Optical Coherence Tomography Features for Identifying Posttreatment Complete Polypoidal Regression in Polypoidal Choroidal Vasculopathy : The Asia-Pacific Journal of Ophthalmology

Secondary Logo

Journal Logo

Original Study

Optical Coherence Tomography Features for Identifying Posttreatment Complete Polypoidal Regression in Polypoidal Choroidal Vasculopathy

Chaikitmongkol, Voraporn MD*; Chaovisitsaree, Thanaphat MD; Patikulsila, Direk MD*; Kunavisarut, Paradee MD*; Phasukkijwatana, Nopasak MD, PhD; Watanachai, Nawat MD*; Choovuthayakorn, Janejit MD, PhD*; Isipradit, Sirawit MD; Boonyot, Pawinee MD§; Sangkaew, Apisara BBA; Ingviya, Thammasin MD, PhD; Bressler, Susan B. MD; Bressler, Neil M. MD

Author Information
Asia-Pacific Journal of Ophthalmology: September/October 2022 - Volume 11 - Issue 5 - p 408-416
doi: 10.1097/APO.0000000000000551

Abstract

Polypoidal choroidal vasculopathy (PCV) is characterized by abnormal branching choroidal vascular network with terminal aneurysmal lesion(s), termed polypoidal lesions.1 PCV has specific phenotypic patterns of choroidal neovascularization on fundus examination, structural optical coherence tomography (OCT), OCT angiography (OCTA), and indocyanine green angiography (ICGA) that may or may not have other characteristics of typical age-related macular degeneration (AMD), for example, drusen.2–4 There are 2 types of clinical presentations of macular PCV including the leakage type [serous maculopathy with fibrovascular pigment epithelial detachment (PED), with or without lipid exudates] and the hemorrhagic type (hemorrhagic maculopathy due to ruptured polypoidal lesion, sometimes with extensive hemorrhage beneath the retina or extending into the vitreous cavity, with varying degree of vision loss).4 Studies suggest that PCV has a higher incidence among aging Asian and African descendants compared with White individuals.5–8 Current treatments for PCV, proven beneficial in relatively large randomized clinical trials, include a combination of intravitreous ranibizumab injection and verteporfin photodynamic therapy (PDT),9 or intravitreous aflibercept monotherapy with rare rescue of PDT.10 Both treatment strategies showed that many PCV eyes had substantial visual improvement from baseline and often demonstrated complete polypoidal regression after receiving treatments.9,10

Complete polypoidal regression characterized by complete disappearance of hypercyanescent polypoidal lesions that were noted at baseline on ICGA can be seen after antivascular endothelial growth factor (anti-VEGF) therapy as well as PDT. Incomplete polypoidal regression has been reported as a factor associated with higher recurrent rates and long-term massive submacular hemorrhage in PCV.11,12 Therefore, in addition to visual acuity outcomes, complete polypoidal regression has been considered as another important treatment outcome in PCV studies and trials.9,10,13–16

At present, ICGA is considered the gold standard for PCV evaluations and diagnosis.17 Two well-accepted PCV diagnostic criteria including the criteria proposed by the Japanese Study Group of PCV18 and the criteria proposed by the EVEREST study2 have included ICGA as an essential part of diagnosing PCV. Subsequently, ICGA has been used to determine the presence or absence of complete polypoidal regression after treatments in PCV trials.9,10,13 However, ICGA is an invasive imaging technique not readily available in many ophthalmology clinics. For example, in Thailand, a nonvalidated survey among retina specialists found that 66% (50 of 76) of retina centers across Thailand did not have ICGA.19 In contrast, structural OCT is a noninvasive and relatively more commonly performed imaging tool available in most centers and can be performed fairly conveniently at each follow-up visit. On OCT, pretreatment polypoidal lesions typically characterize as nonserous PEDs with sharply-peaked border and/or notches and/or hyperreflective rings inside the PEDs.20 Regarding the OCT characteristics of posttreatment polypoidal lesions, there is very limited information in the literature, including a case series from Thailand describing posttreatment OCT features of complete and incomplete polypoidal regressions.21 Based on that previous report, an internal reflectivity of posttreatment PED on OCT seems to help in differentiating complete and incomplete regression. Complete polypoidal regression, confirmed on ICGA, was observed as a PED with internal homogeneous hyperreflectivity, in contrast to incomplete regression that was observed as a PED with internal heterogeneous reflectivity on OCT.21

The current study aimed to evaluate the accuracy of these OCT features for identifying complete or incomplete polypoidal regression in a larger sample size of eyes with PCV.

