Six-Year Outcomes in Subjects with Polypoidal Choroidal Vasculopathy in the PLANET Study

Abstract

Introduction

The PLANET study demonstrated that intravitreal aflibercept 2 mg injection (IAI) monotherapy was non-inferior to IAI with rescue photodynamic therapy (PDT) for polypoidal choroidal vasculopathy (PCV) at 2 years. This follow-up study assessed the long-term outcomes 6 years after initiation of IAI therapy in the PLANET cohort.

Methods

This was a non-interventional cohort study of 72 patients with PCV from six international sites who originally completed the PLANET study. Outcomes were assessed during a recall visit at ~6 years from PLANET enrollment. Data collected included best-corrected visual acuity (BCVA), optical coherence tomography, treatment exposure, and adverse events. Patients were treated according to standard of care after exiting from the PLANET study.

Results

Of 72 participants, 32 were randomized to IAI monotherapy and 40 to IAI with rescue PDT. Of those randomized to IAI with rescue PDT, 11 (15.2%) received rescue PDT during the entire study period over 6 years. Mean ± standard deviation (SD) BCVA improved by 14.6 ± 10.3 letters at 2 years but declined to baseline by 6 years (−0.3 ± 21.8 letters). At year 6, 45.8% of eyes had BCVA ≥ 69 Early Treatment Diabetic Retinopathy Study (ETDRS) letters, while 23.6% had < 36 ETDRS letters. Polypoidal lesions were present in 69.4% at year 2 versus 29.2% at year 6. Patients received a mean ± SD of 10.2 ± 2.9 anti-vascular endothelial growth factor (anti-VEGF) injections between years 2–6. Outcomes were better in reimbursed versus non-reimbursed regions. Multivariable analysis showed that younger age and better mean BCVA at year 2 predicted more favorable long-term vision. PDT received at any time was not associated with overall mean BCVA at year 6, although it was associated with a lower risk of ≥ 15-letter loss.

Conclusion

At 6 years, initial visual gains from aflibercept therapy were not maintained; however, nearly half of the eyes had BCVA ≥ 69 letters, and a large proportion achieved polypoidal lesion regression.

Trial Registration

ClinicalTrials.gov identifier, NCT002120950.

Key Summary Points

Why carry out this study?

While clinical trials such as EVEREST II and PLANET have reported the safety and efficacy of anti-vascular endothelial growth factor (anti-VEGF) therapy with or without photodynamic therapy in polypoidal choroidal vasculopathy (PCV) up to 2 years, this 6-year follow-up study will add important data regarding visual and anatomical outcomes beyond 2 years, and the impact of early disease control.

What was learned from the study?

Almost half of the eyes had good vision (≥ 69 letters) at year 6. The mean number of anti-VEGF re-treatment injections was about 10 over the follow-up period from year 2 to year 6, suggesting undertreatment in the clinical setting, and early disease control and better vision at year 2 were associated with better vision at year 6.

PCV is a chronic, relapsing disease that requires long-term management. Anti-VEGF monotherapy can maintain baseline vision over the long term, but adequate and consistent re-treatment is essential to preserve gains.

Introduction

The optimal management of polypoidal choroidal vasculopathy (PCV) has been evolving over the past two decades [1–5]. Evidence from the EVEREST and EVEREST II trials showed that photodynamic therapy (PDT) with verteporfin in combination with intravitreal ranibizumab conferred superior visual and anatomical outcomes compared to monotherapy with ranibizumab. [6, 7] Combination therapy demonstrated superior regression of polypoidal lesions. A previous follow-up study of the EVEREST II study participants reported that at 6 years, most participants continued to require treatment after exiting the initial trial [8]. We observed no difference in visual outcomes between initial treatment arms, although initial combination therapy was associated with better anatomical outcomes and fewer injections [8]. We emphasized the importance of continuous monitoring and re-treatment in PCV regardless of polyp status.

