AI-Rad Companion Organs RT
Device Facts
| Record ID | K252548 |
|---|---|
| Device Name | AI-Rad Companion Organs RT |
| Applicant | Siemens Healthcare GmbH |
| Product Code | QKB · Radiology |
| Decision Date | Apr 10, 2026 |
| Decision | SESE |
| Submission Type | Traditional |
| Regulation | 21 CFR 892.2050 |
| Device Class | Class 2 |
| Attributes | AI/ML, Software as a Medical Device, Real-World Evidence |
Real-World Evidence
| Submission | Device | Sponsor | RWD Sources | RWE Use Summary | Key Tags |
|---|---|---|---|---|---|
| K252548 · Apr 10, 2026 | AI-Rad Companion Organs RT | Siemens Healthcare GmbH | Retrospective clinical CT and MR image datasets from multiple clinical sites (North America, South America, Asia, Australia, Europe) | Retrospective performance studies were conducted using clinical image data to validate the accuracy of deep-learning-based auto-contouring algorithms for organs at risk and brain metastases. | Retrospective performance study; Clinical image datasets; Auto-contouring validation; Deep learning |
Clinical Evidence
| Study Design | Population | Comparator | Key Endpoints |
|---|---|---|---|
| Retrospective CT performance study; Retrospective performance study | Patients undergoing RT treatment planning; Sample Size: 469; Number of Sites: Multiple clinical sites (North America, South America, Asia, Australia, Europe) | Predicate device (K242745) and reference literature | Dice coefficient, absolute symmetric surface distance (ASSD), fail rate |
| Retrospective MR Pelvis performance study; Retrospective performance study | Patients undergoing RT treatment planning; Sample Size: 153; Number of Sites: Multiple clinical sites (North America, Australia, Europe) | Predicate device (K232899) | Dice coefficient, absolute symmetric surface distance (ASSD) |
| Retrospective MR Brain OAR performance study; Retrospective performance study | Patients undergoing RT treatment planning; Sample Size: 46; Number of Sites: Multiple clinical sites (North America) | Reference literature and cleared devices | Dice coefficient, absolute symmetric surface distance (ASSD) |
| Retrospective MR Brain Metastases performance study; Retrospective performance study | Patients undergoing RT treatment planning; Sample Size: 30; Number of Sites: Multiple clinical sites (North America) | VBrain (K203235) | Lesion-wise dice coefficient, lesion-wise sensitivity, FPR, HD95 |
AI Performance
| Output | Algorithm | Acceptance | Observed | Dev DS | Dev Readers | Test DS | Test Readers |
|---|---|---|---|---|---|---|---|
| CT Organ Segmentation | Deep learning algorithm | Dice score difference < 3% for existing organs; higher value than baseline for new organs | Average Dice score difference < 3% for existing organs; LAD ASSD 4.17 mm, Heart Dice 0.93 | Training: 434 datasets (Thorax) | — | Retrospective performance study: 469 cases (CT) | >1 (radiologists or radiation oncologists) |
| MR Pelvis Organ Segmentation | Deep learning algorithm | Subject device metric higher than baseline value | Anal Canal Dice 0.76, Bladder Dice 0.91, Rectum Dice 0.85, Prostate Dice 0.85 | Training: 219 datasets (T1W VIBE/Dixon), 275 datasets (T2W TSE) | — | Retrospective performance study: 153 cases (MR Pelvis) | >1 (radiologists or radiation oncologists) |
| MR Brain Organ Segmentation | Deep learning algorithm | Subject device metric higher than baseline value | Brainstem Dice 0.90, Optic Chiasm Dice 0.55, Hippocampus Left Dice 0.66, Hippocampus Right Dice 0.71 | Training: 278 datasets (T1W MPRAGE) | — | Retrospective performance study: 46 cases (MR Brain) | >1 (radiologists or radiation oncologists) |
| MR Brain Metastases Segmentation | Deep learning algorithm | Dice ≥ 0.70, Lesion-wise Sensitivity ≥ 0.85, FPR ≤ 5 per MRI, Lesion-wise HD95 ≤ 2.94 mm | Dice 0.72, Lesion-wise Sensitivity 0.86, FPR 1.75, Lesion-wise HD95 1.6 mm | Training: 1931 datasets (T1W MPRAGE) | — | Retrospective performance study: 60 cases (MR Brain Metastases) | >1 (radiologists or radiation oncologists) |
Indications for Use
AI-Rad Companion Organs RT is a post-processing software intended to automatically contour DICOM CT and MR pre-defined structures, including known (diagnosed) brain metastases, using deep-learning-based algorithms. Contours that are generated by AI-Rad Companion Organs RT may be used as input for clinical workflows including radiation therapy treatment planning. AI-Rad Companion Organs RT must be used in conjunction with appropriate software such as Treatment Planning Systems and Interactive Contouring applications, to review, edit, and accept contours generated by AI-Rad Companion Organs RT. The outputs of AI-Rad Companion Organs RT are intended to be used by qualified and trained medical professionals. The software is not intended to be used for diagnostic purposes.
Device Story
Post-processing software for automatic segmentation of anatomical structures and brain metastases from DICOM CT and MR images; intended for radiation therapy treatment planning. Input: CT or MR image series. Operation: Deep-learning algorithms generate contours as DICOM-RT Structure Sets. Workflow: Software processes images; results reviewed in a web-based UI; user confirms or declines contours. Confirmed contours exported to treatment planning systems (TPS) or interactive contouring applications for clinical review and editing. Benefits: Automates time-consuming manual contouring, supporting efficient radiation therapy planning. Used in clinical settings by qualified medical professionals. Deployment: Edge and cloud-based.
Clinical Evidence
No clinical trials conducted. Performance validated via retrospective bench studies using clinical datasets (N=469 for CT; N=153 for MR Pelvis; N=46 for MR Brain OAR; N=60 for MR Brain Metastases). Metrics included Dice coefficient, ASSD, lesion-wise sensitivity, and false positive rates. Results demonstrated performance equivalent or superior to the predicate device and reference literature.
Technological Characteristics
Software-only device; DICOM-compliant. Deep-learning-based segmentation algorithms. Deployment: Edge and cloud-hosted. Web-based UI. Standards: IEC 62366-1 (usability), ISO 14971 (risk management), IEC 62304 (software lifecycle), DICOM PS 3.1-3.20, IEC 82304-1 (health software safety), IEC 81001-5-1 (security).
