Search Results

MAIA is intended for taking digital images of a human retina without the use of a mydriatic agent.

MAIA is intended for taking digital images of a human retina without the use of a mydriatic agent.

The device works with a dedicated software application, operates as a standalone unit, integrates a multitouch display, a push-button and is provided with an external power supply.

MAIA operates in non-mydriatic conditions, i.e. without the need of pharmacological dilation and is intended for prescription use only

The key functional elements of the device are:

• the device base,
• the device optical head,
• the headrest,
• the chinrest,
• the multi-touch display,
• the embedded software,
• patient push-button,
• the power supply and its power cable.

The device acquires confocal retinal images illuminating the retina of the patient's eye, with visible light for imaging purposes and with infrared light for imaging purposes, focusing and retinal tracking.

A specific feature of MAIA is the pupil tracking. The patient's pupils are illuminated and viewed by a different optical system; two cameras track the pupil's movements and allow for automatic alignment of the optical head with the eye position.

From an imaging point of view, an important characteristic of MAIA is that it is a confocal instrument: this means that the illumination of the retina is focused on the same plane of the acquisition (same focus). Compared to traditional imaging, where the entire specimen is flooded evenly in light, a confocal illumination system is able to focus the illuminating light on the same focus plane (conjugate plane) of the acquisition optics.

Another important characteristic of MAIA is the use of white light to illuminate the retina.

The result of the combination of confocal imaging and white light obtained in MAIA is the acquisition of sharp, detailed, naturally colored retinal images.

This medical device product has functions subject to FDA premarket review as well as functions that are not subject to FDA premarket review.

For this application, the product has also functions that are not subject to FDA premarket review; FDA assesses those functions only to the extent that they either could adversely impact the safety and effectiveness of the functions subject to FDA premarket review or they are included as a labeled positive impact that was considered in the assessment of the functions subject to FDA premarket review.

The "other function" of the device is the automatic perimetry (Product Code HPT, 510(K) Exempt), that allows the measurement of retinal threshold sensitivity and the analysis of fixation.

This function consists in projecting light stimuli at different intensities together with a uniform light background; the device records the pressures of the pushbutton by the tested patient when he detects such stimuli, as in Standard Automated Perimetry test.

With its retinal tracking, for all the duration of the exam, MAIA acquires infrared images of the retina and detects its movement, to correct the position for the stimuli, reducing the positioning error that might occur if the patient has poor fixation stability.

The provided text describes the Centervue Spa MAIA (AHMACME001) device, which is an ophthalmoscope intended for taking digital images of a human retina. The 510(k) submission (K243504) seeks to demonstrate substantial equivalence to the predicate device COMPASS (K150320).

However, the provided document explicitly states: "No clinical tests were needed." and "Based on the non-clinical tests (i.e. bench tests), MAIA is safe and performs as intended when used according to its indications for use and in accordance with its labeling. It performs as well as the legally marketed predicate device COMPASS (K150320)." This means that the submission does not contain a study proving the device meets acceptance criteria related to a clinical study involving human performance (e.g., MRMC studies, standalone algorithm performance, expert ground truth).

The acceptance criteria described pertain to design and manufacturing equivalence, and bench testing, rather than clinical performance for a diagnostic device. Since no clinical study was conducted or provided in this submission, it is impossible to populate most of the requested fields regarding performance, sample sizes, expert ground truth, adjudication, or MRMC studies.

Here's how to address the request based only on the provided information, acknowledging the absence of a clinical study:

Acceptance Criteria and Device Performance (Based on Non-Clinical Tests)

Since no clinical studies were performed to establish performance metrics against human-defined acceptance criteria (e.g., sensitivity, specificity, accuracy), the "acceptance criteria" here refer to the successful completion of various engineering and safety tests demonstrating equivalence to the predicate device.

Study Details (None for Clinical Performance)

1. Sample size used for the test set and the data provenance: Not applicable. No clinical test set. The submission relies on "non-clinical tests (i.e. bench tests)" and equivalence to the predicate device.
2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts: Not applicable. No clinical test set requiring expert ground truth for performance evaluation.
3. Adjudication method (e.g. 2+1, 3+1, none) for the test set: Not applicable. No clinical test set.
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: Not applicable. No MRMC study was performed. The device is for "taking digital images," not providing AI assistance for diagnosis.
5. If a standalone (i.e. algorithm only without human-in-the loop performance) was done: Not applicable for performance metrics like sensitivity/specificity for a diagnostic outcome. The device's primary function described is image acquisition, a standalone function that was validated via bench testing and comparison to the predicate's technology.
6. The type of ground truth used (expert consensus, pathology, outcomes data, etc.): Not applicable for clinical performance. The ground truth for the "technical performance" and "safety" of the device (such as image resolution, field of view, and electrical safety) would be established by engineering specifications, calibration standards, and regulatory safety standards.
7. The sample size for the training set: Not applicable. No machine learning training set mentioned or implied for a diagnostic algorithm. The device, an ophthalmoscope, acquires images. While it has components like pupil tracking and automatic alignment, these are presented as inherent functions validated through engineering tests, not trained AI models.
8. How the ground truth for the training set was established: Not applicable. No training set for a diagnostic algorithm.

Ask a specific question about this device

iCare ALTIUS CW is a Medical Device Software indicated for the review, processing and analysis of ophthalmic medical images, for the review of video, clinical and diagnostic data, measurements and reports, generated by ophthalmic medical devices or documentation system through computerized networks, to support trained healthcare professionals in the diagnosis and monitoring of several eye pathologies.

iCare ALTIUS CW is a cloud-based software application with a web-based interface able to:

• review medical ophthalmic images, including videos,
• digitally process images,
• review diagnostic data, clinical information and reports,
  from ophthalmic diagnostic instruments. CW does not perform automated image analysis but provides advanced imaging manipulation tools.
  CW allows to review and process diagnostic data and multiple images with different formats (e.g. PDFS, JPEG, ...) and provides the following features:
• image manipulation filters such as zooming, changing brightness and contrast and gamma, RGB filtering,
• side-by-side image comparison (detached or synchronized mode) with different layouts,
• advanced imaging tools, such as flicker between different pictures and mosaics of several images,
• review and print reports generated by ophthalmic devices.
  CW integrates with PACS software systems, which provide the medical images and reports, to be analysed by the CW. The patient data and medical images exchange between CW and PACS is done through computerized networks using secured network communication.
  The web-based interface of CW is designed to be used through a desktop PC or a laptop using keyboard and mouse (further details in the technical requirements section).
  The User Interface is available in the languages required by the applicable regulatory requirement of the country where the device is placed on the market.

