Notal Vision Home Optical Coherence Tomography (OCT) System

{0}------------------------------------------------ ### DE NOVO CLASSIFICATION REQUEST FOR NOTAL VISION HOME OPTICAL COHERENCE TOMOGRAPHY (OCT) SYSTEM #### REGULATORY INFORMATION FDA identifies this generic type of device as: Home monitoring ophthalmic imaging device. A home monitoring ophthalmic imaging device is a prescription self-imaging device that incorporates imaging system hardware and automated image processing and analysis to enable patients at home to provide measurements that are intended for use by a physician for monitoring ophthalmic diseases or conditions in between regularly scheduled assessments. NEW REGULATION NUMBER: 21 CFR 886.1600 CLASSIFICATION: Class II PRODUCT CODE: SAX #### BACKGROUND DEVICE NAME: Notal Vision Home Optical Coherence Tomography (OCT) System SUBMISSION NUMBER: DEN230043 DATE DE NOVO RECEIVED: June 7, 2023 #### SPONSOR INFORMATION: Notal Vision, Inc. 7717 Coppermine Drive Manassas, VA 20109 #### INDICATIONS FOR USE The Notal Vision Home Optical Coherence Tomography (OCT) System is indicated as follows: The Notal Vision Home Optical Coherence Tomography (OCT) System is an Artificial Intelligence (AI)-based Home Use device indicated for visualization of intraretinal and subretinal hypo-reflective spaces in a 10 by 10-degrees area centered on the point of fixation of eyes diagnosed with neovascular age-related macular degeneration (NV-AMD). In addition, it provides segmentation and an estimation of the volume of hyporeflective spaces. The Notal Home OCT device is intended for imaging at home between regularly scheduled clinic assessments and not intended to be used to make treatment decisions or replace standard-of care regularly scheduled examinations and clinical {1}------------------------------------------------ testing as needed, including in-office imaging and assessments for changes in vision, by an ophthalmologist. #### LIMITATIONS The sale, distribution, and use of the Notal Vision Home Optical Coherence Tomography (NVHO) System are restricted to prescription use in accordance with 21 CFR 801.109. The Notal Home OCT System and Web Viewer results should not be used for diagnosis of any condition. The NVHO device should not be used to delay in-office follow-up or to prolong the interim period between in-office follow-up visits. The NVHO device should not be used on patients with nonneovascular AMD ("dry" AMD) to detect conversion from "dry" to "wet" AMD. Patients should continue self-monitoring for visual changes (e.g., continue self-administration of Amsler grid testing) while using the NVHO device. The Notal Home OCT device shall not be used by patients with Visual Acuity of worse than 20/320. Scans with poor image quality, e.g., below Manufacturer Signal quality Index (MSI) of <2, may be unreliable. MSI values with color indicator are presented under the Web Viewer OCT Bscans. The following conditions may increase the likelihood of poor-quality scans: - Inability to maintain steady fixation . - Unclear ocular media . - . Dementia NOA estimations should be considered in the context of the variability observed across the range of estimations (i.e. larger percent variability for NOA quantification of smaller hypo-reflective spaces); lower notification thresholds in the presence of smaller hypo-reflective spaces will be inherently less reliable. Data on the clinical performance of Notal Home OCT System was limited in the following populations: - . Patients with vision worse than 20/80 - Patients of African and Asian descent and Hispanic/Latino patients . {2}------------------------------------------------ Caution should be exercised when evaluating scans from these patient populations. In addition, the ability of patients with vision worse than 20/80 to successfully self-image and to generate consistently reliable images with the Notal Home OCT System is not well characterized. Participants of the "006" clinical study comprised 21,5% of the safety cohort but 40.3% of those who failed to successfully self-calibrate. Regular and frequent review of all available B-scans, regardless of TRO level, is recommended to evaluate for appropriate scan centration and quality, for the presence of confounding pathologies, and for the presence of hypo-reflective spaces abutting or spanning the edge of the scan area. Regular review of B-scans is also recommended to evaluate for hyper-reflective lesions such as hemorrhage. It is recommended to review and consider the information available over several (at least three) days in the trajectory. No single NOA estimation should be relied upon in making decisions about prompt patient follow-up. Please refer to the "Notal Home Optical Coherence Tomography Instructions for Use" and "Notal Home OCT Web Viewer Instructions for Use" for a complete list of WARNINGS, PRECAUTIONS AND CONTRAINDICATIONS, as well as a description of CLINICAL STUDY OUTCOMES ("Notal Home OCT Web Viewer Instructions for Use" only), ### DEVICE DESCRIPTION The NVHO System is a device that consists of two elements: - . Notal Home OCT device: patients use this to self-image their eyes using Spectral-Domain OCT: At the end of each scanning session the data is transmitted via a secured wireless communication to the Notal Health Cloud. - Notal Health Cloud: cluster of servers and analysis units on which the Notal OCT . Analyzer (NOA) analyzes the scans received from the Notal Home OCT device. Processed data are presented on a dedicated interactive web-application, the Notal Home OCT Web Viewer. Image /page/2/Picture/7 description: The image shows a medical device with a screen and several knobs. The device has a handle on the side and a place to rest an arm. There are some labels and markings on the device, but they are not clear enough to read. The device appears to be designed for some kind of medical testing or treatment. Figure 1: Notal Home OCT Device {3}------------------------------------------------ Image /page/3/Picture/0 description: The image shows a diagram of the Notal Home OCT Device and the Notal Health Cloud. The Notal Home OCT Device is located at the patient's home and is connected to the Notal Health Cloud via a secured internet connection. The Notal Health Cloud consists of a database, a volume generator, and an NOA. The Notal Home OCT Web Viewer is used to view the OCT images. Figure 2: Notal Vision Home OCT System Overview The system overview in Figure 2 shows the overall workflow. Images are captured by the patient at home. An application icon-based Graphical User Interface (GUI) enables NV-AMD patients to self-image. The GUI, presented on an external touch screen, takes the patient step-by-step through the scanning flow. The device contains an internal mini-display that presents an interactive fixation target to the subject. Upon self-imaging, images are