METHODS

This study was a retrospective imaging review approved by the Research Ethics Committee Faculty of Medicine, Chiang Mai University, and adhered to the Declaration of Helsinki. Patients with a diagnosis of PCV who presented at Chiang Mai University Hospital between August 2013 and February 2018 were identified through patient records. Inclusion criteria were subjects who had (1) a diagnosis of treatment-naive PCV in 1 or both eyes based on EVEREST criteria2 (if both eyes were eligible, the right eye was the study eye); (2) performed OCT (at least 49 B scans/macular cube; Spectralis, Heidelberg Engineering, Heidelberg, Germany) and ICGA (Spectralis Heidelberg retina angiograph/HRA; Heidelberg Engineering) at baseline and after treatments (at least once between month 3 and 12 after baseline); (3) received anti-VEGF and/or PDT treatments (any therapeutic regimen) for PCV at baseline; (4) an eligible “study polypoidal lesion” which was an isolated lesion with “well-defined boundaries” on both OCT and ICGA. Exclusion criteria were subjects who had (1) coexisting retinal diseases, for example, diabetic retinopathy, or retinal vein occlusions; (2) poor quality images; (3) polypoidal lesions with unidentified or poor-defined borders on OCT, ICGA, or both, due to subretinal or sub–retinal pigment epithelium (sub-RPE) hemorrhage or other causes. Note that this study included only polypoidal lesions with “well-defined boundaries” on “both OCT and ICGA” to be certain that graders could evaluate characteristics of the whole polypoidal lesions. The location of study polypoidal lesions on OCT and their corresponding lesions on ICGA were marked and labeled by an investigator before sending to 2 groups of standardized retina specialist graders.

Grader group A (D.P., V.C.): independently reviewed ICGA to confirm the PCV diagnosis according to the EVEREST criteria,2 and to determine on posttreatment ICGA whether each study polypoidal lesion had complete or incomplete regression by comparing pretreatment and posttreatment ICGA side by side. Before grading started, graders performed grading standardization with case examples. For the EVEREST criteria, the PCV diagnosis was made when there was a nodular hypercyanescent lesion, seen as a whitish round lesion, on ICGA within the first 6 minutes, along with at least 1 of the following characteristics: subretinal orange nodule on fundus examination corresponding to the nodule seen on ICGA, massive subretinal hemorrhage, hypocyanescent halo around the nodule, nodular appearance on stereoscopic viewing, pulsatile filling of the nodule, or abnormal vasculature supplying the nodule.2 Complete polypoidal regression was defined as complete disappearance of a hypercyanescent nodular lesion that existed at baseline on a posttreatment ICGA. Incomplete polypoidal regression was defined as an incomplete disappearance of a hypercyanescent lesion that previously existed at baseline, with either a decrease in the area of hypercyanescence (partial improvement) or an increase in the area of hypercyanescence (worsening). Any disagreement was managed by open adjudications.

Grader group B (P.K., N.P.): independently reviewed OCT features of each study polypoidal lesion. Note that the pretreatment polypoidal lesion is typically characterized as a PED on OCT. Then, the graders determined if the lesion should be categorized in 1 of the following 5 prespecified OCT features based on the internal reflectivity and RPE lining of the PED (Fig. 1): “A,” no PED; “B,” PED with internal homogeneous hyperreflectivity with “predominant BUN sign” (defined as blended RPE with underlying structure or unidentifiable RPE lining in ≥2/3 of PED border); “C,” PED with internal homogeneous hyperreflectivity with “minimal BUN” sign (<2/3 of PED border); “D,” PED with internal heterogeneous reflectivity; and “E,” PED with internal homogeneous hyporeflectivity. Note that internal homogeneous hyperreflectivity was defined as a uniform whitish appearance inside a PED, and internal heterogeneous reflectivity was defined as a combination of both whitish appearance and blackish space inside a PED. Regarding the BUN sign, given that normal RPE is observed as a hyperreflective (whitish) band at the border of a PED on OCT, this RPE hyperreflective band is no longer observed, or unidentifiable, at the border of the PED when the BUN sign presents.

F1
FIGURE 1:
Five prespecified optical coherence tomography features of posttreatment polypoidal lesions evaluated in this study including feature A—no pigment epithelial detachment (PED) (A); feature B—PED with internal hyperreflectivity (white arrow) with predominant “BUN” sign* (black arrow) (B); feature C—PED with internal hyperreflectivity with minimal “BUN” sign (black arrow) (C); feature D—PED with internal heterogeneous reflectivity (white arrow) (D); and feature E—PED with internal hyporeflectivity (E). *Predominant BUN sign defined as the presence of “blended retinal pigment epithelium (RPE) with underlying structure” or unidentifiable RPE lining in at least 2/3 of PED border.

The OCT graders were standardized by practicing the following 3-step grading process (Fig. 2) with a set of example lesions. Step 1, determine if there is a presence of PED. If no, grade as feature “A”; if yes, proceed to step 2. Step 2, determine the internal reflectivity of PED in each scan. If at least 1 scan shows heterogeneous reflectivity, grade as feature “D.” If all scans show homogeneous hyporeflectivity, grade as feature “E.” If all scans show homogeneous hyperreflectivity, carefully determine the presence of “BUN” sign at PED border. If “BUN” sign presented in ≥2/3 of PED border, grade as feature “B”; if not, grade as feature “C.” Disagreements between graders were managed by open adjudications. Pretreatment and posttreatment OCT images were listed in a random order to mask graders from knowing treatment status when evaluating images.

