Search Results

The CIRRUS™ HD-OCT is a non-contact, high resolution tomographic and biomicroscopic imaging device. It is indicated for in-vivo viewing, axial cross-sectional, and three-dimensional imaging and measurement of anterior and posterior ocular structures, including cornea, retinal nerve fiber layer, ganglion cell plus inner plexiform layer, macula, and optic nerve head.

The CIRRUS™ HD-OCT Reference Database is a quantitative tool used for the comparison of retinal nerve fiber layer thickness, macular thickness, ganglion cell plus inner plexiform layer thickness, and optic nerve head measurements to a database of healthy subjects.

CIRRUS™ HD-OCT AngioPlex angiography is indicated as an aid in the visualization of vascular structures of the retina and choroid.

The CIRRUS™ HD-OCT is indicated 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 subject device is a computerized instrument that acquires and analyses cross-sectional tomograms of anterior ocular structures (including comea, retinal nerve fiber layer, macula, and optic disc). It employs non-invasive, non-contact, low-coherence interferometry to obtain these high-resolution images. CIRRUS 6000 has a 100kHz scan rate for all structural and angiography scans.

The subject device uses the same optical system, and principle of operation as the previously cleared CIRRUS 6000 (K222200) except for the reference database functionality.

The subject device contains a newly acquired reference database which was collected on K222200. This study data compares macular thickness, ganglion cell thickness, optic disc and RNFL measurements to a reference range of healthy eyes as guided by the age of the patient and /or optic disc size. Reference database outputs are available on Macular Cube 200x200, and Optic Disc Cube 200x20 scan patterns. All other technical specifications have remained the same as the predicate K222200.

Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:

Acceptance Criteria and Reported Device Performance

The acceptance criteria are implicitly met by the successful development of the CIRRUS™ HD-OCT Reference Database (RDB) and its ability to provide normative data for comparison. The study aims to establish these reference limits.

Study Details

1. Sample Size and Data Provenance:

   • Test Set (for RDB establishment): 870 subjects had one eye included in the analysis from an initial enrollment of 1000 subjects.
     • Data Provenance: Prospective, multi-site study conducted at eight (8) clinical sites across the USA.

2. Number of Experts and Qualifications for Ground Truth:

   • The document does not specify the number or qualifications of experts used to establish the ground truth for the test set regarding the "healthiness" of the subjects. The eligibility and exclusion criteria (e.g., "presence of any clinicant vitreal, retinal optic nerve, or choroidal disease in the study eye, including glaucoma or suspected glaucoma. This was assessed based on clinical examination and fundus photography.") imply that ophthalmologists or optometrists would have made these clinical judgments, but the specific number or their experience level is not detailed.

3. Adjudication Method for the Test Set:

   • The document does not explicitly describe an adjudication method for determining the "healthy" status of the subjects. It states that inclusion/exclusion was "assessed based on clinical examination and fundus photography" by unnamed personnel at the clinical sites. There is no mention of a consensus process, independent review, or other adjudication for the ground truth.

4. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

   • No MRMC comparative effectiveness study was done to assess how human readers improve with AI vs. without AI assistance. The study focuses solely on establishing the normative reference database for the device's measurements. The RDB itself is a tool to be used by clinicians, but its impact on clinical decision-making or reader performance was not evaluated in this submission.

5. Standalone Performance:

   • This is a standalone performance study in the sense that the device, equipped with the new reference database, generates the normative values and compares patient data to them. It's the performance of the device's RDB calculation and display, not an AI algorithm performing diagnostic tasks without human input.

6. Type of Ground Truth Used:

   • Clinical Ground Truth: The ground truth for defining "healthy subjects" was based on extensive clinical examination, fundus photography, and adherence to strict inclusion/exclusion criteria (e.g., no known ocular disease, specific visual acuity, IOP, refraction limits). This represents a clinically defined healthy population.

7. Sample Size for the Training Set:

   • The term "training set" is not explicitly used in the context of a machine learning model, as the primary objective was to establish a statistical reference database. The entire dataset of 870 subjects (with qualified scans) was used to develop the reference database. So, the sample size for developing the reference database was 870 subjects.

8. How Ground Truth for the Training Set Was Established:

   • The "ground truth" for the subjects included in the reference database was established by defining them as "healthy subjects" through rigorous inclusion and exclusion criteria applied at 8 clinical sites across the USA. These criteria included:
     • Age 18 years and older
     • Best corrected visual acuity (BCVA) of 20/40 or better in either eye
     • IOP

Ask a specific question about this device

CIRRUS™ HD-OCT is a non-contact, high resolution tomographic and biomicroscopic imaging device intended for in vivo viewing, axial cross-sectional, and three-dimensional imaging of anterior ocular structures. The device is indicated for visualizing and measuring and posterior ocular structures, including corneal epithelium, retinal nerve fiber layer, ganglion cell plus inner plexiform layer, macula, and optic nerve head.

CIRRUS' AngioPlex OCT Angiography is indicated as an aid in the visualization of vascular structures of the retina and choroid.

CIRRUS HD-OCT is indicated 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.

This device is Prescription Use (Rx) only.

The CIRRUS™ HD-OCT Model 6000 is indicated for in-vivo viewing, axial cross-sectional, and threedimensional imaging and measurement of anterior and posterior ocular structures. The clinical purpose of this device has not been modified as compared to the predicate.

CIRRUS 6000 uses the same optical system, architecture, and principle of operation as the previously cleared CIRRUS 5000 (K181534). CIRRUS 6000 has a 100 kHz scan rate for all structural and angiography scans. The primary impact of the higher acquisition speed is its impact on signal-to-noise ratio. The signal-to-noise ratio m the subject device is calibrated to match the specifications of the CIRRUS 6000 uses the same segmentation algorithms as the predicate device and therefore the segmentation results will be equivalent.