automatically and securely transmitted to the Notal Health Cloud for viewing by the prescribing user. Raw data are fed into a Volume Generator (VG) module for OCT B-scan reconstruction, Manufacturer Signal Quality Index (MSI) calculation, segmentation, registration, volume scan reconstruction and alignment. Images are fed into the Notal Home OCT Web-viewer for display, and into NOA for analysis. NOA is an artificial intelligence (AI) algorithmic module. NOA runs on the Notal Health Cloud and processes the volume scans generated by the VG. Its purpose is to segment the subretinal hypo-reflective spaces (SRO) and intraretinal hypo-reflective spaces (IRO) in the 2D B-scans. From this, the IRO and SRO volumes in the macular volume scan are calculated. Total retinal hypo-reflective space (TRO) volume information, which is the sum of SRO and IRO, is displayed on the Notal Home OCT Web Viewer. NOA tab (see Figure 4), along with projections of SRO and IRO information. The Notal Home OCT Web Viewer and OCT images are not accessible to patient users. Image /page/3/Picture/5 description: In the image, an elderly woman is seated at a desk, looking into a piece of equipment. The equipment is white and has a black eyepiece that the woman is looking into. The woman has short, white hair and is wearing a black jacket. There is a globe on the desk to the left of the equipment. Figure 3: Patient Using the Notal Home OCT Device {4}------------------------------------------------ Image /page/4/Figure/0 description: This image shows a medical imaging interface with various tools and displays. The interface includes features such as "NOA Movies", a timeline, event indicators, and date selectors. There are also tools for scan playing, AI segmentation overlay, and image locking. The image also displays hypo-reflective spaces volume trajectories and thickness maps. Figure 4: Eyecare Provider's Interactive Viewer Figure 4 shows the Notal Home OCT Web Viewer NOA tab display, which includes an output of TRO volume over time, along with projections and B-scan overlay. An image quality indicator, Manufacturer Signal quality Index (MSI) is displayed, with acceptable scan quality of MSI ≥ 2 indicated by a green color and an unacceptable scan quality MSI < 2 indicated in red. The prescribing user can compare two scans from two different dates, side-by-side. Prescribing-user set notifications are available, based on two notification categories (multi-selection is allowed); 1. TRO - if TRO reaches the threshold, a notification is generated 2. Time - if the selected number of days has elapsed, a notification is generated # Table 1: Notal Home OCT System Technical Specifications {5}------------------------------------------------ | DESIGN CONFIGURATION | NOTAL HOME OCT DEVICE | |---------------------------------|--------------------------------------------------------------------------------| | Methodology | Spectral Domain OCT | | Field of view | 10°X10° measured from the cornea surface | | Scan speed | 10,000 A-scans / sec | | Lateral number of pixels | 500 A-scans / B-scan | | Light source | Super luminescent diode (SLD) | | Center wavelength | 830 ± 10 nm | | Optical resolution | Better than Axial 19 $\mu$ m (in tissue), transverse<br>35 $\mu$ m (in tissue) | | Capture rate | Max: 20 B-scan/sec | | Scan pattern | 88 B-scans over the scanned area of 10°X10° | | Maximal power on the cornea | 710 $\mu$ W | | Maximal daily exposure | 120 sec | | Shut off optical power | 800 $\mu$ W | | SPECTROMETER SPECIFICATIONS | | | Number of pixels in image plane | 512 | | Digital resolution (in tissue) | Better than 4.5 $\mu$ m (in tissue) | | PERFORMANCE MEASURE | | | Sensitivity (in air) | >100 dB | | ELECTRICAL | | | Voltage range | 110-240 V | | Frequency range | 50-60 Hz | | Power consumption | Up to 100 Watts | | Power supply | 90W single output external power medical<br>grade, E-Cap Life of >7 years | | Computer/OS | IMX6 quad core/Linux embedded | | Input device | External touch screen | | External screen | LCD 7" | # SUMMARY OF NONCLINICAL/BENCH STUDIES Non-clinical testing for the NVHO device included device benchtop performance testing, optical radiation safety, software and cybersecurity, electromagnetic compatibility (EMC) testing, wireless testing, electrical safety testing, biological safety evaluation, and human factors validation testing. # BENCH PERFORMANCE TESTING Non-clinical performance testing was conducted to verify the technical specifications, spatial characteristics, device sensitivity and diopter range for the NVHO device on three identical devices. | Test | Purpose | Method | Acceptance Criteria | Results | |--------------------------------------------------------------------|------------------------------------------------------------------------------------------------------|--------|---------------------|---------| | <b>Optical Coherence Tomography Bench Performance Verification</b> | | | | | | Axial resolution | To verify axial resolution | | (b)(4) | Passed | | | | (b)(4) | | | | Lateral resolution | To verify lateral<br>resolution | | Passed | | | Axial range | To verify axial range<br>and axial distance<br>calibration | | Passed | | | Lateral range | To verify the lateral<br>range and lateral<br>distance calibration for<br>each scan pattern | | Passed | | | Device sensitivity<br>testing | To characterize<br>instrument signal-to-<br>noise ratio (SNR) and<br>instrument depth<br>attenuation | | Passed | | | Diopter range | To measure the device<br>diopter range for<br>compensating for<br>refractive error of the | | Passed | | #### Table 2: Verification of NVHO bench performance {6}------------------------------------------------ # OPTICAL RADIATION SAFETY Optical radiation safety testing was provided to ensure the NVHO device provides acceptable light hazard protection, in accordance with the applicable parts of the following standard: - ANSI Z80.36:2021: American National Standard for Ophthalmics Light hazard . protection for ophthalmic instruments {7}------------------------------------------------ The test report for ANSI Z80.36:2021 included descriptions of all the light sources and their optical paths along with functions and technical specifications of these light sources. Measurement procedures, equipment used for measurements, raw data and formulas used for calculations were provided and were found to be acceptable. A justification was provided to explain why the worst-case scenarios were considered in the assessment of optical radiation safety. ### SOFTWARE & CYBERSECURITY The NVHO system is a cyber device per 524B(c) of the Food, Drug & Cosmetic Act. Notal Vision has followed the recommendations in the Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions (fda.gov) (https://www.fda.gov/media/119933/download) guidance document (September 27, 2023). The NVHO System is comprised