In contrast to the EVEREST trials, the PLANET study demonstrated that intravitreal aflibercept monotherapy was non-inferior to aflibercept plus rescue PDT, with more than 85% of eyes achieving favorable outcomes without requiring PDT [9, 10]

The use of anti-vascular endothelial growth factor (anti-VEGF) monotherapy has gained popularity in clinical practice worldwide due to its ease of use and demonstrated efficacy. While the EVEREST and PLANET clinical trials clarified short-term management strategies, uncertainty remains regarding the long-term durability of anti-VEGF therapy in PCV [11, 12]. Chronic re-treatment is often required in neovascular age-related macular degeneration (AMD), and undertreatment has been repeatedly implicated in visual decline observed in clinical practice. Moreover, it is unclear whether the subset of eyes treated with PDT can derive long-lasting benefit in terms of vision, anatomy, or treatment burden reduction.

This study reports the 6-year outcomes of patients originally enrolled in PLANET [9, 10], with comparisons to a similar follow-up study of patients who exited from EVEREST II [8]. We aimed to describe the extent to which early functional and anatomical improvements were maintained, to evaluate the impact of rescue PDT administered during the first 2 years, and to examine the relationship between re-treatment intensity and long-term visual outcomes.

Methods

Study Design

This is a multicenter, non-interventional, cross-sectional recall follow-up study of the long-term outcomes of a cohort of patients originally treated in the PLANET study (ClinicalTrials.gov identifier, NCT02120950) [9, 10]. The present analyses were exploratory and descriptive in nature and were not prespecified in the original randomized trial protocol. The study was conducted in accordance with the tenets of the Declaration of Helsinki and with approval of the respective institutional review boards of the participating centers (SingHealth IRB: R1597/96/2018). Written informed consent was obtained from each participant. This study was investigator-led and funded as part of a grant on Asian AMD from the Ministry of Health of Singapore through a submission by the Singapore Eye Research Institute. Assistance was provided by Bayer AG (Leverkusen, Germany), who compiled the database of completers, which included information on the randomization arm, best-corrected visual acuity (BCVA), and anatomical outcomes at the end of the original study.

Study Cohort

A total of 318 participants were enrolled in the original PLANET study, and 284 completed the end of the original study (year 2). For the purpose of this follow-up study, only sites with a minimum of five completers were invited to participate, in order to limit administrative burden. Sixteen sites with a potential of 181 participants participated in the follow-up study and contributed data on a total of 75 participants. All participating patients provided informed consent for the end of the extension phase (year 6) and for their investigators to access medical records of any additional assessment and treatments they had received after exiting from the original PLANET study.

Examination Procedures

Participants were evaluated during a single study visit at the end of the extension phase (year 6), which took place approximately 6 years (±6 months) after enrollment in the original PLANET study [9, 10]. Patients were interviewed about treatment received in either eye, visits to ophthalmologists, and serious medical events occurring after the end of the original study (year 2). Examination procedures during this visit included BCVA measurement recorded as logarithm of the minimum angle of resolution (logMAR) letters, a dilated eye examination, spectral domain optical coherence tomography (OCT), fundus color photography, fluorescein angiography (FA), and indocyanine green angiography (ICGA). Chart review of the period between the end of the original study (year 2) and the end of the extension phase (year 6) was performed for evidence of re-treatment during this extension phase (years 2–6). Where available, the number of treatments with anti-VEGF agents (ranibizumab, bevacizumab, or aflibercept) or other PCV treatments (PDT, thermal laser) was recorded.

Outcome Measures

The primary aim of the study was to report the long-term BCVA changes from baseline to the end of the original study (year 2) and to the end of the extension phase (year 6). This included the mean BCVA change in logMAR letters and proportion of eyes with relatively good BCVA (≥ 69 letters), worse BCVA (< 36 letters), and those losing 15 letters or more at year 6 relative to baseline. Secondary anatomical outcomes included central subfield thickness (CSFT), choroidal thickness, presence of subretinal fluid (SRF), presence of pigment epithelial detachment (PED), presence of polypoidal lesions (PL), and presence of fibrosis from baseline to the end of the original study (year 2), as well as at the end of extension phase (year 6).