In addition to the acquisition speed change, CIRRUS 6000 also has a wider field of view (FOV) and has increased the number of fixation points to 21.

Here's an analysis of the acceptance criteria and the studies that prove the device meets them, based on the provided text.

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria are implicitly defined by the reported precision summaries (Repeatability %CV and Reproducibility %CV and Limits) for various measurements on the CIRRUS HD-OCT 6000 (C6000) and the qualitative image quality results. The studies aim to demonstrate that the C6000 performs comparably to the predicate device, the CIRRUS HD-OCT 5000 (C5000), and that its image quality is clinically acceptable.

Since explicit numerical acceptance criteria (e.g., "Repeatability %CV must be

Ask a specific question about this device

CIRRUS™ HD-OCT is a non-contact, high resolution tomographic and biomicroscopic imaging device intended for invivo viewing, axial cross-sectional, and three-dimensional imaging of anterior ocular structures. The device is indicated for visualizing and measuring and posterior ocular structures, including corneal epithelium, retinal nerve fiber layer, ganglion cell plus inner plexiform layer, macula, and optic nerve head. The CIRRUS normative databases are quantitative tools indicated for the comparison of retinal nerve fiber layer thickness, macular thickness, ganglion cell plus inner plexiform layer thickness, and optic nerve head measurements to a database of normal subjects.

CIRRUS' AngioPlex OCT Angiography is indicated as an aid in the visualization of vascular structures of the retina and choroid. (Model 5000 only.)

CIRRUS HD-OCT is indicated 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 CIRRUSTM HD-OCT is a computerized instrument that acquires and analyses crosssectional tomograms of anterior and posterior ocular structures (including cornea, retinal nerve fiber layer, macula, and optic disc). It employs non-invasive, non-contact, low-coherence interferometry to obtain these high-resolution images. Using this non-invasive optical technique, CIRRUS HD-OCT produces high-resolution cross-sectional tomograms of the eye without contacting the eye. It also produces images of the retina and layers of the retina from an en face perspective (i.e., as if looking directly in the eye).

The CIRRUS HD-OCT is offered in two models, Model 5000 and 500. In the CIRRUS HD-OCT Model 5000, the fundus camera is a line scanning ophthalmoscope. The CIRRUS HD-OCT Model 500 is similar to the Model 5000 except that it provides the fundus image using the OCT scanner only.

The acquired imaging data can be analyzed to provide thickness and area measurements of regions of interest to the clinician. The system uses acquired data to determine the fovea location or the optic disc location. Measurements can then be oriented using the fovea and/or optic disc locations. The patient's results can be compared to subjects without disease for measurements of RNFL thickness, neuroretinal rim area, average and vertical cup-to-disc area ratio, cup volume, macular thickness and ganglion cell plus inner plexiform layer thickness.

In addition to macular and optic disc cube scans, the CIRRUS HD-OCT 5000 also offers scans for OCT angiography imaging, a non-invasive approach with depth sectioning capability to visualize microvascular structures of the eye.

Anterior segment scans enable analysis of the anterior segment including Anterior Chamber Depth, Angle-to-Angle and automated measurement of the thickness of the cornea with the Pachymetry scan.

The provided text describes the 510(k) summary for the Carl Zeiss Meditec, Inc. CIRRUS HD-OCT with Software Version 10. The main study detailed is for corneal epithelial thickness measurements.

Here's a breakdown of the acceptance criteria and study information:

1. Acceptance Criteria and Reported Device Performance

The acceptance criteria are implied by the statistical analyses performed, primarily focusing on repeatability, reproducibility, and agreement with manual measurements. The performance is reported in terms of these statistical metrics rather than predefined thresholds for acceptance.

Corneal Epithelial Thickness Measurements (Pachymetry Scans)

Note: The table provides data for the "Central" sector as an example. The document provides detailed results for 25 different sectors.

2. Sample Size and Data Provenance

• Test Set Sample Size:
  • Normal Corneas (Group 1): 11 adult participants (one eligible eye per participant).
  • Keratoconus/Post-LASIK (Group 2): 12 participants (one eligible eye per participant).
• Data Provenance: The document does not explicitly state the country of origin. It is a prospective clinical study specifically conducted to determine repeatability, reproducibility, and agreement.

3. Number of Experts and Qualifications for Ground Truth

• Number of Experts: Three masked graders.
• Qualifications of Experts: Not explicitly stated beyond "masked graders."

4. Adjudication Method for the Test Set

The document states: "To generate manually marked corneal epithelial thickness measurements, three masked graders reviewed images and manually performed measurements in the 25 sectors." It does not specify an adjudication method like 2+1 or 3+1; it simply mentions that three graders performed the measurements. The comparison is made between the automated measurement and these manual measurements.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

There is no mention of a multi-reader multi-case (MRMC) comparative effectiveness study comparing human readers with and without AI assistance. The study focuses on the agreement between the device's automated measurements and manual measurements.

6. Standalone (Algorithm Only) Performance

Yes, a standalone performance assessment was done. The study specifically evaluated the "automated corneal epithelial thickness measurements" generated by the device's software. The comparison to manual measurements serves as a validation of this standalone algorithm's performance against human expert measurements.

7. Type of Ground Truth Used

The ground truth for the corneal epithelial thickness measurements was established by expert manual measurements performed by three masked graders. This serves as a reference standard to which the automated measurements were compared.

8. Sample Size for the Training Set

The document does not provide information about the sample size used for the training set of the algorithm. This study focuses on the clinical evaluation of the device in its final form.

9. How the Ground Truth for the Training Set Was Established

The document does not provide information on how the ground truth for the training set (if any) was established. The clinical evaluation described pertains to the performance validation of the already developed algorithm.