of the Notal Vision Home OCT device and the Notal Health Cloud. The device allows the patients to self-image their eyes, and have images securely transmitted from the device to the Notal Health Cloud. The images are then reconstructed on the cloud and made available to the patient's health care provider. A detailed description of all software inputs and outputs was provided, along with description of all software modules and interactions with device hardware. All components of the device are controlled/monitored by software, which is responsible for the functionality, user interface, hardware start-up checks, and head position, pupil detection performance accuracy. A detailed description of data used to train and test the algorithms was provided, including cases, sources, demographics and reference standards. The software was developed and tested according to the following FDA guidance documents and recognized consensus standards: - FDA Guidance, General Principles of Software Validation (January 11, 2002) . - . FDA Guidance, Off-the-Shelf Software Use in Medical Devices (September 27, 2019) - . FDA Guidance, Guidance for the Content of Premarket Submission for Software Contained in Medical Devices (May 11, 2005) - IEC 62304: 2006 /A1:2016 Medical device software Software life-cycle processes . - ISO 14971:2019 Medical devices Application of risk management to medical devices . Software documentation and testing, including cybersecurity information, demonstrates that the software will operate in a manner described in the specifications. The hazard analysis characterized software and cybersecurity risks, including device malfunction, measurementrelated errors, head/pupil position sensors, power/internet/network connectivity and other hardware failures, and unauthorized access by malicious end users. The submission describes verification and validation testing to address the potential hazards with satisfactory results. The cybersecurity documentation included a cybersecurity hazard analysis and mitigation information, an upgrade and maintenance plan, other information for safeguarding the device during manufacturing and upon commercial distribution, and warning and precaution information in the product labeling. {8}------------------------------------------------ Overall, the software documentation contains sufficient detail to provide reasonable assurance that the software will operate in a manner described in the specifications. Software security information demonstrated the device is protected from cyber vulnerability threats. All testing and results were considered to be adequate and met the above standards. # ELECTROMAGNETIC COMPATIBILITY, WIRELESS COEXISTENCE AND ELECTRICAL SAFETY Electrical Safety and Electromagnetic Compatibility (EMC) testing has been performed as per the following recognized consensus standards and the results support electrical safety and electromagnetic compatibility - IEC 60601-1: 2005. A1:2012. A2: 2020 Medical electrical equipment -Part 1: General . requirements for basic safety and essential performance, including US National Deviations. - . IEC 60601-1-11:2015AMD1:2020: Medical electrical equipment - Part 1-11: General requirements for basic safety and essential performance-Collateral Standard: Requirements for medical electrical equipment and medical electrical systems used in the home healthcare environment. - · IEC 60601-1-2:2020: Medical electrical equipment-Part 1-2: General requirements for basic safety and essential performance-Collateral standard: Electromagnetic disturbances-Requirements and tests # BIOCOMPATIBILITY/MATERIALS The NVHO includes five components that are in contact with intact skin: an enclosure (polyurethane), face rest (polyurethane (0(4) handles (thermoplastic polymer), touch screen (silicone dioxide) and volume and height control (polyurethane). A biocompatibility evaluation for each of the five patient-contacting components was conducted according to International Standard Organization (ISO) 10993-1:2018 Biological evaluation of medical devices-Part 1: Evaluation and testing within a risk management process and following FDA Guidance: Use of International Standard ISO 10993-1. "Biological evaluation of medical devices - Part 1: Evaluation and testing within a risk management process" - Guidance for Industry and Food and Drug Administration Staff (fda.gov). This evaluation was found to be acceptable. # HUMAN FACTORS VALIDATION TESTING Two Human Factors Summative Studies were conducted following the FDA guidance: "Applying Human Factors and Usability Engineering to Medical Devices" to evaluate the usability of the Notal Home OCT device and the Notal Home OCT Web Viewer. A NVHO Human Factors Validation Study was conducted for subjects with NV-AMD and their caregivers to establish that the NVHO System patient user interface (NVHO device, and associated labeling) was safe and effective for intended users, uses and use environments by showing that use-related hazards associated with the NVHO System have been adequately {9}------------------------------------------------ mitigated and instructional information for the user (Instructions for User, Packaging Set-up instructions) can be used effectively by representative users under simulated use conditions without producing patterns that could result in harm to patients, caregivers or providers. A total of 54 individuals representing two distinct user groups participated in this human factors validation test: 20 individual patients and 17 pairs of patient and caregivers (34 individuals). Patients went through a built-in, on-screen tutorial session, which provided training on how to self-image; this was provided along with on-package setup instructions and instructions for use. Critical tasks were identified that were related to the use of NVHO device. The testing determined that the Notal Home OCT patient interface and its associated labeling and packaging components are safe and effective for the intended user population in their intended use environment. Human factors validation testing did not yield any results that indicate use errors or patterns of use errors that could result in user harm. No use errors were observed on critical tasks, therefore, the results of the study demonstrate that the device user-interface supports the use-safety and effectiveness of use for the intended use. No additional modifications to the user interface were needed to lower the residual risk of the user interface. A second Human Factors Validation Study was conducted for eyecare providers to establish that the NVHO System physician user interface (Web Viewer and associated labeling) was safe and effective for intended users, uses and use environments by showing that use-related hazards associated with the Notal Home OCT Web Viewer have been adequately mitigated and that instructional