F2
FIGURE 2:
A flow chart demonstrating 3-step grading process that graders used to determine optical coherence tomography features of posttreatment study polypoidal lesions. BUN signs defined as “blended retinal pigment epithelium (RPE) with underlying structure” or unidentifiable RPE lining. PED indicates pigment epithelial detachment.

Results of OCT features of each study lesion were compared with polypoidal status confirmed on ICGA. The main outcome measure was the accuracy of OCT features for identifying complete or incomplete polypoidal regression after treatments. Secondary outcome measures were diagnostic parameters including sensitivity, specificity, positive predictive value, negative predictive value, and relative risk (RR) of OCT features for identifying polypoidal regression with 95% confidence intervals (CI) of the diagnostic parameters calculated based on binomial probabilities.

To account for dependency between lesion from the same eye and same patient, the multiple logistic regression with generalized estimating equation (GEE) was performed. The identification number of patients was used as clustering variable. The correlation structure used was unstructured which yielded the lowest quasi-likelihood information criterion compared to autoregressive and exchangeable correlation structure. For sensitivity analysis and account for small sample size, the RR of each feature was calculated using COPY method as described by Deddens and Petersen.22 To perform the COPY method, a GEE with RR model was fitted to 10,000 copies of our PCV data and 1 copy with RR set to 1−RR. The RRs were then calculated using the modified data.

RESULTS

There were 130 study polypoidal lesions (65 pretreatment and 65 posttreatment lesions) of 39 study eyes (39 subjects; 54% female; mean age±SD: 64.6±8.2 years). These 39 treatment-naive subjects who had complete sets of imaging were identified among 281 subjects who were diagnosed with PCV during the study period. Two hundred forty-two subjects were excluded due to the following reasons: received treatment before baseline ICGA (23%; 56 subjects); number of OCT scans were <49 scans/macular cube (17%; 41 subjects); ICGA not available at both baseline visit and at least 1 follow-up visit (32%; 77 subjects); poor image quality of OCT and/or ICGA images (9%; 22 subjects); and absence of polypoidal lesions with well-defined boundary on OCT or ICGA, or both, due to subretinal or sub-RPE hemorrhage (19%; 46 subjects).

Demographic data are shown in Supplementary Digital Content Table 1, https://links.lww.com/APJO/A159. Of 39 study eyes, 18 (46%) had 1 study polypoidal lesion, 16 (41%) had 2 study polypoidal lesions, and 5 (13%) had 3 study polypoidal lesions per study eye. Thirty-five eyes (90%) received anti-VEGF monotherapy, 2 (5%) received combined anti-VEGF and PDT, and 2 (5%) received PDT monotherapy. Among 35 eyes received anti-VEGF monotherapy, 91%, 6%, and 3% received aflibercept, bevacizumab, and ranibizumab, respectively; the median injection number was 3 injections (interquartile range: 3–3.5 injections). An interval between pretreatment and posttreatment assessments ranged from 3 to 10 months, with the median interval of 4 months (interquartile range: 4–4 months).

Of 65 pretreatment polypoidal lesions, 100% showed early hypercyanescent nodular lesions on ICGA, and 100% showed a PED with internal heterogeneous reflectivity (feature D) on OCT. Of 65 posttreatment polypoidal lesions, ICGA showed complete polypoidal regression in 31 lesions (48%) and partial regression in 34 lesions (52%), while OCT showed no PED in only 10 lesions (15%) with PED presented in 55 lesions (85%). In terms of fluid status, 31 of 31 (100%) posttreatment lesions with complete polypoidal regression on ICGA showed no intraretinal or subretinal fluid on OCT. Regarding the 5 prespecified OCT features of posttreatment lesions evaluated, only 4 features were found among 65 study lesions: 10 (15%) had feature A, 16 (25%) had feature B, 9 (14%) had feature C, 30 (46%) had feature D, and none (0%) had feature E.

Comparing posttreatment polypoidal status on ICGA with OCT features, of 31 lesions with complete polypoidal regression, OCT features A, B, C, D, and E were found in 10 (32%), 14 (45%), 4 (13%), 3 (10%), and 0 (0%), respectively. Among 34 lesions with incomplete regression on ICGA, OCT features A, B, C, D, and E were found in 0 (0%), 2 (6%), 5 (15%), 27 (79%), and 0 (0%), respectively (Table 1, Fig. 3).

TABLE 1 - Pretreatment and Posttreatment Features of Polypoidal Lesions on OCT and ICGA
Pretreatment and Posttreatment OCT Features Posttreatment ICGA Features
OCT Features of Polypoidal Lesions Pretreatment 65 Lesions, n (%) Posttreatment 65 Lesions, n (%) Complete Polypoidal Regression 31 Lesions, n (%) Incomplete Polypoidal Regression 34 Lesions, n (%)
Feature A 10 (15) 10 (32)
Feature B 16 (25) 14 (45) 2 (6)
Feature C 9 (14) 4 (13) 5 (15)
Feature D 65 (100) 30 (46) 3 (10) 27 (79)
Feature E
Feature A defined as complete PED disappearance; feature B defined as PED with intralesional homogeneous hyperreflectivity and poor-defined RPE lining; feature C defined as PED with intralesional homogeneous hyperreflectivity and well-defined RPE lining; feature D defined as PED with intralesional heterogeneous reflectivity (hyporeflectivity-hyperreflectivity); and feature E defined as PED with intralesional homogeneous hyporeflectivity.
ICGA indicates indocyanine green angiography; OCT, optical coherence tomography; PED, pigment epithelial detachment; RPE, retinal pigment epithelium.

