Abstract
Purpose
To evaluate the potential benefits of ultra-widefield (UWF) fluorescein angiography (FA) versus color photography in assessing the risk of diabetic retinopathy (DR) worsening.
Design
Exploratory analysis of a longitudinal observational study (Protocol AA).
Participants
Participants with ≥1 eye having nonproliferative DR at baseline.
Methods
For each UWF-color and UWF-FA image, the extent and severity of DR lesions within the ETDRS 7-field area (“masked”) and the full image area (“unmasked”) were determined. Diabetic retinopathy severity levels were graded for each combination of modality and masking condition using the ETDRS Diabetic Retinopathy Severity Scale (DRSS). Associations between baseline UWF-color and UWF-FA DR severity with disease worsening were evaluated and compared.
Main Outcome Measures
Disease worsening over 4 years, defined as ≥2-step DRSS worsening on masked UWF-color or receipt of DR treatment.
Results
Among the 768 eyes, DRSS agreement between masked UWF-color versus masked or unmasked UWF-FA at each visit was fair to moderate (range for κ: 0.34–0.53). Ultra-widefield FA identified more severe DR than masked UWF-color in 57% of the masked UWF-FA images and 62% of the unmasked. Over 4 years, disease worsening was more common in eyes with a more severe grade on masked UWF-FA than masked UWF-color for hemorrhages and/or microaneurysms (47% vs. 37%; P = 0.04), intraretinal microvascular abnormalities (IRMA) (50% vs. 36%; P = 0.007), and new vessels/neovascularization elsewhere (NVE) (57% vs. 39%; P < 0.003), and on unmasked UWF-FA for IRMA (49% vs. 35%; P = 0.008) and NVE (58% vs. 38%; P < 0.001). A greater risk of disease worsening was also associated with more severe DRSS level on masked (50% vs. 30%; P < 0.001) and on unmasked (50% vs. 28%; P < 0.001) UWF-FA compared with masked UWF-color.
Conclusions
This analysis highlights important differences between UWF-color and UWF-FA in assessing DR severity and predicting disease progression over time. Ultra-widefield FA appears to offer a more accurate risk assessment of future DR worsening. These findings support the integration of UWF-FA findings into the existing grading scales and the development of supplemental DRSS grading criteria that incorporate the additional prognostic information provided by angiographic imaging.
Financial Disclosure(s)
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
Keywords: Diabetic retinopathy severity, Ultra-widefield fluorescein angiography, Color photography.
Diabetic retinopathy (DR) is a common cause of vision impairment.1 Early detection is critical for timely intervention to halt or slow disease progression and vision impairment.2 Accurately assessing disease severity is also critical for guiding treatment decisions and predicting vision loss.3 The ETDRS Diabetic Retinopathy Severity Scale (DRSS), developed from 7-field stereoscopic fundus photography, has served as the gold standard for assessing DR.4 However, advancements in retinal imaging, particularly ultra-widefield (UWF) imaging, allow clinicians to visualize larger portions of the retina, including the peripheral areas that are often excluded in conventional imaging techniques.5
Ultra-widefield color imaging and UWF fluorescein angiography (FA) offer complementary perspectives in evaluating DR. While both modalities provide a larger view of the retina, UWF-FA is valuable for highlighting critical features not seen on color imaging such as retinal nonperfusion and vascular leakage.6, 7, 8 Furthermore, studies have demonstrated that UWF-FA provides superior detection of subtle DR lesions and microvascular abnormalities that may be missed on color images.9 These advantages suggest that UWF-FA could provide a more accurate assessment of the risk of disease progression over time.
The Diabetes Control and Complications Trial and ETDRS studies compared FA with stereoscopic fundus photography, identifying additional lesions and risk factors on FA.10,11 However, these early studies were limited by their focus on a relatively small area of the posterior retina without the ability to standardize the photographic capture of each field. The additional findings on FA evaluated in this manner did not translate into substantial management benefits for patients. Here, we evaluate the potential benefits of UWF-FA versus color photography for assessment of risk of DR worsening.
It remains unclear how DR severity grading on UWF-color compares to that on UWF-FA, or whether UWF-FA offers a more accurate stratification of disease risk. This study addresses these issues by comparing DR evaluations using UWF-color and UWF-FA images, applying the ETDRS color DRSS to both modalities, evaluating the same retinal areas, and determining agreement rates across a broad range of DR severity. This approach will address whether existing color DRSS methodology is sufficient for characterizing DR lesions using UWF-FA and the extent of area needing evaluation to capture prognostic value provided by UWF-FA imaging.
Methods
Study Overview
We performed an exploratory analysis of DRCR Retina Network 4-year multicenter prospective longitudinal observational study Protocol AA. Study participants were ≥18 years old, had type 1 or type 2 diabetes, and had ≥1 eye with nonproliferative DR (NPDR) at baseline based on reading center assessment of modified 7-field standard ETDRS color photographs. Study participants provided written informed consent. The study adhered to the tenets of the Declaration of Helsinki and was approved by multiple ethics boards. The protocol is available on the study website.
Image Acquisition and Grading
Ultra-widefield color images (Optomap plus, Red-Green) were collected at baseline and annual visits, and UWF-FA images were collected at baseline, 1-year, and 4-year visits. The UWF imaging procedures and grading protocol are in prior reports.6,7,12 All UWF images were evaluated by a central reading center, and graders were masked to clinical characteristics and features derived from other imaging modalities. The ETDRS-DRSS, originally designed for color images, was used to determine DR severity in UWF-color and was adapted for use on UWF-FA.4 A detailed description of this grading methodology is illustrated in Figure S1 (available at www.ophthalmologyscience.org) and the accompanying legend. This investigational grading methodology maintained the standard criteria for color images while incorporating modifications to apply the same definitions to UWF-FA images. Each UWF image was first graded with a digital template mask automatically overlaid to reveal only the idealized ETDRS 7-field area, followed by the grading, with the mask template removed, of the full image area (including the retinal periphery). For each UWF image, an ETDRS-DRSS score was derived separately based on the extent and severity of the individual lesions within the standard ETDRS 7 fields (termed “masked”) or the entire visible area (termed “unmasked”), using a 12-level composite score. Sensitivity analyses were performed on the UWF-FA with DR lesion grades assessed within fields 1, 2, and 3.
