(49 days)
The SPECTRALIS is a non-contact ophthalmic diagnostic imaging device. It is intended for:
• viewing the posterior segment of the eye, including two- and three- dimensional imaging
• cross-sectional imaging (SPECTRALIS HRA+OCT and SPECTRALIS OCT)
• fundus imaging
• fluorescence imaging (fluorescein angiography, indocyanine green angiography; SPECTRALIS HRA+OCT, SPECTRALIS HRA)
• autofluorescence imaging (SPECTRALIS HRA+OCT, SPECTRALIS HRA and SPECTRALIS OCT with BluePeak)
• performing measurements of ocular anatomy and ocular lesions.
The device is indicated as an aid in the detection and management of various ocular diseases, including:
• age-related macular degeneration
• macular edema
• diabetic retinopathy
• retinal and choroidal vascular diseases
• glaucoma
The device is indicated for viewing geographic atrophy.
The SPECTRALIS OCT Angiography Module is indicated as an aid in the visualization of vascular structures of the retina and choroid.
The SPECTRALIS HRA+OCT and SPECTRALIS OCT include the following reference databases:
• a retinal nerve fiber layer thickness reference database, which is used to quantitatively compare the retinal nerve fiber layer in the human retina to values of Caucasian normal subjects – the classification result being valid only for Caucasian subjects
• a reference database for retinal nerve fiber layer thickness and optic nerve head neuroretinal rim parameter measurements, which is used to quantitatively compare the retinal nerve fiber layer and neuroretinal rim in the human retina to values of normal subjects of different races and ethnicities representing the population mix of the USA (Glaucoma Module Premium Edition)
The Heidelberg Engineering SPECTRALIS HRA+OCT is a device used to image the anterior and posterior segments of the human eye. The SPECTRALIS HRA+OCT is a combination of a confocal laser-scanning ophthalmoscope (cSLO, the HRA portion) and a spectral-domain optical coherence tomographer (SD-OCT). The confocal laser-scanning part of the device allows for acquisition of reflectance images (with blue, green or infrared light), conventional angiography images (using fluorescein or indocyanine green dye) and autofluorescence images. The different imaging modes can be used either alone or simultaneously. The SD-OCT part of the device acquires cross-sectional and volume images, together with an HRA cSLO image.
A blue laser is used for fluorescein angiography, autofluorescence imaging, and blue reflectance imaging, and two infrared lasers are used for indocyanine green angiography and infrared reflectance imaging. A green laser is used for MultiColor imaging ("composite color images"). MultiColor imaging is the simultaneous acquisition of infrared, green and blue reflectance images that can be viewed separately or as a composite color image. For SD-OCT imaging, an infrared super-luminescent diode and a spectral interferometer are used to create the cross-sectional images.
The following modifications have been applied to the device subject of this 510(k):
- Addition of scan acquisitions for the SPECTRALIS OCT Angiography Module (OCTA) at 250 kHz
- Update of the default contrast display setting from 1:4 to 1:2 for the Superficial Vascular Complex (SVC) and the Deep Vascular Complex (DVC) for the acquisition speeds of 125 kHz and 250 kHz
Acceptance Criteria and Study for SPECTRALIS HRA+OCT
The provided FDA clearance letter for the SPECTRALIS HRA+OCT and variants (K250868) describes a retrospective image grading case study (S-2023-1) performed to demonstrate substantial equivalence for modifications to the device. The modifications include the addition of scan acquisitions for the SPECTRALIS OCT Angiography Module (OCTA) at 250 kHz and an update of the default contrast display setting from 1:4 to 1:2 for the Superficial Vascular Complex (SVC) and the Deep Vascular Complex (DVC) for 125 kHz and 250 kHz acquisition speeds.
The study aimed to show that the investigational SPECTRALIS scan types (with 250 kHz acquisition and updated contrast settings) performed similarly to the predicate SPECTRALIS HRA+OCT with OCTA Angiography Module scan types (HR10 @ 85 kHz, HS20 @ 85 kHz) in terms of image quality, visualization of key anatomical vascular structures, and identification of pathologies.
1. Table of Acceptance Criteria and Reported Device Performance
The FDA clearance letter does not explicitly define specific numerical acceptance criteria in the format of a table with pass/fail thresholds. Instead, it reports performance metrics and concludes on "similarity" and "sufficiency" relative to clinical needs and the predicate device. Based on the provided text, the implied acceptance criteria were that the investigational device's performance should be similar to or sufficient for clinical assessment compared to the predicate device.
