(193 days)
ARGOS is a non-invasive, non-contact biometer based on swept-source optical coherence tomography (SS-OCT). The device is intended to acquire ocular measurements as well as perform calculations to determine the appropriate intraocular lens (IOL) power and type for implantation during intraccular lens placement. ARGOS measures the following 9 parameters: Axial Length, Corneal Thickness, Anterior Chamber Depth, Lens Thickness, K-values (Radii of flattest and steepest meridians), Astigmatism, White (corneal diameter) and Pupil Size. It is intended for use by ophthalmologists, physicians, and other eye-care professionals and may only be used under the supervision of a physician.
Argos is a swept-source based biometer that provides the biometry and keratometry of the eye prior to cataract surgery and aids in the selection of the appropriate IOL.
Argos processes 3 OCT images (B-scans) and 3 CMOS camera images per measurement in seconds - reducing procedure time and minimizing patient discomfort. From the OCT images the biometry parameters are evaluated by segmenting the cornea, iris, lens, and retina. The operator can edit the biometry parameters by manual adjustment: Axial length, Corneal thickness, Anterior chamber depth (or aqueous depth), Lens thickness, White-towhite (corneal diameter), Pupil size. The keratometry values: K-values (flattest and steepest meridians), Astigmatism (angle of flattest meridian), are evaluated from the CMOS camera image, in combination with the OCT information.
All distance/thickness parameters (Axial length, Corneal thickness, Anterior chamber depth, Lens thickness, White, Pupil size) are simultaneously measured from 2dimensional OCT images.
All the boundaries of ocular segments are detected and distances and thicknesses are calculated by taking into account the refractive indices of each medium.
Axial length is the distance from the corneal apex to the fovea or, more specifically, to the ILM (Internal Limiting Membrane). The calculation is performed as the sum of the thicknesses (after refraction correction) of the cornea, aqueous humor, lens and vitreous.
Corneal thickness is the distance between the anterior and the posterior apexes of the cornea divided by the cornea refractive index (1.375).
Aqueous depth is evaluated as the distance between the posterior surface of the cornea and the anterior surface of the lens divided by the refractive index (1.336).
Anterior chamber depth is the sum of corneal and aqueous humor distances.
Lens thickness is the distance between the anterior and the posterior surfaces of the lens divided by its refractive index (1.410).
Pupil size is the lateral distance between the two inner boundaries of the iris region.
White-to-white (Corneal diameter) is the lateral distance between the inner boundaries of the cornea-sclera interface.
K-values (Radii of flattest and steepest meridians) and Astigmatism are evaluated by the size and distance of the reflected images of the infrared LED ring projected onto the cornea.
Safety protocol is well integrated into the device in both hardware and software to ensure the safety of both the patient and the operator.
The control program on the computer runs on Windows. The operation panel on the display attached to the PC provides all the functionalities.
IOL calculation uses widely recognized formulas; Hoffer Q, Haigis, Holladay1, SRK/T, etc.
The provided document describes the ARGOS device, an optical biometer, and its performance evaluation. Here's a breakdown of the acceptance criteria and study details:
1. Table of Acceptance Criteria and Reported Device Performance
The document does not explicitly state formal "acceptance criteria" for the clinical studies in the format of a predefined threshold. Instead, it presents the device's performance in relation to a predicate device (Lenstar LS900) through various statistical measures, implying that performance comparable to the predicate or within acceptable clinical ranges constitutes "acceptance." The "repeatability" values can also be considered a form of performance criteria.