The iCare ALTIUS CW device is a Medical Device Software indicated for the review, processing, and analysis of ophthalmic medical images, video, clinical and diagnostic data, measurements, and reports generated by ophthalmic medical devices or documentation systems. It aims to support trained healthcare professionals in the diagnosis and monitoring of various eye pathologies.

The provided text does not contain detailed acceptance criteria or a comprehensive study report with specific performance metrics and statistical results. It describes the device, its intended use, and states that "Software Verification and Validation Testing" was conducted, and "documentation was provided as recommended by FDA's Guidance for Industry and FDA Staff, 'Content of Premarket Submissions for Device Software Functions.'" However, it does not specify what those acceptance criteria were, what the reported device performance against those criteria was, or provide the specifics of the study methodology (e.g., sample sizes, ground truth establishment, expert qualifications, etc.).

Therefore, I cannot fully answer your request based on the provided input.

However, I can extract the available information and highlight what is missing:

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample size used for the test set and the data provenance

• Sample size for test set: Not specified.
• Data provenance: Not specified.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

• Number of experts: Not specified.
• Qualifications of experts: Not specified.

4. Adjudication method for the test set

• Adjudication method: Not specified.

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

• MRMC study: Not specified. The device description states, "CW does not perform automated image analysis but provides advanced imaging manipulation tools," suggesting it's primarily a viewing and processing platform rather than an AI-driven diagnostic tool in the typical sense that would necessitate an MRMC study comparing AI-assisted vs. unassisted human performance in diagnosis or detection. The capabilities listed (zooming, brightness, contrast, comparison, flickering, mosaic, cup-to-disc ratio annotation) are image manipulation and viewing tools.

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

• Standalone study: Not specified. As noted above, the device is described as a tool to "support trained healthcare professionals" and "does not perform automated image analysis." Thus, a standalone algorithm performance evaluation would not be applicable in the same way it would for an autonomous AI diagnostic system.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)

• Type of ground truth: Not specified.

8. The sample size for the training set

• Sample size for training set: Not specified. Given that the device "does not perform automated image analysis," it's unlikely to have a "training set" in the context of a machine learning algorithm.

9. How the ground truth for the training set was established

• Method for establishing ground truth: Not applicable based on the device description.

Summary of what is known:

• Device Name: iCare ALTIUS CW
• Regulatory Status: K234076, Class II, Product Code NFJ
• Indications for Use: Review, process, and analyze ophthalmic medical images, video, clinical and diagnostic data, measurements, and reports to support healthcare professionals in diagnosis and monitoring of eye pathologies.
• Key Features: Image manipulation (zoom, pan, brightness, contrast, gamma, RGB filtering), side-by-side comparison, advanced imaging tools (flicker, mosaic), review/print reports, cup-to-disc ratio annotation.
• Core Functionality: Cloud-based software providing advanced imaging manipulation tools; it does not perform automated image analysis.
• Performance Data Provided: "Software Verification and Validation Testing" was conducted, and documentation complied with FDA guidance and IEC 62304.
• Conclusion: The device is substantially equivalent to the predicate (FORUM, K213527), and differences (absence of purely database features, measurements only in dimensionless units for cup-to-disc, mosaic and flickering features, different system architecture) have no effect on safety and effectiveness.

What is explicitly missing from the provided text to fully answer the request:

• Specific quantitative or qualitative acceptance criteria.
• Detailed results of the verification and validation testing against those criteria.
• Any information regarding clinical studies, test set sizes, ground truth establishment, expert qualifications, or adjudication methods.
• Information about training sets or AI performance metrics, as the device explicitly states it does not perform automated image analysis.

Ask a specific question about this device

The CenterVue DRSplus is a confocal scanning ophthalmoscope indicated for color and infrared imaging of a human retina without the use of a mydriatic agent.

The CenterVue DRSplus with Software version 2.0 is a modification of the CenterVue DRSplus (K192113). The DRSplus (K192113) is a scanning ophthalmoscope which uses infrared and white light to obtain confocal images of the retina, without pharmacological dilation. With respect to the previous cleared device, DRSplus with software version 2.0 (subject of this submission) provides one additional software feature that enable to acquire, store and review infrared retinal images, in addition to color and red-free photos which were the only imaging modalities of the previously cleared device. This feature does not require any hardware modification, because the infrared illumination is already present in the previously cleared device: in fact, the infrared retinal pictures were automatically taken by the device for alignment and focusing purposes, but they were not presented in the patient's image collection. In the proposed device modification, the user can decide to acquire also infrared pictures that are stored and displayed amongst the other, color and red-free images. The modified device uses the same base technology and maintains the same Intended Use of the previously cleared device; the Indications for Use of the modified device are amended to add the infrared imaging modality. The fundamental scientific technology of the device is unchanged from the previously cleared device and remains confocal, line scanning, LED-based, imaging.

This document describes the CenterVue DRSplus with Software version 2.0, a confocal scanning ophthalmoscope. The key change from the previous version (K192113) is the addition of the ability to acquire, store, and review infrared retinal images.

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

1. Table of Acceptance Criteria and Reported Device Performance:

The document does not present a formal table of acceptance criteria with corresponding performance metrics for the infrared imaging capability specifically. Instead, it focuses on demonstrating
that the device, with its new software feature for infrared imaging, continues to meet existing safety and performance standards relevant to ophthalmoscopes.

The reported device performance, in terms of technical specifications, is available:

The core "acceptance criteria" presented are compliance with various international standards, which is a demonstration of meeting safety and fundamental performance requirements, rather than specific diagnostic accuracy metrics for the new infrared imaging feature.

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

The document explicitly states: "No clinical data required to evaluate the proposed change, only image comparisons are provided." Therefore, there is no clinical test set with a specific sample size of patients. The evaluation relied on technical comparisons and conformance to standards.

The reference to image comparisons between the DRSplus (subject device) and the EIDON (K142047) concerned the infrared imaging modality, as comparing infrared with the predicate's color imaging was not possible. No details are given about the provenance (country of origin, retrospective/prospective) of these comparison images.

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

Given that no clinical data was required and only image comparisons were provided, it is highly probable that no experts were used to establish ground truth for a clinical test set in the traditional sense of diagnostic accuracy evaluation. The evaluation focused on technical performance and image quality for the new infrared modality, likely reviewed by engineers or product specialists.

4. Adjudication Method for the Test Set:

Not applicable, as no clinical test set for diagnostic performance was evaluated or adjudicated using expert consensus.

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

No MRMC comparative effectiveness study was mentioned or performed, as "no clinical data required to evaluate the proposed change." The document does not describe any study comparing human readers with and without AI assistance, or the effect size of such assistance.