Statistical Analysis
The percentages and 95% confidence intervals (CIs) for the agreement between DRSS levels on different imaging modalities were calculated from generalized estimating equations with adjustment for the correlation between eyes from the same study participant. Kappa statistics (κ) and 95% CIs were calculated using quadratic weights. The primary outcome of interest was time to disease worsening over 4 years, defined as DRSS worsening of ≥2 steps from baseline within the ETDRS standard 7 fields on UWF-color images, or receiving DR treatment during follow up as defined in the primary outcome papers.6,12 Because DRSS worsening was defined using masked UWF-color images with an idealized ETDRS 7-field overlay (rather than conventional ETDRS 7-field stereoscopic photography), the absolute progression rates reported here may differ from trials that use more variable standard 7-field photography. The cumulative proportions and 95% CIs for disease worsening over 4 years were estimated with the Kaplan–Meier method. Hazard ratios (HRs) and P values were obtained from a Cox proportional hazards model with adjustment for baseline DRSS level. A robust sandwich estimate of the covariance matrix was used to account for the correlation between eyes of participants with 2 study eyes. All statistical analyses were performed using the observed data without imputation for missing data. All P values were 2-sided. P values <0.05 were considered statistically significant, recognizing all analyses are considered exploratory and unadjusted for multiplicity; therefore, the type I error rate was not controlled at 5%. Analyses were conducted using SAS/STAT 15.1 (SAS Institute).
Results
DRSS Level Agreement for Masked UWF-Color vs. Masked and Unmasked UWF-FA
The distribution of DRSS level on the UWF-color and UWF-FA images within the ETDRS standard 7 fields among study and nonstudy eyes after combining all visits is shown in Table S1 (available at www.ophthalmologyscience.org). Overall, 57% of eyes had more severe DR identified on masked UWF-FA than on masked UWF-color (27% had 1-step more severe and 30% had ≥2-step more severe), while only 12% of eyes had more severe DRSS on masked UWF-color. At baseline, 90% of eyes had a nonzero overall nonperfusion index (NPI), with the proportion increasing with worse baseline DRSS level: 77% among eyes with mild NPDR or better, 93% with moderate NPDR, 97% with moderately severe NPDR, 98% with severe or very severe NPDR, and 100% with proliferative DR. Eyes with worse baseline DR on the masked UWF-FA compared with masked UWF-color images showed a consistent trend of higher NPI (Table S2; Fig S2, available at www.ophthalmologyscience.org). The percentage of agreement in the DRSS levels assessed from the 2 imaging modalities decreased over time but remained moderate at each visit (weighted κ [95% CI] = 0.53 [0.47–0.59] at baseline, 0.44 [0.38–0.50] at 1 year, and 0.40 [0.33–0.47] at 4 years; Fig S3, available at www.ophthalmologyscience.org).
The distribution of DRSS levels assessed from the entire retina visible on the UWF-FA image versus the masked UWF-color is shown in Table S3 (available at www.ophthalmologyscience.org). Combining all visits, DR was more severe on unmasked UWF-FA than on masked UWF-color in 62% of eyes: 1 step more severe in 27% and ≥2 steps more severe in 36%. Conversely, DRSS on masked UWF-color was more severe in 11% of eyes. Eyes with unmasked UWF-FA DRSS more severe than the masked UWF-color DRSS also appeared to have greater nonperfusion at baseline (Table S4; Fig S4, available at www.ophthalmologyscience.org). Over 80% of eyes with mild to moderately severe NPDR on UWF-color were identified as having more severe DRSS on unmasked UWF-FA. The percentage of exact agreement (95% CI) in DRSS levels was 33% (29%–36%) at baseline, 24% (21%–28%) at 1 year, and 20% (17%–24%) at 4 years (Fig S5, available at www.ophthalmologyscience.org). Over time, the agreement in DRSS levels between the 2 imaging modalities was moderate to fair (weighted κ [95% CI] = 0.45 [0.39–0.51] at baseline), 0.38 [0.32–0.44]) at 1 year and 0.34 [0.28–0.41]) at 4 years.
Disease Worsening over 4 Years by Baseline DRSS Levels on Masked, Unmasked, and Field 1 to 3 UWF-FA
Among the 477 study eyes with baseline DRSS level of 35-53 on masked UWF-FA, 91 (19%) had mild NPDR, 74 (16%) had moderate NPDR, 168 (35%) had moderately severe NPDR, and 144 (30%) had severe and very severe NPDR. The cumulative proportion (95% CI) of worsening of ≥2 steps on the DRSS or treatment for DR through 4 years was 21% (13%-33%) in mild NPDR, 50% (38%-63%) in moderate NPDR, 31% (24%-39%) in moderately severe NPDR, and 53% (44%-63%) in severe and very severe NPDR (Fig S6, available at www.ophthalmologyscience.org).
On the unmasked UWF-FA, 68 (15%) of 462 eyes had mild NPDR, 69 (15%) had moderate NPDR, 172 (37%) had moderately severe NPDR, and 153 (33%) had severe and very severe NPDR at baseline. The cumulative proportion (95% CI) of disease worsening over 4 years was 21% (12%-35%) in mild NPDR, 48% (36%-62%) in moderate NPDR, 29% (23%-38%) in moderately severe NPDR, and 50% (42%-60%) in severe and very severe NPDR (Fig S7, available at www.ophthalmologyscience.org).
Stratified by the baseline DRSS level based on only fields 1 to 3 of UWF-FA, disease worsening over 4 years was observed in 33% (27%, 41%) in the mild group (N = 227), 39% (31%, 49%) in the moderate group (N = 148), 51% (42%, 61%) in the moderately severe group (N = 139), and 100% in the severe and very severe group (N = 4; Fig S8, available at www.ophthalmologyscience.org).