Indications for Use
Indicated for adult patients previously selected for radiation therapy requiring automatic contouring of DICOM CT and MR structures, including brain metastases, for treatment planning. Not for diagnostic use.
Regulatory Classification
Identification
A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.
Special Controls
*Classification.* Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).
Predicate Devices
- AI-Rad Companion Organs RT (K242745)
Reference Devices
Submission Summary (Full Text)
{0}
FDA U.S. FOOD & DRUG ADMINISTRATION
April 10, 2026
Siemens Healthcare GmbH
Kira Morales
Regulatory Affairs Managers
Henkestrasse 127
Erlangen, 91052
Germany
Re: K252548
Trade/Device Name: AI-Rad Companion Organs RT
Regulation Number: 21 CFR 892.2050
Regulation Name: Medical Image Management And Processing System
Regulatory Class: Class II
Product Code: QKB
Dated: March 9, 2026
Received: March 9, 2026
Dear Kira Morales:
We have reviewed your section 510(k) premarket notification of intent to market the device referenced above and have determined the device is substantially equivalent (for the indications for use stated in the enclosure) to legally marketed predicate devices marketed in interstate commerce prior to May 28, 1976, the enactment date of the Medical Device Amendments, or to devices that have been reclassified in accordance with the provisions of the Federal Food, Drug, and Cosmetic Act (the Act) that do not require approval of a premarket approval application (PMA). You may, therefore, market the device, subject to the general controls provisions of the Act. Although this letter refers to your product as a device, please be aware that some cleared products may instead be combination products. The 510(k) Premarket Notification Database available at https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm identifies combination product submissions. The general controls provisions of the Act include requirements for annual registration, listing of devices, good manufacturing practice, labeling, and prohibitions against misbranding and adulteration. Please note: CDRH does not evaluate information related to contract liability warranties. We remind you, however, that device labeling must be truthful and not misleading.
If your device is classified (see above) into either class II (Special Controls) or class III (PMA), it may be subject to additional controls. Existing major regulations affecting your device can be found in the Code of
U.S. Food & Drug Administration
10903 New Hampshire Avenue
Silver Spring, MD 20993
www.fda.gov
{1}
K252548 - Kira Morales
Page 2
Federal Regulations, Title 21, Parts 800 to 898. In addition, FDA may publish further announcements concerning your device in the Federal Register.
Additional information about changes that may require a new premarket notification are provided in the FDA guidance documents entitled "Deciding When to Submit a 510(k) for a Change to an Existing Device" (https://www.fda.gov/media/99812/download) and "Deciding When to Submit a 510(k) for a Software Change to an Existing Device" (https://www.fda.gov/media/99785/download).
Your device is also subject to, among other requirements, the Quality Management System Regulation (QMSR) (21 CFR Part 820), which includes, but is not limited to, ISO 13485 clause 7.3 (Design controls), ISO 13485 clause 8.3 (Nonconforming product), ISO 13485 clause 8.5.2 (Corrective action), and ISO 13485 clause 8.5.3 (Preventative action). Please note that regardless of whether a change requires premarket review, the QMSR requires device manufacturers to review and approve changes to device design and production (ISO 13485 clause 7.3 and ISO 13485 clause 7.5) and document changes and approvals in the Medical Device File (ISO 13485 clause 4.2.3).
Please be advised that FDA's issuance of a substantial equivalence determination does not mean that FDA has made a determination that your device complies with other requirements of the Act or any Federal statutes and regulations administered by other Federal agencies. You must comply with all the Act's requirements, including, but not limited to: registration and listing (21 CFR Part 807); labeling (21 CFR Part 801); medical device reporting (reporting of medical device-related adverse events) (21 CFR Part 803) for devices or postmarketing safety reporting (21 CFR Part 4, Subpart B) for combination products (see https://www.fda.gov/combination-products/guidance-regulatory-information/postmarketing-safety-reporting-combination-products); good manufacturing practice requirements as set forth in the Quality Management System Regulation (QMSR) (21 CFR Part 820) for devices or current good manufacturing practices (21 CFR Part 4, Subpart A) for combination products; and, if applicable, the electronic product radiation control provisions (Sections 531-542 of the Act); 21 CFR Parts 1000-1050.
All medical devices, including Class I and unclassified devices and combination product device constituent parts are required to be in compliance with the final Unique Device Identification System rule ("UDI Rule"). The UDI Rule requires, among other things, that a device bear a unique device identifier (UDI) on its label and package (21 CFR 801.20(a)) unless an exception or alternative applies (21 CFR 801.20(b)) and that the dates on the device label be formatted in accordance with 21 CFR 801.18. The UDI Rule (21 CFR 830.300(a) and 830.320(b)) also requires that certain information be submitted to the Global Unique Device Identification Database (GUDID) (21 CFR Part 830 Subpart E). For additional information on these requirements, please see the UDI System webpage at https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance/unique-device-identification-system-udi-system.
Also, please note the regulation entitled, "Misbranding by reference to premarket notification" (21 CFR 807.97). For questions regarding the reporting of adverse events under the MDR regulation (21 CFR Part 803), please go to https://www.fda.gov/medical-devices/medical-device-safety/medical-device-reporting-mdr-how-report-medical-device-problems.
For comprehensive regulatory information about medical devices and radiation-emitting products, including information about labeling regulations, please see Device Advice (https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance) and CDRH Learn
{2}
K252548 - Kira Morales
Page 3
(https://www.fda.gov/training-and-continuing-education/cdrh-learn). Additionally, you may contact the Division of Industry and Consumer Education (DICE) to ask a question about a specific regulatory topic. See the DICE website (https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance/contact-us-division-industry-and-consumer-education-dice) for more information or contact DICE by email (DICE@fda.hhs.gov) or phone (1-800-638-2041 or 301-796-7100).
Sincerely,
Daniel M. Krainak, Ph.D.
Assistant Director
DHT8C: Division of Radiological
Imaging and Radiation Therapy Devices
OHT8: Office of Radiological Health
Office of Product Evaluation and Quality
Center for Devices and Radiological Health
Enclosure
{3}
| Indications for Use | | |
| --- | --- | --- |
| Please type in the marketing application/submission number, if it is known. This
textbox will be left blank for original applications/submissions. | K252548 | ? |
| Please provide the device trade name(s). | | ? |
| AI-Rad Companion Organs RT | | |
| Please provide your Indications for Use below. | | ? |
| AI-Rad Companion Organs RT is a post-processing software intended to automatically contour
DICOM CT and MR pre-defined structures, including known (diagnosed) brain metastases, using deep-
learning-based algorithms.