information for the user (user manual) can be used effectively by representative users under simulated use conditions without producing patterns of failures that could result in harm to users. A total of 15 evecare providers participated in the human factors testing for the OCT Web Viewer. Critical tasks were identified that were related to reviewing of patient scans and use of features in the Web Viewer. The testing determined that the Notal Home OCT Web Viewer and its associated labeling are safe and effective for the intended user population in their intended use environment. Human factors validation testing did not vield any results that indicate use errors or patterns of use errors that could result in user harm. No use errors were observed on critical tasks, therefore, the results of the study demonstrate that the device user-interface supports the use-safety and effectiveness of use for the intended use. No additional modifications to the user interface were needed to lower the residual risk of the user interface. # SUMMARY OF CLINICAL INFORMATION Clinical performance data were collected from two pivotal clinical studies: 1) the C2021.001 study ("001 Study"): A 5-Week "Home OCT Fluid Visualization Agreement Study"; 2) the C2012.006 study ("006 Study"): A cross-sectional in-office study for "The Evaluation of the Agreement and Precision of the Notal Vision Home OCT in the Automatic Fluid Quantification in Patients with NV-AMD." Note: The protocols for these studies used the term "retinal fluid" (e.g., total, sub-retinal and intra-retinal fluid [TRF, SRF, IRF]) to refer to hypo-reflective spaces (HRS; or TRO, SRO, and IRO). However, not all hypo-reflective spaces on macular OCT imaging are retinal fluid, and not all exudative fluids will necessarily be hypo-reflective; the terms TRF, SRF, and IRF are not synonymous or interchangeable with the terms TRO. SRO, and IRO. {10}------------------------------------------------ ### Summary of the 001 Clinical Study #### Overview: The "001 Study" was a prospective, longitudinal study conducted at seven sites in the United States. Adults aged 55 years or older with diagnosed NV-AMD in at least one eligible eve and best-corrected visual acuity of 20/320 or better were enrolled 1 week prior to a previously scheduled, routine clinic visit. Those who required anti-VEGF treatment for NV-AMD in the study eye at the screening visit, those who had any other retinal disease requiring steroidal or anti-VEGF treatment, or those with prior NVHO device use experience were excluded. One NV-AMD eye was determined to be the study eye. The purpose of the study was the following: 1) to evaluate the agreement between in-office OCT macular scans versus the NVHO scans in the visualization of retinal fluid in the central 10 degrees of the macula, as determined by expert graders at a third-party reading center (RC); and 2) to evaluate the ability of participants to successfully self-image with the NVHO device. RC graders were masked to each other's determinations, to the source device, and to the participant ID number. Ordering of the scans was randomized. Disagreements between graders were adjudicated. IRF and SRF were graded as present when definite hypo-reflective space is observed in OCT B-scan images. IRF is anterior to the photoreceptor layer and SRF is posterior to the photoreceptor layer and anterior to the retinal pigment epithelium (RPE). The presence of confounding lesions (epiretinal membrane [ERM], macular hole, pseudocysts, outer retinal tubulations, hemorrhage, pigment epithelial detachments, subretinal hyper-reflective material [SHRM], geographic atrophy [GA], and hyper-reflective retinal spots or foci) was also assessed by the RC. During the enrollment visit. participants underwent imaging with a clinic-based. spectral-domain OCT imaging system (CIRRUS HD-OCT) to establish a "baseline" status of the macula. After verification of eligibility during screening, participants were assigned an NVHO device. The assigned NVHO device was delivered to participants' homes. Participants set up the NVHO device using the onscreen tutorial. To continue in the study, participants must be able to achieve successful initial NVHO calibration. Continuing participants were instructed to self-image at home with the NVHO every day for five consecutive weeks, including on the days of scheduled office visits. Remote telephone technical support ("Notal Vision Diagnostic Clinic," NVDC) was available to participants. In-clinic visits were scheduled at Week 1 and Week 5. At these scheduled visits, CIRRUS HD-OCT imaging was performed and best-corrected visual acuity (BCVA), subjective symptoms, and adverse events were assessed. The daily NVHO scans of participants designated as without retinal fluid at baseline were reviewed by the reading center (RC). The RC triggered an alert for an interim clinic visit if fluid was identified on two consecutive NVHO scans from two consecutive days. CIRRUS HD-OCT images, BCVA, and any symptoms or adverse event information were also collected at these interim visits. The primary effectiveness endpoints were positive and negative percent agreements (PPA, NPA) of central macular (central 3x3-mm area) fluid status between NVHO and RC-graded CIRRUS HD-OCT scans, success rate of initial NVHO setup, and success rate for NVHO self-imaging {11}------------------------------------------------ attempts. The primary analysis of the PPA and NPA of visualizing Total Retinal Fluid (TRF) on NVHO volume scans was based on repeated measurements of the primary eyes using the optimal weighting method (Jung et al.) to test Hg: p ≤ 0.8 vs. Hg: p > 0.8 with normal approximation and a two-sided significance level of 0.05. The secondary endpoints were the success rates of total (completed plus incomplete) and completed self-imaging transmission to the Notal Health Cloud. The safety outcomes of interest were any adverse events occurring during the conduct of the study. # Results: 198 participants were enrolled. Of these, seven (3.5%) were screen failures, eight withdrew consent prior to initiating participation (4.0%), two (1.0%) were exited due to inability to scan, and one (0.51%) was exited because of inability to return for follow-up. Therefore, 180 participants (90.9%) comprise the safety cohort. From the safety cohort, an additional 12 participants were excluded (10 due to device-use: inadequate image quality, self-calibration, and/or self-imaging problems; two who withdrew consent) to form the 168 participants (84.8%) in the "Visualization Analysis Population" (VAP) cohort (those who underwent clinic-based and NVHO imaging at Week-1, interim, or Week-5 visits). From the VAP cohort, another eight participants were excluded (seven due to not having NVHO scans with MSI score >2 and clinicbased OCT scans paired within 24 hours at Week-1, interim and