(17 days)
The Viamo SSA-640 v2.0 Ultrasound System is indicated for the visualization of structures, characteristics, and dynamic processes with the human body using ultrasound and to provide image information for diagnosis in the following clinical applications: fetal, abdominal, pediatric, small organs, trans-vaginal neonatal cephalic, adult cephalic, cardiac, peripheral vascular, and musculoskeletal (both conventional and superficial).
The Viamo is a mobile system. It is a Track 3 device that employs a wide range of probes that include flat linear array, convex array and sector array with a frequency range of approximately 2.5 MHz to 12 MHz.
The provided text describes the Toshiba Viamo SSA-640A Ultrasound System (v2.0), cleared under 510(k) K100067. This submission is for an ultrasound device and its various transducers, and as such, the document does not contain acceptance criteria, detailed study designs (test set, training set, ground truth, expert qualifications, or adjudication methods), or performance metrics typically associated with AI/CADe (Computer-Aided Detection) or CADx (Computer-Aided Diagnosis) devices.
Ultrasound systems are cleared based on substantial equivalence to predicate devices, demonstrating that they are as safe and effective. The clearance process for such devices typically involves:
- Conformance to standards: Demonstrating compliance with recognized electrical safety, electromagnetic compatibility, and acoustic output standards (e.g., IEC 60601-1, AIUM-NEMA UD2, UD3).
- Performance testing: Verifying that the system and its transducers meet specified technical and imaging performance parameters (e.g., image resolution, penetration, uniformity, Doppler accuracy). These tests are usually conducted in-house by the manufacturer and are not usually detailed in summary documents like 510(k) summaries.
- Comparison to predicate device: Providing evidence that the new device has the same intended use and similar technological characteristics as a legally marketed predicate device, or if there are differences, that those differences do not raise new questions of safety and effectiveness.
Therefore, the specific information requested in the prompt (acceptance criteria table, sample sizes, ground truth details, MRMC study, standalone performance) is not typically found in a 510(k) summary for a general diagnostic ultrasound system. This type of information is more relevant for AI/Machine Learning-enabled medical devices or those with specific diagnostic claims where performance against a clinical ground truth is central to the clearance.
If this were an AI/CADe/CADx device, here's how the information would be presented, based on common regulatory expectations (but this information is not in the provided text):
Acceptance Criteria and Device Performance for an AI/CADe/CADx Device (Hypothetical, not from the provided text)
The provided 510(k) summary (K100067) is for a general diagnostic ultrasound system and its transducers. It does not include specific acceptance criteria or performance studies in the context of AI/CADe/CADx devices. Ultrasound systems are typically cleared based on demonstrating substantial equivalence to predicate devices and adherence to relevant industry standards for safety and performance (e.g., image quality, acoustic output).
Therefore, the following sections are hypothetical, illustrating what would be expected for an AI/CADe/CADx device making specific diagnostic claims, rather than a general imaging system.
1. Table of Acceptance Criteria and Reported Device Performance (Hypothetical)
| Performance Metric | Acceptance Criteria (e.g., for specific clinical task) | Reported Device Performance (e.g., from clinical study) |
|---|---|---|
| Sensitivity | ≥ 90% for detecting [condition X] | 92% (95% CI: 89-94%) |
| Specificity | ≥ 80% for detecting [condition X] | 83% (95% CI: 80-85%) |
| AUC | ≥ 0.90 for distinguishing [condition X] from normal | 0.93 (95% CI: 0.91-0.95) |
2. Sample Size and Data Provenance for Test Set (Hypothetical)
- Sample Size (Test Set): Typically, hundreds to thousands of cases are used, depending on the prevalence of the condition and endpoint variability. For example, 500 cases (250 positive, 250 negative).
- Data Provenance: Retrospective, multi-site data from [e.g., multiple hospitals in the United States, Europe, and Asia]. Cases selected to ensure diversity in patient demographics, disease presentation, and image quality.
3. Number and Qualifications of Experts for Ground Truth (Hypothetical)
- Number of Experts: 3-5 independent experts.
- Qualifications: Board-certified radiologists with 5-15 years of experience specializing in [e.g., abdominal, fetal, small organ] ultrasound imaging.
4. Adjudication Method for Test Set (Hypothetical)
- Adjudication Method: 2+1 adjudication. Initial assessment by two experts; if their assessments differed, a third senior expert would resolve the discrepancy. Unanimous agreement (or majority agreement for three experts) established the final ground truth.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study (Hypothetical)
- Was an MRMC study done? Yes.
- Effect Size (Human Readers with vs. without AI Assistance):
- Improvement in AUC: Readers improved their diagnostic accuracy (measured by AUC) by an average of 0.05 (e.g., from 0.85 without AI to 0.90 with AI). This translates to a [e.g., 5.9%] relative improvement in AUC.
- Improvement in Sensitivity: Average sensitivity increased by [e.g., 8 percentage points] (e.g., from 82% to 90%).
- Reduction in False Positives: Average false positive rate decreased by [e.g., 10%] (e.g., from 25% to 15%).
- Reading Time Reduction: Average reading time per case decreased by [e.g., 20%].
6. Standalone Performance Study (Hypothetical)
- Was a standalone study done? Yes.
- Standalone Performance: The algorithm achieved a standalone sensitivity of 91% and a specificity of 82% on the independent test set, with an AUC of 0.92.
7. Type of Ground Truth Used (Hypothetical)
- Type of Ground Truth: Established through expert consensus (as described in section 3 & 4), cross-referenced with confirmatory pathology reports and/or long-term patient outcomes where available.
8. Sample Size for Training Set (Hypothetical)
- Sample Size (Training Set): Typically tens of thousands to hundreds of thousands of images/cases, depending on the complexity of the task and the diversity of the data required. For example, 50,000 cases (25,000 positive, 25,000 negative) from diverse sources.
9. How Ground Truth for Training Set Was Established (Hypothetical)
- Ground Truth Establishment for Training Set: Ground truth was established by a team of trained medical professionals (e.g., sonographers, residents, and attending physicians), often using a single-reader read or two-reader consensus, with a more rigorous multi-expert adjudication applied to a subset of challenging cases or for quality control. Data was also verified against available patient records, pathology, or clinical follow-up.
§ 892.1570 Diagnostic ultrasonic transducer.
(a)
Identification. A diagnostic ultrasonic transducer is a device made of a piezoelectric material that converts electrical signals into acoustic signals and acoustic signals into electrical signals and intended for use in diagnostic ultrasonic medical devices. Accessories of this generic type of device may include transmission media for acoustically coupling the transducer to the body surface, such as acoustic gel, paste, or a flexible fluid container.(b)
Classification. Class II.