(111 days)
The RTA is a computerized laser slitlamp biomicroscope that is intended to provide manual and computerized tomography of the retina in vivo. The RTA scans successive slit images on the fundus, without the need for a contact lens, to determine the thickness and the inner structure of the retina, both by observation of the slit images and by computer analysis of these images. It is indicated for assessing the area and location of retinal thickness abnormalities, such as thickening due to macular edema and atrophy associated with degenerative diseases, and for visualizing other retinal pathologies.
The Talia RTA is a computerized electro-optical system comprised of two primary components, the optical head and the computer system. The main elements of the optical head include laser and conventional light sources, optics, a scanner, and three digital cameras. The optical head has a 0.2 mW helium neon (HeNe) laser that emits green light at a wavelength of 543.5 nm. The beam is focused into a thin slit, 2.2 mm in length and 15 microns in width. The mirror directs the laser beam toward the eye, but permits simultaneous illumination with the conventional slit. The laser scans 10 slits of 2 mm length in each scan and can scan at a rate of 20 msec per slit for retinas. The stereo angle of the scan (the angle between the laser beam on the retina and the detector axis) can be set at 11.5° or 5.7°. The scan density (the distance between slits in the scan) is 200 microns for normal density and has a total scan area of 2 x 2 mm. The scan density for dense density is 100 microns and has a total scan area of 1 x 2 mm. The computer system consists of the computer, peripheral boards, color monitor, printer, keyboard, mouse, footpedal, isolation transformer, and software. The RTA software is composed of two parts, the control software and the analysis software. The control software controls the operation of the RTA. For analysis of images to create thickness maps, it invokes the analysis software.
Here's an analysis of the provided text, focusing on the acceptance criteria and the study details:
Device Name: Talia Technology Ltd. RTA Retinal Thickness Analyzer
1. Table of Acceptance Criteria and Reported Device Performance
The provided text does not explicitly state formal "acceptance criteria" for regulatory approval in the typical sense (e.g., pre-defined thresholds for accuracy, sensitivity, specificity). Instead, the performance data presented describes the device's technical capabilities regarding depth precision, depth resolution, and reproducibility.
| Performance Metric | Acceptance Criteria (Implied/Observed) | Reported Device Performance |
|---|---|---|
| Depth Precision (in vitro) | Not explicitly stated | 5 to 10 µm |
| Depth Resolution (in vitro) | Not explicitly stated | 50 µm |
| Depth Precision (human subjects) | Not explicitly stated | 5 to 10 µm (optimal) |
| Depth Resolution (human subjects) | Not explicitly stated | 50 µm (optimal) |
| Intra-visit Reproducibility | Sufficient resolution and reproducibility | ± 12 µm (based on 3 scans) |
| Inter-visit Reproducibility (single scan) | Sufficient resolution and reproducibility | ± 13 µm |
| Inter-visit Reproducibility (tri-scan average) | Sufficient resolution and reproducibility | ± 10 µm |
| EMC Compliance | Compliance with EN 50081-1 & EN50082-1 | Compliant |
| Laser Radiation Safe Time | Not explicitly stated | 20 minutes |
Note: The phrase "These results indicated that the RTA provides sufficient resolution and reproducibility for retinal thickness mapping" serves as a post-hoc justification that the observed performance meets an unstated, underlying requirement for utility.
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size:
- "In vitro" testing: Sample size not specified.
- Retinal thickness mapping reproducibility and depth precision/resolution in human subjects: 5 human subjects.
- Data Provenance: The document does not explicitly state the country of origin for the human subject data. Given the "Submitter" and "Sponsor" are from the US and Israel respectively, the testing could have occurred in either region, or elsewhere. The study appears to be prospective for the human subject reproducibility testing, as it involved specific visits and scans performed for the study.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
The document does not mention the use of experts to establish ground truth for the performance metrics measured. The "ground truth" for depth precision, resolution, and reproducibility would have been derived directly from the device's measurements and comparisons between those measurements, not from expert interpretation.
4. Adjudication Method for the Test Set
Not applicable. There was no expert-based ground truth or interpretation that would require an adjudication method. The measurements were quantitative and derived from the device's output.
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, a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was not conducted or reported. This document pertains to a device from 1996, which predates the widespread concept of "AI assistance" in medical imaging in the way it's understood today. The device is described as a "computerized laser slitlamp biomicroscope" with "computer analysis of these images," but this refers to algorithmic processing for measurements, not AI in the sense of deep learning or decision support for human readers.
6. If a Standalone (i.e. algorithm only without human-in-the loop performance) Was Done
Yes, in essence. The reported performance metrics (depth precision, depth resolution, reproducibility) are standalone algorithm/device performance metrics. The device's "computer analysis" performs the measurements and thickness mapping, and these results are what were assessed. There is no mention of human-in-the-loop performance influencing these specific quantitative measures.
7. The Type of Ground Truth Used
- In vitro measurements: The ground truth for in vitro depth precision and resolution would likely have been established using precisely calibrated physical phantoms or standards.
- Human subject reproducibility: The "ground truth" here is the individual's actual retinal thickness at the time of measurement, and the study assessed how consistently the device measured that thickness across repeated scans. It's an internal consistency measure rather than comparison to an external gold standard like pathology.
8. The Sample Size for the Training Set
Not applicable. This device is from 1996 and appears to use rule-based algorithmic processing for image analysis and thickness mapping, rather than machine learning models that require a "training set." The software "invokes the analysis software" to create thickness maps, implying pre-programmed algorithms.
9. How the Ground Truth for the Training Set Was Established
Not applicable, as there was no mention of a training set for machine learning.
§ 886.1570 Ophthalmoscope.
(a)
Identification. An ophthalmoscope is an AC-powered or battery-powered device containing illumination and viewing optics intended to examine the media (cornea, aqueous, lens, and vitreous) and the retina of the eye.(b)
Classification. Class II (special controls). The device, when it is an AC-powered opthalmoscope, a battery-powered opthalmoscope, or a hand-held ophthalmoscope replacement battery, is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 886.9.