Ask a specific question about this device

The CIRRUS™ HD-OCT is a non-contact, high resolution tomographic and biomicroscopic imaging device intended for in-vivo viewing, axial cross-sectional, and three-dimensional imaging of anterior ocular structures. The device is indicated for visualizing anterior and posterior ocular structures, including cornea, retina, retinal nerve fiber layer, ganglion cell plus inner plexiform layer, macula, and optic nerve head. The CIRRUS normative databases are quantitative tools indicated for the comparison of retinal nerve fiber layer thickness, ganglion cell plus inner plexiform layer thickness, and optic nerve head measurements to a database of normal subjects. The CIRRUS OCT angiography is indicated as an aid in the visualization of vascular structures of the retina and choroid. The CIRRUS HD-OCT is indicated 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 CIRRUSTM HD-OCT is a computerized instrument that acquires and analyzes crosssectional tomograms of anterior and posterior ocular structures (including cornea, retina, retinal nerve fiber layer, macula, and optic disc). It employs non-invasive, non-contact, low-coherence interferometry to obtain these high-resolution images. Using this noninvasive optical technique. CIRRUS HD-OCT produces high-resolution cross-sectional tomograms of the eye without contacting the eye. It also produces images of the retina and layers of the retina from an en face perspective (i.e., as if looking directly in the eye).

The CIRRUS HD-OCT is offered in four models, Model 4000, 400, 5000 and 500. In the CIRRUS HD-OCT Models 4000 and 5000, the fundus camera is a line scanning ophthalmoscope. The CIRRUS HD-OCT Models 400 and 500 are similar to the Models 4000 and 5000 except that they provide the fundus image using the OCT scanner only.

The acquired imaging data can be analyzed to provide thickness and area measurements of regions of interest to the clinician. The system uses acquired data to determine the fovea location or the optic disc location. Measurements can then be oriented using the fovea and/or optic disc locations. The patient's results can be compared to subjects without disease for measurements of RNFL thickness, neuro-retinal rim area, average and vertical cup-to-disc area ratio, cup volume, macular thickness and ganglion cell plus inner plexiform layer thickness.

In addition to macular and optic disc cube scans, the CIRRUS HD-OCT also offers scans for OCT angiography imaging, a non-invasive approach with depth sectioning capability to visualize microvascular structures of the eye.

Anterior segment scans enable analysis of the anterior segment including Anterior Chamber Depth. Angle-to-Angle and automated measurement of the thickness of the cornea with the Pachymetry scan.

Here's an analysis of the provided text, outlining the acceptance criteria and the study details for the Carl Zeiss Meditec Inc. Cirrus HD-OCT with Software Version 8.

Acceptance Criteria and Device Performance:

The document primarily focuses on demonstrating repeatability and reproducibility of measurements, and comparability between the new device (CIRRUS HD-OCT with Software Version 8, specifically Models 4000 and 5000) and the predicate device (Visante OCT). Specific acceptance criteria are not explicitly stated as numerical targets in quantifiable terms for all parameters. Instead, the study aims to show that the new device's measurements are consistent and comparable to the predicate device. The tables report the demonstrated performance rather than predefined acceptance criteria.

Table of Acceptance Criteria and Reported Device Performance (Inferred from the study's objective to demonstrate comparability, repeatability, and reproducibility):

Study Details:

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

   • Anterior Chamber and Pachymetry Scans:
     • Normal Cornea group: 46 subjects (Group 1)
     • Post-LASIK group: 40 subjects (Group 2)
     • Corneal Pathology group: 45 subjects (Group 3)
     • Age Range (all groups): 25 to 69 years.
     • Data Provenance: Not explicitly stated (e.g., country of origin). The study is described as a "non-significant risk clinical study," implying prospective data collection in a clinical setting.
   • Angle Study:
     • Angle Study: 27 subjects, ranging from 43 to 77 years (mean 62 years).
     • Specific eye counts per measurement type:
       • 26 eyes for Wide Angle-to-Angle scan
       • 27 eyes for HD Angle scan
     • Data Provenance: Not explicitly stated (e.g., country of origin). Described as a "non-significant risk clinical study," implying prospective data collection.
   • OCT Angiography: Series of case studies. No specific sample size is provided beyond "case studies."

2. Number of experts used to establish the ground truth for the test set and their qualifications:

   • Anterior Chamber and Pachymetry Scans: No external experts were used for "ground truth" in the sense of an adjudicated diagnosis. The study focused on comparing measurements between devices and assessing repeatability/reproducibility. The predicate device (Visante OCT) served as the reference for comparability. One operator acquired data on the Visante OCT. Three operators acquired data on the CIRRUS HD-OCT devices.
   • Angle Study: The study focused on device measurement comparison, repeatability, and reproducibility. One operator acquired data on the Visante OCT. Three operators acquired data on the CIRRUS HD-OCT devices. The study population had a variety of angle configurations (Grade II to Grade IV) as assessed by gonioscopy using the Shaffer method, but the gonioscopy results themselves were not used as a direct "ground truth" for individual measurement values from the OCT devices.
   • OCT Angiography: The "findings demonstrate that the CIRRUS OCT Angiography... can give non-invasive three-dimensional information regarding retinal microvasculature." This was compared with "fluorescein angiography images." The number and qualifications of experts interpreting either the OCTA or fluorescein angiography images for these case studies are not specified in the provided text.

3. Adjudication method for the test set:

   • Anterior Chamber and Pachymetry Scans: Not applicable for establishing ground truth, as the study focused on device measurement comparison and repeatability/reproducibility across devices and operators. Measurements were generated by "manual placement of software tools."
   • Angle Study: Similar to the above, not applicable for establishing "ground truth" through adjudication. Measurements were generated by "manual placement of software tools (Angle tool; TISA tool)" by the operators who acquired the data.
   • OCT Angiography: Not specified.

