Risks at 5- and 10-years Post Diagnosis by Age

|  Age | Risk Category | Recurrence Rate at 5 years (95% CI) | Recurrence Rate at 10 years (95% CI) | N (% of subjects)  |
| --- | --- | --- | --- | --- |
|  <50 | High | 9.7% (6.0%, 15.4%) | 24.5% (16.3%, 35.7%) | 179 (27.6%)  |
|   |  Intermediate | 6.6% (3.3%, 12.7%) | 21.6% (12.0%, 37.0%) | 146 (22.5%)  |
|   |  Low | 1.8% (0.7%, 4.2%) | 5.6% (2.7%, 11.1%) | 323 (49.8%)  |
|   |  Overall | 5.1% (3.6%, 7.3%) | 13.8% (10.1%, 18.6%) | 648 (100%)  |
|  ≥50 | High | 9.9% (7.0%, 13.9%) | 26.8% (19.0%, 37.0%) | 356 (24.0%)  |
|   |  Intermediate | 5.4% (3.5%, 8.3%) | 11.0% (7.4%, 16.3%) | 413 (27.9%)  |
|   |  Low | 3.5% (2.3%, 5.3%) | 9.9% (7.2%, 13.6%) | 712 (48.1%)  |
|   |  Overall | 5.5% (4.4%, 6.9%) | 13.4% (11.0%, 16.3%) | 1481 (100%)  |

# c. Subgroup Analysis by Nodal Stage

Clinical performance of the TumorSight Risk device analyzed by nodal stage: (cN0 and cN1) is described below:

Table 8: Subgroup Analysis of Recurrence Risks at 5- and 10-years Post Diagnosis by Nodal Stage

|  Nodal Stage | Risk Category | Recurrence Rate at 5 years (95% CI) | Recurrence Rate at 10 years (95% CI) | N (% of subjects)  |
| --- | --- | --- | --- | --- |
|  cN0 | High | 7.3% (4.6%, 11.3%) | 20.2% (12.4%, 31.8%) | 278 (15.4%)  |
|   |  Intermediate | 5.5% (3.7%, 8.2%) | 13.0% (9.0%, 18.4%) | 504 (27.8%)  |
|   |  Low | 3.0% (2.1%, 4.3%) | 8.7% (6.5%, 11.6%) | 1028 (56.8%)  |
|   |  Overall | 4.4% (3.4%, 5.5%) | 11.1% (9.1%, 13.6%) | 1810 (100%)  |
|  cN1 | High | 12.6% (8.8%, 17.8%) | 32.8% (23.9%, 44.0%) | 257 (80.6%)  |
|   |  Intermediate | 7.9% (3.0%, 19.8%) | 17.3% (7.4%, 37.5%) | 55 (17.2%)  |
|   |  Low | 0.0% (0.0%, 0.0%) | 0.0% (0.0%, 0.0%) | 7 (2.2%)  |
|   |  Overall | 11.4% (8.2%, 15.8%) | 28.6% (21.3%, 37.8%) | 319 (100%)  |

# d. Subgroup Analysis by Treatment

K253561 - Page 13 of 16

{13}

Clinical performance of the TumorSight Risk device analyzed by treatment: (chemo-endocrine therapy and endocrine therapy) is described below.

Table 9: Subgroup Analysis of Recurrence Rates Observed in the Clinical Validation Study at 5- and 10-years Post Diagnosis by Treatment

|  Treatment | Risk Category | Recurrence Rate at 5 years (95% CI) | Recurrence Rate at 10 years (95% CI) | N* (% of subjects)  |
| --- | --- | --- | --- | --- |
|  Chemo-endocrine therapy | High | 6.9% (3.8%, 12.6%) | 20.2% (12.3%, 32.1%) | 163 (37.8%)  |
|   |  Intermediate | 4.8% (2.0%, 11.2%) | 9.7% (4.8%, 19.0%) | 117 (27.1%)  |
|   |  Low | 3.6% (1.5%, 8.4%) | 4.6% (2.1%, 10.0%) | 150 (35.0%)  |
|   |  Overall | 5.2% (3.4%, 7.9%) | 11.2% (7.8%, 15.9%) | 430 (100.0%)  |
|  Endocrine therapy | High | 3.0% (1.1%, 7.8%) | 23.9% (13.5%, 40.3%) | 157 (14.6%)  |
|   |  Intermediate | 4.4% (2.5%, 7.9%) | 12.0% (7.3%, 19.4%) | 293 (27.3%)  |
|   |  Low | 1.8% (1.0%, 3.3%) | 9.0% (6.1%, 13.3%) | 622 (58.0%)  |
|   |  Overall | 2.7% (1.8%, 4.0%) | 11.4% (8.7%, 14.8%) | 1072 (100.0%)  |

*For 627 out of 2129 cases within the clinical validation study cohorts, treatment information was not available.

