The Caption Interpretation Automated Ejection Fraction software is used to process previously acquired transthoracic cardiac ultrasound images, to store images, and to manipulate and make measurements on images using an ultrasound device, personal computer, or a compatible DICOM-compliant PACS system in order to provide automated estimation of left ventricular ejection fraction. This measurement can be used to assist the clinician in a cardiac evaluation.

The Caption Interpretation Automated Ejection Fraction Software is indicated for use in adult patients.

Device Description

The Caption Interpretation Automated Ejection Fraction Software ("AutoEF") applies machine learning algorithms to process echocardiography images in order to calculate left ventricular ejection fraction. The cleared Caption Interpretation AutoEF performs left ventricular ejection fraction measurements using apical four chamber or apical two chamber cardiac ultrasound views, or the parasternal long-axis cardiac ultrasound view in combination with an apical four chamber view. The software selects the image clips to be used, performs the AutoEF calculation, and forwards the results to the desired destination for clinician viewing. The output of the Ejection Fraction estimate stated as a percentage, along with an indication of confidence regarding that estimate.

AI/ML Overview

Here's a breakdown of the acceptance criteria and study details for the Caption Interpretation Automated Ejection Fraction Software, based on the provided FDA 510(k) summary:

1. Acceptance Criteria and Reported Device Performance

Acceptance Criterion (Primary Endpoint)	Predicate Device Performance (K200621)	Subject Device Performance (K210747)	Meets Acceptance?
Root Mean Square Deviation (RMSD) of Ejection Fraction (EF) % below a set threshold compared to reference ground truth EF.	7.94 RMSD EF % (95% CI)	7.21 RMSD EF % (95% CI)	Yes (demonstrated improvement)
Outlier Rate (EF error >15%)	1.61%	1.09%	Yes (comparable performance, slight improvement)
Outlier Rate (EF error >20%)	0%	0.55%	Yes (comparable performance)

Notes on Acceptance Criteria:

The specific "set threshold" for RMSD is not explicitly stated in the provided text. However, the summary indicates that the "primary endpoint for the subject device was met."
The summary highlights that the subject device demonstrated "slightly improved performance" in RMSD compared to the predicate, and "comparable performance" in outlier rates.

2. Sample Size Used for the Test Set and Data Provenance

Sample Size: Over 186 acquired studies.
Data Provenance:
- Country of Origin: Not explicitly stated.
- Retrospective or Prospective: Retrospective, non-interventional validation study.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts

This information is not explicitly stated in the provided text. The text only mentions that the device's measurements were "compared to the biplane method ejection fraction" as the reference. It doesn't detail who performed these biplane measurements or how many experts were involved.

4. Adjudication Method for the Test Set

The text describes the ground truth as the "biplane method ejection fraction." It does not describe an adjudication method for the test set, implying that the biplane measurements served as the direct reference without further expert consensus or adjudication.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size

No, an MRMC comparative effectiveness study was not done. The study described is a standalone performance validation of the algorithm against a declared ground truth (biplane method ejection fraction). There is no mention of human readers assisting or being compared to the AI.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

Yes, a standalone performance study was done. The described study directly compares the Caption Interpretation Automated Ejection Fraction Software's output to the "biplane method ejection fraction" without human intervention in the loop for the device's calculation. The device provides an "automated estimation of left ventricular ejection fraction."

7. The Type of Ground Truth Used

Expert Consensus / Clinical Standard: The ground truth used was the biplane method ejection fraction. This is a widely accepted clinical method for calculating EF, typically performed by trained professionals (e.g., echocardiographers, cardiologists). While the text doesn't specify if it was a "consensus" of multiple experts, it refers to a established clinical method.

8. The Sample Size for the Training Set

The training set included an "additional 30% of training data from three ultrasound devices and two clinical sites" for retraining of algorithms, compared to the predicate device. The absolute sample size of the training set is not explicitly stated.

9. How the Ground Truth for the Training Set Was Established

The text states that "Images and cases used for verification and validation testing were carefully separated from training datasets." While it doesn't explicitly detail how the ground truth for the training set was established, it's generally understood that for machine learning algorithms in medical imaging, the training data would also require some form of expert labeling or ground truth establishment (e.g., manual segmentation and measurement by cardiologists, confirmed by clinical standards like the biplane method). The summary does not provide specific details on this process for the training data.