6. Standalone (Algorithm Only) Performance:

The device is a confocal scanning ophthalmoscope for imaging, not an AI/algorithm-only diagnostic device. The new feature is the ability to acquire, store, and review infrared images. Therefore, the concept of "standalone performance" for an algorithm in a diagnostic capacity is not applicable in this context. The document describes the device's technical capabilities in acquiring images, not an algorithm's diagnostic performance.

7. Type of Ground Truth Used:

For the image comparisons, the "ground truth" implicitly would have been the technical quality and characteristics of the infrared images themselves, possibly judged against expected output or established benchmarks for ophthalmoscope image quality, rather than a clinical ground truth like pathology or patient outcomes. Fundamentally, for standard compliance, the ground truth is adherence to the specifications and requirements outlined in the respective standards (ANSI Z80-26, ISO 15004-1, IEC 62304, ISO 10940).

8. Sample Size for the Training Set:

This information is not provided and is likely not relevant, as the device is not described as involving machine learning or AI that would require a "training set" for diagnostic algorithmic development. The modification is a software feature enabling existing hardware to utilize an already present infrared illumination for image acquisition and storage.

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

Not applicable, as no training set for machine learning/AI is described.

Ask a specific question about this device

The CenterVue EIDON FA is a confocal scanning ophthalmoscope indicated for color, infrared and autofluorescence imaging and fluorescein angiography of a human retina with or without the use of a mydriatic agent.

The CenterVue EIDON FA with Software version 4.0 and Ultra Widefield Lens is a modification of the CenterVue EIDON FA (K180526). The CenterVue EIDON FA is a confocal scanning ophthalmoscope indicated for color, infrared and autofluorescence imaging and fluorescein angiography of a human retina with or without the use of a mydriatic agent. CenterVue EIDON FA (K180526) is a scanning ophthalmoscope white light to obtain color images of the retina, infrared light to obtain infrared-reflectance images of the retina, and blue light to obtain autofluorescence and fluorescence images. The CenterVue EIDON FA is part of a family of devices, which includes three models: EIDON FA, EIDON AF, and EIDON. EIDON is the base model, which features the following imaging modalities: infrared reflectance, color and red-free. EIDON AF adds autofluorescence imaging to the base model. EIDON FA is the fully featured device, which adds fluorescein angiography to the capabilities of the EIDON AF and encompasses the features and functionality of the other models. With respect to the CenterVue EIDON FA, EIDON FA with software version 4.0 provides additional features to acquire retinal images with the Ultra Widefield Lens. The optional Ultra Widefield Lens increases the field of view of the devices from 60° to 80°. All the features related to the acquisition without the Ultra Widefield Lens are unchanged. The modified devices use the same base technology and maintains the same Intended Use and Indications for Use of the predicate device. The fundamental scientific technology of the subject device is unchanged from the predicate and remains confocal, line scanning, LED-based, imaging. The functional differences between EIDON FA and EIDON FA with software version 4.0 are as follows: EIDON FA with software version 4.0 allows to capture retinal picture with an increased field of view, from 60° to 80° to allow the user to view a greater proportion of the posterior eye in one view. EIDON FA with software version 4.0 allows to capture Ultra Widefield images with the same modalities of the unmodified device, namely: color images, infrared-reflectance images, autofluorescence images and fluorescence images. No technological differences between EIDON FA and EIDON FA with SW version 4.0 exist. The modified device can be used with or without the Ultra Widefield Lens. The technological principle of retinal acquisitions remains the same in the modified device. The device software has been modified to implement the acquisition and management of images taken with the Ultra Widefield Lens. The EIDON UWFL – Ultra Widefield Lens accessory is composed of a lens doublet, assembled in a custom lens holder, which can be mounted on the standard objective of the parent device by means of an incorporated threaded ring to increase the field of view of the parent device from 60° (standard objective) to 80° (with UWFL). The CenterVue EIDON UWFL – Ultra Widefield Lens is an optional accessory to extend the field of view of the CenterVue EIDON, EIDON AF and EIDON FA from 60° to 80°. It is indicated for color, infrared and autofluorescence imaging and fluorescein angiography of a human retina with or without the use of a mydriatic agent. The principle of operation of the EIDON UWFL is an increase of the standard objective to achieve a greater field of view. The mechanism of action and the conditions for use are the same of the predicate device. Other than the above, no differences exist between the EIDON FA with Ultra Widefield Lens, in particular the principle of operation, mechanism of action and interaction with the patient are unchanged.

The provided text describes a 510(k) submission for the CenterVue EIDON FA, EIDON, EIDON AF, and EIDON UWFL devices, specifically focusing on the modifications to the EIDON FA with software version 4.0 and an Ultra Widefield Lens (UWFL).

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

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state quantitative acceptance criteria in a structured table. Instead, it focuses on demonstrating substantial equivalence to a predicate device (EIDON FA K180526) by showing that the modified device (EIDON FA with SW 4.0 and UWFL) performs similarly and has no significant effect on safety or effectiveness. The core "acceptance criteria" here are implied to be image quality similarity within the common field of view and conformance to relevant standards.

• ISO 15004-1: Complies with the standard for ophthalmic instruments.
• IEC 62304: Software complies with this standard.
• ISO 10940 (Fundus Cameras): Fully complies without the optional lens, and partially complies with the optional lens mounted. |
  | Safety and Effectiveness | "None of these differences have any significant effect on safety or effectiveness of the Subject Device." Reviewed for optical radiation safety per ANSI Z80-36:2016. |

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

The clinical summary provides a breakdown of the sample size for different imaging modalities:

• Infrared images: 10 eyes without pathology and 8 eyes with pathology. (Total 18 eyes)
• Color images: 11 eyes without pathology and 10 eyes with pathology. (Total 21 eyes)
• Autofluorescence images: 10 eyes with diagnosed pathology. (Total 10 eyes)
• Fluorescein Angiography images: 10 eyes with diagnosed pathology. (Total 10 eyes)

Note: There is overlap in the eyes used across different modalities, as the same patients were likely imaged with various techniques. The total number of unique patients/eyes is not explicitly stated but is at least 21 (from the color images, which has the highest unique count for non-pathology eyes, plus pathology eyes).

Data Provenance:

• Country of Origin: Italy ("The data were collected at four different sites located in Italy.")
• Retrospective or Prospective: The study was conducted prospectively, as patients were "tested according to booked appointments for the day."

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

The document does not explicitly state the number of experts used to establish ground truth or their specific qualifications (e.g., "radiologist with 10 years of experience"). It mentions "diagnosed pathology" and "clinician's decision," implying that clinical diagnosis was used, but the process of establishing the ground truth for comparison images is not detailed.

4. Adjudication Method for the Test Set

The document does not describe an adjudication method (e.g., 2+1, 3+1). It states a "comparison shows that EIDON FA with software version 4.0 and UWFL provides images that are similar in the central 60° to those of the cleared device." This suggests a comparative assessment, but the methodology for agreement or resolution of discrepancies among observers (if more than one was involved) is not provided.

