Trial Outcomes & Findings for PROSPECTIVE EVALUATION OF HEARTFOCUS (NCT NCT05874128)

NCT ID: NCT05874128

Last Updated: 2026-01-29

Results Overview

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the exam had sufficient image quality to allow visual analysis of left ventricular size. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as "Yes" (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as "No", the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Recruitment status

COMPLETED

Study phase

NA

Target enrollment

240 participants

Primary outcome timeframe

Images evaluated by the cardiologists after the acquisition

Results posted on

2026-01-29

Participant Flow

Adult patients (≥18 years) scheduled for a clinically indicated echocardiogram were consecutively recruited at two academic medical centers: University Hospital of Bordeaux (France) and Lenox Hill Hospital (New York, USA). Recruitment occurred between november 2023 and august 2024.

A total of 247 patients were screened. Of these, 246 provided informed consent; among these, 6 participants withdrew before undergoing the study echocardiography. Reasons included withdrawal of consent (n=3), scheduling conflicts (n=2), and technical difficulties (n=1). Thus, 240 patients were finally enrolled in this study, meaning that they completed both novice and expert echocardiography exams and were included in the Full Analysis Set.

Participant milestones

Participant milestones
Measure
Participants
240 patients will be included, 120 in Site 01 (Bordeaux) and 120 in Site 02 (New York). They all received two limited exams: one administered by a novice, and the other by an expert.
Overall Study
STARTED
240
Overall Study
COMPLETED
240
Overall Study
NOT COMPLETED
0