Here's a summary of the reported device performance, interpreted as meeting these implied criteria:
| Performance Metric | Implied Acceptance Criteria | Reported Device Performance (Investigational Scan Types) |
|---|---|---|
| Overall Image Quality (percentage of images graded better than Poor) | Sufficient to assess clinically relevant content | 97.4% for HR10 @ 125 kHz |
| 96.1% for HR10 @ 250 kHz | ||
| 96.2% for HS20 @ 125 kHz | ||
| 93.3% for HS20 @ 250 kHz | ||
| 98.7% for Scout @ 125 kHz | ||
| Conclusion: Overall image quality was sufficient to assess the clinically relevant content. | ||
| Visualization of Key Anatomic Structures (percentage of structures graded better than Unable to Distinguish) | Achieved on a high percentage of assessments | ≥ 92.3% on HR10 @ 125 kHz |
| ≥ 94.8% on HR10 @ 250 kHz | ||
| ≥ 96.2% on HS20 @ 125 kHz | ||
| ≥ 93.3% on HS20 @ 250 kHz | ||
| ≥ 96.2% on Scout @ 125 kHz | ||
| Conclusion: Visualization was achieved on a high percentage of assessments. | ||
| Agreement in Identification of Vascular Abnormalities (vs. Predicate) | High agreement rate, PPA, and NPA | MA: Agreement rate, PPA (with some exceptions), and NPA were ≥ 86.4% |
| Retinal Ischemia: Agreement rate, PPA, and NPA were ≥ 77.4% | ||
| RNV: Agreement rate, PPA, and NPA were ≥ 88.9% | ||
| CNV: Agreement rate, PPA, and NPA were ≥ 88.9% | ||
| Conclusion: Results are similar between investigational and predicate scan types. | ||
| Agreement for Primary Vascular Abnormality of Interest (PVAOI) (vs. Predicate) | High agreement rate, PPA, and NPA | 10x10 HR scan types: Agreement rate, PPA, and NPA was at least 80% |
| 20x20 HS scan types: Agreement rate, PPA, and NPA was at least 83.3% | ||
| Conclusion: Results are similar between investigational and predicate scan types. |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size (Test Set): The effectiveness analysis population from the S-2020-5 study consisted of 79 subjects. All 25 Normal subjects and 54 Pathology subjects were included in this retrospective study. However, the exact count for direct comparison between the predicate and investigational device ranged from 74 to 78 subjects depending on the scan type.
- Data Provenance:
- Country of Origin: United States
- Retrospective/Prospective: The S-2023-1 image grading case study was retrospective, using clinical data that was collected prospectively in a previous study (S-2020-5). Data was collected at a single clinical site.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications
- Number of Experts: Three independent reviewers from a Reading Center.
- Qualifications of Experts: The document does not explicitly state the specific qualifications (e.g., "radiologist with 10 years of experience") of these reviewers. It only identifies them as "independent reviewers from the Reading Center."
4. Adjudication Method for the Test Set
- The document states that the performance metrics for image quality were "summarized based on the percentage of images graded better than Poor (i.e., Good or Average) on consensus." This indicates that some form of consensus method was used for image quality grading. However, the specific adjudication method (e.g., 2+1, 3+1, majority rule, etc.) for achieving this consensus is not detailed. For visualization of key anatomical structures and identification of pathologies, it indicates agreement analysis between predicate and investigational scan types but does not explicitly describe an adjudication method to establish a single "ground truth" before comparison; rather, it assesses agreement between the two device types.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- The study described is an image grading study involving multiple readers (three independent reviewers) evaluating multiple cases, comparing an investigational device's performance to a predicate device. While it aligns with elements of a comparative effectiveness study, it's not explicitly labeled as a "Multi Reader Multi Case (MRMC) comparative effectiveness study" in the statistical sense (e.g., for ROC analysis). Instead, it focuses on agreement rates, PPA, and NPA.
- Effect Size of Human Readers' Improvement with AI vs. without AI: This study does not involve AI assistance for human readers. The device in question is a medical imaging device (OCTA), not an AI-powered diagnostic tool providing automated interpretations or assisting human readers. Therefore, there is no reported effect size for human readers improving with AI vs. without AI assistance. The study compares two versions of the imaging device.
6. Standalone Performance Study
- Yes, a standalone performance assessment was conducted for the investigational scan types. The reported metrics for "Overall Image Quality" and "Visualization of Key Anatomic Structures" are measures of the algorithm's output (images from the investigational scan types) as graded by experts, independent of a human-in-the-loop scenario for diagnostic decision-making. The agreement analysis is essentially comparing the standalone performance of the investigational scans against the standalone performance of the predicate scans.
7. Type of Ground Truth Used
- The "ground truth" for the test set was primarily established through expert consensus/grading by three independent reviewers.
- The "Normal population" was defined by clinical examination showing no retinal conditions or abnormalities.
- The "Pathology population" had specific retinal conditions (e.g., wet age-related macular degeneration, diabetic retinopathy) and abnormalities (e.g., microaneurysm, choroidal neovascularization, retinal neovascularization) that were identified. This implies medical record review and possibly other diagnostic findings contributed to classifying these subjects, but the direct "truth" for the study's performance metrics (image quality, structure visualization, abnormality identification) came from the expert grading of the OCTA scans.
8. Sample Size for the Training Set
- The document does not specify a sample size for a training set. The study primarily focuses on the validation of modifications to an existing device, and the data listed is related to its verification and clinical evaluation (test set). It is possible that the underlying algorithms within the SPECTRALIS were trained on a separate, unmentioned dataset prior to this 510(k) submission, but this information is not provided in the clearance letter.
9. How the Ground Truth for the Training Set Was Established
- As the document does not provide information about a training set, it does not describe how the ground truth for any training set was established.
§ 886.1570 Ophthalmoscope.
(a)
Identification. An ophthalmoscope is an AC-powered or battery-powered device containing illumination and viewing optics intended to examine the media (cornea, aqueous, lens, and vitreous) and the retina of the eye.(b)
Classification. Class II (special controls). The device, when it is an AC-powered opthalmoscope, a battery-powered opthalmoscope, or a hand-held ophthalmoscope replacement battery, is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 886.9.