Contours that are generated by AI-Rad Companion Organs RT may be used as input for clinical
workflows including radiation therapy treatment planning. AI-Rad Companion Organs RT must
be used in conjunction with appropriate software such as Treatment Planning Systems and
Interactive Contouring applications, to review, edit, and accept contours generated by AI-Rad
Companion Organs RT.
The outputs of AI-Rad Companion Organs RT are intended to be used by qualified and trained medical
professionals.
The software is not intended to be used for diagnostic purposes. | | |
| Please select the types of uses (select one or both, as
applicable). | ☑ Prescription Use (Part 21 CFR 801 Subpart D)
☐ Over-The-Counter Use (21 CFR 801 Subpart C) | ? |
{4}
SIEMENS Healthineers
# 510(k) SUMMARY FOR AI-Rad Companion Organs RT K252548
Submitted by:
Siemens Medical Solutions USA, Inc.
40 Liberty Boulevard
Malvern, PA 19355
Date Prepared: March 6, 2026
This summary of 510(k) safety and effectiveness information is being submitted in accordance with the requirements of Safe Medical Devices Act of 1990 and 21 CFR §807.92.
# 1. Submitter
**Importer/Distributor**
Siemens Medical Solutions USA, Inc.
40 Liberty Boulevard
Malvern, PA 19355
Registration Number: 2240869
**Manufacturing Site**
Siemens Healthcare GmbH
Henkestrasse 127
Erlangen, Germany 91052
Registration Number: 3002808157
# 2. Contact Person
Kira Morales
Senior Regulatory Affairs Specialist
Siemens Medical Solutions USA, Inc.
40 Liberty Boulevard
Malvern, PA 19335
Phone: +1 (484) 901 - 9471
Email: kira.morales@siemens-healthineers.com
# 3. Device Name and Classification
**Product Name:** AI-Rad Companion Organs RT
**Common Name:** Medical Imaging Software
**Classification Name:** Medical Image Management and Processing System
AI-Rad Companion Organs RT
{5}
SIEMENS Healthineers
Classification Panel: Radiology
CFR Section: 21 CFR §892.2050
Device Class: Class II
Product Code: QKB
# 4. Predicate Device
Product Name: AI-Rad Companion Organs RT
Common Name: Medical Imaging Software
510(k) Number: K242745
Clearance Date: March 27, 2025
Classification Name: Medical Image Management and Processing System
Classification Panel: Radiology
CFR Section: 21 CFR §892.2050
Device Class: Class II
Primary Product Code: QKB
Recall Information: N/A
# 5. Reference Devices
Product Name: AutoContour Model RADAC V3
Common Name: Medical Image Software
510(k) Number: K230685
Clearance Date: March 14, 2025
Classification Name: Medical Image Management and Processing System
Classification Panel: Radiology
CFR Section: 21 CFR §892.2050
Device Class: Class II
Primary Product Code: QKB
Product Name: AI-Rad Companion Organs RT
Common Name: Medical Imaging Software
510(k) Number: K232899
Clearance Date: April 3, 2024
Classification Name: Medical image management and processing system
Classification Panel: Radiology
CFR Section: 21 CFR §892.2050
Device Class: Class II
Primary Product Code: QKB
Product Name: VBrain
Common Name: Medical Image Segmentation Software
510(k) Number: K203235
Clearance Date: March 19, 2021
Classification Name: Medical image management and processing system
AI-Rad Companion Organs RT
{6}
SIEMENS Healthineers
Classification Panel: Radiology
CFR Section: 21 CFR §892.2050
Device Class: Class II
Primary Product Code: QKB
# 6. Indications for Use
AI-Rad Companion Organs RT is a post-processing software intended to automatically contour DICOM CT and MR pre-defined structures, including known (diagnosed) brain metastases, using deep-learning-based algorithms.
Contours that are generated by AI-Rad Companion Organs RT may be used as input for clinical workflows including radiation therapy treatment planning. AI-Rad Companion Organs RT must be used in conjunction with appropriate software such as Treatment Planning Systems and Interactive Contouring applications, to review, edit, and accept contours generated by AI-Rad Companion Organs RT.
The outputs of AI-Rad Companion Organs RT are intended to be used by qualified and trained medical professionals.
The software is not intended to be used for diagnostic purposes.
# 7. Device Description
AI-Rad Companion Organs RT provides automatic segmentation of pre-defined structures from DICOM CT or MR medical series, prior to dosimetry planning in radiation therapy. AI-Rad Companion Organs RT includes deep-learning algorithms to provide the contouring of organs and structures in the head & neck, thorax, abdomen, and pelvis regions on CT Images, and the contouring of organs and structures in the brain and male pelvis on MR images.
CT or MR series of images serve as input for AI-Rad Companion Organs RT and are acquired as part of a typical scanner acquisition. Once processed by the AI algorithms, contours are created as DICOM Radiotherapy Structure Set (RTSS) objects and can be reviewed in a Result Preview UI, to either confirm or decline the results.
If the results are confirmed, the results are sent to the configured target node, typically a treatment planning system (TPS) or interactive contouring applications. If the results are declined, all contours of this case will be deleted, and the case will be completed.
Optionally, the user may select to directly transfer the contours to a configurable DICOM node, without the need of manual confirmation.
AI-Rad Companion Organs RT must be used in conjunction with appropriate software such as Treatment Planning Systems and Interactive Contouring applications, to review, edit, and accept the automatically generated contours. Then the output of AI-Rad Companion Organs RT must be
AI-Rad Companion Organs RT
{7}
SIEMENS Healthineers
reviewed and, where necessary, edited with appropriate software before accepting generated contours as input to treatment planning steps.
# 8. Substantially Equivalent (SE) and Technological Characteristics
The intended use of the subject device is unchanged from the predicate device. The following modifications have been made in the subject device, compared to the predicate, AI-Rad Companion Organs RT (K242745).
