(72 days)
The Barco Digital Dermatoscope is a non-invasive skin imaging system, which acquires multispectral and white light dermoscopic images and clinical photographs of the skin. These can be stored, retrieved, displayed and reviewed by trained medical practitioners.
The Barco Demetra BDEM-01 is designed to capture images of the skin and optimize the imaging and documentation workflow. The “Barco Demetra BDEM-01” consists of a hardware device and a software application. The hardware device is a portable, battery powered medical device for acquiring and visualizing images of the skin. The device acquires multispectral optical dermoscopic images in a contact mode (device in contact with the skin). In addition, the device can also acquire a clinical close-up image, when it is held at up to 10 to 15 cm from the skin, and a clinical overview image when it is held at a distance of approximately 35-55 cm from the skin. The stand-alone software application is cloud software with a related web application. The cloud software can also generate one or more reports and feedback to the device.
The provided text is a Premarket Notification (510(k)) summary for the Demetra BDEM-01 device, a non-invasive skin imaging system. It focuses on demonstrating substantial equivalence to a predicate device (Microderm - K040171) rather than presenting a performance study against specific acceptance criteria for an AI/algorithm-driven device.
Therefore, this document does not contain the information required to describe acceptance criteria and a study that proves a device meets them in the context of an AI-driven medical device, as it describes a non-algorithm-driven imaging system.
Here's why and what information is missing:
- No AI/Algorithm: The device description clearly states "The Barco Demetra BDEM-01 is designed to capture images of the skin and optimize the imaging and documentation workflow. The 'Barco Demetra BDEM-01' consists of a hardware device and a software application." There is no mention of any artificial intelligence, machine learning, or algorithm that processes these images beyond storage, retrieval, display, and review by trained medical practitioners. The "multispectral images" are acquired, but there's no indication of AI interpretation of these multispectral data for diagnosis or risk assessment.
- No Diagnostic Aid Claim (for AI): The Indications for Use state "These can be stored, retrieved, displayed and reviewed by trained medical practitioners." This implies a tool for human review, not an autonomous diagnostic or assistive AI.
- Performance Testing Focus: The "Performance testing" section (9) focuses on bench tests validating differences from the predicate device related to hardware modifications (Multispectral LEDs illumination, hand-held design, biocompatibility) and basic image acquisition review, not on the performance of a diagnostic algorithm. The criteria listed are simply "PASS," indicating compliance with safety and engineering standards.
To answer your request, if this were an AI-driven device, the following information would typically be found in a 510(k) summary or an associated clinical study report for an AI/ML product:
- A table of acceptance criteria and the reported device performance: This would list metrics like sensitivity, specificity, AUC, F1-score, accuracy, etc., along with the target values and the achieved values from the study.
- Sample sizes used for the test set and the data provenance: Details on the number of cases/images used for testing, where the data came from (e.g., specific hospitals, geographical regions), and whether it was retrospectively collected or prospectively gathered.
- Number of experts used to establish the ground truth for the test set and the qualifications of those experts: Information on the number of clinicians (e.g., dermatologists, radiologists) who provided the reference standard for the test data, and their relevant experience and certifications.
- Adjudication method for the test set: How disagreements among the experts establishing ground truth were resolved (e.g., majority vote, third expert review, consensus meeting).
- If a multi-reader multi-case (MRMC) comparative effectiveness study was done: For AI-assisted reading, this study would compare human performance with and without AI assistance, quantifying the improvement (effect size) in diagnostic accuracy, efficiency, or other relevant metrics.
- If a standalone (i.e., algorithm only without human-in-the-loop performance) was done: Often, an initial study evaluates the AI's performance independently against the ground truth.
- The type of ground truth used: Whether ground truth was established by expert consensus, biopsy/pathology results, long-term patient outcomes, or a combination.
- The sample size for the training set: The number of cases/images used to train the AI model.
- How the ground truth for the training set was established: Similar to the test set, but for the data used during model development.
Because the provided document pertains to a primary imaging device, and not an AI/ML diagnostic aid, it lacks the specific study design elements and performance metrics that would be associated with an AI product's acceptance criteria.
§ 878.4580 Surgical lamp.
(a)
Identification. A surgical lamp (including a fixture) is a device intended to be used to provide visible illumination of the surgical field or the patient.(b)
Classification. Class II (special controls). The device, when it is an operating room lamp, a surgical instrument light, a surgical floor standing light, an endoscopic surgical light, a surgical light connector, a ceiling mounted surgical light, a surgical light carrier, surgical light accessories, a surgical lamp, a remote illuminator, or an incandescent surgical lamp, is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 878.9.