F3
FIGURE 3:
A bar chart demonstrating proportion of each posttreament optical coherence tomography features showing complete or incomplete polypoidal regression on indocyanine green angiography.

Accuracy of OCT Features in Identifying Polypoidal Regression

Predictive accuracy of each prespecified OCT feature in identifying complete and incomplete polypoidal lesions are shown in Table 2. No eyes showed posttreatment feature E; therefore, the analysis was not applicable for this feature.

TABLE 2 - Diagnostic Accuracy of Each OCT Features in Identifying Complete and Incomplete Polypoidal Regression
OCT Features Accuracy Sensitivity Specificity PPV NPV
Diagnostic accuracy in identifying complete polypoidal regression (95% CI)
 Single feature
  Feature A 0.68 (0.55–0.79) 0.32 (0.17–0.51) 1 (0.9–1) 1 (0.69–1) 0.62 (0.48–0.75)
  Feature B 0.71 (0.58–0.81) 0.45 (0.27–0.64) 0.94 (0.8–0.99) 0.88 (0.62–0.98) 0.65 (0.5–0.78)
  Feature C 0.51 (0.38–0.63) 0.13 (0.04–0.3) 0.85 (0.69–0.95) 0.44 (0.14–0.79) 0.52 (0.38–0.65)
  Feature D 0.15 (0.08–0.26) 0.1 (0.02–0.26) 0.21 (0.09–0.38) 0.1 (0.02–0.27) 0.20 (0.08–0.37)
  Feature E NA NA NA NA NA
 Multiple features
  Feature A or B 0.86 (0.75–0.93) 0.77 (0.59–0.9) 0.94 (0.8–0.99) 0.92 (0.75–0.99) 0.82 (0.66–0.92)
  Feature C or D 0.14 (0.07–0.25) 0.23 (0.1–0.41) 0.06 (0.01–0.2) 0.18 (0.08–0.34) 0.08 (0.01–0.25)
OCT features Diagnostic accuracy in identifying incomplete polypoidal regression (95% CI)
 Single feature
  Feature A 0.32 (0.21–0.45) 0 (0–0.1) 0.68 (0.49–0.83) 0 (0–0.31) 0.38 (0.25–0.52)
  Feature B 0.29 (0.19–0.42) 0.06 (0.01–0.2) 0.55 (0.36–0.73) 0.12 (0.02–0.38) 0.35 (0.22–0.5)
  Feature C 0.49 (0.37–0.62) 0.15 (0.05–0.31) 0.87 (0.7–0.96) 0.56 (0.21–0.86) 0.48 (0.35–0.62)
  Feature D 0.85 (0.74–0.92) 0.79 (0.62–0.91) 0.9 (0.74–0.98) 0.9 (0.73–0.98) 0.8 (0.63–0.92)
  Feature E NA NA NA NA NA
 Multiple features
  Feature A or B 0.14 (0.07–0.25) 0.06 (0.01–0.2) 0.23 (0.1–0.41) 0.08 (0.01–0.25) 0.18 (0.08–0.34)
  Feature C or D 0.86 (0.75–0.93) 0.94 (0.8–0.99) 0.77 (0.59–0.9) 0.82 (0.66–0.92) 0.92 (0.75–0.99)
Feature A defined as complete PED disappearance; feature B defined as PED with intralesional homogeneous hyperreflectivity and poor-defined RPE lining; feature C defined as PED with intralesional homogeneous hyperreflectivity and well-defined RPE lining; feature D defined as PED with intralesional heterogeneous reflectivity (hypo-hyperreflectivity); and feature E defined as PED with intralesional homogeneous hyporeflectivity.
CI indicates confidence interval; NPV, negative predictive value; OCT, optical coherence tomography; PED, pigment epithelial detachment; PPV, positive predictive value; RPE, retinal pigment epithelium.

For identifying complete polypoidal regression, an analysis of each OCT feature showed that feature B (71%; 95% CI: 58%–81%) and feature A (68%; 95% CI: 55%–79%) had relatively high accuracy for identifying complete regression; feature C had moderate accuracy (51%; 95% CI: 38%–63%); and feature D had low accuracy (15%; 95% CI: 8%–26%). Given data from Table 1 that a majority of complete polypoidal regression lesions showed either feature A or B on OCT (77%, 24 of 31 lesions), and very few (6%, 2 of 34 lesions) lesions with these 2 features showed incomplete polypoidal regression, multiple features analysis was performed to evaluate the accuracy of “either feature A or B” for identifying complete regression. Results demonstrated that the presence of “either feature A or B” on OCT provided 86% accuracy (95% CI: 75%–93%), 77% sensitivity (95% CI: 59%–90%), 94% specificity (95% CI: 80%–99%), 92% positive predictive value (PPV) (95% CI: 75%–99%), and 82% negative predictive value (NPV) (95% CI: 66%–92%) for identifying complete polypoidal regression.

