(326 days)
The CIRRUS photo (Models 600 and 800) is a non-contact, high resolution tomographic and biomicroscopic imaging device that incorporates a digital camera which is suitable for photographing, displaying and storing the data of the retina and surrounding parts of the eye to be examined under mydriatic and non-mydriatic conditions.
These photographs support the diagnosis and subsequent observation of eye diseases which can be visually monitored and photographically documented. The CIRRUS photo is indicated for in vivo viewing, axial cross sectional, and three-dimensional imaging and measurement of posterior ocular structures, including retina, retinal nerve fiber layer, macula, and optic disc as well as imaging of anterior ocular structures, including the cornea.
It also includes a Retinal Nerve Fiber Layer (RNFL), Optic Nerve Head (ONH), and Macular Normative Database which is a quantitative tool for the comparison of retinal nerve fiber layer, optic nerve head, and the macula in the human retina to a database of known normal subjects. It is intended for use as a diagnostic device to aid in the detection and management of ocular diseases including, but not limited to, macular holes, cystoid macular edema, diabetic retinopathy, age-related macular degeneration, and glaucoma.
The CIRRUS photo is a non-contact, high resolution digital, tomographic and biomicroscopic imaging device that merges fundus imaging and optical coherence tomography into a single device. To optimize the workflow, the system applies the same beam delivery system for imaging and scanning.
Here's a detailed breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:
Device: CIRRUS photo Models 600 and 800 (Fundus Camera and Optical Coherence Tomography)
Acceptance Criteria and Reported Device Performance
The acceptance criteria for the CIRRUS photo are not explicitly listed as specific thresholds in the document (e.g., "accuracy > 90%"). Instead, the study aimed to demonstrate equivalence and repeatability/reproducibility of measurements between the CIRRUS photo and the predicate device (Carl Zeiss Meditec Cirrus HD-OCT Model 4000).
Therefore, the "acceptance criteria" are implicitly that the differences in measurements between the two devices should be small and within acceptable limits (as evidenced by confidence intervals and limits of agreement), and that the CIRRUS photo should demonstrate good repeatability and reproducibility.
The reported device performance section details the findings from these equivalence and repeatability studies.
| Acceptance Criteria Category | Specific Metric (Implicit) | CIRRUS Photo Performance |
|---|---|---|
| Equivalence (Inter-device) | Mean Difference in RNFL, ONH, and Macular Thickness measurements between CIRRUS photo and Cirrus HD-OCT Model 4000. | Normal Eyes (N=33): The mean difference for 31 measurement parameters (17 RNFL, 5 ONH, 9 macular) between the two devices was generally small, with 95% Confidence Intervals for most parameters including or very close to zero. The 95% Limits of Agreement show a range within which differences are expected to fall. For example, Average RNFL Thickness showed a mean difference of 0.6 (1.2) µm with a CI of (0.2, 1.0) and LOA of (-1.7, 2.9). |
| Diseased Eyes (Glaucoma N=17, Macular Disease N=19): Similar to normal eyes, mean differences were generally small, with 95% Confidence Intervals for many parameters including or near zero. For example, Average RNFL Thickness (glaucoma) showed a mean difference of 0.8 (1.3) µm with a CI of (0.1, 1.4) and LOA of (-1.7, 3.2). | ||
| Repeatability | Repeatability Standard Deviation (SD) and Limit | Normal Eyes: Repeatability SDs for RNFL (e.g., Average RNFL Thickness: 1.4634 µm), ONH (e.g., Cup Disc Ratio: 0.0236), and Macular Thickness (e.g., Central Subfield: 1.6398 µm) demonstrate the device's consistency when measurements are repeated under the same conditions. |
| Diseased Eyes: Repeatability SDs for RNFL (e.g., Average RNFL Thickness: 1.4634 µm), ONH (e.g., Cup Disc Ratio: 0.0276), and Macular Thickness (e.g., Central Subfield: 5.6224 µm) were also reported. | ||
| Reproducibility | Reproducibility Standard Deviation (SD) and Limit | Normal Eyes: Reproducibility SDs for RNFL (e.g., Average RNFL Thickness: 2.1899 µm), ONH (e.g., Cup Disc Ratio: 0.0245), and Macular Thickness (e.g., Central Subfield: 2.7756 µm) demonstrate consistency when measurements are repeated under varying conditions (e.g., different operators). |
| Diseased Eyes: Reproducibility SDs for RNFL (e.g., Average RNFL Thickness: 1.8796 µm), ONH (e.g., Cup Disc Ratio: 0.0278), and Macular Thickness (e.g., Central Subfield: 7.6068 µm) were also reported. | ||
