DiaCardio's LVivo Software Application is intended for non-invasive processing of already acquired echocardiographic images in order to detect, measure, and calculate the left ventricular function evaluation. This measurement can be used to assist the clinician in a cardiac evaluation.

Device Description

The LVivo System analyzes echocardiographic patient examination DICOM movies for Global ejection fraction (EF) evaluation. EF is evaluated using two orthogonal planes, four-chamber (4CH) and two-chamber (2CH) views, to provide fully automated analyses of LV function from the echo examination movies. It also has the ability to measure strain

AI/ML Overview

Here's a breakdown of the acceptance criteria and study information for the LVivo Software Application, based on the provided document:

1. Table of Acceptance Criteria and Reported Device Performance

The document focuses on the performance of the LVivoSG module and its correlation with established methods rather than explicit "acceptance criteria" presented as pass/fail thresholds. However, we can infer the desired performance from the study objectives and results.

Metric	Acceptance Criteria (Inferred from Study Objectives)	Reported Device Performance (LVivoSG)
Global Longitudinal Strain (GLS) vs. VVI	Good agreement with correlation coefficient (r) ≥ 0.8	r = 0.85 (p<0.0001)
Inter-observer reliability (GLS) between LVivoSG and VVI	Excellent agreement (Intraclass Correlation, ICC)	ICC = 0.92
Agreement for Normal/Abnormal GLS (LVivoSG vs. VVI)	Good agreement (kappa coefficient)	kappa = 0.77
Sensitivity for Normal/Abnormal GLS (LVivoSG vs. VVI)	High sensitivity	0.95
Specificity for Normal/Abnormal GLS (LVivoSG vs. VVI)	High specificity	0.86
WM Score Index vs. Visual Estimation	Good agreement (ICC)	ICC = 0.86
Accuracy (AUC) for Normal/Abnormal WM Score Index	Good accuracy (AUC)	0.86
Sensitivity for Normal/Abnormal WM Score Index		0.78
Specificity for Normal/Abnormal WM Score Index		0.8
Territories (LAD, RCA, CX) Strain (LVivoSG vs. VVI) - ICC	Good agreement (ICC)	LAD: 0.86, RCA: 0.84, CX: 0.9
Territories (LAD, RCA, CX) WM Scores (LVivoSG vs. Visual Estimation) - ICC	Good agreement (ICC)	LAD: 0.8, RCA: 0.82, CX: 0.83
GLS from LVivoSG vs. WM score index	Very high correlation (r)	r = 0.87
GLS from LVivoSG vs. GLS from LVivoEF	Very high correlation (r)	r = 0.92

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

Test Set Sample Size: 100 subjects (ultrasound clips of 100 subjects).
Data Provenance: Retrospective, single-center study. Ultrasound examinations were collected prospectively according to protocol 100 rev 03 at "Soroka university medical center" (in Israel, based on the manufacturer's address).

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

Number of Experts: At least two distinct groups of experts:
- "Physicians of the echo department in Soroka university medical center" who routinely evaluated segmental wall motion qualitatively.
- An "Additional investigator (expert echocardiologist)" who performed segmental strain evaluation using Syngo® Velocity Vector Imaging (VVI) SW (Siemens) blindly.
- The "PI" (Principal Investigator) who evaluated examinations with impaired global LV function that lacked routine segmental WM scores.
Qualifications of Experts:
- "Physicians of the echo department" (implied to be qualified in echocardiography).
- "Expert echocardiologist."
- "PI" (implied to be an expert in echocardiography, given their role in evaluating complex cases).

4. Adjudication Method for the Test Set

The document describes several comparisons:

LVivoSG WM evaluation was compared against "visual estimation" by physicians. This implies the physicians' visual estimation served as a form of ground truth or benchmark, rather than a formal adjudication of LVivoSG's output.
LVivoSG longitudinal strain was compared against VVI, applied by an "additional investigator (expert echocardiologist)" blindly. This suggests the VVI measurements by this expert served as a reference.

