Predicate Devices

Reference Devices

Not Found

AI/ML (Artificial Intelligence / Machine Learning)

Yes
The intended use and device description explicitly state that the device uses artificial intelligence to automatically segment and quantify cardiac measurements.

Therapeutic

No.
The "Intended Use / Indications for Use" states that the software is "designed to report cardiac function measurements" and that "Its results are not intended to be used on a stand-alone basis for clinical decision-making." This indicates the device provides diagnostic information but does not claim to directly treat or modify a body function or structure.

Diagnostic

No
The software provides cardiac function measurements, but explicitly states "Its results are not intended to be used on a stand-alone basis for clinical decision-making." This indicates it is not a diagnostic device on its own.

SaMD (Software as a Medical Device)

Yes

The device is described as a cloud-hosted service implemented as a plug-in to a third-party DICOM viewer, which serves as the user interface and interface to PACS/scanner. It processes images and provides measurements, with no mention of accompanying hardware.

IVD (In Vitro Diagnostic)

Based on the provided information, AI4CMR software is not an In Vitro Diagnostic (IVD) device.

Here's why:

IVD Definition: In Vitro Diagnostics are devices intended for use in the examination of specimens derived from the human body in order to provide information for diagnostic, monitoring, or compatibility purposes. This typically involves analyzing biological samples like blood, urine, or tissue.
AI4CMR's Function: AI4CMR processes medical images (cardiac MR images) to provide measurements of cardiac function. It does not analyze biological specimens.
Intended Use Statement: The intended use explicitly states that AI4CMR reports cardiac function measurements from MR scanners and its results are not intended to be used on a stand-alone basis for clinical decision-making. This further reinforces that it's a tool for image analysis and quantification, not a diagnostic test performed on a biological sample.

While AI4CMR uses artificial intelligence and provides quantitative measurements, its application is in the realm of medical image analysis, which falls under the category of medical devices, but not specifically In Vitro Diagnostics.

PCCP (Predetermined Change Control Plan) Authorized

No
The letter does not state that the FDA has reviewed and approved or cleared a Predetermined Change Control Plan (PCCP) for this specific device.

Overview

Intended Use / Indications for Use

AI4CMR software is designed to report cardiac function measurements (ventricle volumes, ejection fraction, indices etc.) from 1.5T and 3T magnetic resonance (MR) scanners. AI4CMR uses artificial intelligence to automatically segment and quantify the different cardiac measurements. Its results are not intended to be used on a stand-alone basis for clinical decision-making.

The user incorporating AI4CMR into their DICOM application of choice is responsible for implementing a user interface.

Product codes

QIH

Device Description

AI4CMR v1.0 is a cloud-hosted service used with any third-party DICOM viewer application where the DICOM viewer serves as the user interface and the interface to a PACS or scanner for AI4CMR. AI4CMR is implemented as a plug-in to the DICOM viewer by the user and automatically processes and analyses cardiac MR images received by the DICOM viewer to quantify relevant cardiac function metrics and makes the information available to the user at the user's discretion.

The following are the cardiac function metrics quantified and reported by the software:

Anatomy and tissue segmentation
LV/RV stroke volume
LV/RV cardiac output
LV/RV ejection fraction
LV/RV end-diastolic volume
LV/RV end-systolic volume

The subject device enables the following metrics to be reported as desired by the user:

Metric	Unit	Accuracy
LV/RV stroke volume	ml	n/a1
LV/RV cardiac output	L/min	n/a1
LV/RV ejection fraction (EF)	%	LV bias (std): 3.87 (5.74) LV ICC: 0.96 RV bias (std): 7.80 (7.32) RV ICC: 0.81
LV/RV end-diastolic volume (EDV)	ml	LV bias (std): 8.66 (17.28) LV ICC: 0.99 RV bias (std): 8.36 (15.59) RV ICC: 0.96
LV/RV end-systolic volume (ESV)	ml	LV bias (std): -2.90 (17.52) LV ICC: 0.99 RV bias (std): -9.08 (13.83) RV ICC: 0.95
LV myocardial mass	g	bias (std): 2.45 (18.04) ICC: 0.96
LV/RV end-systolic volume index2	ml/m^2	n/a1
LV/RV end-diastolic volume index2	ml/m^2	n/a1
LV/RV stroke volume index2	ml/m^2	n/a1
Myocardium mass index2	g/m^2	n/a1
Cardiac index 2	L/(min m^2)	n/a 1

Notes:
1 These values are derived by performing simple mathematical operations and are derived from EDV, ESV, EF, and Mass metrics.
2 These values are only provided if the patient's height and weight are included in the DICOM data.