Reasons for withdrawal

Withdrawal data not reported

Baseline Characteristics

PROSPECTIVE EVALUATION OF HEARTFOCUS

Baseline characteristics by cohort

Baseline characteristics by cohort
Measure
France
n=120 Participants
This group consisted of 120 patients enrolled at the participating University Hospital of Bordeaux, France.
United States
n=120 Participants
This group consisted of 120 patients enrolled at Lenox Hill Hospital, New York, United States.
Total
n=240 Participants
Total of all reporting groups
Sex: Female, Male
Female
47 Participants
n=35 Participants
70 Participants
n=4328 Participants
117 Participants
n=8687 Participants
Sex: Female, Male
Male
73 Participants
n=35 Participants
50 Participants
n=4328 Participants
123 Participants
n=8687 Participants
Ethnicity (NIH/OMB)
Hispanic or Latino
0 Participants
n=35 Participants
18 Participants
n=4328 Participants
18 Participants
n=8687 Participants
Ethnicity (NIH/OMB)
Not Hispanic or Latino
0 Participants
n=35 Participants
90 Participants
n=4328 Participants
90 Participants
n=8687 Participants
Ethnicity (NIH/OMB)
Unknown or Not Reported
120 Participants
n=35 Participants
12 Participants
n=4328 Participants
132 Participants
n=8687 Participants
BMI (Body Mass Index)
<25
56 Participants
n=35 Participants
41 Participants
n=4328 Participants
97 Participants
n=8687 Participants
BMI (Body Mass Index)
≤25-30
44 Participants
n=35 Participants
49 Participants
n=4328 Participants
93 Participants
n=8687 Participants
BMI (Body Mass Index)
≥30
20 Participants
n=35 Participants
30 Participants
n=4328 Participants
50 Participants
n=8687 Participants
Race
White
0 Participants
n=35 Participants
52 Participants
n=4328 Participants
52 Participants
n=8687 Participants
Race
Black/African American
0 Participants
n=35 Participants
36 Participants
n=4328 Participants
36 Participants
n=8687 Participants
Race
Asian
0 Participants
n=35 Participants
3 Participants
n=4328 Participants
3 Participants
n=8687 Participants
Race
American Indian/Alaska Native
0 Participants
n=35 Participants
0 Participants
n=4328 Participants
0 Participants
n=8687 Participants
Race
Unknown/not reported
120 Participants
n=35 Participants
11 Participants
n=4328 Participants
131 Participants
n=8687 Participants
Race
Other
0 Participants
n=35 Participants
18 Participants
n=4328 Participants
18 Participants
n=8687 Participants
Age, Customized
64.7 Years
n=35 Participants
60.5 Years
n=4328 Participants
62.6 Years
n=8687 Participants
Prior cardiac diagnoses
Hypertension
28.3 Percentage of participants (%)
n=35 Participants
55.0 Percentage of participants (%)
n=4328 Participants
41.7 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Hyperlipidemia
21.7 Percentage of participants (%)
n=35 Participants
58.3 Percentage of participants (%)
n=4328 Participants
40.0 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Diabetes
15.8 Percentage of participants (%)
n=35 Participants
19.2 Percentage of participants (%)
n=4328 Participants
17.5 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Heart failure
6.7 Percentage of participants (%)
n=35 Participants
11.7 Percentage of participants (%)
n=4328 Participants
9.2 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Atrial fibrillation
20.8 Percentage of participants (%)
n=35 Participants
8.3 Percentage of participants (%)
n=4328 Participants
14.6 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Other arrhythmias
7.5 Percentage of participants (%)
n=35 Participants
7.5 Percentage of participants (%)
n=4328 Participants
7.5 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Coronary artery disease
25.0 Percentage of participants (%)
n=35 Participants
15.8 Percentage of participants (%)
n=4328 Participants
20.4 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Prior heart attack
11.7 Percentage of participants (%)
n=35 Participants
0.0 Percentage of participants (%)
n=4328 Participants
5.8 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Valvular heart disease
25.8 Percentage of participants (%)
n=35 Participants
9.2 Percentage of participants (%)
n=4328 Participants
17.5 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Pulmonary hypertension
3.3 Percentage of participants (%)
n=35 Participants
2.5 Percentage of participants (%)
n=4328 Participants
2.9 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Heart transplant
3.3 Percentage of participants (%)
n=35 Participants
0.8 Percentage of participants (%)
n=4328 Participants
2.1 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Cardiomyopathies
15.0 Percentage of participants (%)
n=35 Participants
5.8 Percentage of participants (%)
n=4328 Participants
10.4 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Congenital heart disease
0.0 Percentage of participants (%)
n=35 Participants
0.8 Percentage of participants (%)
n=4328 Participants
0.4 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Other
4.2 Percentage of participants (%)
n=35 Participants
10.0 Percentage of participants (%)
n=4328 Participants
7.1 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
None
17.5 Percentage of participants (%)
n=35 Participants
17.5 Percentage of participants (%)
n=4328 Participants
17.5 Percentage of participants (%)
n=8687 Participants
Prior cardiac diagnoses
Not reported
0.8 Percentage of participants (%)
n=35 Participants
0.8 Percentage of participants (%)
n=4328 Participants
0.8 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Renal disease
9.2 Percentage of participants (%)
n=35 Participants
10.0 Percentage of participants (%)
n=4328 Participants
9.6 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
COPD/emphysema
5.0 Percentage of participants (%)
n=35 Participants
4.2 Percentage of participants (%)
n=4328 Participants
2.9 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Pulmonary embolus
1.7 Percentage of participants (%)
n=35 Participants
4.2 Percentage of participants (%)
n=4328 Participants
2.9 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Systemic infiltrative disease
0 Percentage of participants (%)
n=35 Participants
0.8 Percentage of participants (%)
n=4328 Participants
0.4 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Cancer
18.3 Percentage of participants (%)
n=35 Participants
21.7 Percentage of participants (%)
n=4328 Participants
20.0 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Underwent chemotherapy
1.7 Percentage of participants (%)
n=35 Participants
10.0 Percentage of participants (%)
n=4328 Participants
5.8 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Underwent radiation
2.5 Percentage of participants (%)
n=35 Participants
2.5 Percentage of participants (%)
n=4328 Participants
2.5 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Other
22.5 Percentage of participants (%)
n=35 Participants
38.3 Percentage of participants (%)
n=4328 Participants
30.4 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
None
48.3 Percentage of participants (%)
n=35 Participants
36.7 Percentage of participants (%)
n=4328 Participants
42.5 Percentage of participants (%)
n=8687 Participants
Prior noncardiac diagnoses
Not reported
0.8 Percentage of participants (%)
n=35 Participants
2.5 Percentage of participants (%)
n=4328 Participants
1.7 Percentage of participants (%)
n=8687 Participants

PRIMARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the exam had sufficient image quality to allow visual analysis of left ventricular size. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as "Yes" (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as "No", the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Left Ventricular Size (Qualitative Visual Assessment)
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

PRIMARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of left ventricular function. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as "Yes" (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as "No", the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Left Ventricular Function (Qualitative Visual Assessment)
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

PRIMARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of right ventricular size. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as "Yes" (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as "No", the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Right Ventricle Size (Qualitative Visual Assessment)
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

PRIMARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound exam had sufficient image quality to visually analyze the presence of non-trivial pericardial effusion. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Presence of Non-trivial Pericardial Effusion (Qualitative Visual Assessment)
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of right ventricular function. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Function of the Right Ventricle (Qualitative Visual Assessment)
99.6 Percentage of scans obtained (%)
Interval 97.7 to 99.9
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of left atrium size. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Left Atrium Size (Qualitative Visual Assessment)
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of right atrium size. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Right Atrium Size (Qualitative Visual Assessment)
98.8 Percentage of scans obtained (%)
Interval 96.4 to 99.6
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of the segmental kinetics of the left ventricle. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Segmental Kinetics of the Left Ventricle (Qualitative Visual Assessment)
95.4 Percentage of scans obtained (%)
Interval 92.0 to 97.4
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of the aortic valve. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Aortic Valve (Qualitative Visual Assessment)
98.8 Percentage of scans obtained (%)
Interval 96.4 to 99.6
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of the mitral valve. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Mitral Valve (Qualitative Visual Assessment)
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0
100 Percentage of scans obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of the tricuspid valve. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Tricuspid Valve (Qualitative Visual Assessment)
95.4 Percentage of scans obtained (%)
Interval 92.0 to 97.4
99.2 Percentage of scans obtained (%)
Interval 97.0 to 99.8

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the ultrasound examination had sufficient image quality to allow visual analysis of the size of the inferior vena cava. Each exam was independently reviewed by five cardiologists. If at least three cardiologists rated the exam as 'Yes' (sufficient quality), the exam was classified as having sufficient image quality; if at least three rated it as 'No', the exam was classified as not sufficient. The results represent the percentage of scans evaluated as having sufficient image quality.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
The Size of Inferior Vena Cava (Qualitative Visual Assessment)
78.3 Percentage of scans obtained (%)
Interval 72.7 to 83.1
98.3 Percentage of scans obtained (%)
Interval 95.8 to 99.4

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Apical-2-Chamber (A2C) clip had sufficient image quality for interpretation, in order to determine the entire endocardial contour of the left ventricle, calculate ejection fraction by Simpson's method, and detect wall motion abnormalities. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Apical-2-Chamber (Diagnostic Quality Clip)
82.5 Percentage of DQC obtained (%)
Interval 77.2 to 86.8
98.3 Percentage of DQC obtained (%)
Interval 95.8 to 99.4

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Apical-3-Chamber (A3C) clip had sufficient image quality for interpretation, in order to determine the endocardial definition of the left ventricle, detect wall motion abnormalities, and ensure that the right ventricle and aortic valve leaflets were correctly visualized. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Apical-3-Chamber (Diagnostic Quality Clip)
90 Percentage of DQC obtained (%)
Interval 85.6 to 93.2
100 Percentage of DQC obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Apical-4-Chamber (A4C) clip had sufficient image quality for interpretation, in order to determine the entire endocardial contour of the left ventricle, calculate ejection fraction by Simpson's method, and detect wall motion abnormalities. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Apical-4-Chamber (Diagnostic Quality Clip)
96.2 Percentage of DQC obtained (%)
Interval 93.0 to 98.0
100 Percentage of DQC obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

"For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Apical-5-Chamber (A5C) clip had sufficient image quality for interpretation, in order to determine the endocardial definition of the left ventricle and ensure proper visualization of the outflow chamber and aortic valve. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Apical-5-Chamber (Diagnostic Quality Clip)
93.3 Percentage of DQC obtained (%)
Interval 89.4 to 95.9
99.2 Percentage of DQC obtained (%)
Interval 97.0 to 99.8

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Parasternal Long-Axis (PLAX) clip had sufficient image quality for interpretation, in order to ensure that the basal segments of the left ventricle were visible with adequate image quality for diameter measurement and detection of wall motion abnormalities, and to confirm that the long axis, aortic valve, and right ventricle were correctly visualized to consider the view as a good PLAX. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Parasternal Long Axis (Diagnostic Quality Clip)
97.5 Percentage of DQC obtained (%)
Interval 94.7 to 98.8
100 Percentage of DQC obtained (%)
Interval 98.4 to 100.0