For identifying incomplete polypoidal regression, results showed that feature D had the highest accuracy (85%; 95% CI: 74%–92%) with 79% sensitivity (95% CI: 62%–91%), 90% specificity (95% CI: 74%–98%), 90% PPV (95% CI: 73%–98%), and 80% NPV (95% CI: 63%–92%). Multiple features analysis of “either feature C or D” did not appear to provide substantially higher accuracy (86%; 95% CI: 75%–93%) than feature D alone (Table 2).

RR of OCT Features in Identifying Polypoidal Regression

Since 16 of 39 (41%) and 5 of 39 (13%) patients had 2 and 3 study lesions per study eye (Supplementary Digital Content Table 1, https://links.lww.com/APJO/A159), the multiple logistic regression with GEE was performed to account for dependency between lesion from the same eye and same patient, and to confirm the associations between each OCT feature and status of polypoidal regression. RR and log odds ratio (LOR) revealed positive associations between feature A or B and complete polypoidal regression (RR: 5.0; 95% CI: 3.5–7.1, P<0.001; LOR 4.0; 95% CI: 2.4–5.7). No significant associations between feature C and complete or incomplete polypoidal regression. Feature D was found to be negatively associated with complete regression and positively associated with incomplete regression (RR: 4.6; 95% CI: 3.0–6.9, P<0.001; LOR: 3.6; 95% CI: 2.0–5.2) (Table 3). Case examples demonstrating each OCT feature of posttreatment polypoidal lesions were shown in Figure 4. Also, a case example with longitudinal serial image demonstrating gradual changes of internal reflectivity of a PED from feature D at baseline to feature C during follow-up, and feature B at 1 year after treatments corresponding to complete polypoidal regression on 1-year ICGA was shown in Figure 5.

TABLE 3 - Log Odds Ratio and Relative Risk of Each OCT Feature in Identifying Complete and Incomplete Polypoidal Regression
SDOCT Features Log Odds Ratio 95% CI Relative Risk* 95% CI P value
Identifying complete polypoidal regression (95% CI)
 Feature A 44.1 43.1–45.0 2.6* 2.1–3.2 <0.001
 Feature B 2.6 0.9–4.2 2.4 1.9–3.1 0.002
 Feature C −0.2 −1.6 to 1.3 0.9 0.5–1.4 0.834
 Feature D −3.6 −5.2 to −2.0 0.1 0.1–0.2 <0.001
 Feature E NA NA NA NA NA
 Feature A or B 4.0 2.4–5.7 5.0 3.5–7.1 <0.001
Identifying incomplete polypoidal regression (95% CI)
 Feature A −44.0 −45.0 to −43.1 0.0 0.0 <0.001
 Feature B −2.6 −4.2 to −0.9 0.2 0.1–0.4 0.002
 Feature C 0.2 −1.3 to 1.6 1.1 0.8–1.6 0.834
 Feature D 3.6 2.0–5.2 4.6 3.0–6.9 <0.001
 Feature E NA NA NA NA NA
 Feature A or B −4.0 −5.7 to −2.4 0.1 0.0–0.2 <0.001
Feature A defined as complete PED disappearance; feature B defined as PED with intralesional homogeneous hyperreflectivity and poor-defined RPE lining; feature C defined as PED with intralesional homogeneous hyperreflectivity and well-defined RPE lining; feature D defined as PED with intralesional heterogeneous reflectivity (hyporeflectivity-hyperreflectivity); and feature E defined as PED with intralesional homogeneous hyporeflectivity.
*BUN sign defined as “blended RPE with underlying structure” or unidentifiable RPE lining.
CI indicates confidence interval; PED, pigment epithelial detachment; RPE, retinal pigment epithelium; SDOCT, spectral-domain optical coherence tomography.