| Parameter | Acceptance Criteria (Implied / Comparator Performance for Agreement) | Reported Device Performance (ARGOS) |
|---|---|---|
| Agreement Study (vs. Lenstar LS900) | ||
| Axial Length (AL) | Similar mean, small mean difference (0.00 mm), small SD (0.05 mm), LoA of [-0.10, 0.09] mm, Deming slope close to 1 (1.03), high Spearman correlation (1.00) compared to predicate (Lenstar: Mean 23.43 (1.01), SD N/A, Spearman 1.00 for comparison, Imp. Repeatability SD: 0.035mm). | Mean: 23.44 (0.98), Mean Difference (Argos-Lenstar): 0.00 (0.05) mm, LoA: [-0.10, 0.09] mm, Deming Regression: Intercept -0.68 [-0.93,-0.43], Slope 1.03 [1.02,1.04], Spearman Correlation: 1.00. Repeatability (Precision Study): SD 0.01 mm, CV 0.02%. Reproducibility: SD 0.02 mm, CV 0.09%. |
| Corneal Thickness (CCT) | Similar mean, small mean difference (0.00 mm), small SD (0.01 mm), LoA of [-0.02, 0.02] mm, Deming slope close to 1 (1.04), high Spearman correlation (0.97) compared to predicate (Lenstar: Mean 0.53 (0.04), SD N/A, Spearman 0.97 for comparison, Imp. Repeatability SD: 2.3um (0.0023mm)). | Mean: 0.53 (0.03), Mean Difference (Argos-Lenstar): 0.00 (0.01) mm, LoA: [-0.02, 0.02] mm, Deming Regression: Intercept -0.02 [-0.05,0.01], Slope 1.04 [0.99,1.10], Spearman Correlation: 0.97. Repeatability (Precision Study): SD 0.00 mm, CV 0.89%. Reproducibility: SD 0.01 mm, CV 1.19%. |
| Aqueous Depth (AD) | Similar mean, small mean difference (0.05 mm), small SD (0.16 mm), LoA of [-0.27, 0.37] mm, Deming slope close to 1 (1.06), high Spearman correlation (0.89) compared to predicate (Lenstar: Mean 2.66 (0.41), SD N/A, Spearman 0.89 for comparison, Imp. Repeatability SD: 0.04mm). | Mean: 2.63 (0.39), Mean Difference (Argos-Lenstar): 0.05 (0.16) mm, LoA: [-0.27, 0.37] mm, Deming Regression: Intercept -0.10 [-0.29,0.09], Slope 1.06 [0.98,1.13], Spearman Correlation: 0.89. Repeatability (Precision Study): SD 0.01 mm, CV 0.37%. Reproducibility: SD 0.03 mm, CV 0.90%. |
| Anterior Chamber Depth (ACD) | Similar mean, small mean difference (0.05 mm), small SD (0.17 mm), LoA of [-0.28, 0.37] mm, Deming slope close to 1 (1.07), high Spearman correlation (0.89) compared to predicate (Lenstar: Mean 3.19 (0.42), SD N/A, Spearman 0.89 for comparison, Imp. Repeatability SD: 0.04mm). | Mean: 3.16 (0.39), Mean Difference (Argos-Lenstar): 0.05 (0.17) mm, LoA: [-0.28, 0.37] mm, Deming Regression: Intercept -0.17 [-0.39,0.06], Slope 1.07 [1.00,1.14], Spearman Correlation: 0.89. Repeatability (Precision Study): SD 0.01 mm, CV 0.27%. Reproducibility: SD 0.03 mm, CV 0.77%. |
| Lens Thickness (LT) | Similar mean, small mean difference (-0.18 mm), small SD (0.22 mm), LoA of [-0.61, 0.24] mm, Deming slope close to 1 (1.09), high Spearman correlation (0.87) compared to predicate (Lenstar: Mean 4.46 (0.45), SD N/A, Spearman 0.87 for comparison, Imp. Repeatability SD: 0.08mm). | Mean: 4.62 (0.41), Mean Difference (Argos-Lenstar): -0.18 (0.22) mm, LoA: [-0.61, 0.24] mm, Deming Regression: Intercept -0.58 [-1.10,-0.05], Slope 1.09 [0.97,1.20], Spearman Correlation: 0.87. Repeatability (Precision Study): SD 0.02 mm, CV 0.38%. Reproducibility: SD 0.06 mm, CV 1.45%. |
| Pupil Size (PS) | Similar mean, small mean difference (-0.26 mm), small SD (0.46 mm), LoA of [-1.16, 0.64] mm, Deming slope close to 1 (0.81), high Spearman correlation (0.88) compared to predicate (Lenstar: Mean 4.10 (0.77), SD N/A, Spearman 0.88 for comparison, Imp. Repeatability SD: not specified). Note: Differences attributed to lack of ambient light control. | Mean: 4.38 (0.91), Mean Difference (Argos-Lenstar): -0.26 (0.46) mm, LoA: [-1.16, 0.64] mm, Deming Regression: Intercept 0.55 [-0.02,1.11], Slope 0.81 [0.68,0.95], Spearman Correlation: 0.88. Repeatability (Precision Study): SD 0.09 mm, CV 1.83%. Reproducibility: SD 1.01 mm, CV 20.31%. |