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:

   • No, a multi-reader multi-case (MRMC) comparative effectiveness study with human readers and AI assistance was not explicitly reported or performed in the provided text. The study primarily focused on the device's technical performance (repeatability, reproducibility, and agreement with a predicate device) rather than its impact on human reader diagnostic accuracy with or without AI. The device itself (CIRRUS HD-OCT with Software Version 8) is an imaging and measurement device, not explicitly an "AI" diagnostic tool in the context of human-in-the-loop performance measurement.

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

   • Yes, in a sense, the primary studies for Anterior Chamber, Pachymetry, and Angle measurements assessed the standalone performance of the device's measurement algorithms by evaluating their repeatability, reproducibility, and agreement with predicate device measurements. These measurements are generated by the device's software (e.g., "manual placement of software tools" by an operator, but the calculation is still algorithmic). The OCT Angiography section also mentions the device's ability to provide information on microvasculature without human "in-the-loop" interpretation for the basic image generation. However, it's not described as an "AI algorithm" in the common sense of diagnosing from images.

6. The type of ground truth used:

   • Anterior Chamber, Pachymetry, and Angle Studies: The "ground truth" was effectively the measurements obtained from the predicate device (Visante OCT). The purpose was to show substantial equivalence and comparability, not to determine diagnostic accuracy against an independent, gold-standard clinical pathology or outcome.
   • OCT Angiography: Comparisons were made with fluorescein angiography images. Fluorescein angiography is a clinical gold standard for visualizing retinal and choroidal vasculature.

7. The sample size for the training set:

   • The document does not provide information on the training set size. The studies described are validation studies for the device's technical performance and comparability to a predicate. The device incorporates "proprietary algorithms" and "normative databases," which would have been developed using some form of training data, but details about this training data are not included in the provided 510(k) summary.

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

   • Not described in the provided text. As mentioned above, the 510(k) summary focuses on the validation studies, not the development or training of any underlying algorithms or normative databases. The "normative databases" would inherently rely on data from "normal subjects," but the specifics of their ground truth establishment are not given.

Ask a specific question about this device

The Cirrus™ HD-OCT with Retinal Nerve Fiber Layer (RNFL), Macular, Optic Nerve Head and Ganglion Cell Normative Databases is indicated for in-vivo viewing, axial cross-sectional, and three-dimensional imaging and measurement of anterior and posterior ocular structures.

The Cirrus™ HD-OCT is a non-contact, high resolution tomographic and biomicroscopic imaging device. It is indicated for in-vivo viewing, axial cross-sectional, and threedimensional imaging and measurement of anterior and posterior ocular structures, including cornea, retinal nerve fiber layer, ganglion cell plus inner plexiform layer, macula, and optic nerve head. The Cirrus normative databases are quantitative tools for the comparison of retinal nerve fiber layer thickness, macular thickness, ganglion cell plus inner plexiform layer thickness, and optic nerve head measurements to a database of normal subjects. The Cirrus HD-OCT 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™ HD-OCT is a computerized instrument that acquires and analyzes crosssectional tomograms of anterior and posterior ocular structures (including cornea, retina, retinal nerve fiber layer, macula, and optic disc). It employs non-invasive, non-contact, low-coherence interferometry to obtain these high-resolution images. Using this noninvasive optical technique, Cirrus HD-OCT produces high-resolution cross-sectional tomograms of the eye without contacting the eye. It also produces images of the retina and layers of the retina from an en face perspective (i.e., as if looking directly in the eye).

The Cirrus HD-OCT is offered in two models, Model 4000 and Model 400. In the Cirrus HD-OCT Model 4000 instrument, the fundus camera is a line scanning ophthalmoscope. The Cirrus HD-OCT Model 400 is similar to the Model 4000 except that it provides the fundus image using the OCT scanner only.

The acquired imaging data can be analyzed to provide thickness and area measurements of regions of interest to the clinician. The system uses acquired data to determine the fovea location or the optic disc location. Measurements can then be oriented using the fovea and/or optic disc locations. The patient's results can be compared to subjects without disease for measurements of RNFL thickness, neuroretinal rim area, average and vertical cup-to-disc area ratio, cup volume, macular thickness and ganglion cell plus inner plexiform layer thickness.

Visit-to-visit comparison of images and measurements is available for the macula. Specifically, change in macular thickness, area and volume of Retinal Pigment Epithelium (RPE) elevations, area of sub-RPE illumination and distance of Sub-RPE illumination to the fovea. Change analysis of multiple visits, up to eight, can be performed for RNFL thickness, neuroretinal rim area, average and vertical cup-to-disc area ratio, cup volume, and macular thickness.

The provided document describes the predicate device and the clinical studies performed to support the substantial equivalence of the "Cirrus HD-OCT with Retinal Nerve Fiber Layer (RNFL), Macular, Optic Nerve Head and Ganglion Cell Normative Databases" (Cirrus HD-OCT). The clinical evaluation focuses on demonstrating the device's measurement capabilities and the establishment of normative databases.

Here's an analysis of the acceptance criteria and study information:

1. Table of Acceptance Criteria and Reported Device Performance:

The document doesn't explicitly state "acceptance criteria" in a pass/fail quantifiable manner for the overall device's performance against a gold standard for disease detection or management. Instead, the studies presented focus on the repeatability and reproducibility of various measurements performed by the device and the comparability of certain automated measurements to expert manual measurements from different imaging modalities.

The tables below summarize the specified repeatability and reproducibility limits (acceptance criteria as per ISO 5725-1 and ISO 5725-6, defined as the upper 95% limit for the difference between repeated results) and the reported performance (SD and limits).