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 was not done. This study is not evaluating the performance of an AI-assisted diagnostic tool or how human readers improve with AI. Instead, it's a submission for a hardware/software modification to an ophthalmoscope, demonstrating that the modified device's image quality for a specific Field of View (FOV) remains substantially equivalent to the predicate device.

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

This is not applicable as the device (EIDON FA with UWFL) is an imaging device, not an algorithm for automated diagnosis or a device requiring standalone performance evaluation of an AI algorithm. It functions to acquire images for clinical interpretation by a human. The "standalone performance" here refers to the device's ability to capture images with certain technical specifications (resolution, FOV, etc.), which are outlined in the technical specifications section.

7. The Type of Ground Truth Used

The ground truth for the comparison appears to be based on:

• Clinical Diagnosis: Indicated by phrases like "eyes which presented without pathology" and "eyes which presented with diagnosed pathology."
• Predicate Device Images: For image quality comparison, the images captured by the predicate EIDON FA device served as a reference for "similarity."

8. The Sample Size for the Training Set

Not applicable. This document describes a 510(k) submission for a medical imaging device and its accessory/software update. It is not an AI/Machine Learning device that requires a separate training set for algorithm development. The "training" for the device would involve engineering and quality assurance tests, not a data-driven model training set in the context of AI.

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

Not applicable, as there is no mention or indication of a training set in the context of an AI/ML algorithm.

Ask a specific question about this device

The CenterVue DRSplus is a confocal scanning ophthalmoscope indicated for color imaging of a human retina without the use of a mydriatic agent.

The DRSplus is a scanning ophthalmoscope which uses infrared and white light to obtain confocal images of the retina, without pharmacological dilation. The DRSplus operates as a standalone unit, running a dedicated software application, is intended for prescription use only, and includes: 1. an optical head; 2. a patient forehead rest; 3. a display; 4. a base; 5. a stand. The CenterVue DRSplus operates based on the following principles: a) An illumination system consisting of infrared (IR) LEDs, white LEDs and a green LED illuminates the patient eye with the following functions: · the IR LED allows the capture of IR photos, which are used for alignment and focusing purposes. The patient's retina is uniformly illuminated by a line in the horizontal direction. Along the optical path there is an oscillating mirror which scans the line in order to illuminate the retina. · Two IR LEDs are seen from the eye in a free viewing system. The two LEDs are equally shifted with respect to the machine's optical axis. The LEDS are switched on during all exams in order to enable pupil tracking. · The white LEDs allow the capture of color photos. The retina is uniformly illuminated by a line in the horizontal direction. Along the optical path an oscillating mirror scans the line in order to illuminate the retina. · The green LED is used as fixation target. b) An imaging system collects back-reflected light from the retina and creates a high-resolution image. A focusing lens is included in the imaging path to achieve optimal retinal focusing on a CMOS camera. c) An anterior segment alignment system is included, using two cameras and the two IR LEDs. The LEDs illuminate the anterior segment by diffusion, whereas the cameras allow a stereoscopic reconstruction of the pupil's position to be obtained with respect to the instrument's frontlens. The DRSplus interacts with the patient by directing infrared, white (for imaging) and green (for fixation) illumination into the patient's eye. The only part of the only part of the device that contacts the patient.

The provided text describes a 510(k) premarket notification for the CenterVue DRSplus, a confocal scanning ophthalmoscope. The majority of the document focuses on regulatory information, device description, technical specifications, and comparison to predicate devices, demonstrating substantial equivalence based on engineering and safety standards.

However, the provided text does not contain the details required to directly answer all parts of your request regarding acceptance criteria and a specific study proving device performance against those criteria as would typically be found in an AI/ML medical device submission.

The document mentions "Performance data - Clinical" but this section is very brief and describes a qualitative comparison ("similar to those of the mentioned predicate device") rather than a quantitative study with defined acceptance criteria and performance metrics. It also does not describe an AI/ML algorithm within the device that would necessitate ground truth establishment, expert adjudication, or MRMC studies.

Given the information, here's what can be extracted and what is missing:

Acceptance Criteria and Device Performance (Based on provided text)

The document implicitly defines acceptance criteria through compliance with various industry standards and demonstrating substantial equivalence to predicate devices, particularly in terms of image quality for a fundus camera. No specific quantitative performance metrics for a disease detection algorithm are provided.

Table of Acceptance Criteria and Reported Device Performance:

Regarding the study proving the device meets acceptance criteria (Clinical Performance Data):

The document describes a very limited qualitative clinical performance assessment, not a robust AI/ML algorithm validation study. Therefore, most of the requested information regarding AI/ML-specific study design is not present in the provided text.

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

   • Test Set Size: "Images from randomly selected normal subjects and subjects with retinal pathologies were acquired." No specific number of subjects or images is provided.
   • Data Provenance: "acquired at two different sites" and "patients were imaged without pharmacological pupil dilation." No country of origin is explicitly stated, but the manufacturer is Italian. The study appears to be "prospective" in the sense that images were acquired specifically for this comparison using both devices.

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

   • Not Applicable / Not Mentioned. The "clinical performance data" section describes a qualitative comparison of images between two devices, not a study evaluating an AI/ML algorithm's diagnostic performance against a clinical ground truth. Hence, there's no mention of experts establishing a ground truth for diagnostic purposes.

3. Adjudication method (e.g., 2+1, 3+1, none) for the test set:

   • Not Applicable / Not Mentioned. As no ground truth for diagnostic performance was established, no adjudication method is described.

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. The provided text does not describe an AI-assisted reading study or an MRMC study. The device is described as an imaging device, not an AI diagnostic algorithm.

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

   • Not Applicable. The device, as described, is an ophthalmoscope that captures images. It's not presented as a standalone diagnostic algorithm.

6. The type of ground truth used (expert consensus, pathology, outcomes data, etc.):

   • Not applicable for diagnostic ground truth. The "ground truth" implicitly used for the clinical comparison was that the DRSplus should produce images similar to the predicate device (EIDON) for both normal and pathological retinas. There's no mention of a ground truth for disease diagnosis itself in this section.

7. The sample size for the training set:

   • Not applicable. This device is an imaging instrument, not an AI/ML algorithm that requires a training set.

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

   • Not applicable. As above.

Summary of what's missing and why:

The provided document (K192113) is a 510(k) clearance letter and summary for an ophthalmoscope. This type of device is an imaging hardware device. It is not an Artificial Intelligence/Machine Learning (AI/ML) diagnostic or assistive device. Therefore, the detailed criteria typically required for AI/ML performance validation (e.g., expert reads, ground truth establishment for disease, MRMC studies, training/testing sets for algorithms) are not present and not relevant to this specific submission as described. The performance data presented focuses on demonstrating imaging quality and electrical/software safety compliance in comparison to a predicate device, which is typical for hardware clearances.