There is no explicit mention of an adjudication method like 2+1 or 3+1 to establish a consensus ground truth amongst multiple experts for the test set itself. Instead, the study compares the device's performance against pre-existing routine evaluations (for WM) and expert-performed measurements using a predicate device (for strain).

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was Done, What Was the Effect Size of How Much Human Readers Improve with AI vs. Without AI Assistance?

No, a MRMC comparative effectiveness study where human readers improve with AI assistance vs. without AI assistance was not done.

The study's primary goal was to compare the performance of the LVivoSG system itself against established methods (manual methods, visual estimation, and a predicate device like VVI), not to evaluate the improvement of human readers using the AI. The device is described as a "decision support tool," implying assistance, but the study design doesn't measure this specific human-in-the-loop improvement.

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

Yes, a standalone performance study was done. The results reported (correlation coefficients, ICC, kappa, sensitivity, specificity, AUC) directly reflect the performance of the LVivoSG algorithms in automatically evaluating segmental wall motion and strain. The comparisons are between the LVivoSG's output and the ground truth/reference methods, without an intermediate human interaction being measured.

7. The Type of Ground Truth Used (Expert Consensus, Pathology, Outcomes Data, etc.)

The ground truth was a combination of:

Expert Visual Estimation: For segmental wall motion evaluation, the "routinely evaluated for segmental wall motion evaluation qualitatively by the physicians of the echo department."
Expert-Applied Predicate Device Measurement: For segmental longitudinal strain, the ground truth was established by an "expert echocardiologist" using the Syngo® Velocity Vector Imaging (VVI) SW (Siemens), a recognized predicate technology.

8. The Sample Size for the Training Set

The document does not explicitly state the sample size for the training set. It only mentions the "100 subjects" used for the clinical trial (test set) for the LVivoSG.

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

Since the training set size and characteristics are not provided, how the ground truth for the training set was established is also not described in this document.

Summary

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

Image /page/0/Picture/1 description: The image shows the logo for the U.S. Department of Health and Human Services. The logo consists of a circular seal with the text "DEPARTMENT OF HEALTH & HUMAN SERVICES - USA" around the perimeter. Inside the circle is a stylized image of three human profiles facing to the right, with flowing lines connecting them.

Food and Drug Administration 10903 New Hampshire Avenue Document Control Center - WO66-G609 Silver Spring, MD 20993-0002

July 28, 2016

Diacardio, Ltd. % Mr. George Hattub Senior Staff Consultant MedicSense, USA 291 Hillside Avenue SOMERSET MA 02726

Re: K161382

Trade/Device Name: LVivo Software Application Regulation Number: 21 CFR 892.2050 Regulation Name: Picture archiving and communications system Regulatory Class: II Product Code: LLZ Dated: May 14, 2016 Received: May 18, 2016

Dear Mr. Hattub:

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. 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); medical device reporting (reporting of medical device-related adverse events) (21 CFR 803); good manufacturing practice requirements as set forth in the quality systems (QS) regulation (21 CFR Part 820); and if applicable, the electronic product radiation control provisions (Sections 531-542 of the Act); 21 CFR 1000-1050.

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

If you desire specific advice for your device on our labeling regulation (21 CFR Part 801), please contact the Division of Industry and Consumer Education at its toll-free number (800) 638 2041 or (301) 796-7100 or at its Internet address

http://www.fda.gov/MedicalDevices/Resourcesfor You/Industry/default.htm. 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

http://www.fda.gov/MedicalDevices/Safety/ReportaProblem/default.htm for the CDRH's Office of Surveillance and Biometrics/Division of Postmarket Surveillance.

You may obtain other general information on your responsibilities under the Act from the Division of Industry and Consumer Education at its toll-free number (800) 638-2041 or (301) 796-7100 or at its Internet address

http://www.fda.gov/MedicalDevices/ResourcesforYou/Industry/default.htm.

Sincerely yours.