Table 1. Repeatability and Reproducibility of Area of Sub-RPE Illumination (Automated Algorithm)

Table 2. Repeatability and Reproducibility of Closest Distance to Fovea (Automated Algorithm)

Table 3. Repeatability and Reproducibility of Area of Sub-RPE Illumination (Manually Edited)

Table 4. Repeatability and Reproducibility of Closest Distance to Fovea (Manually Edited)

Table 5. Repeatability and Reproducibility of Area of RPE Elevations

Table 6. Repeatability and Reproducibility of Volume of RPE Elevations

Table 7. Cirrus Repeatability and Reproducibility of GCA and ONH Parameters - Normal Subjects

Table 8. Repeatability and Visit-to-Visit Variability of ONH Parameters - Glaucomatous Subjects

Table 9. Repeatability of GCA Parameters - Glaucomatous Subjects

Summary of Studies Demonstrating Acceptance:

The studies listed above (Advanced RPE Analysis Study, measurements of Elevated RPE, Repeatability and Reproducibility studies) and the establishment of the normative databases serve as the proof that the device meets its intended use and demonstrates substantial equivalence. The "acceptance criteria" appear to be the demonstrated repeatability and reproducibility limits themselves, calculated according to ISO 5725-1 and ISO 5725-6 standards, or the statistical comparability (e.g., paired t-test showing no significant difference, good correlation R+) in the case of the RPE illumination comparison.

2. Sample Sizes used for the Test Set and Data Provenance:

• Advanced RPE Analysis Study (Areas of Increased Illumination):
  • Test Set Sample Size: 52 eyes from 52 subjects.
  • Data Provenance: Not explicitly stated (e.g., country of origin), but it was a "non-significant risk clinical study" with "Four sites participated in the clinical data collection." The study was conducted on subjects evaluated for dry AMD with geographic atrophy. The study compares Cirrus HD-OCT automated measurements to FAF images (an existing, accepted modality). This implies a prospective data collection for this comparison.
• Advanced RPE Analysis Study (Measurements of Elevated RPE):
  • Test Set Sample Size: 70 eyes from 70 subjects were considered; the number included in final analysis is not explicitly stated but implies a similar number after qualification.
  • Data Provenance: Not explicitly stated (e.g., country of origin), but "Three sites participated in the clinical data collection." Subjects were 50 years or older with dry age-related macular degeneration (AMD) with macular drusen. This implies a prospective data collection for this comparison.
• Measurements of Area of Increased Illumination Under the RPE Repeatability and Reproducibility:
  • Test Set Sample Size: Phase 1: 49 eyes of 37 subjects. Phase 2: 53 eyes of 39 subjects.
  • Data Provenance: "Single-site clinical study." The subjects had dry AMD with geographic atrophy. This suggests prospective data collection.
• Measurements of Elevated RPE Repeatability and Reproducibility:
  • Test Set Sample Size: Phase 1: 26 eyes of 23 subjects. Phase 2: 24 eyes of 21 subjects.
  • Data Provenance: "Single-site clinical study." The subjects had dry AMD with macular drusen. This suggests prospective data collection.
• Optic Nerve Head and Ganglion Cell Analysis Repeatability and Reproducibility (Normal Subjects):
  • Test Set Sample Size: 63 normal subjects.
  • Data Provenance: Not explicitly stated (e.g., multi-site, country). This suggests prospective data collection.
• Optic Nerve Head Parameters Repeatability and Visit-to-Visit Variability (Glaucomatous Subjects):
  • Test Set Sample Size: 55 glaucomatous subjects.
  • Data Provenance: Not explicitly stated (e.g., multi-site, country). This suggests prospective data collection.
• GCA Parameters Repeatability (Glaucomatous Subjects):
  • Test Set Sample Size: 119 subjects with glaucoma enrolled; 94 subjects with two qualified scans each were included in the analysis.
  • Data Provenance: "clinical study conducted at four sites." Not explicitly stated (e.g., country). This suggests prospective data collection.

3. Number of Experts used to Establish the Ground Truth for the Test Set and Qualifications:

• Advanced RPE Analysis Study (Areas of Increased Illumination): Ground truth was established by "expert manual measurements of areas of hypofluorescence typical of geographic atrophy in fundus autofluorescence (FAF) images." The number and specific qualifications of these experts are not specified in the document.
• Advanced RPE Analysis Study (Measurements of Elevated RPE): Ground truth was established by "manually drawn by experts designated as drusen on color fundus photographs (CFPs)." The number and specific qualifications of these experts are not specified in the document.
• For repeatability and reproducibility studies, the ground truth is implicitly the repeated measurements themselves by the device, as the studies aim to quantify the variation within the device's measurements, not compare them against an external gold standard by human experts for classifications.

4. Adjudication Method:

• For the studies comparing automated measurements to expert manual measurements (e.g., Advanced RPE Analysis), the adjudication method is not explicitly stated. It mentions "expert manual measurements" and "manually drawn by experts," implying these were used as reference, but how inconsistencies between experts (if there were multiple) were resolved is not detailed.
• For the repeatability and reproducibility studies, adjudication doesn't apply in the traditional sense, as these studies focus on the intrinsic variation of the device's measurements.

5. If a Multi Reader Multi Case (MRMC) Comparative Effectiveness Study was done, and if so, what was the effect size of how much human readers improve with AI vs without AI assistance:

• No MRMC comparative effectiveness study was described where human readers' performance with and without AI assistance was evaluated. The studies primarily focus on the device's measurement accuracy, repeatability, and reproducibility, and the establishment of normative databases. The comparison for RPE illumination detection was between the device's automated measurements and expert manual measurements from a different imaging modality, not an AI-assisted human vs. human-alone scenario.

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

• Yes, the comparability studies for Advanced RPE Analysis (Areas of Increased Illumination and Elevated RPE) assess the device's "automated measurements" (automated algorithm) against expert manual measurements from other imaging modalities. These can be considered standalone performance assessments of specific algorithms within the device.
• Similarly, all the repeatability and reproducibility studies for various parameters (GCA, ONH, RPE illumination, RPE elevations) assess the intrinsic performance of the device's measurement algorithms in a standalone manner, quantifying their consistency.

