Search Results

True Definition DL is a deep learning based CT reconstruction method intended for high contrast spatial resolution enhancement for bone and lung imaging. True Definition DL may be used for patients of all ages.

Computed Tomography (CT) is an indispensable imaging modality in clinical diagnostics due to its ability to provide detailed cross-sectional images of anatomical structures. However, achieving high spatial resolution remains a persistent challenge, particularly in applications requiring the visualization of fine structural details such as in inner auditory canal imaging, vascular studies, lung imaging, and bone microarchitecture analysis. Reconstruction techniques that attempt to boost spatial resolution typically also would amplify the noise which often result in tradeoffs to be made between resolution and noise.

As part of the continuous innovation to solve the above challenges in CT imaging, GEHC developed a deep learning-based CT reconstruction algorithm specifically designed for high contrast lung and bone imaging, which is the subject of this premarket notification. This reconstruction algorithm, marketed under the name True Definition DL (TDDL), is an additional user-selectable recon option specifically designed for lung and bone imaging. It aims to enhance spatial resolution for both in-plane and cross-plane directions. This algorithm is incorporated into the reconstruction chain of the Revolution CT /Apex Family CT systems including Revolution CT/Revolution CT ES, Revolution Apex Elite, Revolution Apex Plus, Revolution Apex Select all cleared under (K213715) and Revolution Vibe (K250941).

True Definition DL offers three strengths that the user can choose depending on enhancement preferences. The benefits provided by this enhancement include improved spatial resolution measured by MTF.

Here's a breakdown of the acceptance criteria and study details for the "True Definition DL" device, as extracted from the provided FDA 510(k) clearance letter:

Acceptance Criteria and Device Performance

Study Details

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

   • Sample size: Not explicitly stated for the clinical reader study, only referred to as "sample clinical data."
   • Data provenance: Not explicitly stated (e.g., country of origin or retrospective/prospective).

2. 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.
   • Qualifications: "US board-certified radiologists."

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

   • Not explicitly stated. The document mentions "readers confirmed," implying a consensus or majority opinion, but the specific adjudication method is not described.

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 reader study was performed ("clinical testing was carried out in the form of a reader study").
   • The document states that the study "validated that True Definition DL provides diagnostic value by improving high contrast spatial resolution... and improves confidence in the assessment of high contrast structures."
   • However, the document does not explicitly state that it was an MRMC comparative effectiveness study comparing human readers with AI vs. without AI assistance. It primarily focuses on the diagnostic value and confidence with the use of the images enhanced by True Definition DL.
   • No effect size or specific metrics quantifying improvement with AI assistance are provided.

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

   • Yes, a standalone evaluation of the algorithm's performance was indicated through "robustness of the True Definition DL algorithm" testing. This included assessment of hallucination risks, enhancement of anatomical edges, preservation of noise texture, and correlation with input/target images, suggesting an algorithm-level assessment.
   • Bench testing on phantoms also measured the performance of True Definition DL against industry-standard IQ metrics and traditional metrics without direct human interaction.

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

   • For the clinical reader study, the "ground truth" was established by expert opinion/consensus of US board-certified radiologists who evaluated the images and confirmed diagnostic value and lack of artifacts.
   • For bench testing, "well established test methods using industry-standard IQ metrics" provided quantifiable reference points for performance.

7. The sample size for the training set:

   • Not explicitly stated.

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

   • Not explicitly stated. The document only mentions that True Definition DL is a "deep learning-based CT reconstruction algorithm," implying a training phase, but provides no details on how the training data was curated or its ground truth established.

Ask a specific question about this device