Search Results

Bonelogic software is to be used by orthopaedic healthcare professionals for diagnosis and surgical planning in a hospital or clinic environment.

Bonelogic software provides:

• Semi-automatic segmentation with manual or assisted input of bony structure identification from CT imaging input,
• Three-dimensional mathematical models of the anatomical structures of foot and ankle,
• Measurement templates containing radiographic measures of foot and ankle, and tools for manually obtaining linear and angular measurements,
• Surgical planning application for foot and ankle using three-dimensional models of the anatomical structures and radiographic measures.

The three-dimensional models of the anatomical structures combined with the measurements can be used for the diagnosis of orthopaedic healthcare conditions. The surgical planning application containing the three-dimensional structural models combined with the measurements can be used for the planning of treatments and operations to correct orthopaedic healthcare conditions of foot and ankle.

The Bonelogic is a software tool that segments bone anatomy using dedicated semiautomatic tools and fully automatic algorithms. More specifically, Bonelogic is intended to segment foot and ankle bones from computed tomography (CT) images. The segmented structures may then be used to create 3D models of their respective bones and replicate the anatomy of a patient. The semi-automatic tools of the software require a healthcare professional to mark the different bones in an initial 3D rendered model prior to when the segmentation process is initialized. This method is called the semi-automatic workflow with manual input. The software also comprises an optional semiautomatic workflow with assisted input that replaces the required user input with an estimate based on a locked artificial neural network (ANN) model. The fully automatic algorithm processes the final result in the same way based on input generated by the semi-automatic workflow with user input and with ANN model. The user still needs to mark the laterality and as a new step acknowledge the bones discovered by the ANN model.

Here's a summary of the acceptance criteria and the study proving the device meets those criteria, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample Size Used for the Test Set and Data Provenance

• Test Set Sample Size: 82 CT image studies.
• Data Provenance:
  • Country of Origin: Collected from various sites across the USA and Europe, with a minimum of 50% of the images originating from the USA.
  • Retrospective/Prospective: Not explicitly stated, but the description of collected studies from individual patients with varying conditions suggests a retrospective collection.

3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications

• Number of Experts: Three (3).
• Qualifications of Experts: U.S. Orthopedic surgeons (no specific years of experience mentioned).

4. Adjudication Method for the Test Set

• Method: Majority vote of the three experts. Two matching responses out of three were required to establish the ground truth for bone and metal presence in each DICOM series.

5. Multi Reader Multi Case (MRMC) Comparative Effectiveness Study

• No, an MRMC comparative effectiveness study was not conducted. The performance assessment was for the standalone AI algorithms.

6. Standalone (Algorithm Only Without Human-in-the-Loop Performance) Study

• Yes, a standalone performance assessment study was done for the AI algorithms.
  • For bone identification, the algorithm correctly identified bones of the foot and ankle in 100% of the 82 CT image series.
  • For metal identification, the algorithm correctly identified metal in 98.8% of the images, with a specificity of 98% and sensitivity of 100%.

7. Type of Ground Truth Used

• Expert Consensus: The ground truth for bone and metal identification was established by the independent review and majority vote of three U.S. Orthopedic surgeons using a 3rd party software.

8. Sample Size for the Training Set

• Bone Identification AI Algorithm: 145 CT image studies.
• Metal Identification AI Algorithm: 130 CT image studies.

9. How the Ground Truth for the Training Set Was Established

• The document implies that the ground truth for the development of the AI algorithms (training and tuning) was similarly established, as it states: "The AI algorithm for bone identification was developed using 145 CT image studies and metal identification was developed using 130 CT image studies." It then goes on to describe the ground truth establishment for the test data. While not explicitly detailed for the training set itself, it can be reasonably inferred that a similar expert-driven process was used, especially given the emphasis on "independent data sets" for training, tuning, and testing.

Ask a specific question about this device