# e. Subgroup Analysis by MRI Manufacturer/Field Strength

Clinical performance of the TumorSight Risk device analyzed by Manufacturer/Field Strength: (GE/1.5T, GE/3T, Siemens/1.5T, and Siemens/3T) is described below.

Table 10: Subgroup Analysis of Recurrence Risks at 5- and 10-years Post Diagnosis by MRI Manufacturer/magnetic Field Strength

|  MRI Manufacturer/Magnetic Field Strength | Risk Category | Recurrence Rate at 5 years (95% CI) | Recurrence Rate at 10 years (95% CI) | N* (% of subjects)  |
| --- | --- | --- | --- | --- |
|  GE | High | 12.2% (9.0%, 16.4%) | 27.7% (21.1%, 35.8%) | 361 (24.1%)  |
|   |  Intermediate | 6.4% (4.3%, 9.5%) | 11.6% (7.9%, 16.9%) | 388 (25.9%)  |
|   |  Low | 2.9% (1.9%, 4.4%) | 9.2% (6.8%, 12.4%) | 748 (50.0%)  |
|   |  Overall | 6.0% (4.8%, 7.4%) | 13.5% (11.4%, 16.1%) | 1497 (100.0%)  |
|  GE/1.5T | High | 7.0% (3.9%, 12.3%) | 21.8% (14.8%, 35.1%) | 178 (16.8%)  |
|   |  Intermediate | 6.2% (3.8%, 9.9%) | 11.1% (7.3%, 16.8%) | 279 (26.4%)  |
|   |  Low | 3.0% (1.9%, 4.9%) | 9.6% (7.1%, 13.1%) | 601 (56.8%)  |
|   |  Overall | 4.5% (3.4%, 6.0%) | 11.9% (9.6%, 14.6%) | 1058 (100.0%)  |
|  GE/3T | High | 17.3% (12.2%, 24.2%) | 36.6% (21.1%, 58.3%) | 183 (41.7%)  |
|   |  Intermediate | 6.9% (3.3%, 14.0%) | 15.4% (5.1%, 41.4%) | 109 (24.8%)  |
|   |  Low | 2.3% (0.7%, 7.0%) | 3.6% (1.3%, 9.6%) | 147 (33.5%)  |
|   |  Overall | 9.6% | 19.0% | 439  |

K253561 - Page 14 of 16

{14}

|  MRI Manufacturer/ Magnetic Field Strength | Risk Category | Recurrence Rate at 5 years (95% CI) | Recurrence Rate at 10 years (95% CI) | N* (% of subjects)  |
| --- | --- | --- | --- | --- |
|   |  | (7.1%, 13.0%) | (12.2%, 28.9%) | (100.0%)  |
|  Siemens | High | 4.4% (2.1%, 9.1%) | 34.6% (12.6%, 73.9%) | 173 (27.4%)  |
|   |  Intermediate | 4.1% (1.7%, 10.0%) | 27.9% (14.0%, 50.8%) | 171 (27.1%)  |
|   |  Low | 3.2% (1.7%, 6.7%) | 3.2% (1.5%, 6.7%) | 287 (45.5%)  |
|   |  Overall | 3.8% (2.4%, 6.0%) | 17.1% (9.7%, 29.0%) | 631 (100.0%)  |
|  Siemens/1.5T | High | 5.0% (2.3%, 10.9%) | 22.0% (5.7%, 65.5%) | 126 (33.5%)  |
|   |  Intermediate | 4.5% (1.7%, 11.8%) | 35.5% (16.9%, 64.6%) | 106 (28.2%)  |
|   |  Low | 4.0% (1.7%, 11.8%) | 4.0% (1.7%, 9.4%) | 144 (38.3%)  |
|   |  Overall | 4.5% (2.7%, 7.3%) | 17.5% (9.3%, 31.4%) | 376 (100.0%)  |
|  Siemens/3T | High | 2.8% (0.4%, 18.1%) | 61.3% (19.1%, 98.6%) | 47 (18.4%)  |
|   |  Intermediate | 3.8% (0.6%, 24.3%) | 11.9% (2.8%, 42.6%) | 65 (25.5%)  |
|   |  Low | 1.6% (0.4%, 6.4%) | 1.6% (0.4%, 6.4%) | 143 (56.1%)  |
|   |  Overall | 2.5% (0.9%, 7.1%) | 16.6% (5.8%, 42.3%) | 255 (100.0%)  |