Statistical Analysis

Anonymized data were collected on case report forms and sent to the Singapore Eye Research Institute, where data were entered manually into an encrypted online database (REDCap software, 8.1.12). De-identified data were exported for statistical analysis.

Descriptive data included the mean (standard deviation [SD]), median (first and third quartiles [Q1, Q3]) and percentages where appropriate. Statistical tests including Student’s t-test, Wilcoxon rank sum, and Fisher’s tests were used where appropriate to compare baseline, year 2, and year 6 characteristics between the monotherapy and combination groups.

Associations with continuous and categorical visual outcomes were modeled using multivariable regression. Linear and logistic regression was applied for continuous (final visual acuity at the end of extension phase [year 6]) and binary (proportion of patients with good vision [≥ 69 letters] and ≥ 15-letter loss) outcomes, respectively. Two associations were analyzed: (1) clinical predictors of visual outcomes at the end of the extension phase (year 6) using clinical parameters at the start of the extension phase (BCVA, CSFT, choroidal thickness, age, gender, total injections received, and any use of PDT), and (2) associations of visual outcomes with anatomical features at the end of follow-up (presence of fibrosis, atrophy, residual polyps, lesion greatest linear dimension [GLD], and CSFT).

For clinical predictors, models were built including all covariates simultaneously, representing a multivariable analysis adjusted for potential confounders. For structural predictors, models included fibrosis, atrophy, and PL status, as well as an interaction between PL presence and fibrosis to capture their potential synergistic contribution to visual outcomes. Estimates were reported as regression coefficients (β) for linear models and odds ratios (ORs) with 95% confidence intervals (CIs) for logistic models. All p values were two-sided, with p < 0.05 considered statistically significant. Results were visualized using forest plots stratified by continuous and binary outcomes.

To account for potential effects of different healthcare-funding models, outcomes were analyzed by dividing the cohort into countries that received reimbursement for neovascular AMD treatment (Australia, Japan, Korea, Taiwan) and countries that received partial or no reimbursement for neovascular AMD treatment (Singapore, Hong Kong).

The outcomes between (1) treatment groups at 6 years and (2) reimbursement groups were assessed by a mixed effects longitudinal generalized additive model with the therapy group and time as the main predictor variable adjusted for fixed effects (age, baseline vision, CSFT, and PL status), and site as the random effect. A p value of less than 0.05 was considered statistically significant. However, due to the exploratory nature of this extension study, all p values reported are nominal. Analyses were conducted using R version 4.0.0 or SAS 9.4.

Results

Study Population and Characteristics at 6-Year Examination

A total of 318 participants were randomized in the PLANET trial, of whom 284 (89.3%) completed the end of the original study (year 2). At the end of the extension phase (year 6), 72 participants (22.6%) provided data for the final analysis. Participants were from Japan (n = 36), Korea (n = 18), Singapore (n = 8), Taiwan (n = 4), Hong Kong (n = 3), and Australia (n = 3). The most common reason for non-enrollment was loss to follow-up and inability to contact participant.

The characteristics of the participants who provided data to the end of extension phase (6-year) visit are summarized in Table 1. Thirty-two participants (44.4%) were in the aflibercept (IAI) monotherapy arm and 40 participants (55.6%) were in the IAI/PDT arm during the original study. However, 11 patients out of 40 (15.2%) received rescue PDT at any time during the study and the extension phase. Compared to the overall PLANET completers, participants had similar BCVA at baseline (p = 0.80) and at 2 years (p = 0.31). The mean interval ± SD between the end of the original study (year 2) and the end of extension phase (year 6) was 35.05 ± 17.60 months.

Table 1.