- Modified Indications for Use
- Modified Intended Patient Population
- New MR-Brain Metastases contouring algorithm
- New MR Brain OAR contouring algorithm
- Modified CT contouring algorithm
- Modified MR Pelvis contouring algorithm
- Enhanced User Interface
- Multi-arc Support
- Updated Subject device Claims list
AI-Rad Companion Organs RT VA70 and AI-Rad Companion Organs RT VA60 both use a deep learning algorithm to support their AI claims. Additionally, they both process CT and MR data in DICOM format, and create outputs which can be used by any TPS system. The deep learning CT algorithm and MR Pelvis algorithms within AI-Rad Companion Organs RT VA70 has been enhanced from the algorithms in AI-Rad Companion Organs RT VA60 (K242745). Two new algorithms have been added for MR-Brain OAR contouring and MR-Brain Metastases contouring. All models contained within AI-Rad Companion Organs RT VA70 and AI-Rad Companion Organs RT VA60 (K242745) are locked and cannot be modified by the user.
The subject device, AI-Rad Companion Organs RT, is substantially equivalent with regards to the software features, functionalities, and core algorithms. The performance of the enhanced CT contouring algorithm has been validated against FDA/CE cleared devices or from literature.
The risk analysis and non-clinical data support that both devices perform equivalently and do not raise different questions of the safety and effectiveness.
AI-Rad Companion Organs RT
{8}
SIEMENS
Healthineers
| | Subject Device | Predicate Device |
| --- | --- | --- |
| Device Manufacturer | Siemens | Siemens |
| Device Name | AI-Rad Companion Organs RT | AI-Rad Companion Organs RT |
| 510(k) Number | K252548 | K242745 |
| Indications for Use | AI-Rad Companion Organs RT is a post-processing software intended to automatically contour DICOM CT and MR pre-defined structures, including known (diagnosed) brain metastases, using deep-learning-based algorithms.
Contours that are generated by AI-Rad Companion Organs RT may be used as input for clinical workflows including radiation therapy treatment planning. AI-Rad Companion Organs RT must be used in conjunction with appropriate software such as Treatment Planning Systems and Interactive Contouring applications, to review, edit, and accept contours generated by AI-Rad Companion Organs RT.
The outputs of AI-Rad Companion Organs RT are intended to be used by qualified and trained medical professionals. The software is not intended to be used for diagnostic purposes. | AI-Rad Companion Organs RT is a post-processing software intended to automatically contour DICOM CT and MR pre-defined structures using deep-learning-based algorithms.
Contours that are generated by AI-Rad Companion Organs RT may be used as input for clinical workflows including external beam radiation therapy treatment planning. AI-Rad Companion Organs RT must be used in conjunction with appropriate software such as Treatment Planning Systems and Interactive Contouring applications, to review, edit, and accept contours generated by AI-Rad Companion Organs RT.
The outputs of AI-Rad Companion Organs RT are intended to be used by trained medical professionals.
The software is not intended to automatically detect or contour lesions |
| Algorithm | Deep Learning | Deep Learning |
| Segmentation of Organ at Risk in the Anatomic Regions | CT: Head & Neck, Thorax, Abdomen & Pelvis
Head & Neck lymph nodes
MR: Pelvis, Brain & Metastases | CT: Head & Neck, Thorax, Abdomen & Pelvis
Head & Neck lymph nodes
MR: Pelvis |
AI-Rad Companion Organs RT
{9}
SIEMENS
Healthineers
| Compatible Modality | CT & MR Images | CT & MR Images |
| --- | --- | --- |
| Compatible Scanner Models | No Limitation on scanner model for CT. DICOM compliance required. | No Limitation on scanner model for CT. Siemens Healthineers’ data only for MR. DICOM compliance required. |
| Compatible Treatment Planning System | No Limitation on TPS model, DICOM compliance required. | No Limitation on TPS model, DICOM compliance required. |
| Target Population | The intended patient population is not subject to any restrictions. However, the algorithms are only validated for adult populations. In principle, AI-Rad Companion Organs RT is designed for any patient for whom relevant modality scans are available. | AI-Rad Companion Organs RT is designed for use only in adult populations. AI-Rad Companion Organs RT is designed for any patient for whom relevant modality scans are available. |
| Clinical condition the device is intended to diagnose, treat or manage | Limited to patients previously selected for Radiation Therapy. | Limited to patients previously selected for Radiation Therapy. |
| Software Architecture | AI-Rad Companion (Engine) architecture enabling the deployment of AI Rad Companion Organs RT using Edge and in the Cloud. The UI is provided using a web-based interface. | AI-Rad Companion (Engine) architecture enabling the deployment of AI Rad Companion Organs RT using Edge and in the Cloud. The UI is provided using a web-based interface. |
| Deployment Feature | Edge & Cloud Deployment | Edge & Cloud Deployment |
| Organ Templates | Creating, editing and deletion of organ templates. Customize predefined structure database with mapping to international nomenclature schemes. | Creating, editing and deletion of organ templates. Customize predefined structure database with mapping to international nomenclature schemes. |
| Automated workflow | AI-Rad Companion Organs RT automatically processes input image data and sends the results as DICOM-RT Structure Sets to a user-configurable target node. | AI-Rad Companion Organs RT automatically processes input image data and sends the results as DICOM-RT Structure Sets to a user-configurable target node. |
AI-Rad Companion Organs RT
{10}
SIEMENS
Healthineers
| Contour visualization and editing feature | AI-Rad Companion Organs RT provides basic result preview of automatic segmentation results, and no editing feature of the automatic segmented contour. | AI-Rad Companion Organs RT provides basic result preview of automatic segmentation results, and no editing feature of the automatic segmented contour. |
| --- | --- | --- |
| CT Contouring Performance | The target performance was validated using 517 cases distributed to four cohorts. To objectively evaluate the target performance, the DICE coefficient, the absolute symmetric surface distance (ASSD) and the fail rate was evaluated. The segmentation performance of the subject and reference devices & literature were equivalent as well as the overall performance compared to the predicate device. | CT: The target performance was validated using 579 cases distributed to four cohorts. To objectively evaluate the target performance, the DICE coefficient, the absolute symmetric surface distance (ASSD) and the fail rate was evaluated. The segmentation performance of the subject and reference devices & literature were equivalent as well as the overall performance compared to the predicate device |
| MR Pelvis Contouring Performance | The target performance was validated using 153 cases distributed into T2W TSE and T1W DixonW. Different metrics, including Dice coefficient and the absolute symmetric surface distance (ASSD), were determined to quantify the similarity between the automatically contoured OAR and the manually delineated contours (ground truth). | The target performance was validated using 66 cases distributed into T2W TSE and T1W DixonW. Different metrics, including Dice coefficient and the absolute symmetric surface distance (ASSD), were determined to quantify the similarity between the automatically contoured OAR and the manually delineated contours (ground truth). We also introduce the failure rate in this section. |
| MR Brain OAR Contouring Performance | The target performance was validated using 46 cases in one cohort from T1W MPRAGE. Different metrics including DICE coefficient and the absolute symmetric surface distance (ASSD), were determined to quantify the similarity between the automatically contoured OAR and the manually delineated contours (ground truth). | Not applicable – comparison to reference devices |
| MR Brain Metastases | The target performance was validated using 60 cases in one | Not applicable – comparison to reference devices |
AI-Rad Companion Organs RT
{11}
SIEMENS Healthineers
The conclusions from all verification and validation data suggests that these enhancements are equivalent with respect to safety and effectiveness of the predicate device. These modifications do not change the intended use of the product. Siemens is of opinion that AI-Rad Companion Organs RT VA70 is substantially equivalent to the currently marketed device, AI-Rad Companion Organs RT (K242745).