Summary

{0}------------------------------------------------

Image /page/0/Picture/0 description: The image shows the logo of the U.S. Food and Drug Administration (FDA). On the left is the Department of Health & Human Services logo. To the right of that is the FDA logo, with the letters "FDA" in a blue square. To the right of the blue square is the text "U.S. FOOD & DRUG ADMINISTRATION" in blue.

Caption Health % Savannah Hari Quality Systems Manager Manatt, Phelps & Phillips, LLP 1 Embarcadero Center, 30th Floor San Francisco, California 94111

January 19, 2022

Re: K210747

Trade/Device Name: Caption Interpretation Automated Ejection Fraction Software Regulation Number: 21 CFR 892.2050 Regulation Name: Medical Image Management And Processing System Regulatory Class: Class II Product Code: QIH Dated: January 18, 2022 Received: January 18, 2022

Dear Savannah Hari:

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

{1}------------------------------------------------

801); medical device reporting of medical device-related adverse events) (21 CFR 803) 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 mediation-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,

For

Thalia T. Mills, Ph.D. Director Division of Radiological Health OHT7: Office of In Vitro Diagnostics and Radiological Health Office of Product Evaluation and Quality Center for Devices and Radiological Health

Enclosure

{2}------------------------------------------------

Indications for Use

510(k) Number (if known)

K210747

Device Name

Caption Interpretation Automated Ejection Fraction Software

Indications for Use (Describe)

The Caption Interpretation Automated Ejection Fraction software is used to process previously acquired transhoracic cardiac ultrasound images, to store images, and to make measurements on images using an ultrasound device, personal computer, or a compatible DICOM-compliant PACS system in order to provide automation of left ventricular ejection fraction. This measurement can be used to assist the clinician in a cardiac evaluation.

The Caption Interpretation Automated Ejection Fraction Software is indicated for use in adult patients.

Type of Use (Select one or both, as applicable)

🇿 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 Druq Administration Office of Chief Information Officer Paperwork Reduction Act (PRA) Staff PRAStaff@fda.hhs.gov

"An aqency 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}------------------------------------------------

510(k) Summary Caption Health, Inc. Caption Interpretation Automated Ejection Fraction K210747

Submitter

Caption Health, Inc. 2000 Sierra Point Parkway, 8th Floor Brisbane, CA 94005

Contact Person: Savannah Hari, Quality Systems Manager

Phone: (415) 671-4711 Email: savannah@captionhealth.com Date Prepared: January 14, 2022

Device

Proprietary Name: Caption Interpretation Automated Ejection Fraction Software

Common Name: Caption Interpretation Automated Ejection Fraction Software

Classification Name and Number: Medical Image Management and Processing System, 21 CFR 892.2050

Regulatory Class: II

Product Code: QIH, Automated Radiological Image Processing Software

Predicate Device

Caption Health, Inc. Caption Interpretation Automated Ejection Fraction Software (K200621)

Device Description:

{4}------------------------------------------------

Intended Use / Indications for Use:

No differences exist between the subject device and the predicate device with respect to intended use or indications for use. The intended use / indications for use are provided below:

The Caption Interpretation Automated Ejection Fraction Software is indicated for use in adult patients.

Summary of Technological Characteristics:

The Caption Interpretation Automated Ejection Fraction Software is an updated version of the predicate device and features very similar technological characteristics. Specific changes between the subject device and the predicate include:

An additional 30% of training data from three ultrasound devices and two clinical sites for retraining of algorithms
Simplification of network architecture via simple pooling layers and quantization
Simplification of image quality algorithms via linear support vector machines
Optimization of image quality thresholds to maintain performance of the subject device ●
Runtime speed optimization of image preprocessing and removal of vestigial code

These particular changes are intended to:

1. Improve the function of the product across the diversity of ultrasound devices in use and patient types in the population.
1. Reduce the complexity of computations, enabling the algorithm to run faster and allow it potentially to operate on lower computing power hardware platforms.
1. Maintain optimal clinical performance and acceptability due to additional network training.

The technological principle underlying both the current version of Caption Interpretation Automated Ejection Fraction Software and the predicate device remains the same, in that both the subject device and predicate enable the calculation of ejection (EF) on previously acquired cardiac scans using machine learning-based algorithms and biplane apical echocardiographic images. Further details can be found in the comparison table below.

{5}------------------------------------------------

Performance Data:

The Caption Interpretation Automated Ejection Fraction Software was developed and tested in accordance with Caption Health's Design Control processes and has been subjected to extensive safety and performance testing. Verification and validation testing was conducted to demonstrate the substantial equivalence of the subject device to the predicate. The primary success criterion was that the subject device would produce an ejection fraction number with a Root Mean Square Deviation below a set threshold as compared to the reference ground truth EF. The same test protocols, acceptance criteria and endpoints were used between the predicate and the subject device to ensure that performance could be appropriately compared.