Mentions image processing

Yes

Mentions AI, DNN, or ML

AI4CMR uses artificial intelligence to automatically segment and quantify the different cardiac measurements.
AI4CMR uses artificial intelligence to automatically segment and quantify the different cardiac measurements.

Input Imaging Modality

Magnetic Resonance (MR) scanners (1.5T and 3T)

Anatomical Site

Cardiac / Heart (ventricles, myocardium)

Indicated Patient Age Range

Not Found

Intended User / Care Setting

Not Found

Description of the training set, sample size, data source, and annotation protocol

The AI4CMR training was performed on a dataset of 824 anonvmized cases collected retrospectively from Hospital de Braga, Portugal. Acquisition occurred between 2015 to January 2019 and consisted of male (63%) and female (37%) patients ranging in age from 13 to 89 (mean of 58) years old from Siemens acquisition system. This dataset is independent from the clinical validation set. This dataset was split into the 3 sets (training, validation, test). The splitting ratio is 70% for the "training set", 15% for the "validation set" and 15% for the "test set", resulting in 577, 121 and 126 cases each, respectively.

Description of the test set, sample size, data source, and annotation protocol

The AI4CMR training was performed on a dataset of 824 anonvmized cases collected retrospectively from Hospital de Braga, Portugal. This dataset was split into the 3 sets (training, validation, test). The splitting ratio is 70% for the "training set", 15% for the "validation set" and 15% for the "test set", resulting in 577, 121 and 126 cases each, respectively.

Summary of Performance Studies (study type, sample size, AUC, MRMC, standalone performance, key results)

Bench Testing

Study Type: Standalone performance test
Sample Size: 15 CMR cases
Data Source: Society of Cardiac Magnetic Resonance (SCMR) Consensus Contour Data, annotated by seven (7) independent expert readers.
Key Results: Agreement was evaluated and achieved between AI4CMR and the SCMR Consensus data.
For myocardium segmentation, an average dice similarity coefficient (DSC) of 0.72 per image was achieved.
For LVM, EDV, ESV, and EF the intraclass correlation coefficient (ICC) was evaluated and the following performance was obtained:

LV parameter	LOA (± 2 SD)	Bias ± SD	r 2	ICC
EDV	[-58.5912, 45.9097] ml	-6.3407 ± 26.1252 ml	0.85	0.95
ESV	[-30.2173, 22.4768] ml	-3.8703 ± 13.1735 ml	0.97	0.99
EF	[-7.2668, 6.9567] %	-0.155 ± 3.5559 %	0.95	0.99
LVM	[-57.2331, 11.7108] g	-22.7611 ± 17.236 g	0.72	0.78

Clinical Performance Assessment

Study Type: Multi-reader multi-center (MRMC) retrospective study
Sample Size: 146 CMR cases
Patient Characteristics: Patients ranging from 17 to 85 years old (average of 51), predominantly male (77%), with diseased (~60%) and non-diseased (~40%) conditions.
Data Source: CMR cases acquired and balanced across Siemens, GE, and Philips 1.5T scanners.
Annotation Protocol: Two expert readers manually segmented the myocardium for each CMR case per standard of care and manually determined volumes, LV mass, and Ejection Fraction. The dataset was independent of the data used for model training and development.
Primary Objective: To evaluate agreement between the AI4MED device and 2 expert readers who achieved excellent interrater variability (ICC > 0.75).
Key Results (Agreement with Readers' Consensus):

Left Ventricular EDV

	Cronbach's Alpha	Correlation Coef. (ρ)	Bias	ICC	95% Confidence Interval (Lower Bound)	95% Confidence Interval (Upper Bound)
AI4CMR vs consensus	0.992	0.980	8.663	0.990	0.98	0.99

Left Ventricular ESV

	Cronbach's Alpha	Correlation Coef. (ρ)	Bias	ICC	95% Confidence Interval (Lower Bound)	95% Confidence Interval (Upper Bound)
AI4CMR vs consensus	0.992	0.975	-2.893	0.991	0.99	0.99

Left Ventricular Ejection Fraction

	Cronbach's Alpha	Correlation Coef. (ρ)	Bias	ICC	95% Confidence Interval (Lower Bound)	95% Confidence Interval (Upper Bound)
AI4CMR vs consensus	0.969	0.909	3.867	0.956	0.87	0.98