Michael D. O'Hara

For

Robert Ochs, Ph.D. Director Division of Radiological Health Office of In Vitro Diagnostics and Radiological Health Center for Devices and Radiological Health

Enclosure

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

Indications for Use

510(k) Number (if known)

K161382
Device Name

LVivo Software Application

Indications for Use (Describe)

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

rescription 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.qov

"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}------------------------------------------------

510(k) Summary

Pursuant to CFR 807.92, the following 510(k) Summary is provided:

1. (a) Submitter George J. Hattub Address: MedicSense, USA 291 Hillside Avenue Somerset, MA 02726 www.medicsense.com 1. (b) Manufacturer DiACardio, Ltd. Address: HaEnergia Street 77 Be'er Street, Israel Mfa. Phone: Tel.: +972 77 7648318 Contact Person: Mrs. Michal Yaacobi Date: July 21, 2016 2. Picture Archiving Device- classified as Class 2 LLZ, Regulation Number 21 Device & Classification CFR 892.2050 LVivo Software Application Name: Predicate Devices: K072090- Siemens Medical Solution SYNGO Auto Left Heart and VVL 3. Clinical Feature K091286- Siemens Medical Solution SYNGO US Workplace K130779- DiaCardio's LVivoEF Software Application 4. Description: The LVivo System analyzes echocardiographic patient examination DICOM movies for Global ejection fraction (EF) evaluation. EF is evaluated using two orthogonal planes, four-chamber (4CH) and two-chamber (2CH) views, to provide fully automated analyses of LV function from the echo examination movies. It also has the ability to measure strain 5. Intended Use: DiaCardio's L Vivo Software Application is intended for non-invasive processing of already acquired echocardiographic images in order to detect, measure, and calculate the left ventricular wall for left ventricular function evaluation. This measurement can be used to assist the clinician in a cardiac evaluation. 6. Comparison of With respect to technology and intended use, DiaCardio's LVivo Software Technological Application is substantially equivalent to its predicate devices. Based upon the outcomes from clinical trials, DiaCardio believes that their device does Characteristics: not raise additional safety of efficacy concerns. At the end of this summary, a comparison table is provided.

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

1. Clinical Tests: In this study, the performance of LVivoSG was compared with conventional methods used for SG function evaluation in echocardiography, including manual evaluation by sonographers and visual estimation by physicians. In the blinded clinical trial, ultrasound clips of 100 subjects were evaluated with the LVivo System. Average values were calculated for each variable measured by Manual Biplane Method (MBP) and Pearson correlation coefficients were calculated between MBP and LVivoEF results. The primary end point defined for this study was met with a correlation coefficient calculated for biplane EF (r=0.88, p<0001).

	Devices
	Submitted Device	Predicate Device	Predicate Device
Features/Characteristics	LVivo (Diacardio)	LVivoEF (Diacardio)	Syngo (Siemens)
Product Code	LLZ	LLZ	LLZ (K091286)IYN(K072090)
Intended Use	Calculate ofEjection Fractionand measurestrain	Calculate ofEjection Fraction	Calculate EjectionFraction andmeasure strain
Automation	Fully Automated	Fully Automated	Fully Automated
Bi plane EF evaluation	YES	YES	YES
Simultaneous 2CH and4CH evaluation	YES	YES	NO
Off line EF evaluationusing DICOM clips ofany vendor	YES	YES	YES
Automated ED and ESframes selection	YES	YES	YES
Dynamic left ventricularassessment	YES. Frame byframe tracking	YES. Frame byframe tracking	YES. Frame by frametracking

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

Manual editing by user capability
	YES. 7 border points manipulation (dragging) and online contour presentation. Possible to apply to any frame in the clip. Border detection is recalculation is applied to the entire clip.	YES. 7 border points manipulation (dragging) and online contour presentation. Possible to apply to any frame in the clip. Border detection is recalculation is applied to the entire clip.	YES. Click and drag of the contour. Applied to present ED/ES frame.Enable also Manual user input.
Visually confirm EF	YES	YES	YES
Automated rejection of	YES	YES	NO
Volume calculation by	YES	YES	YES
Volume curve	YES	YES	YES
EF results presentation	Displaying full clip	Displaying full clip	Displaying image
Enables presentation EF	YES	YES	YES
Algorithm	Image segmentation	Image segmentation	Adjustment of learned
Calculation speed	Less than 1s per	Less than 1s per cycle	~ 15s for each view
Capability or a part of a	YES	NO	YES
Global Longitudinal	YES	NO	YES
Segmental Longitudinal	YES	NO	YES
Segmental wall motion	YES	NO	Calculation motion
Operating System	Windows	Windows	Unknown
510(k) #	Pending	K130779	K072090 & K091286