Non-clinical verification and validation test results established that the device meets its design requirements and intended use. Extensive algorithm development and software verification testing assessed the performance of the software's image video clip selection function, performance characteristics of the algorithm, including AutoEF accuracy and overall functional performance.

Images and cases used for verification and validation testing were carefully separated from training datasets.

A formal retrospective, non-interventional validation study was conducted using over 186 acquired studies where the biplane method of disks ejection fraction was reported. This patient dataset was constructed to provide a balanced range of gender, ejection fraction values, and body mass index levels. Testing included a wide array of ultrasound system manufacturers to verify that the subject device performs acceptably across multiple scanner platforms. The ejection measurements from the subject device were compared to the biplane method ejection fraction, and a root mean square deviation was calculated and used as the primary endpoint for the subject device was met (results: 7.21 RMSD EF% [95% Cl]), and demonstrated slightly improved performance compared to the predicate device (results: 7.94 RMSD EF % [95% Cl]). Performance improvements between the subject device and the predicate device did not lead to significant differences in the outlier rate, with comparable performance between the subject device [1.09% for EF error >15%, 0.55% for EF error >20%] and the predicate [1.61% for EF error >15%, 0% for EF error >20%]. Based on the clinical performance as documented in this retrospective validation study, the device has a safety and effectiveness profile that is substantially equivalent to the predicate device.

Conclusions:

Performance testing demonstrated that the Caption Interpretation Automated Ejection Fraction Software performs as expected and in a manner that is substantially equivalent to the predicate device. The Caption Interpretation Automated Ejection Fraction software has the same intended use, indications for use, and principles of operation as its predicate device. Thus, the Caption Interpretation Automated Ejection Fraction software is substantially equivalent to its predicate.

{6}------------------------------------------------

Comparison of Features between Proposed Subject Device and Predicate Device
	Caption Health, Inc., CaptionInterpretation AutomatedEjection Fraction ("AutoEF2.5") (K210747) - ProposedDevice	Caption Health, Inc., CaptionInterpretation AutomatedEjection Fraction SoftwareApplication ("AutoEF 2.0")(K200621) - Predicate Device
Product Code	QIH	QIH
Intended Use	The Caption Health, Inc.Caption InterpretationAutomated Ejection Fractionsoftware is used to processpreviously acquiredtransthoracic cardiacultrasound images, to storeimages, and to manipulateand make measurements onimages using an ultrasounddevice, personal computer, ora compatible DICOM-compliant PACS system inorder toprovide automated estimation ofleft ventricular ejection fraction.This measurement can be usedto assist the clinician in a cardiacevaluation.The Caption InterpretationAutomated Ejection FractionSoftware is indicated for use inadult patients.	The Caption Health, Inc. CaptionInterpretation Automated EjectionFraction software is used to processpreviously acquired transthoraciccardiac ultrasound images, to storeimages, and to manipulate and makemeasurements on images using anultrasound device, personalcomputer, or a compatible DICOM-compliant PACS system in order toprovide automated estimation of leftventricular ejection fraction. Thismeasurement can be used to assistthe clinician in a cardiac evaluation.The Caption InterpretationAutomated Ejection FractionSoftware is indicated for use inadult patients.
General Principles of Operation
Machine Learning- BasedAlgorithm	Yes	Yes
Operates on DICOM clips	Yes	Yes
Automation level	Fully automated, including clipselection	Fully automated, including clipselection
User Alert to Use MultipleViews	Yes	Yes
Automated EjectionFraction Calculation	Yes	Yes
Ejection Fraction reported	Whole number estimate	Whole number estimate
Quantitative feedbackto enable clinician toassess EF calculation	• Confidence Metric• Qualitative Bin Likelihood	• Confidence Metric• Qualitative Bin Likelihood
EF Result shown withvideo clip	Yes	Yes
User confirmation/rejection of result	Yes	Yes
Technological Characteristics
Network Architecture	Simple Pooling	Advanced Pooling
Regressor algorithm	Linear Support Vector Machine	Radial-basis-function (RBF) SupportVector Machine
Image Processing	Grayscaling done first and integerresizing.	Floating point resizing and grayscalingdone at the end

{7}------------------------------------------------

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).