F4
FIGURE 4:
Example of a polypoidal choroidal vasculopathy (PCV) case demonstrating 4 optical coherence tomography (OCT) features of posttreatment polypoidal lesions observed in this study. Case 1, pretreatment indocyanine green angiography (ICGA) (A) and B scan of OCT across a polypoidal lesion (B), posttreatment ICGA showed complete disappearance of hypercyanescent lesions (complete polypoidal regression) (C), and posttreatment OCT showed complete disappearance of pigment epithelial detachment (PED) (D). Case 2, pretreatment ICGA (E) and OCT across single polypoidal lesion (F), posttreatment ICGA showed complete polypoidal regression (G), and posttreatment OCT showed PED with internal hyperreflectivity (white arrow) and “blended retinal pigment epithelium (RPE) with underlying structure (BUN)” sign* (black arrow) at least 2/3 of PED border (H). Case 3, pretreatment ICGA (I) and OCT (J) showed a polypoidal lesion, posttreatment ICGA showed partial polypoidal regression (K), and posttreatment OCT showed a PED with internal hyperreflectivity (white arrow) and intact RPE lining (black arrow) (L). Case 4, pretreatment ICGA (M) and OCT across a hemorrhagic PED and polypoidal lesion (N), posttreatment ICGA showed persistent polypoidal lesions (O), and posttreatment OCT showed a PED with internal heterogeneous reflectivity (P). *BUN signs defined as “blended RPE with underlying structure” or unidentifiable RPE lining.
F5
FIGURE 5:
Example of a polypoidal choroidal vasculopathy (PCV) case with complete polypoidal regression following treatments. optical coherence tomography (OCT) shows gradual changes of internal reflectivity of the pigment epithelial detachment (PED), from feature D to feature C and feature B overtime. At baseline, OCT shows a PED with internal heterogeneous reflectivity (feature D) with subretinal fluid (A); indocyanine green angiography (ICGA) demonstrates a polypoidal lesion (B, arrow). Fixed-dosing aflibercept treatments had been initiated (3 consecutive monthly injections then every 8 weeks). OCT taken 1 month after the first (C) and second aflibercept (D) shows internal heterogeneous reflectivity of PED with decreasing subretinal fluid. OCT taken 1 month after the third (E) and fifth aflibercept (F) shows flattening PED with internal homogeneous reflectivity and intact RPE lining (feature C). OCT taken 1 month after seventh (G) and eighth aflibercept (H) shows flattening PED with internal homogeneous reflectivity and a development of blended retinal pigment epithelium with underlying structure (BUN) sign. At 1 year after treatments, OCT reveals PED with internal homogeneous reflectivity with predominant BUN sign (feature B) (H, arrow) and ICGA reveals complete polypoidal regression (I).

DISCUSSION

While visual acuity is the most important outcome in PCV clinical trials, complete polypoidal regression is another important outcome to predict recurrences and massive submacular hemorrhage associated with recurrent PCV in longer term periods.11,12 ICGA has been used as a gold standard to evaluate complete polypoidal regression following PCV treatments. However, ICGA is not available in many ophthalmology clinics, including in the Asia-Pacific region where the prevalence of PCV is high. OCT is a more common retinal imaging that available at most clinics but there is limited information regarding the role of OCT in identifying status of posttreatment polypoidal lesions in PCV. In 2018, we reported a case series which the internal reflectivity of PED seemed to help differentiate complete and incomplete polypoidal regression.21 Subsequently, we further noticed that among PED with internal homogeneous hyperreflectivity, many of them had unidentifiable RPE lining at PED border in a majority of the scans throughout the lesion. We named the new observation as “BUN” sign or “blended RPE with underlying structure.” The “BUN” also implies a similarity between PCV and white steamed buns, which are popular among Asian descendants.

Thus, this study explored an accuracy of 5 prespecified OCT features, based on the internal reflectivity of PED and RPE lining, of posttreatment polypoidal lesions for identifying complete or incomplete polypoidal regression on ICGA following treatments for PCV.

Results of this study suggest that 3 of 5 prespecified OCT features are beneficial in identifying status of posttreatment polypoidal lesions, that is, a presence of either feature A (no PED) or B (PED with internal homogeneous hyperreflectivity with predominant “BUN” sign) had high accuracy in identifying complete polypoidal regression (86%; 95% CI: 75%–93%) and presence of feature D (PED with internal heterogeneous reflectivity) had high accuracy in identifying incomplete polypoidal regression (85%; 95% CI: 74%–92%). Also, the presence of feature D in 100% (65 of 65) of pretreatment polypoidal lesions supports the posttreatment findings that feature D is highly suggesting incomplete polypoidal regression.

Clinical application of study results can be better explained by PPV, NPV, and RR. Presence of either feature A or B had 92% PPV (95% CI: 75%–99%) and 82% NPV (95% CI: 66%–92%) in identifying complete polypoidal regression. This suggests that when physicians evaluate posttreatment polypoidal lesions on OCT and either feature A or B was found, 9 of 10 lesions would reveal complete polypoidal regression on ICGA. If neither OCT feature A nor B was found, 8 out of 10 lesions would not reveal complete polypoidal regression on ICGA. Likewise, presence of feature D had 90% PPV (95% CI: 73%–98%) and 80% NPV (95% CI: 63%–92%) in identifying incomplete regression. This suggests that when feature D was found on posttreatment OCT, 9 of 10 lesions would reveal incomplete regression on ICGA. If feature D was not found, 8 of 10 lesions would not reveal incomplete regression on ICGA. In terms of RR, a lesion with either feature A or B had 5 times greater chance to have complete polypoidal regression on ICGA, comparing to a lesion without these 2 features (RR: 5.0; 95% CI: 3.5–7.1, P<0.001). RR of a lesion with feature D is 4.6 greater chance to have incomplete regression on ICGA (RR: 4.6; 95% CI: 3.0–6.9, P<0.001).

We speculate that the poorly visible RPE of a PED, or the “BUN” sign, noted in the complete polypoidal regression lesions might be hypothesized to result from RPE atrophy due to nutrient deficiency when scarring processes occur within the polypoidal lesions beneath the RPE cells. The well-defined RPE of PEDs might reflect viability of RPE cells due to incomplete scarring process or only partial regression of polypoidal lesions.

