(258 days)
The NeuViz 128 Multi-Slice CT Scanner System can be used as a whole body computed tomography X-ray system featuring a continuously rotating X-ray tube and detector array. The acquired X-RAY transmission data is reconstructed by computer into cross-sectional images of the body from either the same axial plane taken at different angles or spiral planes taken at different angles.
The proposed NeuViz 128 Multi-slice CT Scanner System is composed of a gantry, a patient couch, an operator console and includes image acquisition hardware and software, and associated accessories. It is designed to be a head and whole body X-ray computed tomography scanner which features a continuously rotating tube-detector system and functions according to the fan beam principle.
Here's a summary of the acceptance criteria and the study that proves the device meets them, based on the provided text.
1. Table of Acceptance Criteria and Reported Device Performance
The document describes performance in terms of improvements over Filtered Back Projection (FBP) reconstruction. It does not explicitly state pre-defined "acceptance criteria" with numerical thresholds that the device must meet, but rather reports the observed performance improvements from the testing. The "acceptance criteria" are implied to be achieving demonstrably equivalent or better diagnostic image quality and performance compared to FBP, particularly with dose reduction and improved image characteristics.
| Performance Metric | "ClearView" Performance (vs FBP) | "ClearInfinity" Performance (vs FBP) |
|---|---|---|
| Head Dose Reduction (Same IQ) | 40% to 50% lower | 50% to 60% lower |
| Body Dose Reduction (Same IQ) | 45% to 60% lower | 75% to 85% lower |
| Head Low-Contrast Detectability (Same Dose) | 30% to 35% improvement | 35% to 65% improvement |
| Body Low-Contrast Detectability (Same Dose) | 35% to 45% improvement | 60% to 115% improvement |
| Image Noise Reduction (Same Dose) | Up to 55% reduction | Up to 90% reduction |
| Spatial Resolution (Same Image Noise) | Not specified (implied similar/equivalent) | Up to 1.88X (88%) improvement |
| Diagnostic Quality (Clinical cases) | Not explicitly stated (no specific clinical reader study for ClearView mentioned) | Equivalent or better than FBP |
| Diagnostic Quality (Low-dose clinical cases) | Not explicitly stated | Equivalent or better than FBP |
2. Sample Size for the Test Set and Data Provenance
-
"ClearInfinity" Performance Verification (Quantitative):
- Sample Size: Not explicitly stated as "test set" for the quantitative phantom studies. These are engineering bench tests using raw datasets.
- Data Provenance: Raw datasets obtained on NeuViz 128. Implied to be internal testing.
-
"ClearInfinity" Clinical Image Evaluation (Reader Study):
- Sample Size:
- 30 retrospectively collected clinical cases (for normal dose comparison).
- 20 retrospectively collected clinical cases (for simulated low dose comparison).
- Data Provenance: Retrospectively collected clinical cases.
- For general performance claims, the document states: "US patient-based testing data for ClearInfinity collected from three sites of OK and one site of TX, a total of 445 clinical cases of 297 independent patients." It's unclear if these 445 cases directly constitute the raw data for the specific 30 and 20 cases in the reader study, or a larger pool from which they were selected.
- Sample Size:
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications
- "ClearInfinity" Clinical Image Evaluation (Reader Study):
- Number of Experts: 3 board-certified radiologists.
- Qualifications: "board-certified radiologists." No specific years of experience are provided.
4. Adjudication Method
- "ClearInfinity" Clinical Image Evaluation (Reader Study):
- The document states: "Each image was read by 3 board-certified radiologists who provided an assessment of image quality related to diagnostic use according to a 5-point Likert scale." It does not specify an adjudication method (like 2+1 or 3+1) if there were disagreements among the readers. It implies individual assessments were aggregated or used to determine the overall conclusion of equivalence or betterment.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- Was an MRMC study done? Yes, a reader study was performed for "ClearInfinity."
- Effect Size: The document states that the results "indicate that ClearInfinity is equivalent or better than Filtered Back Projection in diagnostic quality" for both normal and low-dose cases. It does not provide a specific quantitative effect size (e.g., AUC improvement, percentage increase in concordance) for how much human readers improve with AI assistance. It focuses on the comparative diagnostic quality of images produced by ClearInfinity versus FBP.
6. Standalone Performance Study (i.e., algorithm only without human-in-the-loop performance)
- Yes, standalone performance was extensively evaluated through "Engineering bench testing" using phantoms. This included quantitative measurements of:
- Dose reduction capability
- Low-contrast detectability improvement
- Image noise reduction
- Spatial resolution improvement (for ClearInfinity)
7. Type of Ground Truth Used
- Engineering Bench Testing (Phantom Studies): For quantitative metrics, the "ground truth" is established by the known properties of the phantoms (CCT189 and CCT191 MITAIQ LC Phantoms, QA phantom water layer, high-contrast test device) and objective measurements using methods like the channelized Hotelling observer.
- Clinical Image Evaluation (Reader Study): The "ground truth" for diagnostic quality was established by the expert consensus/assessment of 3 board-certified radiologists using a 5-point Likert scale, comparing images reconstructed with FBP and ClearInfinity from the same raw data.
8. Sample Size for the Training Set
- Training Set for ClearInfinity: The document lists "US patient-based testing data for ClearInfinity collected from three sites of OK and one site of TX, a total of 445 clinical cases of 297 independent patients" as performance verification data. It also states ClearInfinity uses a "trained deep learning neural network." However, it does not explicitly state that these 445 cases were used as the training set. It's common for regulatory submissions to mention a larger pool of data used for development/testing, but the specific training set size is not detailed here.
9. How the Ground Truth for the Training Set Was Established
- The document mentions "a trained deep learning neural network" for ClearInfinity but does not describe how the ground truth for this training set was established. This information is typically proprietary and part of the manufacturer's internal development process for the AI algorithm.
§ 892.1750 Computed tomography x-ray system.
(a)
Identification. A computed tomography x-ray system is a diagnostic x-ray system intended to produce cross-sectional images of the body by computer reconstruction of x-ray transmission data from the same axial plane taken at different angles. This generic type of device may include signal analysis and display equipment, patient and equipment supports, component parts, and accessories.(b)
Classification. Class II.