The device is a general purpose ultrasound system intended for use by qualified and trained healthcare professionals. Specific clinical applications remain the same as previously cleared: Fetal/OB; Abdominal (including GYN, pelvic and infertility monitoring/follicle development); Pediatric; Small Organ (breast, testes, thyroid etc.); Neonatal and Adult Cephalic; Cardiac (adult and pediatric); Musculo-skeletal Conventional and Superficial; Vascular; Transvaginal (including GYN); Transrectal
Modes of operation include: B, M, PW Doppler, CW Doppler, Color M Doppler, Color M Doppler, Power Doppier, Harmonic Imaging, Coded Pulse, 3D/4D Imaging mode, Elastography, Shear Wave Elastography and Combined modes: B/M, B/Color, B/PWD, B/Color/PWD, B/Power/ PWD, B/Elastography. The Voluson™ Expert 20, Voluson™ Expert 22 is intended to be used in a hospital or medical clinic.
The systems are full-featured Track 3 ultrasound systems, primarily for general radiology use and specialized for OB/GYN with particular features for real-time 3D/4D acquisition. They consist of a mobile console with keyboard control panel; color LCD/TFT touch panel, color video display and optional image storage and printing devices. They provide high performance ultrasound imaging and analysis and have comprehensive networking and DICOM capability. They utilize a variety of linear, curved linear, matrix phased array transducers including mechanical and electronic scanning transducers, which provide highly accurate real-time three-dimensional imaging supporting all standard acquisition modes.
Here's a breakdown of the acceptance criteria and study details for the AI features of the Voluson Expert Series, based on the provided text:
Acceptance Criteria and Reported Device Performance
| AI Feature | Acceptance Criteria | Reported Device Performance |
|---|---|---|
| SonoPelvic Floor | Success rate of each AI component (MHD plane alignment, LH contour, and measurements) is expected to be 70% or higher on datasets marked as "Good Image Quality." Success rate of each AI component is expected to be 60% or higher on datasets marked as "Challenging Image Quality." | The document states "The success rate of each AI component of the feature (MHD plane alignment, LH contour and measurements) is expected to be 70% or higher. On datasets that are marked as challenging in image quality measure the success rate of each AI component of the feature should be 60% or higher." This implies that the device met these criteria, as there are no reported failures or instances where these thresholds were not achieved. The summary of testing states "Summary test statistics or other test results including acceptance criteria or other information supporting the appropriateness of the characterized performance." and then lists these as expectations. |
| SonoLyst | For SonoLystIR, sorting accuracy is higher than 80%. For SonoLystX, grading accuracy is higher than 80%. For SonoLystLive, accuracy is higher than 80%. | For SonoLystIR, the sorting accuracy is reported as "higher than 80%." For SonoLystX, the grading accuracy is reported as "higher than 80%." For SonoLystLive, the accuracy is reported as "higher than 80%." |
| FetalHS | The success rate of the 4CH view and 3VT view suggestion within cines is expected to be 70% or higher. The success rate of the suggested heart angle measurement is expected to be 80% or higher. | The document states "The success rate of the 4CH view and 3VT view suggestion within cines is expected to be 70% or higher. The success rate of the suggested heart angle measurement is expected to be 80% or higher." Similar to SonoPelvic Floor, this indicates that the device met these criteria. |
Study Details for AI Features
1. SonoPelvic Floor:
- Sample size for the test set: 110 3D/4D Volumes from 70+ individual patients.
- Data provenance: Images collected from Europe, Asia, and South Africa.
- Number of experts used to establish ground truth & qualifications: The results generated by the AI software were verified as "Pass or Fail" by a "certified sonographer/Clinician." The number of such sonographers/clinicians is not specified, but it implies at least one. Specific qualifications beyond "certified sonographer/Clinician" are not detailed.
- Adjudication method: Not explicitly stated, but the process involves a single expert verifying the AI output. There is no mention of multiple experts or a specific consensus method beyond single expert review.
- MRMC comparative effectiveness study: No. This was a standalone AI performance evaluation.
- Standalone (algorithm only) performance: Yes, the study evaluates the performance of the AI algorithm in generating results.
- Type of ground truth used: Expert verification by a "certified sonographer/Clinician."
- Sample size for the training set: Not specified, but it's stated that the volumes used for testing were "completely distinct from the ones used during training process and there is no overlap between the two."
- How the ground truth for the training set was established: Not specified.
2. SonoLyst:
- Sample size for the test set: 40,000+ images derived from 5,000+ exams (individual patients).
- Data provenance: Exams from the United Kingdom, Austria, India, USA. A mix of retrospective and prospective data collection from GE and non-GE scanners.
- Number of experts used to establish ground truth & qualifications:
- Initial curation (sorting and grading) by a "single Sonographer."
- Review panel of "5-sonographer[s]" for instances where SonoLyst differed from the initial ground truth, to determine sorting and grading accuracy. Qualifications beyond "sonographer" are not detailed.
- Adjudication method:
- Initial ground truth established by a single sonographer.
- Discrepancies between the AI and initial ground truth were reviewed by a "5-sonographer review panel." The sorting ground truth was reclassified based on the "majority view of the panel." This implies a form of 5-sonographer consensus, where the majority dictates the revised ground truth.
- MRMC comparative effectiveness study: No. This was a standalone AI performance evaluation.
- Standalone (algorithm only) performance: Yes, the study evaluates the performance of the AI algorithm in sorting and grading images.
- Type of ground truth used: Expert consensus, specifically from a panel of 5 sonographers for ambiguous cases, following initial assessment by a single sonographer.
- Sample size for the training set: Not specified, but it's stated that the exams used for testing were "separated from the ones used during training process and there is no overlap between the two."
- How the ground truth for the training set was established: Not specified.
3. FetalHS:
- Sample size for the test set: 400+ cines from 250+ patients.
- Data provenance: USA, India, Japan, Germany, Austria.
- Number of experts used to establish ground truth & qualifications: The AI software's results were verified as "Pass or Fail" by "certified sonographers/clinicians." The number of such sonographers/clinicians is not specified, but it implies at least one. Specific qualifications beyond "certified sonographers/clinicians" are not detailed.
- Adjudication method: Not explicitly stated, but the process involves a single expert verifying the AI output. There is no mention of multiple experts or specific consensus beyond single expert review.
- MRMC comparative effectiveness study: No. This was a standalone AI performance evaluation.
- Standalone (algorithm only) performance: Yes, the study evaluates the performance of the AI algorithm in suggesting views and measuring heart angles.
- Type of ground truth used: Expert verification by "certified sonographers/clinicians."
- Sample size for the training set: Not specified, but it's stated that the datasets used for testing were "completely distinct from the ones used during training process and there is no overlap between the two."
- How the ground truth for the training set was established: Not specified.
§ 892.1550 Ultrasonic pulsed doppler imaging system.
(a)
Identification. An ultrasonic pulsed doppler imaging system is a device that combines the features of continuous wave doppler-effect technology with pulsed-echo effect technology and is intended to determine stationary body tissue characteristics, such as depth or location of tissue interfaces or dynamic tissue characteristics such as velocity of blood or tissue motion. This generic type of device may include signal analysis and display equipment, patient and equipment supports, component parts, and accessories.(b)
Classification. Class II.