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Parasternal Short-Axis Aortic (PSAX-AV) clip had sufficient image quality for interpretation, in order to ensure adequate image quality to visualize the aortic valve leaflets. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Parasternal Short Axis Aortic (Diagnostic Quality Clip)
89.2 Percentage of DQC obtained (%)
Interval 84.6 to 92.5
99.2 Percentage of DQC obtained (%)
Interval 97.0 to 99.8

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Parasternal Short-Axis Mitral (PSAX-MV) clip had sufficient image quality for interpretation, in order to ensure adequate image quality to detect wall motion abnormalities and visualize valvular movement. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Parasternal Short Axis Mitral (Diagnostic Quality Clip)
90.8 Percentage of DQC obtained (%)
Interval 86.5 to 93.9
99.6 Percentage of DQC obtained (%)
Interval 97.7 to 99.9

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Parasternal Short-Axis Papillary Muscles (PSAX-PM) clip had sufficient image quality for interpretation, in order to ensure adequate image quality to detect wall motion abnormalities, visualize valvular movement, and visualize both papillary muscles. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Parasternal Short Axis Papillary Muscles (Diagnostic Quality Clip)
97.1 Percentage of DQC obtained (%)
Interval 94.1 to 98.6
99.2 Percentage of DQC obtained (%)
Interval 97.0 to 99.8

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Subcostal 4-Chamber (SC-4C) clip had sufficient image quality for interpretation, in order to ensure adequate image quality for detecting a pericardial effusion between the right ventricle and liver. Full visualization of the left and right ventricles was not required. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Sub Costal-4-Chamber (Diagnostic Quality Clip)
89.2 Percentage of DQC obtained (%)
Interval 84.6 to 92.5
98.3 Percentage of DQC obtained (%)
Interval 95.8 to 99.4

SECONDARY outcome

Timeframe: Images evaluated by the cardiologists after the acquisition

Population: All 240 participants received two ultrasound examinations: one performed by novice operators using HeartFocus AI Guidance, and one performed by expert operators using HeartFocus without AI Guidance. Therefore, the 240 individuals included in the 'Novice' and 'Expert' result groups are the same participants. The groups were separated solely to present the performance results obtained by novices versus experts on the same 240 participants, not because they represent distinct populations.

For each participant, one novice and one expert performed a limited echocardiographic exam on the same patient, so that the quality of each exam could be evaluated by cardiologists. Blinded cardiologists evaluated whether the Subcostal Inferior Vena Cava (SC-IVC) clip had sufficient image quality for interpretation, in order to ensure adequate image quality to measure the diameter of the IVC a few centimeters before it reaches the heart. Each clip was independently reviewed by five cardiologists. If at least three cardiologists rated the clip as 'Yes' (sufficient quality), the clip was classified as a Diagnostic Quality Clip (DQC); if at least three rated it as 'No', the clip was classified as not sufficient. The results represent the percentage of Diagnostic Quality Clips (DQC) obtained.

Outcome measures

Outcome measures
Measure
Novice With AI Guidance
n=240 Participants
Nurses without prior ultrasound experience performed limited echocardiograms using the HeartFocus AI-based guidance software (DESKi). After a half-day training and practice on ≤9 patients, each novice acquired echocardiographic clips for 10 standard transthoracic views with the assistance of real-time software guidance and automated recording.
Expert Without AI Guidance
n=240 Participants
Experienced sonographers or cardiologists performed limited echocardiograms using HeartFocus software without AI assistance. Experts acquired echocardiographic clips for the same 10 standard transthoracic views as the novices.
Sub Costal Inferior Vena Cava (Diagnostic Quality Clip)
77.5 Percentage of DQC obtained (%)
Interval 71.8 to 82.3
98.3 Percentage of DQC obtained (%)
Interval 95.8 to 99.4

Adverse Events

Participants

Serious events: 0 serious events
Other events: 0 other events
Deaths: 0 deaths

Serious adverse events

Adverse event data not reported

Other adverse events

Adverse event data not reported

Additional Information

Head of Clinical Operations

DESKi

Phone: xxxxxx

Results disclosure agreements

  • Principal investigator is a sponsor employee
  • Publication restrictions are in place