Search Results
Found 1 results
510(k) Data Aggregation
(255 days)
This product is used in medical institutions, it can carry out 3D reconstruction, segmentation and semi-automatic identification of key points on CT images conforming to DICOM3.0 standard format, and assist physicians to overlay digital templates to assist surgeons with preoperative planning of adult hip and total knee replacement surgery. This product is intended for use by trained and qualified physicians and cannot be used as a basis for clinical diagnosis and treatment decisions alone.
The proposed device, Orthopaedic Surgery Planning Software, is standalone software which is designed to help surgeons' specialists carry out the pre-operative planning for an adult hip and total knee replacement surgery for several surgical procedures, based on the patient's CT images. The Orthopaedic Surgery Planning Software can a carry out 3D reconstruction, segmentation and semi-automated identification of anatomical sites of CT images conforming to DICOM3.0 standard format, and assist physicians to simulate adult hip and total knee replacement surgery.
The provided text describes the Orthopaedic Surgery Planning Software (AIJOINT) and its non-clinical test conclusions, which serve as proof that the device meets acceptance criteria. Here's a breakdown of the requested information:
1. Table of acceptance criteria and the reported device performance
| Acceptance Criteria (Measurement) | Stated Acceptance Criteria | Reported Device Performance |
|---|---|---|
| Key Point Recognition Accuracy | Error distance below 1mm | All points within 95% CI on Bland-Altman plots, actual error distance below 1mm. |
| Image Segmentation Accuracy (Dice Coefficient) | Not explicitly stated but implied to be high for "mature" software comparison | Mean Dice coefficient of 0.9483 |
| Image Segmentation Accuracy (Hausdorff Distance) | Not explicitly stated but implied to be low for "mature" software comparison | Mean Hausdorff Distance of 1.2917 |
| Output Parameter (Length Measurement) Accuracy | Within 1mm | Within 1mm |
| Output Parameter (Angle Measurement) Accuracy | Within 1° | Within 1° |
| Prosthesis Model Data Consistency | Deviations below 0.5mm | Deviations all below 0.5mm |
| Acetabular Cup Coverage Accuracy | Within ±3% | Within ±3% |
2. Sample size used for the test set and the data provenance
- Sample Size: 200 patient cases (100 cases per region for image segmentation testing).
- Data Provenance: Not explicitly stated from the provided text, but inferred to be retrospective as the testing involved existing patient cases. The country of origin of the data is also not explicitly stated; however, given the sponsor's address (Beijing, China), it is highly probable the data originated from China.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts
- Number of Experts: Not explicitly stated how many professional physicians annotated the key points for ground truth.
- Qualifications of Experts: Described as "professional physicians." No further details on their specific qualifications or years of experience are provided.
4. Adjudication method (e.g., 2+1, 3+1, none) for the test set
- The text states the accuracy testing of key point recognition involved "comparing with the key points identified annotated by professional physicians for precision." This implies a single expert or a consensus approach, but no specific adjudication method (e.g., 2+1, 3+1) is mentioned. It is possible a consensus was used, but it's not specified.
5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, if so, what was the effect size of how much human readers improve with AI vs. without AI assistance
- No MRMC comparative effectiveness study was done. The document explicitly states: "No clinical study is included in this submission."
6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done
- Yes, the described non-clinical tests are primarily standalone (algorithm only) performance evaluations. For example, key point recognition accuracy and image segmentation accuracy tests assess the algorithm's performance against established ground truth or mature software, without direct human interaction as part of the measured performance metric. The software is intended to "assist physicians," implying human-in-the-loop for clinical use, but the reported performance metrics are for the software's automated functions.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)
- Expert Consensus/Annotation: For key point recognition, the ground truth was established by "professional physicians" who identified/annotated key points.
- Comparison with Established Software: For image segmentation accuracy, the device's performance was compared to "the segmentation results of the mature software available in the market," implying that the "mature software" results served as a form of ground truth or reference standard.
- Reference Values/Measurements: For output parameter accuracy (length, angle, prosthesis model data, acetabular cup coverage), the ground truth was likely based on calculated reference values, measurements against known standards, or comparison with manufacturer-provided data.
8. The sample size for the training set
- The document does not specify the sample size used for the training set. It only details the test set.
9. How the ground truth for the training set was established
- The document does not provide information on how the ground truth for the training set was established. It focuses solely on the validation/test set.
Ask a specific question about this device
Page 1 of 1