Clinical Summary - LVivoSG

Technology and predicate device

The segmental evaluation by LVivoSG is based on the LVivo decision support platform for fully automated edge detection and tracking of the LV borders. The LVivoSG calculates segmental wall motion scores using a classification system based on wall motion parameters. The wall motion scores by LVivoSG were compared to wall motion scores by visual estimation. The segmental endocardial longitudinal strain is calculated in a way that resembles the VVI technology (Siemens) in which the LV borders are traced in a semi-automated way by initial user input. The strain calculated by LVivoSG was compared to the strain calculated by VVI.

Protocol:

In this study, seqmental wall motion evaluation and segmental strain evaluation by LVivoSG system calculated from 3 apical views (4CH, 2CH and 3CH), were compared with Visual Estimation (done by physicians) and with the semi-automated Velocity Vector Imaging (VVI, Siemens) technology (Applied by a physician).

Study: Retrospective, single center study.
Ultrasound examinations that were collected prospectively according to protocol 100 rev 03 (clinical-protocol-1.4.doc) were used in the LVivoSG clinical trial. These examinations were

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

routinely evaluated for segmental wall motion evaluation qualitatively by the physicians of the echo department in Soroka university medical center.

The WM evaluation by the physician was collected retrospectively using patient number and name initials assigned according to protocol 100 rev 03.
Examinations with impaired global LV function that did not have segmental WM scores from routine evaluation will be evaluated by the PI.
Additional investigator (expert echocardiologist) performed segmental strain evaluation by longitudinal strain using Syngo® Velocity Vector Imaging (VVI) SW (Siemens) blindly.
Segmental WM evaluation by LVivoSG was compared to the Segmental WM evaluation by the visual estimation
Segmental longitudinal strain evaluation by LVivoSG was compared to segmental longitudinal strain evaluation by VVI.

Study Objectives

a. Compare the strain results by LVivoSG to strain evaluated with VVI.

b. Compare the automated wall motion results by LVivoSG to wall motion evaluation by visual estimation.

c. Compare the global strain calculated by LVivoEF to the global strain calculated by LVivoSG.

Since the global longitudinal strain (GLS) is an important parameters of LV function adopted by the Guidelines* the primary end point was to show that there is a good agreement between GLS calculated by both methods with correlation coefficient of r=0.8. Additional goals were to compare wall motion scores by LVivoSG to the wall motion scores by visual estimation and. Results and Conclusions

Global

The results showed that the primary end point was successfully met with a very good correlation between LVivoSG and VVI for GLS (r=0.85, p<0.0001). Excellent inter-observer reliability between methods for GLS, was also demonstrated by intraclass correlation (ICC=0.92). The agreement between LVivoSG and VVI demonstrated by kappa coefficient was calculated from categorical data where the

GLS was divided into two categories Normal/Abnormal. The cutoff value for LVivoSG was -12% and for VVI -15%. The agreement by kappa coefficient was also very good (kappa=0.77) and specificity and sensitivity were high (0.86 and 0.95 respectively) emphasizing the similarity between methods.

Average difference of -3% between VVI and LVivoSG was found. This average difference affects the Normal/Abnormal cutoff value. It is known that different vendors use slightly different methods to evaluate strain, and therefore have different cutoff values for LV function. Even in different labs using the same methods, different cutoff values can be determined.

WM score index was calculated as average of segmental wall motion scores and compared between LVivoSG and Visual estimation. The agreement calculated by ICC was very good (ICC=0.86). Specificity and Sensitivity were calculated by divided the results into two categories Normal/Abnormal. The cutoff value =0.51 for LVivoSG was determined by ROC analysis where the threshold for the visual estimation was zero. The accuracy indicated by AUC=0.86 was very good and the specificity and sensitivity were 0.8 and 0.78 respectively. These results show very good agreement comparing WM score index calculated by LVivoSG to WM score index calculated by visual estimation where WM score index<=0.51 by LVivoSG indicates Normal LV function.