Along with OCT feature B, an absence of PED (feature A) also had high accuracy in identifying complete polypoidal regression in the current study. Both features showed similar findings on ICGA (complete regression) but different OCT appearances, these OCT features might represent an appearance of complete polypoidal regression in a different time point. The height of PED lesions corresponded to complete polypoidal regression can be decreased overtime.19 Therefore, it may be possible that feature B occurred before feature A, and eventually progressed to feature A (no PED) overtime.

Recently, the Asia-Pacific Ocular Imaging Society (APOIS) PCV workgroup has evaluated the non-ICGA criteria to identify PCV among neovascular AMD cases which showed suboptimal treatment response after 3 anti-VEGF injections. Results showed that a combination of 3 non–ICGA-based criteria (sharp-peaked PED on OCT, sub-RPE ring-like lesion on OCT, and orange nodule on fundus photography) could differentiate PCV from neovascular AMD eyes. However, the OCT characteristics of complete or incomplete polypoidal regression were not evaluated in this study.23

Strengths of this study include use of a standardized grading protocol and well-trained image graders who are retina specialists with extensive clinical experience managing PCV. Limitations of this study include a relatively small sample size resulting in moderately sized confidence intervals around the point estimate results and inclusion criteria that included only polypoidal lesions with well-defined boundary on both OCT and ICGA, to be certain that the entire polypoidal lesions were evaluated by graders. Given the 2 types of PCV presentations, leakage and hemorrhagic types, the leakage type typically presents with well-defined boundary on both OCT and ICGA. The hemorrhagic type often presented with relatively poor-defined border of polypoidal lesions on either OCT or ICGA due to subretinal blood. Therefore, the results of this study could be applied to most cases with leakage type, but should be applied with cautions in hemorrhagic type. Also, this study evaluated OCT images with a minimum number of 49 scans per macular cube (interscan distance 120 µm). Thus, some polypoidal lesions might have been missed if the lesion size was smaller than the interscan distance. This study evaluated PCV lesions that received either anti-VEGF therapy or PDT, or both. It is unknown whether the results can be applicable to PCV eyes receiving other treatment modalities, for example, laser photocoagulation. Note that this study evaluated the association between internal reflectivity of PED seen on OCT and polypoidal status on ICGA. However, it is unknown whether the internal reflectivity of the PED truly represents polypoidal status. OCT angiography that provides flow signal or thin slab beneath the RPE or histopathologic study may be warranted in the future.

Additionally, even though this study proposed OCT characteristics of complete polypoidal regression, which sometimes may be a treatment goal in the management of PCV, there is limited information whether physicians should discontinue treatments right after complete polypoidal regression, that is, either feature A or B on OCT, is identified. Studies with relatively long-term data are warranted to answer this question. However, for eyes with incomplete polypoidal regression, that is, feature D on OCT, several studies have suggested that these eyes have a relatively high risk for recurrent, large subretinal hemorrhage.11,12 Thus, recognizing this feature may have clinical relevance wherein this feature suggests that discontinuing treatment should be done cautiously, since a vision-threatening condition, namely large subretinal hemorrhage, might have an increased risk of developing in the absence of continued treatment for eyes with incomplete polypoidal regression.24

CONCLUSIONS

Despite limited information regarding characteristics of posttreatment polypoidal lesions on OCT images, this study proposed 3 OCT features which seem to be helpful in evaluating status of posttreatment polypoidal lesion in PCV. Presence of either feature A (no PED) or B (PED with internal homogeneous hyperreflectivity with predominant “BUN” sign) provided high accuracy in identifying complete polypoidal regression, and presence of feature D (PED with internal heterogeneous reflectivity) provided high accuracy in identifying incomplete polypoidal regression. When ICGA is unavailable or contraindicated, these OCT features might be helpful in identifying the status of posttreatment polypoidal lesions as completely or incompletely regressed in PCV eyes.