Week-5 visits; one exited early due to difficulty with self-imaging) to form 160 participants (80.8%) in the "Modified Visualization Analysis Population" (mVAP) cohort. Primary analyses were based on the mVAP cohort. Demographics, and relevant baseline clinical characteristics are shown in Tables 3 and 4. The majority of participants were White and not Hispanic or Latino. | Demographics | | Safety<br>Population<br>N = 180 | Modified<br>Visualization<br>Analysis<br>Population<br>N = 160 | |--------------|------------------------------|---------------------------------|----------------------------------------------------------------| | Age | N | 180 | 160 | | | Mean ± SD | 77.1 ± 7.2 | 76.8 ± 7.2 | | | Median | 77.5 | 77.0 | | | Min, Max | 55, 92 | 55, 92 | | Gender | Male | 78 (43.3%) | 70 (43.8%) | | | Female | 102 (56.7%) | 90 (56.3%) | | Race | Asian | 1 (0.6%) | 1 (0.6%) | | | Black or African American | 4 (2.2%) | 3 (1.9%) | | | White | 174 (96.7%) | 155 (96.9%) | | | Not Reported | 1 (0.6%) | 1 (0.6%) | | Ethnicity | Not Hispanic or Latino | 178 (98.9%) | 158 (98.8%) | | | Not Reported | 2 (1.1%) | 2 (1.3%) | | Education | Less than High School Degree | 8 (4.4%) | 7 (4.4%) | | | High School Degree | 43 (23.9%) | 38 (23.8%) | | Table 3: Demographics of Safety Population and mVAP - 001 Study | | | |-----------------------------------------------------------------|--|--| | | | | {12}------------------------------------------------ | | Some college (no degree) | 47 (26.1%) | 44 (27.5%) | |-----------|-------------------------------------------------|------------|------------| | | College Degree (Associate or Bachelor's Degree) | 51 (28.3%) | 45 (28.1%) | | | Graduate Degree | 25 (13.9%) | 21 (13.1%) | | | Other¹ | 6 (3.3%) | 5 (3.1%) | | Study Eye | OD | 93 (51.7%) | 82 (51.3%) | | | OS | 87 (48.3%) | 78 (48.8%) | % = n / N × 100%. 1 Including some graduate school, trade school, and tech school | | | Safety Population | | Modified Visualization<br>Analysis Population | | |-------------------------------------------------------------------------------------------|-----------------------------------------------------------|-------------------------|------------------------------|-----------------------------------------------|-----------------------------| | Baseline Characteristics | | Primary Eye¹<br>N = 180 | Secondary<br>Eye¹<br>N = 137 | Primary Eye<br>N = 160 | Secondary<br>Eye<br>N = 123 | | AMD diagnostic² | AMD - Early AMD | 0 (0.0%) | 10 (7.3%) | 0 (0.0%) | 9 (7.3%) | | | NV-AMD - non active<br>(no fluid present) | 92 (51.1%) | 35 (25.5%) | 84 (52.5%) | 27 (22.0%) | | | AMD - Intermediate<br>AMD | 0 (0.0%) | 61 (44.5%) | 0 (0.0%) | 58 (47.2%) | | | NV-AMD - active<br>(fluid present) | 88 (48.9%) | 31 (22.6%) | 76 (47.5%) | 29 (23.6%) | | Lens status | Phakia (cataract<br>present) | 65 (36.1%) | 47 (34.3%) | 61 (38.1%) | 46 (37.4%) | | | Pseudophakia | 115 (63.9%) | 90 (65.7%) | 99 (61.9%) | 77 (62.6%) | | Ocular Media<br>Assessment | Main vessels and the<br>small vessels are<br>clearly seen | 179 (99.4%) | 136 (99.3%) | 159 (99.4%) | 122 (99.2%) | | | Both main and small<br>vessels cannot be seen | 1 (0.6%) | 1 (0.7%) | 1 (0.6%) | 1 (0.8%) | | Visual Distortions | Present | 17 (9.4%) | 7 (5.1%) | 15 (9.4%) | 7 (5.7%) | | | Absent | 163 (90.6%) | 130 (94.9%) | 145 (90.6%) | 116 (94.3%) | | Blurry Vision | Present | 57 (31.7%) | 28 (20.4%) | 54 (33.8%) | 26 (21.1%) | | | Absent | 123 (68.3%) | 109 (79.6%) | 106 (66.3%) | 97 (78.9%) | | Scotoma | Present | 57 (31.7%) | 28 (20.4%) | 54 (33.8%) | 26 (21.1%) | | | Absent | 123 (68.3%) | 109 (79.6%) | 106 (66.3%) | 97 (78.9%) | | Prior total # of | N | 180 | 137 | 160 | 123 | | Injections | Mean ± SD | 26.4 ± 26.5 | 12.1 ± 21.4 | 25.3 ± 26.6 | 12.0 ± 21.9 | | | Median | 17.0 | 0.0 | 16.0 | 0.0 | | | Min, Max | 0, 128 | 0, 125 | 0, 128 | 0, 125 | | Manifest<br>Refraction<br>Spherical<br>Equivalent | N | 151 | 113 | 137 | 104 | | | Mean ± SD | 0.066 ± 1.924 | 0.044 ± 1.810 | 0.109 ± 1.881 | 0.135 ± 1.672 | | | Median | 0.000 | 0.000 | 0.000 | 0.000 | | | Min, Max | -8.000, 5.000 | -8.000, 5.000 | -7.000, 5.000 | -7.000, 5.000 | | Best Corrected | N | 180 | 137 | 160 | 123 | | Visual Acuity | Mean logMAR<br>(Snellen) | 0.301<br>(20/40.0) | 0.234<br>(20/34.3) | 0.281<br>(20/38.2) | 0.211<br>(20/32.5) | | Visual Acuity | SD logMAR | 0.251 | 0.324 | 0.226 | 0.309 | | | Median logMAR<br>(Snellen) | 0.220<br>(20/33.2) | 0.120<br>(20/26.4) | 0.220<br>(20/33.2) | 0.120<br>(20/26.4) | | | Min logMAR<br>(Snellen) | -0.10<br>(20/16.0) | -0.10<br>(20/16.0) | -0.10<br>(20/16.0) | -0.10<br>(20/16.0) | | | Max logMAR<br>(Snellen) | 1.20<br>(20/320.0) | 1.90 (CF) | 1.04<br>(20/219.3) | 1.90 (CF) | | Best Corrected<br>Visual Acuity<br>category | 20/40 or Better | 110 (61.1%) | 106 (77.4%) | 102 (63.8%) | 100 (81.3%) | | | 20/41 to 20/80 | 50 (27.8%) | 20 (14.6%) | 44 (27.5%) | 16 (13.0%) | | | 20/81 to 20/200 | 14 (7.8%) | 9 (6.6%) | 12 (7.5%) | 6 (4.9%) | | | 20/201 to 20/320 | 6 (3.3%) | 1 (0.7%) | 2 (1.3%) | 1 (0.8%) | | | Worse than 20/320 | 0 (0.0%) | 1 (0.7%) | 0 (0.0%) | 0 (0.0%) | | Principal<br>Investigator's<br>retinal fluid<br>assessment based<br>on<br>commercial OCT2 | Both IRF and SRF | 10 (5.6%) | 3 (2.2%) | 8 (5.0%) | 3 (2.4%) | | | SRF only | 40 (22.2%) | 13 (9.5%) | 36 (22.5%) | 13 (10.6%) | | | IRF only | 25 (13.9%) | 14 (10.2%) | 20 (12.5%) | 11 (8.9%) | | | No IRF nor SRF | 105 (58.3%) | 107 (78.1%) | 96 (60.0%) | 96 (78.0%) | Table 4: Baseline Characteristics of Safety Population and mYAP - 001 Study {13}------------------------------------------------ % = n / N × 100%. Manifest refraction was not recorded at the initial phase of the study. 1 Primary Eye = study eye. Secondary Eye = the AMD or NV-AMD fellow eye of the study eye AMD diagnostic findings were collected from participants' medical record. OCT fluid status is based upon review by PI of the OCT taken during the enrollment visit. For the safety cohort (N=180), the success rate of initial NVHO setup with completion of the non-qualifying tutorial was 86.7% (95% CI 80.8% - 91.3%; 156/180). 24 participants (13.3%) who still performed NVHO self-imaging did not successfully complete tutorials for either the primary or secondary eyes. The success rate of NVHO self-imaging (i.e., completing the selfimaging regardless of completing imaging data transmission to Notal Health Cloud) was 96.1% (95% CI 92.2% - 98.4%). Seven participants (3.9%) did not self-image successfully. The rate of successful transmission of any self-imaging being transmitted to the Notal Health Cloud was 97.2% (95% CI, 93.6% - 99.1%) in study eyes and 94.9% (95% CI, 89.8% - 97.9%) in fellow eyes. The rate of successful transmission of completed self-imaging to the Notal Health Cloud was 96.7% (95% CI, 92.9% - 98.8%) in study eyes and 94.2% (95% CI, 88.8% - 97.4%) in fellow eyes. 31 of 180 participants (17.2%) encountered device errors and/or malfunctions that precluded self-imaging and necessitated a device exchange. 