# 9. Nonclinical Tests
Non-clinical tests were conducted to test the functionality of AI-Rad Companion Organs RT. Software validation and bench testing have been conducted to assess the performance claims as well as the claim of substantial equivalence to the predicate device.
AI-Rad Companion has been tested to meet the requirements of conformity to multiple industry standards. Non-clinical performance testing demonstrates that AI-Rad Companion Organs RT complies with the FDA guidance document, "Guidance for the Content of Premarket Submissions for Device Software Functions" (June 2023) as well as with the following voluntary FDA recognized Consensus Standards listed in Table 2.
Table 1: Indications for Use and Segmentation Feature Comparison
| Recognition Number | Product Area | Title of Standard | Reference Number and Date | Standards Development Organization |
| --- | --- | --- | --- | --- |
| 5-129 | General | Medical Devices – Application of usability engineering to medical devices | 62366-1 Ed 1.1 2020-06 CV | IEC |
| 5-125 | General | Medical Devices – application of risk | 14971:2019-12 | ISO |
AI-Rad Companion Organs RT
{12}
SIEMENS Healthineers
| | | management to medical devices | | |
| --- | --- | --- | --- | --- |
| 13-79 | Software/ Informatics | Medical device software – software life cycle processes [Including Amendment 1 (2016)] | 62304 Ed 1.1
2015-06 CV | AAMI ANSI IEC |
| 12-352 | Radiology | Digital Imaging and Communications in Medicine (DICOM) Set | PS 3.1 – 3.20
2023e | NEMA |
| 5-134 | General | Medical devices – symbols to be used with information to be supplied by the manufacturer – Part 1: General Requirements | 15223-1 Fourth edition
2021-07 | ISO IEC |
| 13-97 | Software/ Informatics | Health software – Part 1: General requirements for product safety | 82304-1 Edition 1.0
2016-10 | IEC |
| 13-122 | Software/ Informatics | Health software and Health IT system safety effectiveness and security | 81001-5-1 Edition 1.0
2021-12 | IEC |
| 5-135 | General | Medical devices – Information to be supplied by the manufacturer | 20417 First edition 2021-04 | ISO |
Table 2: List of recognized standards
## Verification and Validation
Software documentation level, per FDA’s Guidance Document “Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices” issued on June 14, 2023, is also included as part of this submission. The performance data demonstrates continued conformance with special controls for medical devices containing software. Non-clinical tests were conducted on the subject device during product development.
Software bench testing in the form of Unit, System and Integration tests were performed to evaluate the performance and functionality of the new features and software updates. All testable requirements in the Requirement Specifications and the Risk Analysis have been successfully verified and traced in accordance with the Siemens Healthineers DH product development process. Human factor usability validation is addressed in system testing and usability validation test records. Software verification and regression testing have been performed successfully to meet their previously determined acceptance criteria as stated in the test plans.
Siemens Healthineers adheres to the cybersecurity recommendations as defined the FDA Guidance “Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions” (September 2023) by implementing a process of preventing unauthorized access, modifications, misuse or denial of use, or the unauthorized use of information that is stored, accessed, or transferred from a medical device to an external recipient.
AI-Rad Companion Organs RT
{13}
SIEMENS Healthineers
# 10. Performance Software Validation
To validate the AI-Rad Companion Organs RT software from clinical perspective, the auto-contouring algorithms underwent a scientific evaluation. The results of clinical data-based software validation for the subject device AI-Rad Companion Organs RT (SW VA70A) demonstrated equivalent performance in comparison to the predicate device (SW VA60A, K242745).
# CT Contouring Algorithm Performance
The performance of the AI-Rad Companion Organs RT CT contouring algorithm has been validated in a retrospective performance study on CT data previously acquired for RT treatment planning (N= 469, data from multiple clinical sites across the North American, South American, Asia, Australia and Europe). Ground truth annotations were established following RTOG and clinical guidelines using manual annotation. The mean and standard deviation Dice coefficients, along with the lower 95th percentile confidence bound, were calculated for each organ in the subject device. The results of subject device were equivalent or had better performance than the predicate device. To encountered for different datasets, variation in annotation, we first calculate the average of multiple references or the average of anatomical region for the specific organ or anatomical region. We then define the baseline value by subtracting the reference value using $5\%$ error margin in case of Dice and $0.1 \mathrm{~mm}$ in case of ASSD.
The performance results of the subject device for the new CT organs are comparable to the reference literature & cleared devices. Here equivalence for the new organs is defined such that the selected reference metric has a higher value than the defined baseline. For existing organs, the average (AVG) Dice score difference between the subject device and predicate device is smaller than $3\%$ .
| Validation Testing Subject | Acceptance Criteria |
| --- | --- |
| Existing Organs | ·The subject device in the selected reference metric has a higher value than the defined baseline value |
| New Organs | ·DICE and ASSD reference & baseline value comparison |
Table 3: Acceptance Criteria of AIRC Organs RT VA70
The performance of the existing organs segmented using AI (excluding LAD and Heart) were tested and all organs passed the acceptance criteria.