7. The Type of Ground Truth Used:

• Advanced RPE Analysis Study (Areas of Increased Illumination): Ground truth was "expert manual measurements of areas of hypofluorescence typical of geographic atrophy in fundus autofluorescence (FAF) images." This is a form of expert consensus/manual delineation using another imaging modality.
• Advanced RPE Analysis Study (Measurements of Elevated RPE): Ground truth was "manually drawn by experts designated as drusen on color fundus photographs (CFPs)." This is also a form of expert consensus/manual delineation using another imaging modality.
• Normative Databases (RNFL, Macular, Optic Nerve Head, Ganglion Cell): The ground truth for these databases is implicitly the classification of subjects as "normal" based on clinical criteria (not specified in detail, but implied by the selection of normal subjects). The measurements are then statistically analyzed to establish reference values.
• For repeatability and reproducibility studies, the ground truth is the repeated measurements, not an external gold standard.

8. The Sample Size for the Training Set:

The document doesn't explicitly mention "training sets" as it would for a typical machine learning algorithm development (e.g., deep learning). The device is likely based on established image processing algorithms.

• The normative databases serve a similar function to a reference set, providing "normal" ranges for comparison in diagnosis.
  • Optic Nerve Head Normative Database: Derived from a "post-hoc analysis" of 282 eyes from 284 subjects (aged 19-84 years) included in a previous RNFL normative database (K083291).
  • Ganglion Cell Normative Database: Utilized the "same 282 subjects, aged 19-84 years that were deemed representative of a normal population" as the original macula normative database (K083291).

9. How the Ground Truth for the Training Set was Established:

• For the normative databases, the "ground truth" for inclusion was that the subjects were considered "normal." The document states the subjects (n=282) were "deemed representative of a normal population." However, the specific criteria or methods (e.g., expert clinical review, exclusion criteria) used to define and confirm these subjects as "normal" are not detailed in this summary. The data were collected from "seven sites."

Ask a specific question about this device

The Cirrus™ HD-OCT with Retinal Nerve Fiber Layer and Macular Normative Databases is indicated for in-vivo viewing, axial cross-sectional, and threedimensional imaging and measurement of anterior and posterior ocular structures.

The Cirrus™ HD-OCT is a non-contact, high resolution tomographic and biomicroscopic imaging device. It is indicated for in-vivo viewing, axial crosssectional, and three-dimensional imaging and measurement of anterior and posterior ocular structures, including cornea, retinal nerve fiber layer, macula, and optic disc. The Cirrus HD-OCT with Retinal Nerve Fiber Layer (RNFL) and Macular Normative Database is a quantitative tool for the comparison of retinal nerve fiber layer 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™ HD-OCT is a computerized instrument that acquires and analyzes crosssectional tomograms of anterior and posterior ocular structures (including cornea, retina, retinal nerve fiber layer, macula, and optic disc). It emplovs non-invasive, non-contact, low-coherence interferometry to obtain these high-resolution images. Using this non-invasive optical technique. Cirrus HD-OCT produces high-resolution cross-sectional tomograms of the eye without contacting the eye.

Here's a breakdown of the acceptance criteria and study details for the Cirrus HD-OCT device, based on the provided 510(k) summary:

Acceptance Criteria and Device Performance

The device's performance was evaluated across several key areas: Retinal Nerve Fiber Layer (RNFL) repeatability and reproducibility, retinal segmentation accuracy, retinal segmentation precision, agreement of retinal thickness measurements with a predicate device (Stratus OCT), and central corneal thickness measurements (CCT) repeatability, reproducibility, and agreement with ultrasound pachymetry.

Table 1: Acceptance Criteria and Reported Device Performance

• AMD: 85.7% (60/70)
• Diabetic Retinopathy: 95.2% (40/42)
• VRI Disorder: 96.4% (27/28)
• Other Retinal Disease: 86.3% (44/51)
• Macular Edema: 96.4% (27/28)
• No Retinal Disease: 100.0% (37/37)
  512x128 Scans:
• AMD: 86.1% (62/72)
• Diabetic Retinopathy: 97.6% (41/42)
• VRI Disorder: 89.3% (25/28)
• Other Retinal Disease: 88.5% (46/52)
• Macular Edema: 93.1% (27/29)
• No Retinal Disease: 100.0% (40/40) |
  | Retinal Segmentation Accuracy (ILM Layer) | Software and hand-segmentations agreed for 100% of A-scans within specified tolerances. | 200x200 Scans:
• AMD: 97.1% (68/70)
• Diabetic Retinopathy: 95.2% (40/42)
• VRI Disorder: 92.9% (26/28)
• Other Retinal Disease: 98.0% (50/51)
• Macular Edema: 100.0% (28/28)
• No Retinal Disease: 100.0% (37/37)
  512x128 Scans:
• AMD: 98.6% (73/74)
• Diabetic Retinopathy: 95.2% (40/42)
• VRI Disorder: 96.3% (26/27)
• Other Retinal Disease: 98.1% (51/52)
• Macular Edema: 96.6% (28/29)
• No Retinal Disease: 100.0% (40/40) |
  | Retinal Segmentation Precision (CSMT Repeatability) | Low standard deviation for central subfield macular thickness measurements. | Repeatability SD (µm) with Cirrus 4.0 MTA with Fovea Placement:
• AMD: 6.3
• DR: 9.8
• VRI Disorder: 5.4
• Other: 7.5
• ME: 7.9
• No Disease: 2.2
  Significantly improved with Fovea Placement and Registration compared to Cirrus 3.0 MTA alone. |
  | Agreement of Retinal Thickness with Stratus OCT | Demonstrate and explain the mean difference, acknowledge non-interchangeability. | Mean Difference Cirrus – Stratus (µm) for Central Subfield:
• AMD: 53.6 (SD: 35.0)
• Diabetic Retinopathy: 40.0 (SD: 47.1)
• VRI Disorder: 43.8 (SD: 35.9)
• Other: 41.7 (SD: 47.1)
• Macular Edema: 45.5 (SD: 45.3)
• Normal: 59.4 (SD: 11.7)
  Non-interchangeable; better for qualitative comparison. |
  | CCT Repeatability and Reproducibility | Not explicitly stated as acceptance criteria, but demonstrating low variability. | Repeatability SD: 4.08 µm
  Repeatability Limits: 11.42 µm
  Reproducibility SD: 4.23 µm
  Reproducibility Limits: 11.84 µm |
  | CCT Agreement with Ultrasound Pachymetry | Demonstrate and explain the mean difference and consistency with other OCT devices. | Mean difference (Cirrus CCT - Ultrasound pachymetry CCT): -9.06 µm (SD: 5.63).
  Consistent with other OCT devices (e.g., Visante OCT is ~15.1 µm thinner than ultrasound). |

