Can a Smartphone Listen to Your Heart? A Performance Study on Detecting Abnormalities in Your Heart Sounds
Study Details
Study Description
Brief Summary
This observational study aims to assess the performance of the software called ausculto™. ausculto™ is a collection of computer algorithms that intend to analyse heart sounds recorded from the built-in microphone of a smartphone for abnormal sounds.
Participants will have their heart sounds recorded during their regular clinic appointment after consenting to participate in this study.
Researchers will manually annotate the recorded heart sounds to create a database for use in future training and testing of artificial intelligence (AI) intended for medical uses.
| Condition or Disease | Intervention/Treatment | Phase |
|---|---|---|
|
Study Design
Arms and Interventions
| Arm | Intervention/Treatment |
|---|---|
| Abnormal Echocardiography Diagnosis Participants with abnormal echocardiography diagnosis within three years. |
Diagnostic Test: Computer algorithms
Collection of lightweight computer algorithms called ausculto™ that is designed to perform real-time heart sound analysis to detect heart murmur.
Other Names:
|
| Normal Echocardiography Diagnosis Participants with normal echocardiography diagnosis within three years. |
Diagnostic Test: Computer algorithms
Collection of lightweight computer algorithms called ausculto™ that is designed to perform real-time heart sound analysis to detect heart murmur.
Other Names:
|
Outcome Measures
Primary Outcome Measures
- Sensitivity of ausculto™ Using Auscultation Paired with Echocardiography as Ground Truth [Day 0]
Sensitivity is defined as the percentage of true positives out of all positive cases from the ground truth. Cardiologist electronic stethoscope heart sound recording annotations paired with echocardiography diagnosis will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
- Specificity of ausculto™ Using Auscultation Paired with Echocardiography as Ground Truth [Day 0]
Specificity is defined as the percentage of true negatives out of all negative cases from the ground truth. Cardiologist electronic stethoscope heart sound recording annotations paired with echocardiography diagnosis will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
Secondary Outcome Measures
- Sensitivity of ausculto™ Using Auscultation as Ground Truth [Day 0]
Sensitivity is defined as the percentage of true positives out of all positive cases from the ground truth. Cardiologist electronic stethoscope heart sound recording annotations will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
- Specificity of ausculto™ Using Auscultation as Ground Truth [Day 0]
Specificity is defined as the percentage of true negatives out of all negative cases from the ground truth. Cardiologist electronic stethoscope heart sound recording annotations will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
- Sensitivity of ausculto™ Using Echocardiography as Ground Truth [Day 0]
Sensitivity is defined as the percentage of true positives out of all positive cases from the ground truth. Echocardiography diagnosis will be used as the ground truth.
- Specificity of ausculto™ Using Echocardiography as Ground Truth [Day 0]
Specificity is defined as the percentage of true negatives out of all negative cases from the ground truth. Echocardiography diagnosis will be used as the ground truth.
- Positive Predictive Value of ausculto™ Using Auscultation Paired with Echocardiography as Ground Truth [Day 0]
Positive predictive value is defined as the percentage of true positives out of all positive results from ausculto™. Cardiologist electronic stethoscope heart sound recording annotations paired with echocardiography diagnosis will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
- Negative Predictive Value of ausculto™ Using Auscultation Paired with Echocardiography as Ground Truth [Day 0]
Negative predictive value is defined as the percentage of true negatives out of all negative results from ausculto™. Cardiologist electronic stethoscope heart sound recording annotations paired with echocardiography diagnosis will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
- Positive Predictive Value of ausculto™ Using Auscultation as Ground Truth [Day 0]
Positive predictive value is defined as the percentage of true positives out of all positive results from ausculto™. Cardiologist electronic stethoscope heart sound recording annotations will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
- Negative Predictive Value of ausculto™ Using Auscultation as Ground Truth [Day 0]
Negative predictive value is defined as the percentage of true negatives out of all negative results from ausculto™. Cardiologist electronic stethoscope heart sound recording annotations will be used as the ground truth. Heart sound recordings from electronic stethoscope will be rated by two independent cardiologists for presence or absence of murmur, with a third cardiologist consulted in case of disagreement.
- Positive Predictive Value of ausculto™ Using Echocardiography as Ground Truth [Day 0]
Positive predictive value is defined as the percentage of true positives out of all positive results from ausculto™. Echocardiography diagnosis will be used as the ground truth.
- Negative Predictive Value of ausculto™ Using Echocardiography as Ground Truth [Day 0]
Negative predictive value is defined as the percentage of true negatives out of all negative results from ausculto™. Echocardiography diagnosis will be used as the ground truth.
- Number of Adverse Events Reported by Participants Deemed to be Caused by the Observational Procedures [Day 0 to Day 1]
Eligibility Criteria
Criteria
Inclusion Criteria:
-
Age: ≥22 years (adult)
-
Outpatients who have undergone echocardiography within 3 years at their normal non-research follow-up clinic visit appointment
Exclusion Criterion:
- Implanted active medical devices in the torso, such as pacemakers and defibrillators
Contacts and Locations
Locations
| Site | City | State | Country | Postal Code | |
|---|---|---|---|---|---|
| 1 | Queen Mary Hospital | Hong Kong | Hong Kong SAR | China |
Sponsors and Collaborators
- Wong Chun Ka
- Laboratory of Data Discovery for Health
- The University of Hong Kong
Investigators
- Principal Investigator: Chun Ka Wong, MBBS, The University of Hong Kong
Study Documents (Full-Text)
None provided.More Information
Publications
None provided.- AUSC1