120 of 180 participants (66.7%) contacted the NVDC for technical support. The NVDC contacted 47 of 180 participants (26.1%) with low adherence to device use to remind them to perform self-imaging and 84 of 180 (46.7%) for technical support. The NVDC also contacted 173 of 180 participants (96.1%) to remind participants of an upcoming scheduled in-clinic study visit. Of those in the safety cohort who did not discontinue after completion of the initial tutorial and device calibration and performed self-imaging (N=165), the mean MSI score of the firstcompleted, study-eye self-images ranged from 4.388 to 4.557 during the first week (study days 1 to 7), 4.47 to 4.57 during the second week (study days 8 to 14), 4.28 to 4.49 during the third week (study days 15 to 21), 4.32 to 4.45 during the fourth week (study days 22 to 28), and 3.64 to 4.35 during the fifth week (days 29 to 35). The proportion of participants who obtained a first {14}------------------------------------------------ completed NVHO study-eye scan with MSI <2 ranged from 0.6% to 4.3% during the first week (study days 1 to 7), zero to 3.7% during the second week (study days 8 to 14), zero to 6.4% during the third week (study days 15 to 21), 0.8% to 4.0% during the fourth week (study days 22 to 28), and 2.1% to 6.5% during the fifth week (days 29 to 35). The following table (Table 5) shows the primary analysis of NVHO Positive Percent Agreement (PPA) and Negative Percent Agreement (NPA) for Retinal Hypo-reflective Spaces Visualization based on Retinal Hypo-reflective Spaces Visualized on the Cirrus 3mm Area Modified Visualization Analysis Population (including subjects with NVHO and Cirrus scans within 24 Hours apart). The PPA was 0.864 (95% CI 0.802, 0.926; p=0.043) and NPA was 0.849 (95% CI 0.792, 0.907; p=0.094). | | Table 5: Primary Analysis of NVHO PPA and NPA for Retinal Hypo-reflective Spaces | | | | |--------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------|-----------------------------------------------------------------|-------------------------------------------------------|-----------------------------------------------------------------| | Visualization | | | | | | | Positive Percent Agreement<br>(PPA) | | Negative Percent Agreement<br>(NPA) | | | Total Retinal Hypo-reflective spaces<br>(TRO) | Rate<br>(95% CI) | Two-sided<br>p-value for<br>H0: Pp ≤ 0.8<br>vs.<br>Ha: Pp > 0.8 | Rate<br>(95% CI) | Two-sided<br>p-value for<br>H0: PN ≤ 0.8<br>vs.<br>Ha: PN > 0.8 | | # pairs of scans of # subjects | 163 pairs of scans of 105 eyes<br>with Cirrus = TRO | | 146 pairs of scans of 95 eyes<br>with Cirrus = no TRO | | | # NVHO scans with TRO | 143 | | 20 | | | # NVHO scans without TRO | 19 | | 124 | | | # NVHO scans not gradable | 1 | | 2 | | | Pragmatic estimation per repeated<br>measurements using optimal weighting<br>method, Pw1,2 | 0.864<br>(0.802,<br>0.926) | 0.043 | 0.849<br>(0.792,<br>0.907) | 0.094 | | Point estimates and 95% CIs per 2000<br>Cluster Bootstrap resampling* for PPA<br>and NPA | 0.877 [143/163]<br>(95% CI, 0.819-0.929) | | 0.849 [124/146]<br>(95% CI, 0.802-0.894) | | The NVHO and Cirrus scan pairs with a scan time difference of > 24 hours were excluded from the analysis. : The mean and (2.5th, 97.5th) percentiles of 2000 Cluster Bootstrap samples with re-sampling of participants. - 1 Article. Sin-Ho Jung. Seung-Ho Kang and Chul Ahn (2001). Sample size calculations for clustered binary data, Statist. Med. 2001; 20:1971-1982 - 2 Pragmatic estimation: PPA is based on all pairs of Cirrus and NVHO scans with Cirrus graded as with fluid (i.e., + or present) regardless of whether NVHO cannot be graded; NPA is based on all pairs of Cirrus and NVHO with Cirrus graded as without fluid (i.e., - or absent) regardless of whether NVHO cannot be graded. Colin B. Begg. Robert A. Greenest. and Boris Iglewicz (1986). {15}------------------------------------------------ The Influence of Uninterpretability on the Assessment of Diagnostic Tests. J Chron Dis Vol. 39. No. 8. pp. 575-584. 51 interim visits for 51 of 180 participants (28.3%) were conducted in response to the daily RC review. An NV-AMD treatment was administered at 26 of the 51 interim visits (51.0%). (Note: The "001 Study" was not designed to demonstrate the ability of the NVHO device to serve as an "early detection tool.") 10 adverse events (AEs) were reported for six participants. Of these 10. four were considered serious (SAEs). All SAEs were non-ocular in nature (myocardial infarction, pulmonary edema, fall, COVID-19 infection). Two AEs in two participants were ocular in nature (ocular pain and redness at the site of an intravitreal injection; eyelid stye). No ocular AEs involving vision loss were reported. None of the AEs were considered related to the NVHO device, and all AEs resolved prior to study termination. As noted above, the presence of confounding lesions was assessed by the RC. ### Summary of the 006 Clinical Study ### Overview: The "006 Study" was a prospective, cross-sectional, observational, single-visit study conducted at six sites in the United States. Participants with diagnosed NV-AMD in at least one eligible eve and best-corrected visual acuity of 20/320 or better in the study eve were enrolled. There was no minimum or maximum age requirement for eligibility. Those who had any other retinal disease requiring steroidal or anti-VEGF treatment, CIRRUS HD-OCT scan on screening visit with a signal strength <6, or those with non-neovascular (i.e., "dry") AMD in the study eve were excluded. The presence of confounding lesions (epiretinal membrane [ERM], macular hole, pseudocysts, outer retinal tubulations, hemorrhage, pigment epithelial detachments, subretinal hyper-reflective material [SHRM], geographic atrophy [GA], and hyper-reflective retinal spots or foci) was not exclusionary. One NV-AMD eve was determined to be the study eve. The purpose of the study was the following: 1) to evaluate the agreement in estimated retinal fluid volume between manually segmented CIRRUS HD-OCT macular scans versus the Notal OCT Analyzer (NOA) algorithm analyzing NVHO scans: 2) to estimate the repeatability and reproducibility of the TRO parameter: 3) to evaluate the amount of overlap in segmentation of IRF and SRF between NOA and manual graders. Demographic information and medical history were collected from enrolled participants. Manifest refraction, best-corrected Snellen visual acuity (BCVA) assessment, and assessment of media opacity were performed. Initial macular scaming with the CIRRUS HD-OCT was performed. One study eve per participant was selected. Participants then received a general overview on how to self-operate the NVHO device. Imaging without pharmacologic pupil dilation using CIRRUS HD-OCT and two NVHO devices