Study Details

The provided document describes several clinical evaluations rather than a single overarching study. Here's a breakdown for each:

1. RNFL Repeatability and Reproducibility Study

• Sample Size (Test Set): 32 normal subjects.
• Data Provenance: In-house study, likely US (not explicitly stated, but common for domestic manufacturers). Retrospective/Prospective not specified, but the nature of the study (inter-visit, inter-instrument) points to prospective data collection for this specific evaluation.
• Number of Experts & Qualifications: Not applicable or specified, as this focuses on objective measurement variability.
• Adjudication Method: Not applicable.
• MRMC Comparative Effectiveness Study: No.
• Standalone Performance: Yes (algorithm/device performance in measurement).
• Type of Ground Truth: N/A for repeatability/reproducibility; focuses on consistency of device measurements.
• Training Set Sample Size: N/A (this study evaluates the device's measurement consistency, not an AI model).
• Training Set Ground Truth: N/A.

2. RNFL Agreement with Stratus OCT Study

• Sample Size (Test Set): 130 subjects (normal and patients).
• Data Provenance: Not explicitly stated, but the reference to "a recent study" and the nature of the publication suggests a clinical research setting, likely academic and possibly multi-site. The reference mentions "Poster 4628, ARVO 2008," indicating an academic presentation.
• Number of Experts & Qualifications: Not applicable for establishing ground truth as it's a comparison of device measurements.
• Adjudication Method: Not applicable.
• MRMC Comparative Effectiveness Study: No.
• Standalone Performance: Yes (comparing two devices' standalone measurements).
• Type of Ground Truth: N/A; comparing measurements from two different OCT systems.
• Training Set Sample Size: N/A.
• Training Set Ground Truth: N/A.

3. Retinal Segmentation Accuracy and Precision Study

• Sample Size (Test Set): Both eyes of 370 subjects. One eye chosen as the "study eye." Subjects classified into 6 pathology groups (AMD, DR, VRI, Other Retinal, Macular Edema, No Retinal Pathology).
• Data Provenance: Four sites, location not specified (likely US, given the sponsor). Retrospective/Prospective not explicitly stated, but the controlled scanning protocol suggests prospective data collection for this evaluation.
• Number of Experts & Qualifications: Not explicitly stated, but "hand-segmentations" imply expert manual delineation. No specific qualifications are listed.
• Adjudication Method: Not explicitly stated, but since agreement was defined as software and hand-segmentations being within a certain micron range, there was a defined metric for evaluating agreement.
• MRMC Comparative Effectiveness Study: No.
• Standalone Performance: Yes (algorithm's ability to segment layers).
• Type of Ground Truth: "Hand-segmentations" by experts. This serves as the reference standard for evaluating the automatic segmentation algorithm's accuracy.
• Training Set Sample Size: N/A (this study evaluates the segmentation algorithm, not AI training specific to this device).
• Training Set Ground Truth: N/A.

4. Retinal Thickness Measurements: Agreement with Stratus Study

• Sample Size (Test Set): N=63 (AMD), N=39 (Diabetic Retinopathy), N=45 (VRI Disorder), N=53 (Other), N=35 (Macular Edema), N=48 (Normal) - total not explicitly summed but likely combined from the larger 370 subject cohort, or a subset.
• Data Provenance: Implied same as the Retinal Segmentation study (four sites, likely US, prospective-like data collection).
• Number of Experts & Qualifications: Not applicable for establishing ground truth, as it's a comparison of device measurements.
• Adjudication Method: Not applicable.
• MRMC Comparative Effectiveness Study: No.
• Standalone Performance: Yes (comparing two devices' standalone measurements).
• Type of Ground Truth: N/A; comparing measurements from two different OCT systems.
• Training Set Sample Size: N/A.
• Training Set Ground Truth: N/A.

5. Central Corneal Thickness (CCT) Measurements Study

• Sample Size (Test Set): Phase I: 28 subjects; Phase II: 22 subjects (different subjects). For CCT vs. Ultrasound: 50 eyes.
• Data Provenance: In-house study (likely US). Prospective data collection for both phases.
• Number of Experts & Qualifications: Not applicable for establishing ground truth for CCT measurements.
• Adjudication Method: Not applicable.
• MRMC Comparative Effectiveness Study: No.
• Standalone Performance: Yes (device measurement performance).
• Type of Ground Truth: N/A for repeatability/reproducibility. For CCT vs. Ultrasound, ultrasound pachymetry serves as a comparative reference, but not explicitly stated as "ground truth" for CCT itself in the same vein as expert pathology.
• Training Set Sample Size: N/A.
• Training Set Ground Truth: N/A.