Ask a specific question about this device

The CenterVue EIDON FA is a confocal scanning ophthalmoscope indicated for color, infrared and autofluorescence imaging and fluorescein angiography of a human retina with or without the use of a mydriatic agent.

The CenterVue EIDON FA has been derived from the CenterVue EIDON, a retinal imaging device cleared under K142047. EIDON FA is a scanning ophthalmoscope which uses LED light to capture confocal images of the retina. In particular, EIDON FA uses infrared light to obtain infraredreflectance images, white light to obtain color images and blue light to obtain auto-fluorescence and fluorescence images. EIDON FA can be used with or without pharmacological dilation.

The CenterVue EIDON FA is part of a family of devices, which includes three models: EIDON FA, EIDON AF, and EIDON. EIDON is the base model, which features the following imaging modalities: infrared reflectance, color and red-free. EIDON AF adds autofluorescence imaging to the base model. The EIDON FA is the fully featured device, which adds fluorescein angiography to the capabilities of the EIDON AF and encompasses the features and functionality of the other models.

EIDON FA operates as a standalone unit, running a dedicated software application, is intended for prescription use only, and includes:

1. an optical head, including a removable lens cap;
2. a patient head-rest, including a removable front-rest;
3. a patient chin rest;
4. a base, including a touch-screen device (tablet with magnetic holder and USB cable), USB joystick and an external power supply.

EIDON FA operates based on the following principles:
a) An illumination system consisting of an infrared (IR) LED (825-875 nm and 940 nm), a white LED (440-650 nm), a blue LED (440-475 nm) and a green LED, illuminates the patient eye with the following functions:

• the IR LED with central wavelength at 850 nm allows the capture of IR photos. The patient's retina is uniformly illuminated by a line in a horizontal direction. Along the optical path there is an oscillating mirror which scans the line in order to illuminate the retina with a field of view of 60°.
• Two IR LEDs with central wavelength at 940 nm are seen from the eye in a free viewing system. The two LEDs are equally shifted with respect to the machine's optical axis. The LEDS are switched on during all exams in order to enable pupil tracking.
• The white and blue LEDs allow the capture of color photos. The retina is uniformly illuminated by a line in the horizontal direction. Along the optical path an oscillating mirror scans the line in order to illuminate the retina with a field of view of 60°.
• The blue LED is also used to capture auto-fluorescence and fluorescence retinal images;
• The green LED is used as fixation target.
  b) An imaging system including a barrier filter (high-pass with a cutoff at 500 nm) stops back reflected light from the retina and allows fluorescent light to be detected for imaging. A focusing lens is included in the imaging path to achieve optimal focusing on a CMOS camera.
  c) An anterior segment alignment system is included, using two cameras and the two IR LEDs. The LEDs illuminate the anterior segment by diffusion, whereas allow a stereoscopic reconstruction of the pupil's position to be obtained with respect to the instrument's front lens.

EIDON FA interacts with the patient by directing infrared, white, blue (for imaging) and green (for fixation) illumination into the patient's eye. The chin-rest and head-rest are the only parts of the device that contact the patient. The chin-rest includes a patient proximity sensor and is motorized for height adjustment.

The provided text describes the regulatory information and technical specifications of the CenterVue EIDON FA device, focusing on its substantial equivalence to predicate devices for FDA clearance. However, it does not contain specific acceptance criteria, comprehensive study details, or performance data in the structured format requested. The clinical performance data section is very brief and high-level.

Therefore, I cannot fully complete the requested table and answer all points accurately based solely on the provided text. I will extract and present the available information, and explicitly state where information is missing.

Acceptance Criteria and Device Performance Study for CenterVue EIDON FA

The provided documentation describes the CenterVue EIDON FA as a confocal scanning ophthalmoscope indicated for color, infrared, autofluorescence imaging, and fluorescein angiography of the human retina. The substantial equivalence determination relies on comparisons to predicate devices (CenterVue EIDON, Kowa VX-20, and Heidelberg Spectralis HRA²).

1. Table of Acceptance Criteria and Reported Device Performance

Based on the provided text, the acceptance criteria are not explicitly stated as numerical thresholds or specific performance metrics. Instead, the focus is on demonstrating "substantial equivalence" to predicate devices, particularly regarding imaging capability and safety. The performance is reported qualitatively.

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

• Sample Size: The document states, "Images of eyes with and without diagnosed pathology were included in the comparison." However, it does not specify the number of images or patients used in the clinical comparison.
• Data Provenance: Not explicitly stated (e.g., country of origin). The study implicitly appears to be retrospective, as it involves comparing EIDON FA images with existing images from predicate devices.

3. Number of Experts Used to Establish Ground Truth and Qualifications

• Number of Experts: This information is not provided in the document.
• Qualifications of Experts: This information is not provided in the document. The assessment is described as a "comparison" of images, but who performed this comparison (e.g., a panel of ophthalmologists/radiologists or internal staff) and their qualifications are not detailed.

4. Adjudication Method for the Test Set

• The document does not describe any specific adjudication method (e.g., 2+1, 3+1) for the clinical image comparison. The assessment seems to be a general comparison.

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

• No, a MRMC study comparing human readers with and without AI assistance was not mentioned or performed. This submission is for an imaging device, not an AI-powered diagnostic algorithm designed to assist human readers. The clinical performance data section describes a comparison of image quality between the EIDON FA device and predicate devices.

6. Standalone (Algorithm Only) Performance

• This question is not applicable as the EIDON FA is an imaging device, not an AI algorithm. Its "performance" refers to the quality of the images it produces, which are then interpreted by a human clinician.

7. Type of Ground Truth Used

• The "ground truth" for the clinical comparison was implicitly the diagnosed pathology of the eyes. The study included "images of eyes with and without diagnosed pathology." The comparison was about the similarity of images captured by EIDON FA to those from predicate devices, rather than a direct diagnostic accuracy assessment against a gold standard for a specific condition.

8. Sample Size for the Training Set

• This question is not applicable, as the device is an imaging system and not an AI algorithm that requires a training set in the typical machine learning sense. The "training" of the device is its manufacturing and calibration process, not data-driven machine learning.

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

• This question is not applicable for the same reason as point 8.

Ask a specific question about this device

The CenterVue RETIA 2 is a confocal scanning ophthalmoscope indicated for infrared and auto-fluorescence imaging of a human retina with or without the use of a mydriatic agent.