| Corneal Diameter (CD) | Similar mean, small mean difference (-0.28 mm), small SD (0.69 mm), LoA of [-1.62, 1.06] mm, Deming slope close to 1 (0.57 - lower than 1 but still within CI), high Spearman correlation (0.48 - lower correlation acknowledged due to different measurement principle) compared to predicate (Lenstar: Mean 11.86 (0.49), SD N/A, Spearman 0.48 for comparison, Imp. Repeatability SD: 0.04mm). Note: Accuracy not well confirmed due to different measurement principle. | Mean: 12.11 (0.68), Mean Difference (Argos-Lenstar): -0.28 (0.69) mm, LoA: [-1.62, 1.06] mm, Deming Regression: Intercept 4.91 [1.79,8.04], Slope 0.57 [0.31,0.83], Spearman Correlation: 0.48. Repeatability (Precision Study): SD 0.06 mm, CV 0.46%. Reproducibility: SD 0.10 mm, CV 0.77%. |
| Flattest Radius of Curvature (R1) | Similar mean, small mean difference (-0.01 mm), small SD (0.07 mm), LoA of [-0.15, 0.13] mm, Deming slope close to 1 (1.01), high Spearman correlation (0.97) compared to predicate (Lenstar: Mean 7.79 (0.30), SD N/A, Spearman 0.97 for comparison, Imp. Repeatability SD: 0.02mm). | Mean: 7.80 (0.29), Mean Difference (Argos-Lenstar): -0.01 (0.07) mm, LoA: [-0.15, 0.13] mm, Deming Regression: Intercept -0.09 [-0.45,0.26], Slope 1.01 [0.96,1.06], Spearman Correlation: 0.97. Repeatability (Precision Study): SD 0.01 mm, CV 0.14%. Reproducibility: SD 0.03 mm, CV 0.34%. |
| Steepest Radius of Curvature (R2) | Similar mean, small mean difference (0.00 mm), small SD (0.07 mm), LoA of [-0.14, 0.14] mm, Deming slope close to 1 (0.97), high Spearman correlation (0.98) compared to predicate (Lenstar: Mean 7.61 (0.27), SD N/A, Spearman 0.98 for comparison, Imp. Repeatability SD: 0.03mm). | Mean: 7.61 (0.28), Mean Difference (Argos-Lenstar): 0.00 (0.07) mm, LoA: [-0.14, 0.14] mm, Deming Regression: Intercept 0.22 [-0.05,0.48], Slope 0.97 [0.94,1.01], Spearman Correlation: 0.98. Repeatability (Precision Study): SD 0.02 mm, CV 0.21%. Reproducibility: SD 0.04 mm, CV 0.48%. |
| Average Radius of Curvature (Rav) | Similar mean, small mean difference (-0.01 mm), small SD (0.06 mm), LoA of [-0.12, 0.11] mm, Deming slope close to 1 (0.99), high Spearman correlation (0.97) compared to predicate (Lenstar: Mean 7.70 (0.28), SD N/A, Spearman 0.97 for comparison, Imp. Repeatability SD: N/A). | Mean: 7.70 (0.28), Mean Difference (Argos-Lenstar): -0.01 (0.06) mm, LoA: [-0.12, 0.11] mm, Deming Regression: Intercept 0.06 [-0.17,0.30], Slope 0.99 [0.96,1.02], Spearman Correlation: 0.97. Repeatability (Precision Study): SD 0.01 mm, CV 0.13%. Reproducibility: SD 0.03 mm, CV 0.35%. |
| Astigmatism (Cyl. 1D) | Similar mean, small mean difference (4.36 degrees), small SD (9.71 degrees), LoA of [-14.67, 23.39] degrees, Deming slope close to 1 (1.05), high Spearman correlation (0.97) compared to predicate (Lenstar: Mean 81.44 (46.96), SD N/A, Spearman 0.97 for comparison, Imp. Repeatability SD: 11deg). | Mean: 78.40 (46.34), Mean Difference (Argos-Lenstar): 4.36 (9.71) Degrees, LoA: [-14.67, 23.39] Degrees, Deming Regression: Intercept 0.07 [-6.02,6.15], Slope 1.05 [0.96,1.15], Spearman Correlation: 0.97. Repeatability (Precision Study, All AST): SD 4.39 degrees, CV 5.59%. Reproducibility (All AST): SD 11.13 degrees, CV 14.17%. |
| Enhanced Retinal Visualization (ERV) mode - Cataractous Eyes (vs. Lenstar for AL) | ||
| Axial Length (AL) (ERV) | Statistically insignificant differences (p-value 1D): 36 eyes |
* **Data Provenance:** A single U.S. clinical site; collected prospectively for the agreement study.
