(153 days)
DS Core Diagnosis is a software product for use in dentistry for viewing and interacting with digital or digitized patient media as well as supporting the diagnosis of digital or digitized patient media. Intended users will be able to view and analyze different media types in one viewer.
The Panoramic Curve Proposal feature is intended for patients aged 12 years and older with permanent dentition.
DS Core Diagnosis is a cloud-based dental imaging software that provides access to dental media and images via an active internet connection. It is a medical device that can only be used in combination with the DS Core Platform. Classified as a Class II dental imaging software, DS Core Diagnosis allows users to view a variety of media types, including 2D and 3D images (e.g., bitewings, periapicals, panoramic, CBCT, etc.), as well as photos and documentation.
The software includes standard image viewing tools such as annotation, flip, rotate, contrast adjustment, brightness, and magnification.
DS Core Diagnosis also features a Machine Learning enabled panoramic curve proposal algorithm for a proposed panoramic curve of a CBCT scan along the dental arch, which can be adjusted and reviewed by the user. The architecture is a Convolutional Neural Network (CNN) model.
DS Core Diagnosis does not interact directly with patients or control any life-sustaining devices. The software does not perform diagnoses; instead, it provides tools to assist qualified clinicians in interpreting the displayed images and making informed decisions.
Here's an analysis of the acceptance criteria and the study proving the device meets them, based on the provided FDA 510(k) clearance letter for DS Core Diagnosis:
Acceptance Criteria and Reported Device Performance
| Acceptance Criteria | Reported Device Performance |
|---|---|
| Panoramic Curve Proposal Success Rate (PCPSR): Percentage of cases where the proposed panoramic curve projection layer includes all teeth in the scan. | Dentsply Sirona Scans: PCPSR of 98.81%, with the lower bound of the 95% CI at 94.57%. |
| All Scans (Dentsply Sirona & Third Party): PCPSR of 99%, with 99 of 100 scans considered "successful" and the lower bound of the 95% CI at 95.42%. | |
| Various Subgroups: PCPSR range of 97.78-100%, all passing the acceptance criteria. | |
| Clinical Acceptability (Pass/Fail) of Default Panoramic Image: Determine the acceptability of the CBCT reconstructed panoramic images (without user modification) with the panoramic curve proposal generated by the PCP algorithm, to begin working per the standard of care. | Clinical Evaluation Pass Rate: 92.5% pass rate overall for the primary endpoint. |
| System validation and usability testing for the DS Core Diagnosis device. | All tests successfully passed. |
Study Details
Based on the provided text, the following details about the studies can be extracted:
2. Sample size used for the test set and the data provenance:
- PCPSR Evaluation:
- Test Set Size (Dentsply Sirona & Third Party): 100 scans (specifically mentioned "99 of 100 scans considered 'successful'").
- Data Provenance: A mix of "Dentsply Sirona Scans" and "Third Party" scans. Specific countries of origin are not detailed, but the inclusion of third-party scans suggests a broader dataset. The text does not explicitly state if it was retrospective or prospective, but given the nature of evaluating existing scans, it's highly likely to be retrospective.
- Clinical Evaluation:
- Test Set Size: "a dataset made up of CBCT images". The exact number of cases or images in this dataset is not explicitly stated.
- Data Provenance: Not explicitly stated, but would likely be retrospective CBCT images used for evaluation.
3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:
- PCPSR Evaluation: The text doesn't explicitly state that experts established ground truth for the PCPSR calculation directly. The success rate is described as whether "the proposed panoramic curve projection layer includes all teeth in the scan," which likely refers to an objective assessment based on the image content itself, possibly verified by a technical process rather than subjective expert consensus.
- Clinical Evaluation: "clinical experts assessed a dataset". The number of experts is not specified. Their qualifications are described as "clinical experts," implying dental professionals (e.g., dentists, oral and maxillofacial radiologists) but their specific experience levels (e.g., "radiologist with 10 years of experience") are not detailed.
4. Adjudication method for the test set:
- The text does not explicitly mention an adjudication method (like 2+1, 3+1). For the clinical evaluation, it states "experts assessed," which could imply independent assessment or consensus, but no specific method is described. For the PCPSR evaluation, the metric seems more objective and less prone to requiring adjudication in the same way.
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 where human readers' performance with and without AI assistance is compared is not explicitly described. The clinical evaluation focuses on the acceptability of the default AI-generated curve, not a comparison of expert performance with and without AI. It emphasizes that users retain full autonomy to adjust the curve.
6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:
- Yes, a standalone performance evaluation was done for the "Panoramic Curve Proposal" feature. The PCPSR evaluation directly assesses the algorithm's ability to propose a curve that includes all teeth, without human intervention in the proposal generation or acceptance. The "clinical evaluation" also assesses the default AI-generated panoramic image "without user modification."
7. The type of ground truth used:
- PCPSR Evaluation: The ground truth appears to be an objective determination of whether "all teeth in the scan" are included within the proposed panoramic curve. This would likely be established by a clear definition and potentially manual verification of tooth inclusion. It is closer to expert consensus if human verification was involved, but the description sounds more like an objective technical check.
- Clinical Evaluation: The ground truth for this aspect was established by expert consensus/assessment from "clinical experts" on the "clinical acceptability" (pass/fail) of the reconstructed panoramic images.
8. The sample size for the training set:
- The document does not provide any information about the sample size used for the training set for the Convolutional Neural Network (CNN) model.
9. How the ground truth for the training set was established:
- The document does not provide any information about how the ground truth for the training set was established. It only describes that the architecture is a Convolutional Neural Network (CNN) model.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).