Left Ventricular Myocardial Mass

	Cronbach's Alpha	Correlation Coef. (ρ)	Bias	ICC	95% Confidence Interval (Lower Bound)	95% Confidence Interval (Upper Bound)
AI4CMR vs consensus	0.956	0.936	2.452	0.955	0.94	0.97

Right Ventricular EDV

	Cronbach's Alpha	Correlation Coef. (ρ)	Bias	ICC	95% Confidence Interval (Lower Bound)	95% Confidence Interval (Upper Bound)
AI4CMR vs consensus	0.972	0.924	8.355	0.964	0.92	0.98

Right Ventricular ESV

	Cronbach's Alpha	Correlation Coef. (p)	Bias	ICC	95% Confidence Interval (Lower Bound)	95% Confidence Interval (Upper Bound)
AI4CMR vs consensus	0.967	0.888	-9.083	0.953	0.87	0.98

Right Ventricular Ejection Fraction

	Cronbach's Alpha	Correlation Coef. (ρ)	Bias	ICC	95% Confidence Interval (Lower Bound)	95% Confidence Interval (Upper Bound)
AI4CMR vs consensus	0.902	0.712	7.802	0.814	0.15	0.93

Key Metrics (Sensitivity, Specificity, PPV, NPV, etc.)

Bench Testing

Dice Similarity Coefficient (DSC): 0.72 per image for myocardium segmentation.
Intraclass Correlation Coefficient (ICC):
- LV EDV: 0.95
- LV ESV: 0.99
- LV EF: 0.99
- LV LVM: 0.78
Bias and Standard Deviation (SD) for LV parameters:
- LV EDV: -6.3407 ± 26.1252 ml
- LV ESV: -3.8703 ± 13.1735 ml
- LV EF: -0.155 ± 3.5559 %
- LV LVM: -22.7611 ± 17.236 g

Clinical Performance Assessment (Agreement with Readers' Consensus)

Cronbach's Alpha, Correlation Coefficient (ρ), Bias, and ICC for:
- Left Ventricular EDV: Cronbach's Alpha: 0.992, Correlation Coef. (ρ): 0.980, Bias: 8.663, ICC: 0.990
- Left Ventricular ESV: Cronbach's Alpha: 0.992, Correlation Coef. (ρ): 0.975, Bias: -2.893, ICC: 0.991
- Left Ventricular Ejection Fraction: Cronbach's Alpha: 0.969, Correlation Coef. (ρ): 0.909, Bias: 3.867, ICC: 0.956
- Left Ventricular Myocardial Mass: Cronbach's Alpha: 0.956, Correlation Coef. (ρ): 0.936, Bias: 2.452, ICC: 0.955
- Right Ventricular EDV: Cronbach's Alpha: 0.972, Correlation Coef. (ρ): 0.924, Bias: 8.355, ICC: 0.964
- Right Ventricular ESV: Cronbach's Alpha: 0.967, Correlation Coef. (p): 0.888, Bias: -9.083, ICC: 0.953
- Right Ventricular Ejection Fraction: Cronbach's Alpha: 0.902, Correlation Coef. (ρ): 0.712, Bias: 7.802, ICC: 0.814

Predicate Device(s)

K203256

Reference Device(s)

Not Found

Predetermined Change Control Plan (PCCP) - All Relevant Information

Not Found

Regulation Number and Section

§ 892.2050 Medical image management and processing system.

(a)
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.(b)
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).

Summary

0

Image /page/0/Picture/0 description: The image shows the logo for the U.S. Food and Drug Administration (FDA). The logo consists of two parts: on the left, there is the Department of Health & Human Services logo, which features an abstract image of a human figure. To the right of this is the FDA logo, which has the letters "FDA" in a blue square, followed by the words "U.S. FOOD & DRUG ADMINISTRATION" in blue text.

AI4MedImaging Medical Solutions S.A. % Carla Almeida Regulatory Affairs and Quality Manager Rua do Parque Poente, Lote 35 Braga, Minho 4705-002 PORTUGAL

July 22, 2022

Re: K220624

Trade/Device Name: AI4CMR v1.0 Regulation Number: 21 CFR 892.2050 Regulation Name: Medical image management and processing system Regulatory Class: Class II Product Code: QIH Dated: June 17, 2022 Received: June 23, 2022

Dear Carla Almeida:

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 (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 located 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 Federal Regulations, Title 21, Parts 800 to 898. In addition, FDA may publish further announcements concerning your device in the Federal Register.