The CenterVue RETIA 2 is a modification of the CenterVue EIDON (K142047). RETIA 2 is a scanning ophthalmoscope which uses infrared light to obtain infrared-reflectance images of the retina and blue light to obtain auto-fluorescence images of the retina. RETIA 2 operates as a standalone unit, running a dedicated software application, intended for prescription use only, and includes: 1. Optical head, including a removable lens cap; 2. Patient head-rest, including removable front-rest; 3. Patient chin rest; 4. Base, including touch-screen device (tablet with magnetic holder and USB cable), USB joystick and an external power supply. RETIA 2 operates based on the following principles: a) An illumination system consisting of infrared LEDs (850nm, 940nm), a blue LED (450nm) and a green LED illuminates the patient eye with the following functionality: - The IR LED with a centroid wavelength of 850 nm allows the capture of IR photos. The patient's retina is uniformly illuminated by a line in a horizontal direction. An oscillating mirror scans the line in order to illuminate the retina with a field of view of 60°. - Two IR LEDs with a centroid wavelength of 940 nm are seen from the eye in a free viewing system. The two LEDs are equally shifted with respect to the machine's optical axis. The LEDS are switched on during all exams in order to enable pupil tracking. - . The blue LED (450 nm) is used to illuminate the retina and capture fluorescence retinal images. The retina is uniformly illuminated by a line in a horizontal direction. An oscillating mirror scans the line in order to illuminate the retina with a field of view of 60°. - The green LED is used as fixation target. b) An imaging system including a barrier filter (high-pass with a cutoff at 500 nm) stops back-reflected light from the retina and only allows fluorescent light to be detected for imaging. A focusing lens is included in the imaging path to achieve optimal retinal focusing on a CMOS camera having a resolution of 5 megapixels. c) An anterior segment alignment system is included, using two cameras and the two IR LEDs. The LEDs illuminate the anterior segment by diffusion, whereas the cameras allow a stereoscopic reconstruction of the pupil's position to be obtained with respect to the instrument's front lens. RETIA 2 interacts with the patient by directing infrared- and blue-wavelength illumination for imaging purposes and green illumination for fixation purposes into the patient's eye. The chin-rest are the only parts of the device that contact the patient. The chin-rest includes a patient proximity sensor and is motorized for height adjustment.

The document describes the CenterVue RETIA 2, a confocal scanning ophthalmoscope, and its comparison to predicate devices, specifically the CenterVue EIDON (K142047) and Heidelberg SPECTRALIS (K172649). The study primarily focuses on demonstrating that the RETIA 2 provides infrared and autofluorescence images that are similar to those of the predicate devices.

Here's a breakdown of the requested information based on the provided text:

1. A table of acceptance criteria and the reported device performance

The document does not explicitly state numerical acceptance criteria in terms of sensitivity, specificity, or other quantitative metrics for image quality or diagnostic performance. Instead, it relies on a qualitative comparison.

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

The document does not specify the exact sample size for the test set (number of images or patients). It states: "Images of eyes with and without diagnosed pathology were included in the comparison."
The data provenance (country of origin, retrospective/prospective) is not mentioned.

3. 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 document does not provide this information. It only states that images "with and without diagnosed pathology were included," implying that some form of diagnosis or ground truth was available, but it doesn't describe how it was established or by whom.

4. Adjudication method (e.g. 2+1, 3+1, none) for the test set

The document does not specify the adjudication method used for the comparison.

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

There is no mention of a multi-reader multi-case (MRMC) comparative effectiveness study, nor is there any discussion of AI assistance or its impact on human reader improvement. The study described is a device-to-device comparison based on image similarity.

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

The device itself is a standalone imaging system. The "performance data - Clinical" section describes a comparison of images generated by the RETIA 2 against images from predicate devices. This represents the standalone performance of the device in generating images, as there's no mention of a human in the loop for interpreting the images for the purpose of this comparison study. However, it's not an "algorithm only" study as the device is a piece of hardware.

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)

The document mentions "images of eyes with and without diagnosed pathology." This suggests that the ground truth for image classification (e.g., presence or absence of pathology) was based on pre-existing diagnoses, likely from clinicians or medical records. However, the exact nature of this "diagnosis" (e.g., expert consensus, pathology reports, clinical outcomes) is not further detailed.

8. The sample size for the training set

The document does not mention a training set, as the study described is a device comparison for image similarity, not an AI model requiring training data.

9. How the ground truth for the training set was established

Not applicable, as no training set for an AI model is described.

Ask a specific question about this device

The Centervue MAIA is intended for:
• measuring macular sensitivity,
• measuring fixation stability and the locus of fixation,
• providing infrared retinal imaging, and
• aiding visual rehabilitation.
It contains a reference database that is a quantitative tool for the comparison of macular sensitivity to a database of known normal subjects.

A previous version of the CenterVue MAIA, a device for macular integrity assessment, has been cleared by FDA under K133758 on 23 April 2014. The present submission relates to a revised version of the MAIA device in which the only difference between the subject device and the MAIA device cleared under K133758 is in the software, where a new function called "Fixation Training" (FT) has been introduced to aid visual rehabilitation of patients with unstable fixation. The FT is independent from the functions available in the device cleared under K133758 and it does not interfere or modify the original functions in any way. No other design changes are being introduced by this revision to the MAIA device.
The FT is intended for visual rehabilitation, to help Vision Rehabilitation Specialists train patients with unstable fixation to improve their fixation stability.
A FT session consists of asking the patient to move his/her gaze according to the trainer's instructions and to an audible signal, so to attempt fixation of the internal visual target using a specific retinal area, which is identified by the trainer ahead of the training session. The center of such area is called Fixation Training Target (FTT).
During the FT session, the MAIA retinal tracker continuously determines the position of the fixation point and provides an audible feedback to the patient in the form of pulses of a certain repetition frequency. The number of pulses / sec (i.e. the repetition frequency) is inversely proportional to the distance between the patient's fixation point at that time and the FTT; when such distance falls below one degree, the sound becomes continuous. Optionally, before starting the FT session, operators are able to replace the continuous sound with an MP3 audio file.
The MAIA device interacts with the patient by directing illumination into the patient's eye. The chin-rest and head-rest are the only parts of the device that contact the patient. The chin-rest includes a patient proximity sensor and is motorized for height adjustment. The biocompatibility of the patient-contacting materials, which are the same as used in the previous version of the subject device (K133758) has been established.
The MAIA device operates as a 'stand-alone' device and does not need to interface with other medical devices.

Here's an analysis of the provided text regarding the MAIA device, focusing on acceptance criteria and study details:

The document is a 510(k) premarket notification for the CenterVue MAIA device, which is an ophthalmoscope and perimeter. The submission is for a revised version of a previously cleared device (K133758), with the only change being the introduction of a new "Fixation Training" (FT) software function.