A qualitative improvement was done to the Left Anterior Descending Artery (LAD) and Heart segmentation models by adding additional landmarks. The detailed performance evaluation of the updated LAD and Heart organs is below:
AI-Rad Companion Organs RT
{14}
SIEMENS Healthineers
| Structure Name | No. | Metric | AIRC Organs RT VA60x (predicate device) | AIRC Organs RT VA70x (Subject device) | Pass |
| --- | --- | --- | --- | --- | --- |
| LAD | 57 | ASSD (mm) | 4.12 | 4.17 | Yes |
| | | MSD (mm) | 2.8 | 2.9 | Yes |
| Heart | 57 | Dice | 0.92 | 0.93 | Yes |
Table 4: Detailed Performance evaluation of LAD & Heart in the subject device
| Organ Name | NO. | Dice | | | | ASSD (mm) | | | |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| | | Avg. | Std. | Med. | 95%CI | Avg. | Std. | Med. | 95%CI |
| LCX | 48 | 0.28 | 0.13 | 0.29 | [0.24, 0.32] | 3.8 | 1.5 | 3.8 | [3.4, 4.3] |
| RCA | 45 | 0.27 | 0.13 | 0.26 | [0.23, 0.31] | 4.7 | 2.4 | 4.1 | [4.0, 5.4] |
Table 5: Detailed Performance evaluation of new organs in the subject device
| Testing Data | |
| --- | --- |
| Anatomical Region | Cardiac |
| Annotated Organs | LCX, RCA |
| # Datasets | 48 |
| Data Origin | USA: Chicago (21), TCIA (LIDC-IDRI) TCIA (LCTSC) Australia: Brisbane (21) |
| Manufacturer | Siemens Healthineers (10) GE MEDICAL SYSTEMS (5) Philips (20) Others/unknown (13) |
| Gender | Male (14) Female (27) unknown (7) |
| Slice Thickness | ≤ 3 mm |
Table 6: Validation Testing Data Information for new OARs
| Training Data | |
| --- | --- |
| Anatomical Region | Thorax |
| Annotated Structure | LCX, RCA |
| # Datasets | 434 |
| Data Origin | USA: Medical University of South Carolina (90), |
AI-Rad Companion Organs RT
{15}
SIEMENS Healthineers
| | Princeton Radiology (254) |
| --- | --- |
| | USA and EU: Maastricht University Medical Center & Harvard Medical School (90) |
| Manufacturer | Siemens Healthineers (417) unknown/other (17) |
| Slice Thickness | 1.0 - 3.6 mm |
| Pixel Spacing | 0.6 - 0.9 mm |
Table 7: Training Dataset Characteristics
# MR Pelvis Contouring Algorithm Performance
The performance of the AI-Rad Companion Organs RT MR Pelvis contouring algorithm has been validated in a retrospective performance study on MR data previously acquired for RT treatment planning (N= 153, data from multiple clinical sites across the North American, Australia and Europe). Compared to the previous version of the MR Pelvis algorithm (K232899), landmark detection and retraining with new dataset has been implemented for T2W model and testing datasets have been enlarged for T1 and T2 models. Ground truth annotations were established following RTOG and clinical guidelines using manual annotation. The dice coefficient and the absolute symmetric surface distance (ASSD) were determined to quantify the similarity between the automatically contoured OAR and the manually delineated contours (ground truth). The results of subject device were equivalent or had better performance than the predicate device.
The acceptance criteria is defined as the subject device in the selected reference metric has a higher value than the defined baseline value.
| Organ Name | Ref. Standard | Comment | Avg. Ref. Value (Dice) | Baseline Value (Dice) | Avg. Ref. Value ± Std. (ASSD) | Baseline Value (ASSD) |
| --- | --- | --- | --- | --- | --- | --- |
| Anal Canal | AI-Rad Companion Organs RT VA50 (K232899) | Avg. Dice reported in FDA report summary | 0.76 | 0.66 | 2.09 ± 0.9 | 2.99 |
| Bladder | | | 0.91 | 0.81 | 1.44 ± 0.89 | 2.33 |
| Rectum | | | 0.85 | 0.75 | 2.34 ± 2.17 | 4.51 |
| Penile Bulb | | | 0.82 | 0.72 | 0.79 ± 0.75 | 1.54 |
| Seminal Vesicle | | | 0.66 | 0.56 | 2.43 ± 2.61 | 5.04 |
| Prostate | | | 0.85 | 0.75 | 1.56 ± 0.54 | 2.1 |
| Left Femur Head | | | 0.94 | 0.84 | 0.85 ± 0.59 | 1.44 |
| Right Femur Head | | | 0.94 | 0.84 | 0.88 ± 0.54 | 1.42 |
| Body | | | 0.98 | 0.88 | 2.07 ± 1.81 | 3.88 |
AI-Rad Companion Organs RT
{16}
SIEMENS Healthineers
Table 8: Detailed Performance evaluation of new organs in the subject device
| Testing Data | | |
| --- | --- | --- |
| Anatomical Region | Male Pelvis Organs at Risk | |
| Sequence | T1W Dixon | T2W TSE |
| Annotated Structures | Body, Femoral Head Right, Femoral Head Left | Anal Canal, Prostate, Rectum, Penile Bulb, Seminal Vesicle, Bladder |
| # Datasets | 55 | 98 |
| Data Origin | USA: Wisconsin (15) EU: Germany site 1 (6), Germany site 2 (9), Eastern EU site1 & site 2 (25)1 | USA: Wisconsin (16), Princeton Radiology (9), Site 3 (18)1, Site 4 (20)1 EU: France (4), Romania (1), Germany (3), Switzerland (1), Spain (1), Eastern EU site 1 & site 2 (24)1 Australia: Site 1 (1) |
| Manufacturer | Siemens Healthineers (30) Philips1 (25) | Siemens Healthineers (36) Philips1 (44) GE1 (18) |
| Field Strength | 1.5T (19) 3.0T (36) | 1.5T (32) 3.0T (66) |
| Slice Thickness | (1, 2] mm (30) (2, 3] mm (25) | ≤ 1 mm (1) (2, 3] mm (72) (3, 3.5] mm (25) |
| Pixel Spacing | < 2 mm | < 1.25 mm |
| Age | 45 years and older (34) unknown (21) | 23 years and older (73) unknown (25) |
Table 9: Validation Testing Data for Improved MR Pelvis Algorithm
| Training Data | |
| --- | --- |
| Anatomical Region | Male Pelvis Organs at Risk |
Al-Rad Companion Organs RT
{17}
SIEMENS Healthineers
| Sequence | T1W VIBE/Dixon | T2W TSE |
| --- | --- | --- |
| Annotated Structures | Body, Femoral Head Right, Femoral Head Left | Anal Canal, Rectum, Penile Bulb, Seminal Vesicle, Bladder |
| # Datasets | 219 | 275 |
| Data Origin | USA: Wisconsin (59) EU: Austria (160) | USA: Wisconsin (60), Princeton Radiology (165), OHSU (50)² |
| Manufacturer | Siemens Healthineers (219) | Siemens Healthineers (225) Philips² (50) |
| Field Strength | 1.5T (59) 3.0T (160) | 1.5T (89) 3.0T (186) |
| Slice Thickness | < 2 mm | < 4 mm |
| Pixel Spacing | < 2 mm | < 1.25 mm |
| Age³ | 30 years and older | 43 years and older |
Table 10: Validation Testing Data for Improved MR Pelvis Algorithm
# MR Brain OAR Contouring Algorithm Performance
The performance of the AI-Rad Companion Organs RT MR Brain OAR contouring algorithm has been validated in a retrospective performance study on MR data previously acquired for RT treatment planning (N= 46, data from multiple clinical sites across North American). The newly added algorithm computes segmentation masks of predefined anatomic structures for a given T1w MPRAGE, post-contrast (T1wPost) MR Image. The segmentation is performed on the region of interest (ROI) of the organs instead of the entire image volume. Ground truth annotations were established following RTOG and clinical guidelines using manual annotation. The dice coefficient and the absolute symmetric surface distance (ASSD) were determined to quantify the similarity between the automatically contoured OAR and the manually delineated contours (ground truth).