Comparison of 6-year follow-up participants with all year 2 completers

	All			IAI monotherapy			IAI/PDT
	6-year follow-up N = 72	Year 2 completers N = 284	P value (6-year versus 2-year completers)	6-year follow-up N = 32	Year 2 completers N = 137	P value (6-year versus 2-year completers)	6-year follow-up N = 40	Year 2 completers N = 147	P value (6-year versus 2 year completers)
Age at exit of PLANET, mean (SD)	68.3 (8.2)	70.4 (8.2)	0.06	68.7 (8.7)	70.6 (8.4)	0.27	68.1 (7.9)	70.2 (8.0)	0.14
Gender, male/female (%)	52/22 (72.2%/27.8%)	198/86 (69.7%/30.3)	0.79	24/8 (75.0%/25%)	97/40 (70.8%/29.2%)	0.79	28/12 (70.0%/30%)	101/46 (68.7%/31.3%)	1.00
Baseline characteristics
BCVA, letters, mean (SD)	58.1 (12.2)	58.5 (11.4)	0.80	57.6 (12.1)	58.1 (11.2)	0.83	58.5 (12.4)	58.9 (11.6)	0.86
Proportion with ≥ 69 letter (%)	17 (23.6%)	68 (23.9%)	1.00	6 (18.8%)	29 (21.2%)	0.95	11 (27.5%)	39 (26.5%)	1.00
Central subfield thickness, µm, mean (SD)	361.7 (127.3)	343.0 (114.1)	0.26	337.4 (96.7)	339.7 (109.3)	0.91	381.1 (145.5)	346.0 (118.6)	0.17
Characteristics at year 2
BCVA, letters, mean (SD)	72.6 (11.5)	69.1 (14.1)	0.31	73.1 (10.4)	69.3 (13.7)	0.09	72.2 (12.4)	68.9 (14.5)	0.16
Proportion with ≥ 69 letter (%)	48 (66.7%)	176 (62.0%)	0.55	23 (71.9%)	88 (64.2%)	0.54	25 (62.5%)	88 (59.9%)	0.91
Central subfield thickness, µm, mean (SD)	204.0 (46.2)	205.6 (77.3)	0.82	205.7 (43.3)	198.2 (42.3)	0.17	202.5 (48.9)	212.4 (98.8)	0.38
Polypoidal lesions present (%)	50 (69.4%)	183 (64.4%)	0.51	24 (75%)	84 (61.3%)	0.21	26 (65.0%)	99 (67.3%)	0.93
Presence of active polypoidal lesions (%)	15 (20.8%)	60 (21.1%)	1.00	5 (15.6%)	25 (18.2%)	0.93	10 (25.0%)	35 (23.8%)	1.00

BCVA best-corrected visual acuity, IAI intravitreal aflibercept injection, PDT photodynamic therapy, SD standard deviation

BCVA Change Baseline Through Year 6

Visual acuity outcomes are summarized in Table 2. The mean BCVA improved in both monotherapy and combination groups at year 2 but dropped to baseline at year 6. The mean ± SD change in the 6-year follow-up cohort (n = 72) from baseline in mean BCVA was an increase of 14.6 ± 10.3 in year 2, and a decrease of 0.3 ± 21.8 letters in year 6. Since only 11 patients (15.3%) received rescue PDT at any point during the study and the extension phase, we report the BCVA outcomes of the overall group, followed by outcomes excluding the 11 patients who had received PDT. Results were similar in the analysis limited to the 66 eyes that did not receive PDT at any time during the original PLANET study and the extension phase and the entire cohort (57.8 ± 24.0 versus 56.8 ± 23.9 letters, p = 0.81).

Table 2.

Visual and anatomical features at baseline, year 2, and year 6 for all 6-year completers (n = 72)