-
Precision Study (Repeatability and Reproducibility - Table 17.2.2):
- 43 healthy eyes (for both repeatability and reproducibility).
- Data Provenance: Not explicitly stated, but generally part of the same prospective clinical study.
-
Enhanced Retinal Visualization (ERV) mode Validation Study (Table 17.2.3 and 17.2.4):
- Cataractous eyes: 45 eyes (23 right eyes) for comparison with Lenstar.
- Healthy eyes: 43 consecutive healthy eyes for repeatability of ERV mode.
- Data Provenance: Not explicitly stated, but generally part of the same prospective clinical study.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
The document does not mention the use of experts to establish ground truth for the test set beyond the clinical measurements themselves. The study primarily relies on the agreement between the ARGOS device and the predicate device (Lenstar LS900) as the reference for performance, and the precision of the ARGOS itself.
4. Adjudication Method for the Test Set
No adjudication method is described for the test set. The clinical studies compare measurements between devices and evaluate the precision of the ARGOS device.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done, and the Effect Size
No MRMC comparative effectiveness study involving human readers' improvement with or without AI assistance was done or described. This device is an optical biometer for objective measurements, not an image interpretation AI system for diagnosis where human reader studies are typical.
6. If a Standalone (Algorithm Only Without Human-in-the-Loop Performance) Was Done
Yes, the studies described are standalone performance evaluations of the ARGOS device. The ARGOS device is an automated measurement tool. The "Agreement Study" and "Precision Study" assess the device's inherent measurement capabilities. While an operator initiates measurements, the performance metrics (Axial Length, CCT, etc.) are generated solely by the device's algorithms and hardware.
7. The Type of Ground Truth Used
The "ground truth" for the clinical studies is implicitly established by:
- Predicate Device Measurements: The Lenstar LS900, an already legally marketed optical biometer, serves as the primary comparator. The assumption is that its measurements are an acceptable reference.
- Internal Consistency/Precision: The repeatability and reproducibility studies establish the device's consistency in its own measurements, which is a form of internal ground truth for precision.
There is no mention of "pathology" or external "outcomes data" being used as ground truth for parameter measurements in this context.
8. The Sample Size for the Training Set
The document does not specify a training set or its sample size. This is typical for medical devices that perform direct physical measurements rather than AI algorithms trained on large datasets for pattern recognition. The algorithms for biometry are based on physical principles (e.g., swept-source OCT) and established calculations, not machine learning model training in the conventional sense.
9. How the Ground Truth for the Training Set Was Established
As no training set is described for an AI model, there is no information on how its ground truth was established. The device relies on physical measurement principles and established formulas (e.g., Hoffer Q, Haigis, Holladay1, SRK/T) for IOL power calculation.
§ 886.1850 AC-powered slitlamp biomicroscope.
(a)
Identification. An AC-powered slitlamp biomicroscope is an AC-powered device that is a microscope intended for use in eye examination that projects into a patient's eye through a control diaphragm a thin, intense beam of light.(b)
Classification. Class II (special controls). The device, when it is intended only for the visual examination of the anterior segment of the eye, is classified as Group 1 per FDA-recognized consensus standard ANSI Z80.36, does not provide any quantitative output, and is not intended for screening or automated diagnostic indications, is exempt from the premarket notification procedures in subpart E of part 807 of this chapter subject to the limitations in § 886.9.