Search Results
Found 1 results
510(k) Data Aggregation
(113 days)
EchoPAC Software Only / EchoPAC Plug-in is intended for diagnostic review and analysis of ultrasound images, patient record management and reporting, for use by, or on the order of a licensed physician. EchoPAC Software Only / EchoPAC Plug-in allows post-processing of raw data images from GE ultrasound scanners and DICOM ultrasound images.
Ultrasound images are acquired via B (2D), M, Color M modes, Color, Power, Pulsed and CW Doppler modes, Coded Pulse, Harmonic,3D, and Real time (RT) 3D Mode (4D).
Clinical applications include: Fetal/Obstetrics; Abdominal (including renal and GYN); Urology (including prostate); Pediatric; Small organs (breast, testes, thyroid); Neonatal and Adult and pediatric); Peripheral Vascular; Transesophageal (TEE); Musculo-skeletal Conventional; Musculo-skeletal Superficial; Transrectal (TR); Transvaginal (TV): Intraoperative ( vascular); Intra-Cardiac; Thoracic/Pleural and Intra-Luminal.
EchoPAC Software Only / EchoPAC Plug-in provides image processing, annotation, analysis, measurement, report generation, communication, storage and retrieval functionality to ultrasound images that are acquired via the GE Healthcare Vivid family of ultrasound systems, as well as DICOM images from other ultrasound systems. EchoPAC Software Only will be offered as SW only to be installed directly on customer PC hardware and EchoPAC Plug-in is intended to be hosted by a generalized PACS host workstation. EchoPAC Software Only / EchoPAC Plug-in is DICOM compliant, transferring images and data via LAN between systems, hard copy devices, file servers and other workstations.
The provided FDA 510(k) summary for GE Medical Systems Ultrasound and Primary Care Diagnostics, LLC's EchoPAC Software Only/EchoPAC Plug-in includes an "AI Summary of Testing" section for the Easy Auto EF and Easy AFI LV algorithms. This section provides information relevant to acceptance criteria and study details.
Here's a breakdown of the requested information based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The acceptance criteria are implied by the reported performance figures, as they state the accuracy achieved.
| Acceptance Criteria (Implied) | Reported Device Performance (Accuracy) |
|---|---|
| ≥ 92% average Dice score (general) | 92% or higher |
| ≥ 91% average Dice score (different scanning views) | 91% or higher |
| ≥ 92% average Dice score (different left ventricle volumes) | 92% or higher |
Note: The document only provides Dice score for "accuracy." Other common performance metrics like sensitivity, specificity, or F1-score are not explicitly stated.
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size for Test Set:
- Individual patients' images: 45 exams from assumed 45 patients (exact number of patients unknown due to anonymization).
- Number of samples (images): 135 images extracted from the 45 exams.
- Data Provenance: Europe, Asia, US (retrospective, as indicated by anonymization and collection for testing).
3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications
- Number of Experts: Two certified cardiologists initially, with a panel of experienced experts for adjudication.
- Qualifications of Experts:
- Two certified cardiologists (for initial manual delineation and review).
- A panel of experienced experts (for reviewing annotations that the two cardiologists could not agree on). Specific years of experience are not mentioned beyond "experienced."
4. Adjudication Method
The adjudication method used was a 2+1 process (consensus followed by expert panel review):
- Consensus Reading: Two certified cardiologists performed manual delineation and then reviewed each other's annotations. They discussed disagreements to reach a consensus.
- Expert Panel Review: If the two cardiologists could not agree on an annotation, a panel of experienced experts further reviewed those annotations to establish the final ground truth.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
The document does not mention a Multi-Reader Multi-Case (MRMC) comparative effectiveness study. It focuses on the standalone performance of the AI algorithm. Therefore, no effect size of human readers improving with AI vs. without AI assistance is provided.
6. Standalone Performance Study
Yes, a standalone (i.e., algorithm-only without human-in-the-loop performance) study was done. The reported Dice scores directly evaluate the algorithm's accuracy in segmenting regions of interest, independent of human interaction during the measurement process.
7. Type of Ground Truth Used
The type of ground truth used was expert consensus. It was derived from manual delineations by certified cardiologists, with a further review and consensus by an expert panel for disagreements.
8. Sample Size for the Training Set
The document does not explicitly state the sample size for the training set. It only mentions that "datasets from different clinical sites for testing as compared to the clinical sites for training" were used.
9. How the Ground Truth for the Training Set Was Established
The document does not explicitly state how the ground truth for the training set was established. It only describes the process for the test set's ground truth. However, it is generally assumed that similar expert-driven annotation processes would have been used for training data.
Ask a specific question about this device
Page 1 of 1