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly present a table of acceptance criteria with corresponding performance results for the new Fixation Training (FT) software function. Instead, it relies on demonstrating that the new software does not alter the safety or effectiveness of the previously cleared device and meets relevant software and risk management standards.

However, the comparison table (Table 1) provides performance specifications for various features of the MAIA device and its predicates. While not "acceptance criteria" in the sense of a numerical pass/fail for the specific new feature, these are the performance characteristics being maintained or compared.

Here's a table based on the provided predicate comparison, focusing on relevant aspects mentioned:

The crucial "study" for the new FT function is its compliance with software development and risk management standards, rather than a clinical performance study with numerical criteria. The document states: "the Fixation Training software meets the requirements of: ISO 62304: 2006, ISO 14971: 2007." This is the primary demonstration of its acceptability.

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

The document does not describe a specific clinical test set with a sample size for the new Fixation Training (FT) software. The submission focuses on the software's compliance with standards and its technological similarity to a feature in a secondary predicate device (Nidek MP1).

The core of the submission is that the FT function is independent, does not interfere with existing functions, and no other design changes were made. Therefore, the detailed studies for the underlying hardware and existing functions (macular sensitivity, fixation stability measurement, retinal imaging) from the K133758 clearance are implicitly reused.

There is no mention of country of origin for new data or whether any hidden data was retrospective or prospective, as no new clinical data is presented for the FT feature.

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

Not Applicable. Since no new clinical test set is described for the FT function, there is no mention of experts establishing a ground truth for such a set. The acceptance of the FT function hinges on its compliance with international software and risk management standards and its functionality being similar to an existing predicate device.

4. Adjudication Method for the Test Set

Not Applicable. No new clinical test set requiring adjudication is described for the FT function.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was Done

No. The document does not describe an MRMC comparative effectiveness study for the new Fixation Training (FT) software. The submission is centered on substantial equivalence to predicate devices and software standard compliance, not on demonstrating improved human reader performance with or without AI assistance. The FT function itself is for patient rehabilitation, not for aiding human readers in diagnosis.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was Done

The Fixation Training (FT) is a software feature that provides real-time feedback to the patient based on fixation point tracking. While the tracking itself is an algorithm, the FT functionality is described with the patient in the loop, acting on the audible feedback. The document states: "the MAIA retinal tracker continuously determines the position of the fixation point and provides an audible feedback to the patient." This is an algorithm-driven feature intended for use with a human (patient) in the loop (for visual rehabilitation). It is not a standalone diagnostic algorithm for interpretation by an expert.

7. The Type of Ground Truth Used

For the new Fixation Training (FT) software:

• Software Design/Functionality: The "ground truth" for accepting the FT feature is its compliance with ISO 62304: 2006 (Medical device software – Software life cycle processes) and ISO 14971: 2007 (Medical devices – Application of risk management to medical devices). This represents a ground truth that the software is developed safely and effectively according to recognized standards.
• Functional Equivalence: The comparison to the Nidek MP1 (PD2) for the visual rehabilitation/fixation training aspects also acts as a form of "ground truth" for functional equivalence, showing that similar technology is already marketed and cleared.

For the core device functions (macular sensitivity, fixation stability measurement, retinal imaging), the ground truth for their original clearance (K133758) would have been established through clinical data, expert consensus, and comparison to other cleared devices, but these details are not provided in this document as it's a submission for an updated feature, not the initial clearance.

8. The Sample Size for the Training Set

Not Applicable / Not Provided. The document does not describe any machine learning or AI algorithm that would require a "training set" in the conventional sense for the new Fixation Training (FT) feature. The FT function appears to be based on deterministic algorithms for tracking and feedback, rather than a learned model.

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

Not Applicable / Not Provided. As no training set is described, there's no mention of how its ground truth was established.

Ask a specific question about this device

The CenterVue COMPASS is intended for taking digital images of a human retina without the use of a mydriatic agent and for measuring retinal sensitivity, fixation stability and the locus of fixation. It contains a reference database that is a quantitative tool for the comparison of retinal sensitivity to a database of known normal subjects.

The CenterVue COMPASS is a scanning ophthalmoscope combined with an automatic perimeter that allows the acquisition of images of the retina, as well as the measurement of retinal threshold sensitivity and the analysis of fixation. The device works with a dedicated software application, operates as a standalone unit, integrates a dedicated tablet, a joystick, a push-button and is provided with an external power supply. COMPASS operates in non-mydriatic conditions, i.e. without the need of pharmacological dilation and is intended for prescription use only.

The Centervue COMPASS device operates on the following principles:

• An anterior segment alignment system is included, which uses two infrared LEDs with a centroid wavelength of 940 nm and two cameras, whereas the former illuminate the external eye by diffusion and the latter allow a stereoscopic reconstruction of the pupil's position, used for automated alignment purposes via pupil tracking;
• An infrared imaging system captures live monochromatic images of the central retina over a circular field of view of 60° in diameter, by an horizontal line from an infrared LED with a centroid wavelength of 850 nm and by an oscillating mirror which scans the line to uniformly illuminate the retina; such images are in turn used for auto-focusing purposes and to track eye movements, providing a measure of a patient's fixation characteristics and allowing active compensation of the position of perimetric stimuli;
• A concurrent color imaging system allows the capture of color images of the central retina over a circular field of view of 60° in diameter, using a white LED and a blue LED combined to obtain a white light illuminating the retina by the same scan mechanism;
• A fixation target projecting onto the retina a fixation target obtained from a green LED;
• A stimuli projector, projecting onto the retina white light Goldmann stimuli at variable intensity and allowing measurements of threshold sensitivity at multiple locations, according to a patient's subjective response to the light stimulus projected at a certain location.

The COMPASS device interacts with the patient by directing infrared, white, blue and green wavelength illumination into the patient's eye and by recording a patient's confirmation that a certain light stimulus has been perceived or not.

Here's an analysis of the acceptance criteria and the supporting study for the CenterVue COMPASS device, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The FDA clearance letter (K150320) primarily focuses on establishing "substantial equivalence" to predicate devices, rather than explicit numerical acceptance criteria for clinical performance that might be found in a performance goal document for a novel device. However, the clinical study serves to demonstrate this equivalence. The key performance comparison is between the CenterVue COMPASS and the Humphrey HFA-II.

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

• Test Set Sample Size:
  • 200 normal subjects
  • 120 subjects with pathology affecting the visual field (specifically glaucoma)
  • Total: 320 subjects
• Data Provenance: The document does not explicitly state the country of origin. It indicates the manufacturer is in Padova, Italy, and the study was conducted to support FDA clearance in the USA, suggesting the study likely occurred in conjunction with the manufacturer's operations or clinical sites. The study is presented as prospective clinical testing ("Measurements have been obtained...").