was performed. The order of CIRRUS vs. NVHO imaging for each participant and the order of NVHO device use were randomized. Independent, masked graders from a third-party reading center (RC) performed manual segmentation of hypo-reflective spaces on the central 3×3-mm area of acceptable CIRRUS macular scans. Graders were masked to each others' determinations and to the participant ID {16}------------------------------------------------ number. Ordering of the scans was randomized. SRF was defined as a hypo-reflective space located beneath the retina between the ISE (integrity of the inner segment ellipsoid band, or inner segment/outer segment [IS/OS] border) and RPE layers and IRF was defined as a hypo-reflective space located in the retina between the internal limiting membrane (ILM) and ISE layers. The measurement variability of NOA-based volume estimates under repeatability conditions (i.e., within the same person using/operating the same NVHO device repeatedly) and reproducibility conditions (i.e., the same person using/operating different NVHO devices repeatedly) were determined using a random-effects analysis of variance (ANOVA) model. The agreement between NOA-based and CIRRUS-based volume estimates was calculated using Bland-Altman 95% limits of agreement (LOAs) and Deming regression analyses. The segmentation overlap analysis included calculation of the device-grader Dice coefficient for each grader averaged over cases and the calculation of the grader Dice coefficient for each pair of graders averaged over cases. Since the Dice coefficient is undefined when both methods in the comparison do not provide a segmentation, those cases were excluded from the calculation of the average of Dice coefficients. To account for these excluded cases, the device-grader negative percent agreement (NPA) was calculated for each expert. Uncertainty of results was characterized by 95% confidence intervals representing case variability. # Results: 398 participants were enrolled. 11 (2.8%) were screen failures; therefore, the safety cohort (all eligible participants who underwent CIRRUS or NVHO scanning) is comprised of 387 participants. 78 (19.6%) exited from the study early (due mainly to inability to successfully calibrate the NVHO [N=57]). Participants who could not successfully complete either NVHO or CIRRUS imaging (or both) or whose scans did not have eligible fluid measurements were excluded from precision, agreement, and Dice coefficient analysis cohorts. The "Fluid Precision Analysis Population" and "Fluid Agreement Analysis Population" cohorts were comprised of 331 participants and the "Dice Analysis Population" was comprised of 336 participants. Demographics, and relevant baseline clinical characteristics are shown in Tables 6 and 7. The majority of the participants were White and not Hispanic or Latino. | Table 6: Demographics of Safety Population and mVAP – 006 Study<br>Demographics | | Safety<br>Population<br>N = 387 | Fluid<br>Marking<br>(DICE)<br>Analysis<br>Population<br>N = 336 | Fluid<br>Agreement<br>Analysis<br>Population1<br>N = 331 | |---------------------------------------------------------------------------------|----------------------------------|---------------------------------|-----------------------------------------------------------------|----------------------------------------------------------| | Age | N | 387 | 336 | 331 | | | Mean ± SD | 76.2 ± 7.5 | 75.4 ± 7.4 | 75.5 ± 7.3 | | | Median | 77.0 | 76.0 | 76.0 | | | Min, Max | 53, 91 | 53, 91 | 53, 91 | | Gender | Male | 158 (40.8%) | 140 (41.7%) | 137 (41.4%) | | | Female | 229 (59.2%) | 196 (58.3%) | 194 (58.6%) | | Race | American Indian or Alaska Native | 2 (0.5%) | 1 (0.3%) | 1 (0.3%) | Table 6: Domographics of Safaty Ropulation and mVAD . 006 Study {17}------------------------------------------------ | | Asian | 2 (0.5%) | 1 (0.3%) | 1 (0.3%) | |-----------|----------------------------------------------------|-------------|-------------|-------------| | | Black or African American | 6 (1.6%) | 5 (1.5%) | 5 (1.5%) | | | Native Hawaiian or Other Pacific Islander | 1 (0.3%) | 1 (0.3%) | 1 (0.3%) | | | White | 370 (95.6%) | 322 (95.8%) | 317 (95.8%) | | | Not Reported | 6 (1.6%) | 6 (1.8%) | 6 (1.8%) | | Ethnicity | Hispanic or Latino | 9 (2.3%) | 9 (2.7%) | 9 (2.7%) | | | Not Hispanic or Latino | 375 (96.9%) | 324 (96.4%) | 319 (96.4%) | | | Not Reported | 3 (0.8%) | 3 (0.9%) | 3 (0.9%) | | Education | Less than High School Degree | 13 (3.4%) | 11 (3.3%) | 11 (3.3%) | | | High School Degree | 109 (28.2%) | 87 (25.9%) | 85 (25.7%) | | | Some college (no degree) | 80 (20.7%) | 71 (21.1%) | 70 (21.1%) | | | College Degree (Associate or Bachelor's<br>Degree) | 125 (32.3%) | 112 (33.3%) | 111 (33.5%) | | | Graduate Degree | 52 (13.4%) | 49 (14.6%) | 48 (14.5%) | | | Trade School | 6 (1.6%) | 5 (1.5%) | 5 (1.5%) | | | Other2 | 2 (0.5%) | 1 (0.3%) | 1 (0.3%) | | Study Eye | OD | 181 (46.8%) | 160 (47.6%) | 157 (47.4%) | | | OS | 206 (53.2%) | 176 (52.4%) | 174 (52.6%) | % = n / N × 100%. 1 Fluid Agreement Analysis Population = Fluid Precision Analysis Population for this study 2 Including some graduate school, trade school, and tech school | Table 7: Baseline Characteristics Safety, Fluid Marking and Fluid Agreement | |-----------------------------------------------------------------------------| | Analysis Populations | | | | Safety<br>Population | Fluid<br>Marking<br>(DICE)<br>Analysis<br>Population | Fluid<br>Agreement<br>Analysis<br>Population¹ | |----------------------------|----------------------------------------------------------------|----------------------|------------------------------------------------------|-----------------------------------------------| | Baseline Characteristics | | N = 387 | N = 336 | N = 331 | | AMD Diagnosis | AMD - Early AMD | 1 (0.3%) | 1 (0.3%) | 1 (0.3%) | | Based on<br>Participants' | AMD - Intermediate AMD | 6 (1.6%) | 4 (1.2%) | 4 (1.2%) | | Medical Record | NV-AMD - active (fluid present) | 372 (96.1%) | 323 (96.1%) | 319 (96.4%) | | | NV-AMD - non active (no fluid<br>present) | 8 (2.1%) | 8 (2.4%) | 7 (2.1%) | | Lens | Phakia (cataract absent) | 6 (1.6%) | 5 (1.5%) | 5 (1.5%) | | Status | Phakia (cataract present) | 123 (31.8%) | 119 (35.4%) | 116 (35.0%) | | | Pseudophakia | 258 (66.7%) | 212 (63.1%) | 210 (63.4%) | | Ocular Media<br>Assessment | Main vessels and the small vessels<br>are clearly seen. | 357 (92.2%) | 307 (91.4%) | 305 (92.1%) | | | Small vessels are invisible while<br>main vessels can be seen. | 30 (7.8%) | 29 (8.6%) | 26 (7.9%) | | Prior | N | 359 | 313 | 310 | | Total # of<br>Injections | Mean ± SD | 25.9 ± 25.6 | 25.7 ± 25.2 | 26.0 ± 25.2 | | | Median | 17.0 | 18.0 | 18.0 | {18}------------------------------------------------ | | Min, Max | 1, 142 | 1, 134 | 1, 134 | |----------------------|----------------------------|----------------|----------------|----------------| | | N (Unavailable) | 28 | 23 | 21 | | Medication | Aflibercept (Eylea) | 190 (49.1%) | 167 (49.7%) | 166 (50.2%) | | Most | Bevacizumab (Avastin) | 97 (25.1%) | 85 (25.3%) | 84 (25.4%) | | Recently | Brolucizumab (Beovu) | 5 (1.3%) | 4 (1.2%) | 4 (1.2%) | | Administered | Other Investigational Drug | 13 (3.4%) | 12 (3.6%) | 11 (3.3%) | | | Ranibizumab (Lucentis) | 47 (12.1%) | 39 (11.6%) | 39 (11.8%) | | | Vabysmo (Faricimab-svoa) | 7 (1.8%) | 6 (1.8%) | 6 (1.8%) | | | Unavailable | 28 (7.2%) | 23 (6.8%) | 21 (6.3%) | | Spherical | N | 387 | 336 | 331 | | Equivalent | Mean ± SD | -0.131 + 1.686 | -0.166 ± 1.745 | -0.172 ± 1.749 | | | Median | 0.000 | 0.000 | 0.000 | | | Min, Max | -7.500, 4.750 | -7.500, 4.750 | -7.500, 4.750 | | Best | N | 387 | 336 | 331 | | Corrected | Mean logMAR (Snellen) | 0.370 | 0.350 | 0.344 | | Visual | | (20/46.9) | (20/44.8) | (20/44.2) | | Activity | SD logMAR | 0.293 | 0.283 | 0.277 | | | Median logMAR (Snellen) | 0.300 | 0.300 | 0.300 | | | | (20/39.9) | (20/39.9) | (20/39.9) | | | Min logMAR (Snellen) | -0.10 | -0.10 | -0.10 | | | | (20/16.0) | (20/16.0) | (20/16.0) | | | Max logMAR (Snellen) | 1.20 | 1.20 | 1.20 | | | | (20/320.0) | (20/320.0) | (20/320.0) | | Best | 20/40 or Better | 194 (50.1%) | 178 (53.0%) | 177 (53.5%) | | Corrected | 20/41 to 20/80 | 110 (28.4%) | 92 (27.4%) | 92 (27.8%) | | Visual | 20/81 to 20/200 | 68 (17.6%) | 55 (16.4%) | 52 (15.7%) | | Activity | 20/201 to 20/320 | 15 (3.9%) | 11 (3.3%) | 10 (3.0%) | | Principal | Both IRF and SRF | 82 (21.2%) | 73 (21.7%) | 73 (22.1%) | | Investigator's | | | | | | Retinal Fluid | SRF only | 123 (31.8%) | 108 (32.1%) | 109 (32.9%) | | Assessment Based | IRF only | 142 (36.7%) | 121 (36.0%) | 120 (36.3%) | | on<br>Commercial OCT | No IRF nor SRF | 35 (9.0%) | 29 (8.6%) | 28 (8.5%) | | | Unavailable | 5 (1.3%) | 5 (1.5%) | 1 (0.3%) | %=n/N×100%. 1 Fluid Agreement Analysis Population = Fluid Precision Analysis Population for this study The repeatability and reproducibility percent coefficient of variation (%CV) ranged from 24.6% to 436.4% and from 26.2% to 475.2%, respectively, for TRO <10 VU. For TRO >10 VU, repeatability %CVs ranged from 5.9% to 25.0%, and reproducibility %CVs ranged from 11.4% to 33.4%. The following table (Table 8) show the device-grader Dice coefficient and NPA for each grader acting as the "reference standard". # Table 8: Descriptive Statistics of Eye-Level DICE Between NOA versus Graders and Graders versus Graders {19}------------------------------------------------ | Statistics | NOA vs.<br>Grader 1 | NOA vs.<br>Grader 2 | NOA vs.<br>Grader 3 | Grader 1 vs.<br>Grader 2 | Grader 1 vs.<br>Grader 3 | Grader 2 vs.<br>Grader 3 | |---------------------------------------------------------|---------------------|---------------------|---------------------|--------------------------|--------------------------|--------------------------| | Comparing NOA-Grader TRO DICE to Grader-Grader TRO DICE | | | | | | | | N* | 278 | 289 | 299 | 279 | 297 | 298 | | Dice Mean ± SD | 0.5819 ±<br>0.2958 | 0.5655 ±<br>0.3203 | 0.5196 ±<br>0.3182 | 0.6222 ±<br>0.2754 | 0.5453 ±<br>0.3018 | 0.6000 ±<br>0.2918 | | 95% CI of DICE<br>Mean | 0.5470,<br>0.6168 | 0.5284,<br>0.6026 | 0.4834,<br>0.5558 | 0.5897,<br>0.6546 | 0.5108,<br>0.5797 | 0.5668,<br>0.6333 | | Dice Median | 0.6802 | 0.6844 | 0.6123 | 0.7221 | 0.6274 | 0.6964 | | NPA | 0.734<br>(58/79) | 0.746<br>(47/63) | 0.787<br>(37/47) | 0.722<br>(57/79) | 0.494<br>(39/79) | 0.603<br>(38/63) | | 95% CI of NPA¹ | 0.623,<br>0.827 | 0.621,<br>0.847 | 0.643,<br>0.893 | 0.609, 0.817 | 0.379, 0.609 | 0.472, 0.724 | | Comparing NOA-Grader SRO DICE to Grader-Grader SRO DICE | | | | | | | | N* | 213 | 223 | 241 | 205 | 232 | 235 | | Dice Mean ± SD | 0.5379 ±<br>0.3495 | 0.5394 ±<br>0.3550 | 0.4951 ±<br>0.3612 | 0.5670 ±<br>0.3283 | 0.4934 ±<br>0.3530 | 0.5558 ±<br>0.3403 | | 95% CI of DICE<br>Mean | 0.4907,<br>0.5852 | 0.4925,<br>0.5862 | 0.4492,<br>0.5409 | 0.5218,<br>0.6122 | 0.4477,<br>0.5390 | 0.5120,<br>0.5995 | | Dice Median | 0.6811 | 0.6799 | 0.6536 | 0.7161 | 0.6129 | 0.6956 | | NPA | 0.764<br>(123/161) | 0.801<br>(113/141) | 0.856<br>(95/111) | 0.814<br>(131/161) | 0.646<br>(104/161) | 0.716<br>(101/141) | | 95% CI of NPA¹ | 0.691,<br>0.827 | 0.726,<br>0.864 | 0.776,<br>0.915 | 0.745, 0.871 | 0.567, 0.720 | 0.634, 0.789 | | Comparing NOA-Grader IRO DICE to Grader-Grader IRO DICE | | | | | | | | N* | 160 | 172 | 224 | 180 | 231 | 229 | | Dice Mean ± SD | 0.4594 ±<br>0.3139 | 0.4316 ±<br>0.3263 | 0.2972 ±<br>0.3107 | 0.5100 ±<br>0.3312 | 0.3631 ±<br>0.3358 | 0.4227 ±<br>0.3444 | | 95% CI of DICE<br>Mean | 0.4104,<br>0.5084 | 0.3825,<br>0.4807 | 0.2563,<br>0.3381 | 0.4613,<br>0.5587 | 0.3196,<br>0.4067 | 0.3779,<br>0.4676 | | Dice Median | 0.5429 | 0.5133 | 0.1970 | 0.6519 | 0.4059 | 0.5467 | | NPA | 0.946<br>(176/186) | 0.965<br>(164/170) | 0.966<br>(112/116) | 0.839<br>(156/186) | 0.565<br>(105/186) | 0.629<br>(107/170) | | 95% CI of NPA¹ | 0.903,<br>0.974 | 0.925,<br>0.987 | 0.914,<br>0.991 | 0.778, 0.888 | 0.490, 0.637 | 0.552, 0.702 | N* is the number of cases that have a segmentation from NOA or the Grader 1 Exact CI per binomial distribution The following table (Table 9) shows the device-grader PPA for each grader PPA for each grader acting as the "reference standard." Table Error! No text of specified style in document.9: Descriptive Statistics of Pixel-level Positive Percent Agreement of Study Eyes on NVHO Scan {20}------------------------------------------------ | "True" | | Grader 1 | | | Grader 2 | | | Grader 3 | | | |--------|-------------------|-----------------------|-----------------------|-----------------------|-----------------------|-----------------------|-----------------------|-----------------------|-----------------------|-----------------------| | "Test" | | NOA | Grader<br>2 | Grader<br>3 | NOA | Grader<br>1 | Grader<br>3 | NOA | Grader<br>1 | Grader<br>2 | | TRO | N1 | 273 | 273 | 273 | 257 | 257 | 257 | 289 | 289 | 289 | | | Mean ±<br>SD | 0.5315<br>±<br>0.2958 | 0.6147<br>±<br>0.2676 | 0.7007<br>±<br>0.2346 | 0.5919<br>±<br>0.2686 | 0.7511<br>±<br>0.2197 | 0.7262<br>±<br>0.2140 | 0.4832<br>±<br>0.3162 | 0.6278<br>±<br>0.2989 | 0.5356<br>±<br>0.3054 | | | 95% CI of<br>Mean | 0.4963,<br>0.5668 | 0.5828,<br>0.6466 | 0.6727,<br>0.7286 | 0.5589,<br>0.6249 | 0.7241,<br>0.7781 | 0.6999,<br>0.7525 | 0.4466,<br>0.5198 | 0.5932,<br>0.6624 | 0.5002,<br>0.5709 | | | Median | 0.6011 | 0.7049 | 0.7793 | 0.6600 | 0.8158 | 0.7926 | 0.5143 | 0.7207 | 0.6262 | | SRO | N1 | 195 | 195 | 195 | 175 | 175 | 175 | 225 | 225 | 225 | | | Mean ±<br>SD | 0.5584<br>±<br>0.3147 | 0.5640<br>±<br>0.3066 | 0.6727<br>±<br>0.2478 | 0.6462<br>±<br>0.2853 | 0.7580<br>±<br>0.2502 | 0.7133<br>±<br>0.2287 | 0.5011<br>±<br>0.3615 | 0.6277<br>±<br>0.3544 | 0.5019<br>±<br>0.3567 | | | 95% CI of<br>Mean | 0.5139,<…