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

Territories

Good agreement was demonstrated comparing territories of coronary arteries between strain by LVivoSG and strain by VVI with ICC =0.86, 0.84 and 0.9 for LAD, RCA and CX respectively. The best agreement was for CX with kappa=0.71 and sensitivity 0.8 and 0.94 respectively.

Good results were also demonstrated comparing average of wall motion scores by LVivoSG to wall motion scores by Visual Estimation over territories of coronary arteries. The ICC comparing LAD, RCA and CX was 0.8, 0.82 and 0.83 respectively. Normal/Abnormal cutoff values for the results of the LVivoSG were calculated by ROC analysis for each territory. The accuracy by AUC for LAD, RCA and CX was 0.86, 0.82 and 0.81 respectively. The highest agreement was obtained for LAD territory using cutoff value=0.34 (kappa=0.65). The level of agreement was 0.83 and specificity and sensitivity were 0.86 and 0.81 respectively. The lowest agreement was obtained for the CX territory using cutoff value=0.51). The level of agreement was 0.76 and specificity and sensitivity were 0.75 and 0.76 respectively. The cutoff value for WM scores by visual estimation was zero.

It is interesting to note that findings from studies in the literature showed that the inter-observer reliability for visual estimation (physicians) was highest for segments in the left anterior descending artery territory (ICC, 0.73) and lowest in the circumflex territory (ICC, 0.61). In our study the WM scores by LVivoSG were compared to WM scores by visual estimation of different physicians from the echo department and the highest agreement was for the LAD territory and lowest for the CX as well, indicating that, the results of the current study reflect the "real life" agreement between physicians.

Individual segments

In the current study, good inter-observer reliability between LVivoSG and VVI was for apical and mid segments where the best was for Mid-lateral (ICC=0.83) and Midanterior (ICC=0.79) and the lowest for the basal segments. It was reported in the literature that the highest intra-observer correlation (R > 0.8) was for mid segments of all walls, while low correlation (R<=0.65) was basal lateral, basal anterior and apical anterior segments, implying that mid segment are easier to evaluate than basal segments. The results of the current study show that the agreement between LVivoSG and VVI is higher in segments for which the diagnosis is more conclusive for physicians in "real life".

Wall motion scores were compared between individual seqments and the separation error between normal and akinetic segments was calculated. For most segments, the separation error was <=15%. For the segments Apical Septal. Mid Lateral. Basal Lateral. Mid Anterior. Basal Anterior and Mid Inferolateral the Normal/Akinetic separation error was <=5%.

GLS vs WM score index

To show the connection between strain evaluation by LVivoSG and segmental wall motion evaluation, comparison between WM score index by LVivoSG and GLS by LVivoSG was made. Very high correlation (r=0.87) was obtain between methods, showing that both GLS and WM score index calculated by LVivoSG are comparable.

GLS from LVivoSG vs GLS from LVivoEF

Finally, due to the addition of GLS to the LVivoEF module, GLS by LVivoEF, calculated as average of the strain of the walls from two views was compared to GLS by LVivoSG calculated as average of segmental strain. The results showed very high correlation between methods (r=0.92). This result indicates that users of the LVivoEF module can benefit from the addition of the GLS calculation and obtain important information about the global state of the left ventricle.

The present study has demonstrated that the LVivoSG system provides accurate measurements of segmental LV function. The performance of LVivoSG demonstrated high agreement between

{8}------------------------------------------------

strain results in compare to the strain calculated by VVI and segmental wall motion evaluation in compare to segmental scores by visual estimation. Therefore, the LVivoSG system can be used as a decision support tool for segmental wall motion evaluation and segmental strain.

Lang RM, Badano LP et al. Recommendations for Cardiac Chamber Quantification by Echocardiography in Adults:An Update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. JASE 2015;28:1-39.

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).