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 and Part 809); medical device reporting of medical device-related adverse events) (21 CFR

1

for devices or postmarketing safety reporting (21 CFR 4, Subpart B) for combination products (see https://www.fda.gov/combination-products/guidance-regulatory-information/postmarketing-safety-reportingcombination-products); good manufacturing practice requirements as set forth in the quality systems (QS) regulation (21 CFR Part 820) for devices or current good manufacturing practices (21 CFR 4, Subpart A) for combination products; and, if applicable, the electronic product radiation control provisions (Sections 531-542 of the Act); 21 CFR 1000-1050.

Also, please note the regulation entitled, "Misbranding by reference to premarket notification" (21 CFR Part 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-device-safety/medical-device-reportingmdr-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/medicaldevices/device-advice-comprehensive-regulatory-assistance) and CDRH Learn (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-device-advice-comprehensive-regulatoryassistance/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,

Jessica Lamb. Ph.D. Assistant Director Imaging Software Team DHT8B: Division of Radiological Imaging Devices and Electronic Products OHT8: Office of Radiological Health Office of Product Evaluation and Quality Center for Devices and Radiological Health

Enclosure

2

Indications for Use

510(k) Number (if known) K220624

Device Name AI4CMR v1.0

Indications for Use (Describe)

AI4CMR software is designed to report cardiac function measurements (ventricle volumes, ejection fraction, indices etc.) from 1.5T and 3T magnetic resonance (MR) scanners. AI4CMR uses artificial intelligence to automatically segment and quantify the different cardiac measurements. Its results are not intended to be used on a stand-alone basis for clinical decision-making.

The user incorporating AI4CMR into their DICOM application of choice is responsible for implementing a user interface.

Type of Use (Select one or both, as applicable)
X Prescription Use (Part 21 CFR 801 Subpart D)	☐ Over-The-Counter Use (21 CFR 801 Subpart C)

CONTINUE ON A SEPARATE PAGE IF NEEDED.

This section applies only to requirements of the Paperwork Reduction Act of 1995.

DO NOT SEND YOUR COMPLETED FORM TO THE PRA STAFF EMAIL ADDRESS BELOW.

The burden time for this collection of information is estimated to average 79 hours per response, including the time to review instructions, search existing data sources, gather and maintain the data needed and complete and review the collection of information. Send comments regarding this burden estimate or any other aspect of this information collection, including suggestions for reducing this burden, to:

Department of Health and Human Services Food and Drug Administration Office of Chief Information Officer Paperwork Reduction Act (PRA) Staff PRAStaff(@fda.hhs.gov

"An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB number."

3

Image /page/3/Picture/1 description: The image shows the logo for AI4MedImaging. The logo consists of a circular graphic on the left and the text "AI4MedImaging" on the right. The circular graphic is made up of a series of blue lines that radiate out from the center, creating a spiral effect. The text is also blue and is written in a sans-serif font.

Section 5. 510(k) Summary

1. General Information

510(k) Sponsor	AI4MedImaging Medical Solutions S.A.
Address	Rua do Parque Poente, It 32
4705-002 Sequeira, Braga Portugal
Correspondence Person	Rory A. Carrillo
Quality and Regulatory Consultant
Cosm
Contact Information	Email: rory@cosmhq.com
Phone: 562-533-7010
Date Prepared	June 17, 2022

2. Subject Device

Proprietary Name	AI4CMR v1.0
Common Name	AI4CMR
Classification Name	System, Image Processing, Radiological
Regulation Number	21 CFR 892.2050
Regulation Name	Medical Image Management and Processing System
Product Code	OIH
Regulatory Class	II

3. Predicate Device

Proprietary Name	Imbio RV/LV Software
Premarket Notification	K203256
Classification Name	System, Image Processing, Radiological
Regulation Number	21 CFR 892.2050
Regulation Name	Medical Image Management and Processing System
Product Code	QIH
Regulatory Class	II

Device Description 4.

AI4CMR v1.0 is a cloud-hosted service used with any third-party DICOM viewer application where the DICOM viewer serves as the user interface and the interface to a PACS or scanner for AI4CMR. AI4CMR is implemented as a plug-in to the DICOM viewer by the user and automatically processes and analyses cardiac MR images received by the DICOM viewer to quantify relevant cardiac function metrics and makes the information available to the user at the user's discretion.

The following are the cardiac function metrics quantified and reported by the software:

AI4CMR v1.0 Traditional 510(k)

4

Image /page/4/Picture/0 description: The image shows the logo for AI4MedImaging. The logo consists of a circular graphic on the left and the text "AI4MedImaging" on the right. The circular graphic is made up of many small blue lines that form a spiral shape. The text "AI4MedImaging" is also in blue and is written in a sans-serif font.