Disease Worsening over 4 Years by Baseline DR Lesions on Masked, Unmasked, and Field 1 to 3 UWF-FA as Compared with Masked UWF-Color
Compared with the lesion-level grades on baseline masked UWF-color, a more severe grade on baseline masked UWF-FA was observed in 176 (32%) of 542 eyes for hemorrhages and/or microaneurysms (H/Ma), 169 (31%) of 539 for intraretinal microvascular abnormalities (IRMA), 57 (11%) of 542 for new vessels/neovascularization elsewhere (NVE), and 45 (8%) of 539 for venous beading (VB). A greater risk of disease worsening over 4 years was reported among eyes with a more severe lesion grade on the masked UWF-FA for H/Ma (47% vs. 37%; HR: 1.34; 95% CI, 1.01–1.79; P = 0.04), IRMA (50% vs. 36%; HR: 1.55; 95% CI, 1.13–2.14; P = 0.007) and NVE (57% vs. 39%; HR: 1.91; 95% CI, 1.24–2.95; P = 0.003), but not for VB (49% vs. 39%; HR: 1.34; 95% CI, 0.84–2.13; P = 0.22; Table 5).
Table 5.
Disease Worsening over 4 Yrs by the Presence of a More Severe Baseline Lesion Grade on Masked UWF-FA Compared with Masked UWF-Color
| Lesion Type | More Severe Lesion Grade on Masked UWF-FA than Masked UWF-Color∗ | N | Cumulative Proportion (95% CI) | Hazard Ratio (95% CI)† | P Value† |
|---|---|---|---|---|---|
| H/Ma | No | 366 | 37% (31%, 43%) | Reference | 0.04 |
| Yes | 176 | 47% (40%–55%) | 1.34 (1.01–1.79) | ||
| IRMA | No | 370 | 36% (31%–42%) | Reference | 0.007 |
| Yes | 169 | 50% (42%–58%) | 1.55 (1.13–2.14) | ||
| NVE | No | 485 | 39% (34%–44%) | Reference | 0.003 |
| Yes | 57 | 57% (44%–71%) | 1.91 (1.24–2.95) | ||
| VB | No | 494 | 39% (35%–44%) | Reference | 0.22 |
| Yes | 45 | 49% (34%–66%) | 1.34 (0.84–2.13) |
CI = confidence interval; FA = fluorescein angiography; H/Ma = hemorrhages and/or microaneurysms; IRMA = intraretinal microvascular anomalies; NVE = new vessels/neovascularization elsewhere; UWF = ultra-widefield; VB = venous beading.
Based on the most severe lesion-level grade in field 3-7 for each imaging modality.
Hazard ratio and P value were obtained from a Cox proportional hazards model with adjustment for baseline Diabetic Retinopathy Severity Scale level. A robust sandwich estimate of the covariance matrix was used to account for the correlation between 2 eyes from the same participants.
On baseline unmasked UWF-FA, a lesion-level grade more severe than the masked UWF-color grade was identified in 227 (42%) of 542 eyes for H/Ma, 182 (34%) of 539 for IRMA, 72 (13%) of 542 for NVE, and 48 (9%) of 539 for VB. Representative images showing ≥2-step more severe disease on UWF-FA than on masked UWF-color are shown in Figure 9. The risk of disease worsening was higher among eyes with a lesion more severe on unmasked UWF-FA than on masked UWF-color for IRMA (49% vs. 35%; HR: 1.57; 95% CI, 1.12–2.18; P = 0.008) and NVE (58% vs. 38%; HR: 1.99; 95% CI, 1.36–2.93; P < 0.001), but not for H/Ma (43% vs. 38%; HR: 1.08; 95% CI, 0.81–1.45; P = 0.60) or VB (50% vs. 39%; HR: 1.41; 95% CI, 0.91–2.18; P = 0.13; Table 6). Compared with the field 1-3 UWF-FA, NVE was the only lesion type associated with an increased risk of disease worsening over 4 years when a more severe baseline lesion grade was identified in the unmasked area (56% vs. 38%; HR: 1.90; 95% CI, 1.29–2.81; P = 0.001; Table S7, available at www.ophthalmologyscience.org).
Figure 9.
Representative paired UWF-FA and UWF-color images illustrating ≥2-step more severe DR detected on UWF-FA compared with masked UWF-color images. A, An eye with featureless retinal characteristics on UWF-color imaging but exhibiting significant retinal nonperfusion on UWF-FA, with a measured nonperfusion area of 196 mm2 (nonperfusion index 0.24). Peripheral retinal nonperfusion is evident on UWF-FA but not detectable on UWF-color imaging. B, An eye with mild nonproliferative DR observed on masked UWF-color imaging but graded as early proliferative DR on UWF-FA, demonstrating a new vessel in peripheral field 5 (white circle) and extensive intraretinal microvascular abnormalities more severe than standard 8A in the periphery (white dashed line, magnified in-set). C, An eye with mild nonproliferative DR (ETDRS level 35) on masked UWF-color and severe nonproliferative DR (ETDRS level 53) on unmasked UWF-FA. DR = diabetic retinopathy; FA = fluorescein angiography; UWF = ultra-widefield.
Table 6.
Disease Worsening over 4 Yrs by Presence of a More Severe Baseline Lesion Grade on Unmasked UWF-FA Compared with Masked UWF-Color
| Lesion Type | More Severe Lesion Grade on Unmasked UWF-FA than Masked UWF-Color∗ | N | Cumulative Proportion (95% CI) | Hazard Ratio (95% CI)† | P Value† |
|---|---|---|---|---|---|
| H/Ma | No | 315 | 38% (33%–45%) | Reference | 0.60 |
| Yes | 227 | 43% (36%–50%) | 1.08 (0.81–1.45) | ||
| IRMA | No | 357 | 35% (30%–41%) | Reference | 0.008 |
| Yes | 182 | 49% (41%–57%) | 1.57 (1.12–2.18) | ||
| NVE | No | 470 | 38% (33%–43%) | Reference | <0.001 |
| Yes | 72 | 58% (46%–70%) | 1.99 (1.36–2.93) | ||
| VB | No | 491 | 39% (35%–44%) | Reference | 0.13 |
| Yes | 48 | 50% (36%–66%) | 1.41 (0.91–2.18) |
CI = confidence interval; FA = fluorescein angiography; H/Ma = hemorrhages and/or microaneurysms; IRMA = intraretinal microvascular anomalies; NVE = new vessels/neovascularization elsewhere; UWF = ultra-widefield; VB = venous beading.