| Qualitative Assessment | Overall conclusion regarding performance | "The mean values of the 31 thickness parameters were very similar for the two devices." |
| "Cirrus photo showed good repeatability and reproducibility for both normal and diseased eyes." |
Study Details
-
Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)
- Normal Eyes Study:
- Phase 1 (Inter-operator variability): 30 subjects
- Phase 2 (Inter-device variability): 33 subjects
- Diseased Eyes Study:
- Retinal disease (inter-device): 19 subjects
- Retinal disease (inter-operator): 19 subjects
- Glaucoma (inter-device): 17 subjects
- Glaucoma (inter-operator): 18 subjects
- Total Test Set: 63 normal subjects, 73 diseased subjects (total of 136 subjects, though there's some overlap in eyes examined across phases/devices, but subjects did not participate in both phases of a study)
- Data Provenance: The document does not specify the country of origin of the data. The studies were prospective.
- Normal Eyes Study:
-
Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)
- The study design focused on comparing measurements from the new device (CIRRUS photo) to an existing, cleared device (Cirrus HD-OCT Model 4000), which itself presumably has established accuracy.
- The "ground truth" here is the measurement obtained from the predicate device. Therefore, no external experts were used to establish a separate "ground truth" beyond the measurements themselves.
- For the inter-operator variability phases, four operators were involved, implying expertise in operating OCT devices. However, their specific qualifications (e.g., years of experience, medical degree) are not detailed.
-
Adjudication method (e.g. 2+1, 3+1, none) for the test set
- Not applicable. This was a quantitative measurement comparison study, not a diagnostic classification study requiring adjudication of expert opinions. The method involved calculating mean differences, confidence intervals, and limits of agreement between device measurements.
-
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 MRMC comparative effectiveness study was not done. This study focuses on the technical equivalence and repeatability/reproducibility of a diagnostic imaging device (OCT) rather than evaluating human reader performance with or without AI assistance.
- The term "AI" is not mentioned in this document.
-
If a standalone (i.e. algorithm only without human-in-the-loop performance) was done
- Yes, the primary studies described are standalone performance evaluations of the device, focusing on the measurements generated by the CIRRUS photo itself compared to a predicate device. While human operators are involved in acquiring the images, the analysis of the measurement data (31 parameters of RNFL, ONH, and macular thickness) is performed by the device's algorithms. The document does not describe a human-in-the-loop study to modify or interpret the device's measurements.
-
The type of ground truth used (expert consensus, pathology, outcomes data, etc)
- The "ground truth" in this context is the measurement data obtained from the predicate device (Cirrus HD-OCT Model 4000). The study aimed to show that the CIRRUS photo provides comparable measurements. It's a device-to-device comparison rather than an evaluation against a clinical ground truth like pathology or patient outcomes.
-
The sample size for the training set
- The document implies that the CIRRUS photo uses a "normative database" (RNFL, ONH, and Macular Normative Database) which comes from the Cirrus HD-OCT Model 4000 and is adjusted based on regression analysis. However, it does not specify the sample size of this underlying normative database or detail a separate "training set" for the CIRRUS photo's algorithms in this context. The study described focuses on testing a new device's performance against an established device, not on training a new algorithm.
-
How the ground truth for the training set was established
- As noted above, a distinct "training set" for the CIRRUS photo's algorithms (other than the inherited normative database from the predicate device, K083291; K111157) is not explicitly described. For the normal subjects in the "Normative Database," the implicit "ground truth" would be their classification as "normal" based on clinical criteria and measurements from the original predicate device (Cirrus HD-OCT 4000). The method for establishing this original normative database's "ground truth" is not explained in this document.
§ 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.