Quantitative Analysis

The subject device performs the following anatomical measurements:

Anatomy and tissue segmentation ●
LV/RV stroke volume 0
LV/RV cardiac output
LV/RV ejection fraction ●
LV/RV end-diastolic volume
LV/RV end-systolic volume

Reporting

The subject device enables the following metrics to be reported as desired by the user:

Metric	Unit	Accuracy
LV/RV stroke volume	ml	n/a1
LV/RV cardiac output	L/min	n/a1
LV/RV ejection fraction (EF)	%	LV bias (std): 3.87 (5.74)
LV ICC: 0.96
RV bias (std): 7.80 (7.32)
RV ICC: 0.81
LV/RV end-diastolic volume (EDV)	ml	LV bias (std): 8.66 (17.28)
LV ICC: 0.99
RV bias (std): 8.36 (15.59)
RV ICC: 0.96
LV/RV end-systolic volume (ESV)	ml	LV bias (std): -2.90 (17.52)
LV ICC: 0.99
RV bias (std): -9.08 (13.83)
RV ICC: 0.95
LV myocardial mass	g	bias (std): 2.45 (18.04)
ICC: 0.96
LV/RV end-systolic volume index2	ml/m^2	n/a1
LV/RV end-diastolic volume index2	ml/m^2	n/a1
LV/RV stroke volume index2	ml/m^2	n/a1
Myocardium mass index2	g/m^2	n/a1

AI4CMR v1.0 Traditional 510(k)

5

Image /page/5/Picture/0 description: The image contains the logo for AI4MedImaging. The logo consists of a circular design on the left, resembling a stylized spiral or a series of concentric circles with varying shades of blue. To the right of the circular design, the text "AI4MedImaging" is written in a sans-serif font, with each letter in a light blue color. The text is aligned horizontally and appears to be the primary identifier for the organization or product.

Metric	Unit	Accuracy
Cardiac index $2$	L/(min m^2)	n/a $1$

Notes:

1 These values are derived by performing simple mathematical operations and are derived from EDV, ESV, EF, and Mass metrics.

2 These values are only provided if the patient's height and weight are included in the DICOM data.

Training Dataset

The AI4CMR training was performed on a dataset of 824 anonvmized cases collected retrospectively from Hospital de Braga, Portugal. Acquisition occurred between 2015 to January 2019 and consisted of male (63%) and female (37%) patients ranging in age from 13 to 89 (mean of 58) years old from Siemens acquisition system. This dataset is independent from the clinical validation set. This dataset was split into the 3 sets (training, validation, test). The splitting ratio is 70% for the "training set", 15% for the "validation set" and 15% for the "test set", resulting in 577, 121 and 126 cases each, respectively.

5. Indications for Use

AI4CMR software is designed to report cardiac function measurements (ventricle volumes, ejection fraction, indices etc.) from 1.5T and 3T magnetic resonance (MR) scanners. AI4CMR uses artificial intelligence to automatically segment and quantify the different cardiac measurements. Its results are not intended to be used on a stand-alone basis for clinical decision-making.

The user incorporating AI4CMR into their DICOM application of choice is responsible for implementing a user interface.

Substantial Equivalence & Technical Characteristics 6.

| | Subject Device
AI4CMR v1.0 | Predicate Device:
Imbio RV/LV (K203256) |
|--------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| Intended Use | AI4CMR software is designed to
report cardiac function
measurements (ventricle volumes,
ejection fraction, indices etc.) from
1.5T and 3T magnetic resonance
(MR) scanners. AI4CMR uses
artificial intelligence to
automatically segment and
quantify the different cardiac
measurements. Its results are not
intended to be used on a | The Imbio RV/LV Software device
is designed to measure the
maximal diameters of the right and
left ventricles of the heart from a
volumetric CTPA acquisition and
report the ratio of those
measurements. RV/LV analyzes
cases using an artificial
intelligence algorithm to identify
the location and measurements of
the ventricles. The RV/LV software |

6

Image /page/6/Picture/0 description: The image shows the logo for AI4MedImaging. The logo consists of a circular graphic on the left and the text "AI4MedImaging" on the right. The circular graphic is made up of several blue lines that radiate out from the center. The text is also blue and is in a sans-serif font.