REFERENCES

1. Yannuzzi LA, Sorenson J, Spaide RF, et al. Idiopathic polypoidal choroidal vasculopathy (IPCV). 1990. Retina Phila Pa. 2012;32(suppl 1):1–8.
2. Tan CS, Ngo WK, Chen JP, et al. EVEREST study report 2: imaging and grading protocol, and baseline characteristics of a randomised controlled trial of polypoidal choroidal vasculopathy. Br J Ophthalmol. 2015;99:624–628.
3. Invernizzi A, Pellegrini M, Cornish E, et al. Imaging the choroid: from indocyanine green angiography to optical coherence tomography angiography. Asia Pac J Ophthalmol (Phila). 2020;9:335–348.
4. Chaikitmongkol V, Cheung CMG, Koizumi H, et al. Latest developments in polypoidal choroidal vasculopathy: epidemiology, etiology, diagnosis, and treatment. Asia Pac J Ophthalmol (Phila). 2020;9:260–268.
5. Yannuzzi LA, Wong DW, Sforzolini BS, et al. Polypoidal choroidal vasculopathy and neovascularized age-related macular degeneration. Arch Ophthalmol Chic Ill 1960. 1999;117:1503–1510.
6. Liu Y, Wen F, Huang S, et al. Subtype lesions of neovascular age-related macular degeneration in Chinese patients. Graefes Arch Clin Exp Ophthalmol. 2007;245:1441–1445.
7. Maruko I, Iida T, Saito M, et al. Clinical characteristics of exudative age-related macular degeneration in Japanese patients. Am J Ophthalmol. 2007;144:15–22.
8. Bhoomibunchoo C, Yospaiboon Y, Thoongsuwan S, et al. Idiopathic polypoidal choroidal vasculopathy in Thai patients with clinical and angiographic choroidal neovascularization. Clin Ophthalmol Auckl NZ. 2017;11:317–322.
9. Koh A, Lai TYY, Takahashi K, et al. Efficacy and safety of ranibizumab with or without verteporfin photodynamic therapy for polypoidal choroidal vasculopathy: a randomized clinical trial. JAMA Ophthalmol. 2017;135:1206–1213.
10. Lee WK, Iida T, Ogura Y, et al. Efficacy and safety of intravitreal aflibercept for polypoidal choroidal vasculopathy in the PLANET study: a randomized clinical trial. JAMA Ophthalmol. 2018;136:786–793.
11. Morizane-Hosokawa M, Morizane Y, Kimura S, et al. Impact of polyp regression on 2-year outcomes of intravitreal aflibercept injections: a treat-and-extend regimen for polypoidal choroidal vasculopathy. Acta Med Okayama. 2018;72:379–385.
12. Cho JH, Park YJ, Cho SC, et al. Posttreatment polyp regression and risk of massive submacular hemorrhage in eyes with polypoidal choroidal vasculopathy. Retina Phila Pa. 2020;40:468–476.
13. Koh A, Lee WK, Chen L-J, et al. EVEREST study: efficacy and safety of verteporfin photodynamic therapy in combination with ranibizumab or alone versus ranibizumab monotherapy in patients with symptomatic macular polypoidal choroidal vasculopathy. Retina Phila Pa. 2012;32:1453–1464.
14. Gomi F, Oshima Y, Mori R, et al. Initial versus delayed photodynamic therapy in combination with ranibizumab for treatment of polypoidal choroidal vasculopathy: The Fujisan Study. Retina Phila Pa. 2015;35:1569–1576.
15. Oishi A, Kojima H, Mandai M, et al. Comparison of the effect of ranibizumab and verteporfin for polypoidal choroidal vasculopathy: 12-month LAPTOP study results. Am J Ophthalmol. 2013;156:644–651.
16. Wong TY, Ogura Y, Lee WK, et al. Efficacy and safety of intravitreal aflibercept for polypoidal choroidal vasculopathy: two-year results of the aflibercept in polypoidal choroidal vasculopathy study. Am J Ophthalmol. 2019;204:80–89.
17. Expert PCV Panel, Koh AHC, Chen L-J, et al. Expert PCV Panel. Polypoidal choroidal vasculopathy: evidence-based guidelines for clinical diagnosis and treatment. Retina Phila Pa. 2013;33:686–716.
18. Japanese Study Group of Polypoidal Choroidal Vasculopathy. Criteria for diagnosis of polypoidal choroidal vasculopathy. Nippon Ganka Gakkai Zasshi. 2005;109:417–427.
19. Data from an unofficial survey by Bayer Thailand. March 24, 2018.
20. Chaikitmongkol V, Kong J, Khunsongkiet P, et al. Sensitivity and specificity of potential diagnostic features detected using fundus photography, optical coherence tomography, and fluorescein angiography for polypoidal choroidal vasculopathy. JAMA Ophthalmol. 2019;137:661–667.
21. Chaikitmongkol V, Upaphong P, Patikulsila D, et al. OCT vs indocyanine green angiography features of polypoidal choroidal vasculopathy. Ophthalmol Retina. 2018;2:170–172.
22. Deddens JA, Petersen MR. Re: “Estimating the relative risk in cohort studies and clinical trials of common outcomes”. Am J Epidemiol. 2004;159:213–214; author reply: 214–215.
23. Chong Teo KY, Sadda SR, Gemmy Cheung CM, et al. Non ICGA treatment criteria for suboptimal anti VEGF response for polypoidal choroidal vasculopathy: APOIS PCV Workgroup Report 2. Ophthalmol Retina. 2021;5:945–953.
24. Chaikitmongkol V, Sakong M, Lai TYY, et al. Treat-and-Extend regimens for the management of neovascular age-related macular degeneration and polypoidal choroidal vasculopathy: consensus and recommendations from the Asia-Pacific Vitreo-retina Society. Asia Pac J Ophthalmol (Phila). 2021;10:507–518.
Keywords:

age-related macular degeneration; polypoidal choroidal vasculopathy; complete polypoidal regression; optical coherence tomography; indocyanine green angiography

Supplemental Digital Content

Copyright © 2022 Asia-Pacific Academy of Ophthalmology. Published by Wolters Kluwer Health, Inc. on behalf of the Asia-Pacific Academy of Ophthalmology.