6. RNFL and Macula Normative Databases

• Sample Size (Test Set): RNFL: 284 subjects (aged 19-84); Macula: 282 subjects (aged 19-84).
• Data Provenance: Seven sites, location not specified (likely multi-center within the US or potentially international, but typical for normative data collection). Prospective.
• Number of Experts & Qualifications: Not explicitly stated, but the subjects would have been confirmed "normal" by qualified clinicians, typically ophthalmologists or optometrists.
• Adjudication Method: N/A for normative database collection directly.
• MRMC Comparative Effectiveness Study: No.
• Standalone Performance: Not applicable, as this is the collection and analysis of normative data, not an algorithm's performance against it.
• Type of Ground Truth: Clinically determined healthy/normal status of subjects.
• Training Set Sample Size: N/A, this is the data from which normative ranges are derived or "trained" for comparison.
• Training Set Ground Truth: "Known normal subjects" based on clinical assessment.

Summary of AI Specifics:

The provided document describes the performance of an optical coherence tomography (OCT) device and its inherent algorithms for retinal layer segmentation and thickness measurement. It does not explicitly mention the use of "AI" in the modern sense of machine learning or deep learning models, nor does it present an "AI vs. no AI assistance" comparative effectiveness study for human readers. The algorithms described are more in line with traditional image processing and segmentation techniques. Therefore, questions related to AI effect size, AI training set specifics, and AI ground truth are not directly answered by this 2009 submission.

Ask a specific question about this device

The Cirrus™ HD-OCT is a non-contact, high resolution tomographic and biomicroscopic imaging device. It is indicated for in vivo viewing, axial cross-sectional, and threedimensional imaging and measurement of posterior ocular structures, including retina, retinal nerve fiber layer, macula, and optic disc. 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™ HD-OCT is a computerized instrument that acquires and analyzes crosssectional tomograms of posterior ocular structures (including retina, retinal nerve fiber layer, macula, and optic disc.). It employs non-invasive, non-contact, low-coherence interferometry to obtain these high-resolution images. Using this non-invasive optical technique, Cirrus HD-OCT produces high-resolution cross-sectional tomograms of the eye without contacting the eye.

The document provided is a 510(k) Summary for the Carl Zeiss Meditec Inc. Cirrus™ HD-OCT device, submitted to the FDA. It focuses on demonstrating substantial equivalence to predicate devices rather than providing specific acceptance criteria and detailed study results for device performance.

Therefore, the requested information elements related to specific acceptance criteria, detailed device performance metrics, sample sizes for test/training sets, ground truth establishment methods, expert qualifications, and comparative effectiveness studies cannot be fully extracted from the provided text.

Here's a breakdown of what can be extracted or inferred:

1. Table of Acceptance Criteria and Reported Device Performance:

This information is not provided in the document. The 510(k) summary asserts that "all testing deemed necessary was conducted" and that the device "demonstrates the device is substantially equivalent to the predicate devices and does not raise new questions regarding safety and effectiveness with respect to ophthalmoscopes." However, no specific performance metrics or acceptance criteria (e.g., accuracy, sensitivity, specificity, measurement tolerances) are detailed.

2. Sample Size Used for the Test Set and Data Provenance:

• Sample Size for Test Set: Not explicitly stated. The document mentions "Clinical data was collected and evaluated."
• Data Provenance (e.g., country of origin, retrospective/prospective): Not explicitly stated. It only mentions "Clinical data was collected."

3. Number of Experts Used to Establish the Ground Truth for the Test Set and the Qualifications of Those Experts:

This information is not provided in the document.

4. Adjudication Method (e.g., 2+1, 3+1, none) for the Test Set:

This information is not provided in the document.

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:

This information is not provided and is unlikely to be relevant. The Cirrus™ HD-OCT is described as an imaging device for "in vivo viewing, axial cross-sectional, and three-dimensional imaging and measurement of posterior ocular structures." It's a diagnostic device to aid in detection and management of diseases. The description does not suggest an AI-driven interpretation or assistive role for human readers in the way an MRMC study typically assesses. The primary goal of the 510(k) is to demonstrate substantial equivalence of the device itself as an ophthalmoscope, not an AI-powered diagnostic tool.

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

This information is not provided. As noted above, the device's function is imaging and measurement, not solely algorithmic diagnosis.

7. The Type of Ground Truth Used (expert consensus, pathology, outcomes data, etc.):

This information is not provided. The document only states "Clinical data was collected and evaluated."

8. The Sample Size for the Training Set:

This information is not provided. The Cirrus™ HD-OCT is an imaging device, and the summary focuses on its hardware and measurement capabilities rather than an AI model that would require a distinct training set. If there are normative databases, their sample sizes are not mentioned.

9. How the Ground Truth for the Training Set was Established:

This information is not provided.

Summary of available information regarding the clinical evaluation (as described in the 510(k) summary):

The document states:

• Study Purpose: "Clinical data was collected and evaluated to support the indications for use statement for the Cirrus™ HD-OCT and to demonstrate substantial equivalence to the Stratus OCT."
• Conclusion: "As described in this 510(k) Summary, all testing deemed necessary was conducted on the Cirrus™ HD-OCT to ensure that the device is safe and effective for its intended use when used in accordance with its Instructions for Use."

Crucially, this 510(k) summary is for an ophthalmoscope, an imaging device. The focus is on its ability to produce images and measurements comparable to existing predicate devices, rather than a novel AI diagnostic algorithm with specific performance criteria against a detailed ground truth. The document does not provide the kind of detailed clinical study results (e.g., sensitivity, specificity, AUC, reader performance) that would typically be found for an AI-powered diagnostic aid. The substantial equivalence argument often relies on demonstrating that the new device operates similarly and yields comparable output to predicate devices, without needing to re-prove the clinical utility of the imaging modality itself.

Ask a specific question about this device