| Subject Device
AI4CMR v1.0 | Predicate Device:
Imbio RV/LV (K203256) |
|---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| stand-alone basis for clinical
decision-making.
The user incorporating AI4CMR
into their DICOM application of
choice is responsible for
implementing a user interface. | provides the user with annotated
images showing ventricular
measurements. Its results are not
intended to be used on a
stand-alone basis for clinical
decision-making or otherwise
preclude clinical assessment of
CTPA cases. |

| Feature/
Function | Subject Device:
AI4CMR v1.0 | Predicate Device:
Imbio RV/LV
(K203256) | Substantially
Equivalent? |
|---------------------------------|-------------------------------------------------------------------------------------|-----------------------------------------------|------------------------------|
| Indication for Use | See table above | See table above | Yes |
| Input Data
Requirements | Cardiovascular images:
multi-phase,
multi-slice acquired from
MRI scanners | Non-gated, CT Pulmonary
Angiography images | Yes¹ |
| DICOM Compliant | Yes | Yes | Yes |
| LV Segmentation | Yes | Yes | Yes |
| RV Segmentation | Yes | Yes | Yes |
| Diameter Measurements | Yes | Yes | Yes |
| Fully Automated
Segmentation | Yes | Yes | Yes |
| Interface | 3rd party Viewer as a
plug-in | Command line | Yes |
| Outputs | Report only | Report, DICOM
Secondary Capture Series | Yes |

See discussion below

The subject device and predicate device have similar indications for use and technological characteristics. Differences with the input data requirement do not raise questions of safety or effectiveness as the underlying technology is similar with similar risks that are mitigated by the same general and special controls.

7. Performance Data

Safety and performance of the AI4CMR v1.0 has been evaluated and verified in accordance with software specifications and applicable performance standards through software verification and validation testing. Additionally, the software validation activities were performed in accordance

7

Image /page/7/Picture/0 description: The image contains the logo for AI4MedImaging. The logo consists of a circular graphic to the left of the company name. The graphic is made up of several blue lines that form a spiral shape. The company name, "AI4MedImaging", is written in blue, sans-serif font to the right of the graphic.

with ANSI AAMI IEC 62304:2006/41:2016 - Medical device software - Software life cycle processes, in addition to the FDA Guidance documents, "Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices" and "Content of Premarket Submission for Management of Cybersecurity in Medical Devices."

7.1 Bench Testing

AI4MED performed a standalone performance test on the Society of Cardiac Magnetic Resonance (SCMR) Consensus Contour Data' which consists of a total of 15 CMR cases annotated by seven (7) independent expert readers from various core laboratories. The dataset consisted of male and female patients with an age range of 42 to 77 (average of 61) years old across various 1.5T and 3T scanners (Siemens, GE, Phillips). Readers performed myocardial segmentation and quantified End-diastolic volume (EDV), End-systolic volume (ESV), LV mass (LVM), and Ejection Fraction (EF). Agreement was evaluated and achieved between AI4CMR and the SCMR Consensus data. For myocardium segmentation, an average dice similarity coefficient (DSC) of 0.72 per image was achieved. For LVM, EDV, ESV, and EF the intraclass correlation coefficient (ICC) was evaluated and the following performance was obtained:

LV parameter	LOA (± 2 SD)	Bias ± SD	r 2	ICC
EDV	[-58.5912, 45.9097] ml	-6.3407 ± 26.1252 ml	0.85	0.95
ESV	[-30.2173, 22.4768] ml	-3.8703 ± 13.1735 ml	0.97	0.99
EF	[-7.2668, 6.9567] %	-0.155 ± 3.5559 %	0.95	0.99
LVM	[-57.2331, 11.7108] g	-22.7611 ± 17.236 g	0.72	0.78

7.2 Clinical Performance Assessment

AI4MED performed a multi-reader multi-center (MRMC) retrospective study consisting of 146 CMR cases with patients ranging from 17 to 85 years old (average of 51) that were predominantly male (77%) - consistent with cardiovascular disease incidence". Patient data consisted of diseased (~60%) and non-diseased (~40%) where the diseased was spread across the prevalent cardiovascular diseases worldwide'. CMR cases was acquired and balanced across Siemens, GE, and Philips 1.5T scanners with slice thickness of 8mm and slice diameter ranging from 8 to 10.5mm.

The primary objective was to evaluate agreement between the AI4MED device and 2 expert readers who achieved excellent interrater variability (ICC > 0.75). Readers manually segmented the myocardium for each CMR case per standard of care and manually determined volumes, LV mass, and Ejection Fraction. The dataset was independent of the data used for model training and development.

https://www.cardiacatlas.org/studies/scmr-consensus-data.