Based on the most severe lesion-level grade in field 3-7 for each imaging modality.
Hazard ratio and P value were obtained from a Cox proportional hazards model with adjustment for baseline Diabetic Retinopathy Severity Scale level. A robust sandwich estimate of the covariance matrix was used to account for the correlation between 2 eyes from the same participants.
Disease Worsening over 4 Years by Baseline DRSS between Masked, Unmasked, and Field 1 to 3 UWF-FA as Compared with Masked UWF-Color
Among the 542 study eyes with gradable DRSS on both imaging modalities, 275 (51%) had a worse baseline DRSS level on the masked UWF-FA than on masked UWF-color, among whom an increased risk of disease worsening over 4 years was found (50% vs. 30%; HR: 2.06; 95% CI, 1.47–2.88; P < 0.001; Fig 10). This trend was consistent within each baseline DRSS subgroup (P value for interaction = 0.83; Fig S11, available at www.ophthalmologyscience.org). Similarly, 307 (57%) of 542 eyes had a more severe grade on unmasked UWF-FA than on masked UWF-color, who also had a significantly higher risk of disease worsening over 4 years (50% vs. 28%; HR: 2.37; 95% CI, 1.61–3.48; P < 0.001; Fig 12). This relationship was retained within baseline DRSS subgroups (P value for interaction = 0.73; Fig S13, available at www.ophthalmologyscience.org).
Figure 10.
Disease worsening over 4 years by baseline DRSS level on masked UWF-FA compared with masked UWF-color. The cumulative proportion of disease worsening was 30% (95% CI, 25%–37%) for eyes with the same or better baseline DRSS level on masked UWF-FA, and 50% (95% CI, 43%–56%) for eyes with worse baseline DRSS on masked UWF-FA compared with masked UWF-color. The hazard ratio was 2.06 (95% CI, 1.47–2.88; P < 0.001), estimated from a Cox proportional hazards model with adjustment for baseline ETDRS DR severity level within the ETDRS fields on UWF-color images and the correlation between the 2 study eyes from the same participant. CI = confidence interval; DR = diabetic retinopathy; DRSS = Diabetic Retinopathy Severity Scale; FA = fluorescein angiography; UWF = ultra-widefield.
Figure 12.
Disease worsening over 4 years by baseline DRSS level on unmasked UWF-FA compared with masked UWF-color. The cumulative proportion of disease worsening was 28% (95% CI, 22%–35%) for eyes with the same or better baseline DRSS level on unmasked UWF-FA, and 50% (95% CI, 44%–56%) for eyes with worse baseline DRSS on unmasked UWF-FA compared with masked UWF-color. The hazard ratio was 2.37 (95% CI, 1.61–3.48; P< 0.001), estimated from a Cox proportional hazards model with adjustment for baseline ETDRS DR severity level within the ETDRS fields on UWF-color images and the correlation between the 2 study eyes from the same participant. CI = confidence interval; DR = diabetic retinopathy; DRSS = Diabetic Retinopathy Severity Scale; FA = fluorescein angiography; UWF = ultra-widefield.
When comparing DRSS levels on the field 1-3 UWF-FA versus masked UWF-color, a more severe level from field 1-3 UWF-FA was observed in 109 (20%) of 541 eyes, whose risk of disease worsening over 4 years was significantly higher than those with a same or better DRSS level (59% vs. 36%; HR: 1.78; 95% CI, 1.31–2.43; P < 0.001; Fig S14, available at www.ophthalmologyscience.org). Compared with the field 1-3 FA DRSS level, a more severe level was observed in 352 (68%) eyes on the masked UWF-FA and 385 (74%) eyes on the unmasked UWF-FA (N = 518). The cumulative proportion of disease worsening over 4 years was similar regardless of the comparison of field 1-3 FA DRSS with the masked (42% vs. 37%; HR: 1.38; 95% CI, 0.97–1.96; P = 0.08; Fig S15, available at www.ophthalmologyscience.org) or unmasked (42% vs. 35%; HR: 1.39; 95% CI, 0.94–2.06; P = 0.10; Fig S16, available at www.ophthalmologyscience.org) UWF-FA DRSS.
Discussion
This exploratory analysis of Protocol AA data highlights important differences between UWF-color and UWF-FA in assessing DR severity and ability to predict disease progression over time. A summary of key findings and implications are provided in Table S8 (available at www.ophthalmologyscience.org). Ultra-widefield FA identified more severe DRSS grades in >50% of eyes compared with UWF-color, largely due to its ability to detect subtle DR and peripheral lesions that may not be visible on color imaging. This finding is consistent with prior studies with more limited data sets.9 Ultra-widefield FA images consistently identified more extensive disease in eyes with DR severity levels below ETDRS level 53 (severe NPDR) compared with UWF-color images. Furthermore, the agreement between DRSS levels on UWF-color and UWF-FA is moderate, indicating that these 2 imaging modalities are not interchangeable for DR severity grading. More severe baseline DRSS level on UWF-FA was also associated with more than twofold increased risk of disease worsening over 4 years, particularly for lesions such as H/Ma, IRMA, and NVE. This trend was consistent across baseline DRSS subgroups, supporting the hypothesis that UWF-FA provides better risk stratification for DR progression compared with UWF-color. The presence of more severe grades for individual lesion types on UWF-FA was also linked to a significantly greater risk of disease worsening.
