Search Results

The Praezis Plus system is computer software intended to be used in planning intracranial stereotactic surgeries such as biopsies and treatment of lesions. The intended user of Praezis Plus is a medical doctor - neurosurgeon.

Praezis Plus is a software tool intended to help the surgeon plan stereotactic treatments. It enables the user to select optimally fitting stereotactic trajectory positions by digitizing the entry (trephination) and target points of the trajectory. If both points define a trajectory which can be set on the operating device, the corresponding device settings are calculated and displayed in the trajectory window. Any trajectory can be defined and displayed in CT and/or MRI slices or reconstructions or angiographic images if they are available. Trajectories are always stored in stereotactic coordinates and can be digitized and displayed in available patient's modalities (CT, MR or Angiographic). The software can project the selected trajectory to any slices or sections and generate surgeon's eye views. For any possible needle position, the angle settings for either the Riechert/Mundinger (RM) or Zamorano/Dujovny (ZD) stereotactic operating devices can be displayed. Praezis Plus supports the transfer of stereotactic coordinates between angiographic, CT and MR images. Regions of Interest ("ROIs") and Volumes of Interest ("VOIs") can be defined, transferred and displayed between angiographic, CT and MR images. Their area and volume can be calculated. Global stereotactical transformation is available to improve the planning. Image fusion, based on various methods of correlation, allows a combination of valuable information from different modalities. Several types of reconstructed sections such as arbitrary, along and perpendicular to trajectory together with 3D reconstructions are also available. It is possible to print out all the images and protocols and together with patient data archive them on a CD. The output from Praezis Plus is a printed stereotactical plan defining the trajectory and the setup of the stereotactical device. The treatment plans may be used to administer treatments after review and approval by the neurosurgeon.

The provided text describes the regulatory clearance of a medical device software, Praezis Plus, and its comparison to a predicate device, Stereoplan Plus. The information focuses on substantial equivalence based on intended use, technological characteristics, and performance data. However, the document does not contain specific acceptance criteria, a detailed quantitative study of device performance against these criteria, or information regarding AI/ML components.

Therefore, I cannot provide a table of acceptance criteria and reported device performance from the provided text, nor can I answer questions related to AI/ML specific study designs such as MRMC studies, standalone algorithm performance, or training/test set details for an AI model.

The document primarily discusses:

• System Validation: This involved internal testing by the manufacturer.
• Benchmark Tests: Comparing Praezis Plus to the predicate device (Stereoplan Plus) quantitatively for computation accuracy.
• Beta Site Testing: Qualitative evaluation in a "near-clinic (virtual) environment" at two hospitals to confirm it "fulfills the declared requirements and is consistently safe and accurate comparing to the bench mark device."

Here's what can be extracted and inferred from the document based on the information it does provide:

1. A table of acceptance criteria and the reported device performance

The document does not explicitly present a table of "acceptance criteria" and "reported device performance" in the quantitative sense typically seen for AI/ML device studies. Instead, it refers to:

2. Sample size used for the test set and the data provenance (e.g., country of origin of the data, retrospective or prospective)

• Sample Size for Test Set: Not explicitly stated. The "System Validation" involved a "set of test cases" using "test data," and "benchmark tests" compared calculations against the Stereoplan Plus. The "Beta Site Testing" was conducted at two hospitals, but the number of cases or operations tested is not specified.
• Data Provenance: Not explicitly stated. The "Beta Site Testing" was done at "Beth Israel Deaconess Medical Center - Harvard Medical School, Boston" and "University of Chicago Medical Center - Section Of Neurosurgery, Chicago," implying the data, if collected during these beta tests, would be from the USA. However, the specific origin (retrospective or prospective) of the clinical scenarios or data used for testing is not detailed. The "test data" for system validation and benchmark tests is also not described.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g., radiologist with 10 years of experience)

• Number of Experts: Not specified.
• Qualifications of Experts: The intended user is "a medical doctor - neurosurgeon." It can be inferred that neurosurgeons were involved in the beta site testing to evaluate the device in a "near-clinic (virtual) environment" by following "existing relevant procedures." However, their specific role in establishing "ground truth" (as typically defined for AI/ML studies) or their qualifications are not detailed. For the benchmark tests, "reference values were measured digitally using the benchmark software (Stereoplan Plus)" and considered "ideally exact," implying no human expert "ground truth" was established for these specific numerical comparisons.

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

• Adjudication Method: Not described. The beta site testing involved a comparison to the benchmark device, rather than a formal adjudication process for interpreting results or reaching a consensus on ground truth.

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, an MRMC comparative effectiveness study was not done and is not mentioned in the document. This device is described as "software intended to help the surgeon plan stereotactic treatments," not an AI-assisted diagnostic tool that would typically involve a comparative effectiveness study with human readers. The comparison is primarily between the new software and an existing predicate software.

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

• The document implies that the device's computational accuracy was tested in a standalone manner during system validation and benchmark tests, as it aimed for "absolute equality" or "error less than 1mm" compared to the predicate software's calculations. However, this is not an "algorithm only performance" in the context of an AI/ML algorithm making a decision or interpretation independently. Praezis Plus is a "planning system" used by a "neurosurgeon," indicating its role is always human-in-the-loop.

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

• For benchmark tests, the "ground truth" was the "reference values measured digitally using the benchmark software (Stereoplan Plus)," considered "ideally exact." This is a software-generated computational ground truth rather than a clinical or pathological ground truth.
• For beta site testing, the "ground truth" was effectively the performance and experience with the predicate device, against which the Praezis Plus was compared in a "virtual" clinical environment. No specific clinical outcome data or pathology was used as ground truth for evaluating the software's planning capabilities directly within the scope of this submission.
• The output of the software (stereotactical plan) still requires "review and approval by the neurosurgeon" for actual treatment.

8. The sample size for the training set

• This information is not applicable and not provided. Praezis Plus is described as a "software tool" for planning, not an AI/ML system that would require a "training set" in the context of machine learning. Its functionality seems to be based on computational geometry and image processing principles, rather than learning from data.

9. How the ground truth for the training set was established

• This information is not applicable and not provided, as the device is not described as an AI/ML system requiring a training set.

Ask a specific question about this device