1 SCMR. Cardiac Atlas Project - SCMR Consensus Contour Data [Internet]. Available from:

2 Walli-Attaei, Marjan, et al. "Variations between women in risk factors, treatments, cardiovascular disease incidence, and death in 27 high-income, middle-income, and low-income countries (PURE); a prospective cohort study: " The Lancet 396.10244 (2020): 97-109 3 Ischemic heart disease, cardiomyopathis, pericardial abnormality, valve disease, cardiac mass tumor and others

AI4CMR v1.0 Traditional 510(k)

8

Image /page/8/Picture/0 description: The image contains the logo for AI4MedImaging. The logo consists of a circular graphic on the left and the text "AI4MedImaging" on the right. The circular graphic is made up of several concentric circles, each of which is composed of a series of small, blue lines. The text "AI4MedImaging" is written in a sans-serif font and is also blue.

A summary of the agreement for Left and Right Ventricular EDV, Left and Right Ventricular ESV, Left and Right Ventricular Ejection Fraction, LV Myocardial Mass is provided below:

Left Ventricular EDV

| | Cronbach's
Alpha | Correlation
Coef. (ρ) | Bias | ICC | 95% Confidence Interval | |
|---------------------|---------------------|--------------------------|-------|-------|-------------------------|-------------|
| | | | | | Lower Bound | Upper Bound |
| AI4CMR vs consensus | 0.992 | 0.980 | 8.663 | 0.990 | 0.98 | 0.99 |

Image /page/8/Figure/4 description: The image contains two scatter plots, labeled A and B. Plot A shows the difference between AI4CMR and Readers' Consensus on the y-axis, and the mean of AI4CMR and Readers' Consensus on the x-axis. The plot includes a mean line at 8.66, and lines for +1.96 SD at 42.65 and -1.96 SD at -25.33. Plot B shows AI4CMR on the y-axis and Readers' Consensus on the x-axis, with a regression line described by the equation Y = 14.49 + 0.90 * X and an R-squared value of 0.98.

[see next page]

AI4CMR v1.0 Traditional 510(k)

9

Image /page/9/Picture/0 description: The image shows the logo for AI4MedImaging. The logo consists of a circular design on the left and the text "AI4MedImaging" on the right. The circular design is made up of several blue lines that form a spiral shape. The text is also in blue and is written in a clear, sans-serif font.

Left Ventricular ESV

Image /page/9/Figure/2 description: The image contains two scatter plots, labeled A and B, and a table of statistical values. Plot A shows the difference between AI4CMR and Readers' Consensus against the mean of AI4CMR and Readers' Consensus, with a mean difference of -2.89 and limits of agreement at +1.96 SD (31.56) and -1.96 SD (-37.35). Plot B shows AI4CMR values plotted against Readers' Consensus values, with a regression line equation of Y = 14.38 + 0.91 * X and an R^2 value of 0.97. The table presents statistical measures such as Cronbach's Alpha (0.992), Correlation Coefficient (0.975), Bias (-2.893), ICC (0.991), and a 95% Confidence Interval with a lower and upper bound of 0.99.

Left Ventricular Ejection Fraction

| | Cronbach's
Alpha | Correlation
Coef. (ρ) | Bias | ICC | 95% Confidence Interval | |
|---------------------|---------------------|--------------------------|-------|-------|-------------------------|-------------|
| | | | | | Lower Bound | Upper Bound |
| AI4CMR vs consensus | 0,969 | 0,909 | 3,867 | 0,956 | 0,87 | 0,98 |

Image /page/9/Figure/5 description: The image contains two scatter plots, labeled A and B. Plot A shows the difference between AI4CMR and Readers' Consensus on the y-axis, plotted against the mean of AI4CMR and Readers' Consensus on the x-axis. The plot includes a mean line at 3.87 and lines indicating +1.96 SD at 15.15 and -1.96 SD at -7.42. Plot B shows AI4CMR on the y-axis plotted against Readers' Consensus on the x-axis, along with a regression line described by the equation Y = 1.36 + 0.89 * X, with an R^2 value of 0.89.

10

Image /page/10/Picture/0 description: The image contains the logo for AI4MedImaging. The logo consists of a circular graphic on the left and the text "AI4MedImaging" on the right. The circular graphic is made up of several blue lines arranged in a circular pattern. The text "AI4MedImaging" is written in a blue sans-serif font.