The findings from this study are consistent with the idea that UWF-FA lesions better predict diabetic eye disease progression than UWF-color.6 Eyes with worse DRSS on UWF-FA compared with UWF-color images also had more nonperfusion than those with the same or better DRSS, suggesting that the DR lesions identified on UWF-FA may more accurately reflect the severity of retinal ischemia. The distribution of NPI was right-skewed across DRSS levels, reflecting that most eyes had relatively low levels of ischemic burden while a smaller subset demonstrated extensive areas of nonperfusion. Because NPI is bounded at zero, values of zero occurred in eyes without detectable nonperfusion within the gradable retinal area, contributing to the observed zero inflation. The wide ranges observed in box plots reflect variability among eyes with more advanced ischemia. Nonperfusion index analyses in this report were descriptive and intended to provide context regarding ischemic burden rather than to establish independent prognostic thresholds. These results may support the need for a more updated approach to grading DR with UWF-FA because the current color-based DRSS does not capture the additional prognostic value that these studies show UWF-FA provides. By incorporating additional UWF-FA findings into DR grading criteria, clinicians and researchers may better predict eyes at risk for disease progression and thus improve risk stratification and outcomes. If subsequent studies show UWF-FA can capture subtle and peripheral DR lesions and thereby drive earlier and more targeted interventions, the use of UWF-FA could improve long-term outcomes for patients with DR.
Ultra-widefield FA identifies areas of nonperfusion, vascular leakage, and subtle microvascular changes often missed in color imaging.7,8 Previous studies demonstrated that increasing nonperfusion is prevalent in more severe stages of DR and is associated with future disease worsening.6,12 The ability to identify the extent and severity of nonperfusion on UWF-FA may be a key to its more accurate assessment of risk of DR worsening, particularly in eyes with less severe DRSS on color imaging. The development of vision-threatening complications, such as proliferative DR, is closely associated with increasing nonperfusion, highlighting the potential prognostic value of UWF-FA in identifying DR severity and predicting disease worsening and potentially vision-threatening complications over time.7 A better understanding of how severity and location of retinal nonperfusion or ischemia are related to future progression of DR could impact the development of future treatment strategies.
Additional analysis was performed using an idealized area encompassing the nonwide 3-field FA, which is an important metric to be comparable with prior analysis in landmark studies such as Diabetes Control and Complications Trial and ETDRS. Over the past 35 years since these studies were performed, retinal FA image resolution has significantly improved, allowing for greater ability to capture subtle retinal abnormalities that may have been missed in the Diabetes Control and Complications Trial and ETDRS, supporting refinement of grading criteria to fully leverage the additional information provided by FA imaging.
We found similar rates of disease progression when applying the current grading scales based on the ETDRS criteria for color photography to UWF masked (ETDRS 7-field color), 3-field FA, and UWF-FA. In this AA cohort, unmasked UWF-FA identified more severe DR in up to 62% of eyes compared with color imaging. This difference was particularly pronounced in eyes with mild to moderately severe NPDR, where up to 80% of eyes had more severe disease. Identification of more severe disease on UWF-FA is clinically significant, given that it was associated with a 2.4-fold increased risk for progression over 4 years across all DR severity levels.
The longitudinal nature of Protocol AA allows for robust observation of DR progression, providing valuable insights into the relative performance of UWF-color and UWF-FA in assessing the risk of DR worsening. However, the widespread clinical use of UWF-FA remains constrained by its invasive nature, the need for contrast agents, and associated risks such as allergic reactions or adverse events. These factors limit feasibility of routine use in clinical practice. Ultra-widefield FA findings in this report are best viewed as providing risk enrichment and stratification for identifying a higher-risk subset and are not intended to replace established color-based DRSS endpoints. Future research should focus on the development of noninvasive imaging alternatives, including optical coherence tomography angiography or advanced computational algorithms that may provide findings similar to FA without dye injections.
Additionally, this study used DRSS change based on masked UWF-color imaging as the primary outcome measure and may have been affected by inherent differences in contrast and resolution of UWF-color and standard 7-field photography that were not addressed. The use of scanning laser imaging technology is influenced by laser wavelength and image processing algorithms that may affect lesion detection. These implications are important in clinical programs outside of reading center environments as image characteristics and display features may influence DR severity assessments.
These exploratory analyses suggest that UWF-FA is superior to UWF-color in detecting subtle and peripheral DR lesions, potentially leading to more accurate DR severity grading and risk stratification for disease progression. These findings highlight the need for updated DR grading criteria to fully leverage the prognostic value of UWF-FA.
Manuscript no. XOPS-D-25-00629.
Footnotes
Supplemental material available atwww.ophthalmologyscience.org.
The full list of group members in the DRCR Retina Network is available at www.ophthalmologyscience.org.
Disclosure(s):
The Article Publishing Charge (APC) for this article was paid by DRCR Retina Network.
All authors have completed and submitted the ICMJE disclosures form.
The author(s) have made the following disclosure(s):
P.S.S.: Grants – Boehringer Ingelheim GmbH, Optos plc; Honoraria – Optos; Receipt of equipment, materials, drugs, medical writing, gifts or other services – Optos plc; Other financial or nonfinancial interests – Optos plc.
D.L.: Grants – Breakthrough T1D supported a portion of the study, Optos provided some equipment for the study.
L.P.A.: Consultant – Ceramedix; Travel expenses – Optos, Novo Nordisk; Participation on a Data Safety Monitoring Board or Advisory Board – Novo Nordisk, DRCR Retina Network, Optos.
S.G.: Financial support – DRCR Retina Network, NEI, National Institute of Diabetes and Digestive and Kidney Diseases, Breakthrough T1D; Travel expenses – DRCR Retina Network.
D.F.M.: Consultant – Ocular Therapeutix.
M.M.: Grants – Breakthrough T1D supported a portion of the study, Optos provided some equipment for the study.
C.Y.W.: Financial support – Breakthrough T1D, Optos; Grants – AGTC, Alimera Sciences, DRCR Retina Network; Consultant – Allergan/AbbVie, Alcon, Apellis Pharmaceuticals, Alimera Sciences, Zeiss/DORC, Ocular Therapeutix, Genentech, Regeneron, REGENXBIO, Iveric Bio/Astellas, EyePoint, Frontera, BVI; Royalties or licenses – Springer Publishers; Leadership or fiduciary role in other board, society, committee or advocacy group, paid or unpaid – American Society of Retina Specialists, Women in Ophthalmology.
