Search Results

The Bolt Navigation System assists in the accurate placement of pedicle screws when used in conjunction with an intraoperative fluoroscope. It utilizes intraoperative fluoroscopic and pre- operative MRI or CT axial images to provide surgical planning and navigational telemetry relative to gravity, based on a fixed entry point ascertained by the user and validated by intraoperative fluoroscopic imaging. It is not intended to track patient position. The System is indicated for open and minimally invasive pedicle screw placement using a posterior approach in the thoracolumbar and sacral spine (T-9 to S1) where the patients' relevant rigid anatomical structures can be clearly identified on the imaging.

The BNS is comprised of the Bolt Navigation Unit (BNU) (an iPod touch® mobile digital device with the Bolt navigation software loaded on it), the Bolt single use case, and sterile drape. The BNS is intended to provide navigational guidance during spine surgery. The system uses preand perioperative imaging data, and input from the surgeon via the BNU touchscreen to construct the proper angular position of the instrumentation and implants relative to gravity, and communicates this information to the surgeon via the BNU screen attached to the instrument allowing the surgeon to look at both the surgical site and the navigation data at the same time, thus attenuating the risk of attention shift. The BNS provides guidance data by displaying the angular orientation of a surgical instrument (such as a pedicle probe or awl) relative to a surgeon selected entry point on the patient and gravity. Angular orientation of the instruments is linked to the imaging data via the BNS. The system is intended to be used for both image fusion and navigation for spine surgery applications where reference to relevant rigid structures can be identified relative to a perioperative image data of the anatomy and the gravity vector.

Here's an analysis of the acceptance criteria and study details for the Bolt Navigation System ("BNS"), based on the provided FDA 510(k) summary:

Acceptance Criteria and Device Performance

Study Details

1. Sample size used for the test set and the data provenance:

   • Test Set (Cadaveric Trajectory Accuracy Study): The document doesn't explicitly state the number of cadavers or individual screw placements used in the cadaveric study. However, the results are presented as overall accuracy, testing levels T9 to S1 (thoracolumbar and sacral spine).
   • Data Provenance: The cadaveric study is a non-clinical test, implying it was conducted in a controlled lab setting, likely in the country of the manufacturer or test facility. It is a prospective study in the sense that the data was generated specifically for this validation.

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

   • The document does not provide information on the number of experts or their qualifications used to establish ground truth for the cadaveric study.
   • For the cadaveric study, "Planned trajectory vs actual placement accuracy" was assessed, but the method for determining "actual placement" (e.g., post-insertion imaging, dissection measurement) and who evaluated it is not detailed.

3. Adjudication method (e.g., 2+1, 3+1, none) for the test set:

   • The document does not specify an adjudication method for the cadaveric study.

4. 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:

   • A formal MRMC comparative effectiveness study, comparing human readers with and without AI assistance, was not explicitly detailed in the provided summary.
   • A "multi-surgeon clinical study" was conducted, but its results are only summarized as supplementary information to show the device performs as intended, not as a comparative effectiveness study against a non-AI control group with quantifiable improvement.

5. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:

   • The "Clinical phantom accuracy results" represent a standalone performance assessment of the system against a known phantom, demonstrating the system's inherent accuracy in a controlled environment. The mean accuracy was 0.35 degrees.
   • The "Cadaveric accuracy results" also reflect the algorithm's performance in a more realistic surgical simulation, where surgical instruments guided by the BNS were evaluated for accuracy of placement against a planned trajectory. While a human surgeon uses the system, the
   • accuracy measurement itself quantifies the system's guidance.

6. The type of ground truth used (expert consensus, pathology, outcomes data, etc):

   • Clinical Phantom Accuracy: The ground truth for the phantom study would be the known, precisely manufactured dimensions and orientations of the phantom, which the device is designed to measure.
   • Cadaveric Trajectory Accuracy Study: The ground truth for this study is described as "Planned trajectory vs actual placement accuracy." The "planned trajectory" would be established pre-procedure (likely by a surgeon using imaging), and the "actual placement" would be measured post-procedure using a precise method (e.g., high-resolution CT, C-arm imaging, or physical measurement) to determine the deviation from the plan. The specific method for determining "actual placement" (which serves as the "truth") is not detailed.

7. The sample size for the training set:

   • The document does not provide information on the sample size of any training set used for the BNS software. The BNS utilizes "pre- and perioperative imaging data" and "intraoperative fluoroscopic and pre-operative MRI or CT axial images," suggesting it may rely on existing imaging data for its functionality, but details about supervised machine learning training, if any, are absent.

8. How the ground truth for the training set was established:

   • Since the document does not specify a training set or the use of machine learning that would require a distinct training set with ground truth, this information is not available. The system appears to be more of a navigation system based on physical principles and image registration rather than a deep learning model requiring extensive annotated training data.

Ask a specific question about this device