Left Ventricular Myocardial Mass

Image /page/10/Figure/2 description: This image shows a table comparing AI4CMR vs consensus. The table includes Cronbach's Alpha, Correlation Coef. (p), Bias, ICC, and 95% Confidence Interval. The Cronbach's Alpha is 0.956, the Correlation Coef. (p) is 0.936, the Bias is 2.452, the ICC is 0.955, and the 95% Confidence Interval is 0.94 to 0.97.

Image /page/10/Figure/3 description: The image contains two scatter plots, labeled A and B. Plot A shows the difference between AI4CMR and Readers' Consensus on the y-axis, and the mean of AI4CMR and Readers' Consensus on the x-axis. The plot includes a mean line at 2.45, and lines indicating +1.96 SD at 37.94 and -1.96 SD at -33.03. Plot B shows AI4CMR on the y-axis and Readers' Consensus on the x-axis, with a regression line and the equation Y = 22.18 + 0.82 * X, and R^2 = 0.85.

Right Ventricular EDV

| | Cronbach's
Alpha | Correlation
Coef. (ρ) | Bias | ICC | 95% Confidence Interval | |
|---------------------|---------------------|--------------------------|-------|-------|-------------------------|-------------|
| | | | | | Lower Bound | Upper Bound |
| AI4CMR vs consensus | 0,972 | 0,924 | 8,355 | 0,964 | 0,92 | 0,98 |

Image /page/10/Figure/6 description: The image contains two scatter plots, labeled A and B. Plot A shows the difference between AI4CMR and Readers' Consensus on the y-axis, and the mean of AI4CMR and Readers' Consensus on the x-axis. The plot includes a mean line at 8.35, and lines indicating +1.96 SD at 39.02 and -1.96 SD at -22.31. Plot B shows AI4CMR on the y-axis and Readers' Consensus on the x-axis, with a regression line and the equation Y = 10.70 + 0.88 * X, with R^2 = 0.90.

AI4CMR v1.0 Traditional 510(k)

11

Image /page/11/Picture/0 description: The image shows the logo for AI4MedImaging. The logo consists of a circular graphic on the left and the text "AI4MedImaging" on the right. The circular graphic is made up of several blue lines that form a spiral shape. The text "AI4MedImaging" is also blue and is written in a sans-serif font.

Right Ventricular ESV

| | Cronbach's
Alpha | Correlation
Coef. (p) | Bias | ICC | 95% Confidence Interval | |
|---------------------|---------------------|--------------------------|--------|-------|-------------------------|-------------|
| | | | | | Lower Bound | Upper Bound |
| AI4CMR vs consensus | 0,967 | 0,888 | -9,083 | 0,953 | 0,87 | 0,98 |

Image /page/11/Figure/3 description: The image contains two scatter plots, labeled A and B. Plot A shows the difference between AI4CMR and Readers' Consensus on the y-axis, and the mean of AI4CMR and Readers' Consensus on the x-axis. The plot includes a mean line at -9.08, and lines for +1.96 SD at 18.12 and -1.96 SD at -36.28. Plot B shows AI4CMR on the y-axis and Readers' Consensus on the x-axis, with a regression line and the equation Y = 11.47 + 0.96 * X and R^2 = 0.88.

Right Ventricular Ejection Fraction

| | Cronbach's
Alpha | Correlation
Coef. (ρ) | Bias | ICC | 95% Confidence Interval | |
|---------------------|---------------------|--------------------------|-------|-------|-------------------------|-------------|
| | | | | | Lower Bound | Upper Bound |
| AI4CMR vs consensus | 0,902 | 0,712 | 7,802 | 0,814 | 0,15 | 0,93 |

Image /page/11/Figure/6 description: The image contains two scatter plots, labeled A and B. Plot A shows the difference between AI4CMR and Readers' Consensus on the y-axis, and the mean of AI4CMR and Readers' Consensus on the x-axis. The plot includes a mean line at 7.80, and lines indicating +1.96 SD at 22.20 and -1.96 SD at -6.60. Plot B shows AI4CMR on the y-axis and Readers' Consensus on the x-axis, with a regression line described by the equation Y = 0.93 + 0.86 * X and an R-squared value of 0.68.

12

Image /page/12/Picture/0 description: The image contains the logo for AI4MedImaging. The logo consists of a circular graphic on the left and the text "AI4MedImaging" on the right. The circular graphic is made up of several blue lines arranged in a circular pattern. The text "AI4MedImaging" is written in a blue sans-serif font.

8. Conclusion

Based on the information submitted in this premarket notification, and based on the indications for use, technological characteristics and performance testing, the AI4CMR v1.0 raises no new questions of safety and effectiveness and is substantially equivalent to the predicate device in terms of safety and effectiveness.