J.K.S.: Grants – Massachusetts Lions Eye Research Fund, Mary Tyler Moore Vision Initiative (through Mary Tyler Moore and S. Robert Levine, MD Charitable Foundation), Boehringer Ingelheim, Roche, Breakthrough T1D; Travel expenses – Alcon, Boehringer Ingelheim, Novo Nordisk; Receipt of equipment, materials, drugs, medical writing, gifts or other services – Adaptive Sensory Technology, Optovue, Konan, LKC Technologies, Boehringer Ingelheim, Genentech; Other financial or nonfinancial interests – American Diabetes Association, American Medical Association.
Research reported in this publication was supported by the National Eye Institute and the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under Award Number UG1EY014231. The DRCR Retina Network had complete control over the design of the protocol, ownership of the data, all editorial content of presentations and publications related to the protocol, and the decision to submit for publication. The funding organization (National Institutes of Health) participated in oversight of the conduct of the study and review of the manuscript but not directly in the design or conduct of the study, nor in the collection, management, analysis, or interpretation of the data, or the preparation of the manuscript. Additional grants to the Jaeb Center for Health Research from Breakthrough T1D supported a portion of the study. Optos provided some equipment for the study.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The DRCR Retina Network had complete control over the design of the protocol, ownership of the data, all editorial content of presentations and publications related to the protocol, and the decision to submit for publication.
HUMAN SUBJECTS: Human subjects were included in this study. Study participants provided written informed consent. The study adhered to the tenets of the Declaration of Helsinki and was approved by multiple ethics boards (Jaeb Center for Health Research Institutional Review Board, Henry Ford Health System Institutional Review Board, Medical College of Wisconsin & Froedtert Hospital Institutional Review Board, Joslin Diabetes Center Committee on Human Studies, Rush University Institutional Review Board, The Johns Hopkins School of Medicine Institutional Review Board, The University of Oklahoma Institutional Review Board, Institutional Review Board of the University of Pennsylvania, The University of North Carolina at Chapel Hill Institutional Review Board, Western Institutional Review Board, University of Illinois at Chicago Institutional Review Board, University at Buffalo Health Sciences Institutional Review Board, Weill Cornell Medical College Institutional Review Board, University Health Network Research Ethics Board, University of British Columbia Clinical Research Ethics Board, and Nova Scotia Health Authority Research Ethics Board). The protocol is available on the study website.
No animal subjects were used in this study.
Author Contributions:
Conception and design: Silva, Liu, Aiello, Grigorian, Grover, Jampol, Martin, Melia, Weng, Sun
Analysis and interpretation: Silva, Liu, Aiello, Grigorian, Grover, Jampol, Martin, Melia, Weng, Sun
Data collection: Silva, Liu, Aiello, Grigorian, Grover, Jampol, Melia, Weng, Sun
Obtained funding: Jampol, Martin, Sun
Overall responsibility: Silva, Liu, Aiello, Grigorian, Grover, Jampol, Martin, Melia, Sun
Contributor Information
Danni Liu, Email: drcrstat2@jaeb.org.
DRCR Retina Network:
Abla M. Harara, Angela N. Palacios, Brian B. Berger, Boris Corak, Bianca Luong, Chirag D. Jhaveri, Daniela Mariel Wilson, Gowtham Jonna, Ivana Gunderson, Kimberly Hosein, Ryan M. Reid, Saradha Chexal, Tori Moore, Tina A. Seidu, Valerie Gatavaski, Yong Ren, Bradley A. Stern, Celia E. Benvenutti, Dinah S. Oude-Reimerink, Jenny Shaheen, John Grybas, Julianne Vitale-Kuhn, Jessica L. Staffne, Katie M. Ventimiglia, Megan Allis, Mary K. Monk, Marc E. Thomas, Nicole M. Massu, Paul Andrew Edwards, Tracy A. Troszak, Amber N. Irons, Brittany Rego, Dennis P. Han, Eleanor Dorsey, Erika Nelson, Hannah Sheppard, Joseph R. Beringer, Judy E. Kim, Kristy L. Keller, Krissa L. Packard, Marriner L. Altmann, Mara Goldberg, Nickolas Chen, Pat A. Winter, Shay Bourgeois, Samantha Jacobo, Stephanie J. Moebius, Thomas B. Connor, Jr., Vicki Barwick, Vesper V. Williams, William J. Wirostko, A. Thomas Ghuman, Anita H. Leslie, Ashish G. Sharma, Cheryl Kiesel, Danielle Dyshanowitz, Eileen Knips, Glenn Wing, Joseph P. Walker, Paul A. Raskauskas, Raymond K. Kiesel, Deborah K. Schlossman, Elizabeth S. Weimann, George S. Sharuk, Hanna Kwak, Jerry D. Cavallerano, Jae W. Rhee, Konstantina Sampani, Katie V. Tran, Leila Bestourous, Linette Miranda, Michael N. Krigman, Margaret E. Stockman, Paul G. Arrigg, Robert W. Cavicchi, Rita K. Kirby, Shireen Glynn, Steve L. Papaconstantinou, Sabera T. Shah, Timothy J. Murtha, William Carli, Autumn K. Finch, Angella K. Gentile, Angela