# f. Subgroup analysis by Bins

Using TSR Score numeric values, patients with output “Low” were divided into two bins using: [0, 11) and [11, 21) and patients with output High” were divided into 2 bins: [36, 58) and [58, 100]. The recurrence risks and 95% confidence intervals at 5- and 10-years were estimated from the clinical study data (presented in Table 11 below) and plotted against TSR Score (Presented in Figure 3).

Table 11: Observed 5- and 10-year recurrence rate based on binned TSR Score

|  Risk Category | Device output (TSRS Range) | Device Output (Mean TSRS) | Bin | N | TSRS Range | Observed Recurrence Rate at 5 years (95% CI) | Observed Recurrence Rate at 10 years (95% CI) | Percentage of patients in the bin  |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
|  High | [58, 100] | 78.6 | 5 | 218 | [58, 100] | 13.6% (9.3%, 19.5%) | 34.2% (24.4%, 46.4%) | 10.2%  |
|  High | [36, 58) | 44.2 | 4 | 316 | [36, 57] | 7.4% (4.9%, 11.1%) | 21.1% (13.7%, 31.7%) | 14.8%  |
|  Intermediate | [21, 36) | 27.3 | 3 | 559 | [21, 35] | 5.7% (4.0%, 8.2%) | 13.4% (9.6%, 18.5%) | 26.3%  |
|  Low | [11, 21) | 15.0 | 2 | 521 | [11, 20] | 3.7% (2.3%, 5.9%) | 11.1% (7.8%, 15.8%) | 24.5%  |

K253561 - Page 15 of 16

{15}

|  Risk Category | Device output (TSRS Range) | Device Output (Mean TSRS) | Bin | N | TSRS Range | Observed Recurrence Rate at 5 years (95% CI) | Observed Recurrence Rate at 10 years (95% CI) | Percentage of patients in the bin  |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
|  Low | [0, 11) | 5.2 | 1 | 515 | [0, 10] | 2.2% (1.2%, 4.1%) | 6.2% (3.7%, 10.3%) | 24.2%  |
|   | Total | 26.7 |  | 2,129 | [0, 100] | 5.4% (4.5%, 6.5%) | 13.5% (11.4%, 15.9%) | 100.0%  |

![img-2.jpeg](img-2.jpeg)

Figure 3: Observed 5- and 10-year risk of recurrence based on binned TSR scores

The empirically determined recurrence probability at 5-year and 10-years is plotted on the y-axis along with 95% CIs, versus TSR score on the x-axis. Hatched boxes and black error bars show 5-year recurrence probability. Open boxes and grey error bars show 10-year recurrence probability.

The analysis of TSR bins demonstrated that a monotonic increase is observed for 5-year and 10-year risks of recurrence.

### VIII Proposed Labeling:

The labeling supports the finding of substantial equivalence for this device.

### IX Conclusion:

The submitted information in this premarket notification is complete and supports a substantial equivalence decision.

K253561 - Page 16 of 16

---

**Source:** [https://fda-staging.innolitics.com/device/K260023](https://fda-staging.innolitics.com/device/K260023)

**Published by [Innolitics](https://innolitics.com)** — a medical-device software consultancy. We help companies design, build, and clear FDA-regulated software and AI/ML devices. If you're preparing [a 510(k)](https://innolitics.com/services/510ks/), [a De Novo](https://innolitics.com/services/regulatory/), [a SaMD](https://innolitics.com/services/end-to-end-samd/), [an AI/ML medical device](https://innolitics.com/services/medical-imaging-ai-development/), or [an FDA regulatory strategy](https://innolitics.com/services/regulatory/), [get in touch](https://innolitics.com/contact).

**Cite:** Innolitics at https://innolitics.com