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

The document does not specify the number of experts or their qualifications for establishing the "ground truth" for the test set (i.e., whether subjects were truly "normal" or had "glaucoma"). It only states that subjects were categorized as "normal" or with "pathology affecting the visual field (in particular glaucoma)." This implies a clinical diagnosis was used, but the specific process or number of experts for this diagnosis is not detailed.

4. Adjudication Method for the Test Set

The document does not describe an adjudication method for the test set in terms of expert review or consensus. The study compares the performance of the COMPASS directly to the predicate device (Humphrey HFA-II) on the same subjects, rather than assessing the COMPASS's ability to classify against a pre-established ground truth determined by multiple experts.

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 was not done. This device is primarily a diagnostic instrument for measuring retinal sensitivity and imaging, not an AI-assisted diagnostic aid for interpretation by human readers. The clinical study compares the device's measurements to another device, not human performance.

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

Yes, the clinical study presents data on the standalone performance of the CenterVue COMPASS device in measuring retinal threshold sensitivity. It directly compares the measurements obtained by the COMPASS to those obtained by the Humphrey HFA-II. The device operates as a standalone unit for acquiring images and measuring retinal sensitivity.

7. The Type of Ground Truth Used

The "ground truth" in this context is the measurement of retinal threshold sensitivity as determined by the accepted standard, the Humphrey HFA-II. The study aims to demonstrate that the COMPASS's measurements are "equivalent" to those of the HFA-II, specifically that the mean differences in thresholds are comparable to known differences within the HFA-II platform (SITA Standard vs. full threshold). The classification of subjects as "normal" or with "glaucoma" would have been based on clinical diagnosis, implicitly serving as a form of "expert consensus" or "clinical diagnosis" ground truth for subject selection, but not for the specific performance metric being evaluated (threshold sensitivity differences).

8. The Sample Size for the Training Set

The document describes a "reference database" that was developed to serve as a quantitative tool for comparison of retinal sensitivity to known normal subjects.

• Reference Database Sample Size: 200 eyes of 200 normal subjects.
• The age range of this population was 20 - 86 years (50.6 ± 15.2).

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

The ground truth for the "training set" (referred to as the "reference database" in the document) was established by obtaining threshold sensitivity data from 200 subjects confirmed to be "normal." The specific criteria or expert qualifications for determining "normalcy" are not detailed in this summary, but it implies a clinical assessment of individuals free from visual field pathology. The perimetric settings used to gather this data are listed (24-2 grid, 4-2 strategy, Goldmann III stimulus, etc.).

Ask a specific question about this device

The CenterVue EIDON is intended for taking digital images of a human retina without the use of a mydriatic agent.

The CenterVue EIDON scanning ophthalmoscope operates as a standalone unit, running a dedicated software application, intended for prescription use only, and includes:

1. Optical head, including a removable lens cap;
2. Patient head-rest, including removable front-rest;
3. Patient chin rest;
   Base, including touch-screen device (tablet with magnetic holder and USB cable), USB joystick 4. and an external power supply.
   EIDON uses infrared and visible light to obtain colored confocal digital images of a human retina without a mydriatic agent.
   The CenterVue EIDON device operates on the following principles:
   a) An illumination system consisting of infrared (IR) LEDs (850nm, 940nm), a white LED, a blue LED and a green LED illuminates the patient eye with the following functionality:

• The IR LED with a centroid wavelength of 850 nm allows the capture of IR photos. The patient retina is uniformly illuminated by a line in a horizontal direction. Along the optical path there is an oscillating mirror which scans the line in order to illuminate the retina with a field of view of 60°.
• Two IR Leds with a centroid wavelength of 940 nm are seen from the eye in a free viewing system. The two LEDs are equally shifted with respect to the machine optical axis. The LEDS are switched on during all exams in order to enable pupil tracking.
• The white and blue LEDs allow the capture of color photos. The retina is uniformly illuminated by a line in a horizontal direction. Along the optical path there is an oscillating mirror which scans the line in order to illuminate the retina with a field of view of 60°.
• The green LED is used as fixation target.
  b) An imaging system collects back-reflected light from the retina and creates a high resolution image. A focusing lens is included in the imaging path to achieve optimal retinal focusing on a CMOS camera having a resolution of 14 megapixels.
  c) An anterior segment alignment system is included, using two cameras and the two IR LEDs: the LEDs illuminate the anterior segment by diffusion, whereas allow a stereoscopic reconstruction of the pupil's position to be obtained with respect to the instrument front lens.
  The EIDON device interacts with the patient by directing infrared, white, blue and green wavelength illumination into the patient's eye. The chin-rest and head-rest are the only parts of the device that contact the patient. The chin-rest includes a patient proximity sensor and is motorized for height adjustment.

The CenterVue EIDON Ophthalmoscope is intended for taking digital images of a human retina without the use of a mydriatic agent. The device was evaluated through bench tests against a set of performance criteria.

Here's a breakdown of the acceptance criteria and performance data:

1. Acceptance Criteria and Reported Device Performance

Based on the "Test results" column, the CenterVue EIDON Ophthalmoscope met all of its stated performance requirements.

2. Sample Size and Data Provenance

The document describes "Bench tests" for the device's technical specifications. This typically implies testing on the manufactured device itself rather than human subjects or a dataset of images. Therefore, questions of sample size for a test set, data provenance (country of origin, retrospective/prospective), and ground truth establishment methods for a test set are not applicable in this context. The evaluation focuses on the engineering and optical performance of the device.

3. Number and Qualifications of Experts for Ground Truth

Not applicable for bench testing of device specifications.

4. Adjudication Method

Not applicable for bench testing of device specifications.

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

No MRMC study was mentioned in the provided text. The submission focuses on the technical specifications and substantial equivalence to predicate devices, not on human reader performance with or without AI assistance.

6. Standalone (Algorithm Only) Performance Study

The description indicates the device operates as a "standalone unit, running a dedicated software application", suggesting standalone performance of the image acquisition. However, the provided information focuses on the hardware's technical specifications rather than the performance of any image analysis algorithms that might be embedded within the software. There's no mention of a standalone study specifically evaluating algorithmic performance in areas like disease detection or image quality assessment by AI.

7. Type of Ground Truth Used

For the performance criteria listed (e.g., sensor resolution, optical resolution), the ground truth is based on engineering measurements and specifications of the device's optical and imaging components.

8. Sample Size for the Training Set

Not applicable. This submission describes bench testing of a medical device's physical and optical performance, not the development or training of an AI algorithm.

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

Not applicable, as there is no mention of an AI algorithm training set.

Ask a specific question about this device