K. Price, Brittany A. Murphy, Beverly O. Rowland, Christina J. Fleming, Courtney Mahr, Carol A. Shore, David Browning, Donna McClain, Erica Breglio, Gina M. Lester, Jenna T. Herby, Kayla A. Bratcher, Loraine M. Clark, Lisa A. Jackson, Lynn Watson, Michael D. McOwen, Omar S. Punjabi, Swann J. Bojaj, Sarah A. Ennis, Sherry L. Fredenberg, Taylor S. Jones, Teneisha A. Ragin, Uma M. Balasubramaniam, Blanca Ornelas, Brenda Rodriquez, Carla Edwards, Danielle R. Carns, Eileen E. Tonner, Kisung Woo, Len Richine, Mathew W. MacCumber, Pauline Townsend Merrill, Sarah Kociborski, Ashley M. Harless, Charlotte Harris, Lorraine White, Raj K. Maturi, Julie Asher, Justin Walsh, Jeff Wheeler, Katie Milstead, Kristina Oliver, Lisa Lovelady, Nicholas G. Anderson, Patricia Coppola, Raul E. Lince, R. Keith Shuler, Jr., Steve Morris, Sarah M. Oelrich, Brandon S. Gardner, Bob Moore, Dennis Cain, Deborah Donohue, David Emmert, Kemi Adeyemo, Lisa K. Levin, Mary Frey, Nick Rhoton, Susan Bressler, Sharon D. Solomon, Amy L. Ford, Ashley Hughes, Alisha N. Brewer, JoAnn T. Booth, Keven W. Lunsford, Lauren D. Ukleya, Russ Burris, Ronald M. Kingsley, Shannon R. Almeida, Sonny Icks, Vinay A. Shah, Vanessa A. Bergman, Alessandro A. Castellarin, Aimee H. Shook, Aimee Walker, Dante J. Pieramici, Gina Hong, Kelly Avery, Kate M. McKee, Matthew Giust, Marco A. Munoz, Sarah Fishbein, Alecia B. Camp, Carl W. Baker, Jil D. Baker, Kylie S. Sedberry, Lynnette F. Lambert, Margaret J. Orr, Sonya L. Alcaraz, Samantha Kettler, Tracey M. Caldwell, Abigail Miller, Christine M. Dorr, G. Robert Hampton, Jamin S. Brown, Jeffrey P. Barker, Kevin I. Rosenberg, Lynn M. Kwasniewski, Laurie J. Sienkiewycz, Lisa Spuches, Michelle L. Manley, Nicole E. Robarge, Stefanie R. DeSantis, Teresa M. DeForge, Alexander J. Brucker, Benjamin J. Kim, Jim M. Berger, Joan C. DuPont, Sheri Drossner, Sara Freeman, Ashley Studebaker, John F. Payne, John A. Wells, III, Robbin Spivey, Tiffany N. Ogbuewu, Tiffany R. Swinford, Adrienne Guillory, Amy Hutson, Amy C. Schefler, Ankoor R. Shah, Belinda A. Almanza, Brenda Dives, Beau A. Richter, Cary A. Stoever, David M. Brown, Danee Foerster, David Garcia, Diana Rodriguez, Daniel Park, Eric Chen, Eric N. Kegley, Elizabeth Quellar, Garret L. Twining, Heather Koger-Grifaldo, Ilsa Ortega, Jolene Carranza, James C. Major, Jr., Kimberly Williamson, Lindsay Burt, Luis R. Salinas, Lisa M. Wolff, Matthew S. Benz, Maura A. Estes, Miranda F. James, Meredith Berry, Melina Vela, Nubia Landaverde, Nina A. Webb, Richard H. Fish, Rosa Y. Kim, Rebecca Yee, Sadia Y. Karani, Stacy M. Supapo, Tamara L. Dodel, Tyneisha McCoy, Tien P. Wong, Veronica A. Sneed, Cassandra J. Barnhart, Debra Cantrell, Elizabeth L. DuBose, Houston P. Sharpe, Jan Niklas Ulrich, Kanika A. Bhansali, Rona Lyn Esquejo, Seema Garg, Sean Grout, Allen McKinney, Brenda J. Bobbitt, Ceara L. Wendel, Damanda F. Fagan, Jacqueline Andrews, Krystal Nikki Holmes, Karen L. Seyez, Kimberly A. Williamson, Nader Moinfar, Paige N. Walters, Steve Carlton, Shannon M. Rehling, Shana E. Williams, Tiara L. Reed, Amber R. VandeVelde, Frank T. Yeager, Gregory M. Fox, Ivan R. Batlle, Kiersten Bruce, Katherine Pippin, Lexie R. Ainley, Ravi S.J. Singh, Ashley M. Adamo, Adrian Guardado, Apurva K. Patel, Brian S. Puckett, Christine Hoerner, Colin Ma, David J. Clark, Inessa M. Flato, Joshua Cohen, Margaret E. Charpentier, Marcia Kopfer, Mark A. Peters, Pualani Smith, Paul S. Tlucek, Stephen Hobbs, Stephanie L. Ho, Ashley M. Metzger, Alesia K. McCalla, Amy Thompson, Christine Ringrose, Dallas R. Sandler, Henry A. Leder, Jennifer L. Belz, JoAnn Starr, Jennifer L. Simmons, Peggy R. Orr, Peter Sotirakos, Pamela V. Singletary, Terri Cain, Teresa Coffey, Tiffany M. Carter, Twyla J. Robinson, Chirag P. Shah, Dominique Cammarata, Jennifer L. Kruger, Lindsey Colegrove, Margie Graham, Shane T. Gleason, Bryan Noel, Catherine Damron, Diana M. Holcomb, Edward A. Slade, Jeanne Van Arsdall, Lisa Bicknell, Michelle Buck, Thomas W. Stone, Amina Farooq, Brook Parsons, Harinderjit Singh, Ken Ivey, Lindsay Allison Foster, Michele Woodward, Siobhan O. Ortiz, Thomas Bailey, Bharani Krishna Mynampati A, Cheryl L. White, Ghulam Shabbir Hamdani, Jazzmin N. Smith, Kakarla V. Chalam, Kumar Sambhav, Romesh Babaria, Sandeep Grover, Catherine Carroll, Felix Y. Chau, Jennifer I. Lim, Lauren A. Talasnik, Mark Janowicz, Natasa Stankovic, Sarah L. Berlatsky, Marcia Niec, Jie Sun, Tametha Johnson, Yesenia Ovando, Brenda Nakoski, Calvin E. Mein, Christopher Sean Wienecke, Elaine Castillo, Jaynee Baker, Jonathan San Roman, Lydia Adams, Lita Kirschbaum, Moises A. Chica, Sara L. Cloudt, Tori R. Moore, Felix N. Sabates, Sr., Gary S. Gallimore, Yin C. Chen, Adrienne C. Swann, Deborah M. Cadwell, Kenneth R. Diddie, Taryn F. Boisvert, Carrie D. Tessau, Jack Bowers, Jared S. Nielsen, Jay Rostvold, Jamie Spillman, Kyle J. Alliman, Lisa M. Boender, Marilyn A. Johnson, Marianne Parker, Paula L. Bix, Spencer