The acceptance criteria is defined as the subject device in the selected reference metric has a higher value than the defined baseline value defined in the reference devices or literature.
| Organ Name | Reference Standard (Dice) | Comment | Ref. Value (Dice) | Baseline Value (Dice) | Reference Standard (ASSD) | Comment | Avg. Ref. Value ± Std. (ASSD) | Baseline Value (ASSD) |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Brainstem | | Avg. Dice based on auto | 0.90 | 0.80 | | Avg. MD value based on auto | 0.9±0.16 | 1.06 |
| Optic Chiasm | | | 0.55 | 0.45 | | | 0.69±0.80 | 1.49 |
2 Newly added data in the training cohort.
3 Due to data anonymization, the age information was not available for all data and therefore the provided age information is only based on available entries.
AI-Rad Companion Organs RT
{18}
SIEMENS
Healthineers
| Optic Nerve Left | Turcas, Andrada et al. 2023 | contouring solution MVision GBS™ device, Version 1.2.2 | 0.49 | 0.39 | Turcas, Andrada et al. 2023 | contouring solution MVision GBS™ device, Version 1.2.2 | 1.47±2.17 | 3.64 |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Optic Nerve Right | | | 0.56 | 0.46 | | | 0.95±0.65 | 1.6 |
| Lacrimal Gland Left | | | 0.46 | 0.36 | | | 1.39±0.97 | 2.36 |
| Lacrimal Gland Right | | | 0.55 | 0.45 | | | 1.31±0.95 | 2.26 |
| Pituitary Gland | | | 0.61 | 0.51 | | | 0.94±0.26 | 1.2 |
| Hippocampus Left | | | 0.66 | 0.56 | | | 0.75±0.75 | 1.5 |
| Hippocampus Right | | | 0.71 | 0.61 | | | 0.67±0.47 | 1.14 |
| Cornea Left | Radformation Inc. (K230685) | Passing threshold reported in FDA summary report | 0.50 | 0.40 | Al-Rad Companion VAS0 (K232899) | Avg. ASSD reported in FDA summary report | 0.60±0.24 | 0.84 |
| Cornea Right | | | 0.50 | 0.40 | | | 0.62±0.17 | 0.79 |
| Retina Left | | | 0.50 | 0.40 | | | 0.60±0.24 | 0.84 |
| Retina Right | | | 0.50 | 0.40 | | | 0.62±0.17 | 0.79 |
| Cochlea Left | | | 0.50 | 0.40 | n.a. | n.a. | n.a. | n.a. |
| Cochlea Right | | | 0.50 | 0.40 | n.a. | n.a. | n.a. | n.a. |
| Spinal Cord | Al-Rad Companion VAS0 (K232899) | Avg. Dice reported in FDA summary report | 0.68 | 0.58 | Al-Rad Companion VAS0 (K232899) | Avg. ASSD reported in FDA summary report | 1.70±0.72 | 2.42 |
| Eye Globe Left | | | 0.90 | 0.80 | | | 0.60±0.24 | 0.84 |
| Eye Globe Right | | | 0.89 | 0.79 | | | 0.62±0.17 | 0.79 |
| Lens Left | | | 0.68 | 0.58 | | | 0.61±0.56 | 1.17 |
| Lens Right | | | 0.67 | 0.57 | | | 0.60±0.25 | 0.85 |
Table 11: Detailed Performance evaluation of new organs in the subject device
| Testing Data | |
| --- | --- |
| Anatomical Region | Brain/Head |
| Sequence | T1W MPRAGE |
| Annotated Structures | Brainstem, Cochlea Left/ Right Cornea Left/ Right Eye Left/ Right Hippocampus Left/Right Optic Chiasm Lens Left/Right Lacrimal Gland Left/Right Optic Nerve Left/ Right Retina Left/ Right |
AI-Rad Companion Organs RT
{19}
SIEMENS Healthineers
Table 12: Validation Testing Data for MR Brain OAR contouring algorithm
| Training Data | |
| --- | --- |
| Anatomical Region | Brain/Head |
| Sequence | T1W MPRAGE |
| Annotated Structures | Brainstem,
Cochlea Left/ Right
Cornea Left/ Right
Eye Left/ Right
Hippocampus Left/Right
Optic Chiasm
Lens Left/Right
Lacrimal Gland Left/Right
Optic Nerve Left/ Right
Retina Left/ Right
Pituitary
Spinal Cord |
| # Datasets | 278 |
| Data Origin | USA:
TCIA - GammaKnife-Hippocampal (110),
TCIA – GLIS-RT (135),
Ohio (25)
Asia:
India (8) |
| Manufacturer | Siemens Healthineers (238) |
AI-Rad Companion Organs RT
{20}
SIEMENS Healthineers
| | GE (32) unknown (8) |
| --- | --- |
| Field Strength | 1.5T (184) 3.0T (86) unknown (8) |
| Slice Thickness | 0.9 – 4 mm |
| Pixel Spacing | 0.4 – 1.2 mm |
| Age | 19 years and older |
# MR Brain Metastases Contouring Algorithm Performance
The performance of the AI-Rad Companion Organs RT MR Brain OAR contouring algorithm has been validated in a retrospective performance study on MR data previously acquired for RT treatment planning (N= 30, data from multiple clinical sites across North American). The newly added algorithm computes segmentation masks of parenchymal brain metastases for a given T1w MPRAGE near-1mm-isotropic, post-contrast (T1wPost) MR Image. The segmentation is performed on the region of interest (ROI) of the organs instead of the entire image volume. Ground truth annotations were established following RTOG and clinical guidelines using manual annotation. Lesion-wise dice coefficient and lesion-wise sensitivity were used as acceptance criteria for the performance of the algorithm. The baseline values are defined by subtracting the reference value using $5\%$ error margin. The performance was also assessed on lesion-wise $95^{\text{th}}$ percentile Hausdorff distance, false positive rate, false positive per metastasis and average sensitivity. The performance of the algorithm was compared to the reference device, VBrain (K203235).