D. Ridgway, Tami Jo Woehl, Whitney Stonewall, Christopher M. Brown, Gareth MC. Lema, Luann Wiechelt, Pradeepa Yoganathan, Sandra L. Boglione, Chris A. Montesclaros, Cory Mangham, Gopal Karsaliya, Phillip V. Le, Robert W. Wong, Anne Marie Godfrey, Aleksandra Kuzmanovic, Andrew William Kirker, Bryan Harrison, Farzin Forooghian, Garnet Louise, Elvena, Laura J. Hall, Bilgin Turhal, Ian Brown, Isaac A. Kotei, Lina Chen, Michael Henry Brent, Michelle Moon, Olivera Sutakovic, Angela Chang, Anne-Marie Godfrey, Andrew William Kirker, Aleksandra Kuzmanovic, Bryan Harrison, David Albiani, David A.L. Maberley, Eduardo Vitor Navajas, Kelly Grant, Khoi A. Tran, Laura J. Hall, Mira Jovanovic, Sijia Cao, Theresa Wiens, Andrzej Kozbial, Anton Orlin, Courtney Nichole Lenane, Susan P. Herder, Szilard Kiss, Tom Reeves, Alan F. Cruess, Andrea Dean, Ann Hoskin-Mott, Christine Morrison, Meggie D. Caldwell, Mitzi Hynes, R. RishiGupta, Stacey Durling, Trina MacDonnell, Alyssa Baptista, Lauren Brandt, Claire T. Calhoun, Sharon R. Constantine, Brian B. Dale, Dara Dombrowski, Simone S. Dupre, Crystal A. Franklin, Sandra Galusic, Adam R. Glassman, Meagan Huggins, Brenda L. Hunter, Paula A. Johnson, Brittany Kelly, Danni Liu, Britney Meadows, James Miller, Carin M. Preston, Cynthia R. Stockdale, Wei Tian, Carmen M. Valenzuela-Silva, Thu P. Vu, Daniel F. Martin, Lee M. Jampol, Frederick L. Ferris, III, Raj K. Maturi, Emily Y. Chew, Sharon D. Solomon, Darrell Baskin, Mathew W. MacCumber, Negin Atri, Glenn J. Jaffe, Barbara A. Blodi, Andrew J. Barkmeier, Christina Y. Weng, Mohamed A. Elmasry, Leanne T. Labriola, Thomas W. Gardner, and Roy W. Beck
Supplementary Data
References
- 1.Teo Z.L., Tham Y.-C., Yu M., et al. Global prevalence of diabetic retinopathy and projection of burden through 2045: systematic review and meta-analysis. Ophthalmology. 2021;128:1580–1591. doi: 10.1016/j.ophtha.2021.04.027. [DOI] [PubMed] [Google Scholar]
- 2.Solomon S.D., Chew E., Duh E.J., et al. Diabetic retinopathy: a position statement by the American Diabetes Association. Diabetes Care. 2017;40:412–418. doi: 10.2337/dc16-2641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Early Treatment Diabetic Retinopathy Study Research Group Fundus photographic risk factors for progression of diabetic retinopathy. ETDRS report number 12. Ophthalmology. 1991;98:823–833. [PubMed] [Google Scholar]
- 4.Early Treatment Diabetic Retinopathy Study Research Group Grading diabetic retinopathy from stereoscopic color fundus photographs - an extension of the modified Airlie House classification. ETDRS report number10. Ophthalmology. 1991;98:786–806. [PubMed] [Google Scholar]
- 5.Ashraf M., Shokrollahi S., Salongcay R.P., et al. Diabetic retinopathy and ultrawide field imaging. Semin Ophthalmol. 2020;35:56–65. doi: 10.1080/08820538.2020.1729818. [DOI] [PubMed] [Google Scholar]
- 6.Silva P.S., Marcus D.M., Liu D., et al. Association of ultra-widefield fluorescein angiography–identified retinal nonperfusion and the risk of diabetic retinopathy worsening over time. JAMA Ophthalmol. 2022;140:936–945. doi: 10.1001/jamaophthalmol.2022.3130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Silva P.S., Liu D., Aiello L.P., et al. Diabetic retinopathy lesion types and distribution on ultrawide field imaging and the risk of disease worsening over time. Retina. 2022;10:1097. doi: 10.1097/IAE.0000000000004263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ehlers J.P., Josic K., Yordi S., et al. Assessment of baseline Ultrawidefield fluorescein angiographic quantitative leakage parameters with ultrawidefield fundus features and clinical parameters in diabetic retinopathy in protocol AA. Ophthalmol Retina. 2024;9:243–252. doi: 10.1016/j.oret.2024.08.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ashraf M., AbdelAl O., Shokrollahi S., et al. Evaluation of diabetic retinopathy severity on ultrawide field colour images compared with ultrawide fluorescein angiograms. Br J Ophthalmol. 2023;107:534–539. doi: 10.1136/bjo-2022-322163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Early Treatment Diabetic Retinopathy Study Research Group Fluorescein angiographic risk factors for progression of diabetic retinopathy. ETDRS report number 13. Early Treatment Diabetic Retinopathy Study Research Group. Ophthalmology. 1991;98:834–840. [PubMed] [Google Scholar]
- 11.Trial C. Color photography vs fluorescein angiography in the detection of diabetic retinopathy in the diabetes control and complications trial. Arch Ophthalmol. 1987;105:1344–1351. doi: 10.1001/archopht.1987.01060100046022. [DOI] [PubMed] [Google Scholar]
- 12.Marcus D.M., Silva P.S., Liu D., et al. Association of predominantly peripheral lesions on ultra-widefield imaging and the risk of diabetic retinopathy worsening over time. JAMA Ophthalmol. 2022;140:946–954. doi: 10.1001/jamaophthalmol.2022.3131. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.