Table 13: Validation Testing Data for Improved MR Brain OAR Algorithm
| Organ Name | Acceptance Criteria | Subject Device 95%CI Lower | Pass |
| --- | --- | --- | --- |
| Dice | ≥70 | 0.72 | Yes |
| Lesion-wise Sensitivity | ≥85 | 0.86 | Yes |
| FPR | ≤5 false positive lesions per MRI | 1.75 | Yes |
| Lesion-wise HD95 | ≤2.94 mm | 1.6 mm | Yes |
Table 14: Baseline to Subject device value comparison
| Organ Name | Additional Metrics | Subject Device Value |
| --- | --- | --- |
| Brain Metastases (parenchymal) | FPR (FP/case) | 1.93 |
| | FPPM (FP/metastasis) | 0.74 |
| | Avg. Sensitivity (%) | 91 |
| | Avg. Lesion-wise HD95 (mm) | 1.67 |
Table 15: Performance evaluation of MR Brain metastases on additional metrics
AI-Rad Companion Organs RT
{21}
SIEMENS
Healthineers
| Testing Data | |
| --- | --- |
| Anatomical Region | Brain/Head |
| Sequence | MR T1WMPRAGE |
| Annotated Structures | Intraparenchymal Metastases |
| # Datasets | 60 |
| # Metastases | 266 |
| Data Origin | USA:TCIA - GammaKnife-Hippocampal (11)EU:Switzerland (30)unknown (19) |
| Manufacturer | Siemens Healthineers (33)GE (27) |
| Gender | Male (24)Female (17)unknown (19) |
| Field Strength | 1.5T (48)3.0T (12) |
| Slice Thickness | ≤1 mm (41)(1, 2] mm (18)(2-2.2) mm (1) |
| Pixel Spacing | ≤1 mm (41)(1, 2] mm (18)(2, 2.2] mm (1) |
| Age | 17 years and older |
| Metastasis size | (0.01-0.25] cm3 (183)(0.25-1.00] cm3 (41)>1 cm3 (42) |
| Contrast | With Contrasted (53)No Contrast (7) |
| # Metastases per case | [1]: 23[2 - 5]: 25[6 - 9]: 8[10-20]: 3[63]: 1 |
| Edema Status | Edema: 29No Edema: 176Unknown: 61 |
| Enhancement Pattern: | Ring-Enhancing: 56Non-Ring-Enhancing: 208Unknown: 2 |
| Morphology | Primary Cystic: 24Primary Solid: 201Mixed Cystic-Solid: 41 |
| Nodularity | Nodular: 261 |
AI-Rad Companion Organs RT
{22}
SIEMENS Healthineers
| | Non-Nodular:1
Unknown: 4 |
| --- | --- |
| Hemorrhagic | Likely Hemorrhagic: 25
Unlikely Hemorrhagic: 202
Unknown: 39 |
Table 16: Validation Testing Data for MR Brain Metastases contouring algorithm
| Training Data | |
| --- | --- |
| Anatomical Region | Brain/Head |
| Sequence | T1W MPRAGE |
| Annotated Structures | Intraparenchymal Metastases |
| # Datasets | 1931 |
| Data Origin | USA:
Michigan (765),
NewYork (538),
North Carolina (350),
NewYork2 (178),
NewJersey (32)
Asia:
Taiwan (10),
India (59) |
| Manufacturer | Siemens Healthineers (1656)
GE (138)
Philips (100)
unknown (31) |
| Field Strength | 1.5T (858)
3.0T (294)
unknown 1.5T or 3.0T (779) |
| Slice Thickness | 0.4 – 6 mm |
| Pixel Spacing | 0.2 – 1.5 mm |
| Age | 18 years and older |
| Metastasis size | 0.01-2 cm³ (8765)
> 2 cm³ (1084) |
Table 17: Validation Testing Data for Improved MR Brain Metastases Algorithm
# Standard Annotation Process
In both the annotation process for the training and validation testing data, the annotation protocols for the OAR were defined following the applicable guidelines. The ground truth annotations were drawn manually by a team of experienced annotators mentored by radiologists or radiation oncologists using an internal annotation tool. Additionally, a quality assessment
AI-Rad Companion Organs RT
{23}
SIEMENS Healthineers
including review and correction of each annotation was done by a board-certified radiation oncologist using validated medical image annotation tools.
## Validation Testing & Training Data Independence
The training data used for the training of the algorithm is independent of the data used to test the algorithm.
## 11. Clinical Tests
No clinical tests were conducted to test the performance and functionality of the modifications introduced within AI-Rad Companion Organs RT. Verification and validation of the enhancements and improvements have been performed and these modifications have been validated for their intended use. The data from these activities were used to support the subject device and the substantial equivalence argument. No animal testing has been performed on the subject device.
## 12. Safety and Effectiveness
The device labeling contains instructions for use and any necessary cautions and warnings to ensure safe and effective use of the device as compared to the predicate.
Risk management is ensured via ISO 14971:2019 compliance to identify and provide mitigation of potential hazards in a risk analysis early in the design phase and continuously throughout the development of the product. These risks are controlled via measures realized during software development, testing and product labeling.
## 13. Conclusion
Based on the discussion and validation testing and performance data above, the proposed device is determined to be as safe and effective as its predicate device, AI-Rad Companion Organs RT VA60 (K242745).
AI-Rad Companion Organs RT