	Baseline (BL)	Year 2 (Y2)	Year 6 (Y6)	P value BL vs Y2	P value Y2 vs Y6	P value BL vs Y6
BCVA, mean (SD) letters	58.1 (12.1)	72.6 (11.5)	57.8 (23.7)	< 0.01	< 0.01	0.92
> 69 letters, n (%)	17 (23.6)	48 (66.7)	33 (45.8)	< 0.01	0.01	< 0.01
36–68 letters, n (%)	52 (72.2)	24 (33.3)	22 (30.6)	< 0.01	0.5	< 0.01
< 36 letters, n (%)	3 (4.2)	0 (0)	17 (23.6)	0.25	< 0.01	< 0.01
CSFT, mean (SD), μm	361.7 (127.3)	204.0 (46.2)	261.2 (129.8)	< 0.01	< 0.01	< 0.01
Polypoidal lesions present	NA	50 (69.4%)	21 (29.2%)	NA	< 0.01	NA
Polypoidal activity present	NA	15 (20.8%)	19 (26.4%)	NA	0.43	NA

BCVA best-corrected visual acuity, CSFT central subfield thickness, SD standard deviation

Change in Anatomical Outcomes from Baseline Through Year 6

Anatomical outcomes are summarized in Table 2 and include CSFT change and the proportion of eyes with the presence of polypoidal lesions and polypoidal lesion activity at baseline, year 2, and year 6. The proportion of eyes with polypoidal lesions at year 6 was similar in the 66 eyes that did not receive PDT (27.2% compared to the overall cohort, p = 0.85). Mean ± SD choroidal thickness was 198.5 ± 126.5 µm at year 2 and 165.7 ± 124.9 µm at year 6. Fellow-eye involvement was uncommon overall, noted in six (8.3%) cases, with only one fellow eye diagnosed with PCV.

End Stage Phenotypes

At the end of the extension phase (year 6), 7seven (9.7%) and 21 (29.2%) of 72 eyes developed macular atrophy and fibrosis, respectively. Visual acuity at year 6 was significantly worse in the eyes with fibrosis than those that had neither fibrosis nor atrophy (39.4 ± 25.8 letters vs. 65.4 ± 17.2 letters, p < 0.01). The mean BCVA in eyes with macular atrophy was similarly lower compared to those that had neither fibrosis nor atrophy, but the difference was not statistically significant (45.9 ± 29.2 letters vs. 65.4 ± 17.2 letters, p = 0.13).

Treatment Exposure Between Year 2 and Year 6

On a yearly basis during the follow-up period, 39–52 eyes (54–72%) received additional anti-VEGF injections each year. Patients received a mean ± SD of 10.2 ± 2.9 anti-VEGF injections between year 2 and year 6 (median 7; range 0–41), corresponding to an average of 2.6 ± 2.3 injections per year. The mean number of injections given during the extension phase was 3.5 (year 3), 2.5 (year 4), 2.2 (year 5), and 2.0 (year 6). PDT and focal laser were rarely used treatments during the extension phase, with only six (8.3%) and three (4.2%) eyes receiving these treatments, respectively, after the end of the original study (year 2). Treatment intensity over time is summarized in Fig. 1.

Fig. 1.

Re-treatments received by the follow-up cohort between exiting the PLANET study and the 6-year follow-up visit. PDT photodynamic therapy

Outcome and Treatments by Geographic Location and Reimbursement

Variation in re-treatments in the extension phase was noted among the six countries, with the proportion of eyes receiving additional treatments ranging from 50% to 100% (Australia [n = 3], 100%; Hong Kong [n = 4], 50%; Japan [n = 36], 86%; Korea [n = 20], 65%; Singapore [n = 8], 88%; Taiwan [n = 5], 100%). The mean number of additional treatments ranged from 1.5 to 10.5 (Australia, 6.0 ± 0.0; Hong Kong, 1.5 ± 1.7; Japan, 10.4 ± 9.0; Korea, 10.5 ± 9.7; Singapore, 8.0 ± 6.3; Taiwan, 5.2 ± 4.1).

Participants in the reimbursed group received a higher mean number of anti-VEGF injections than those in the non-reimbursed group (9.1 vs. 6.1, p < 0.01). However, there was no significant difference in visit numbers over the extension period between the two groups (mean 14.5 visits in reimbursed vs. 12.1 in non-reimbursed countries, p = 0.53).

In the extension phase, there was a numerically greater decline in mean BCVA in the non-reimbursed group than in the reimbursed group of countries (with change in BCVA in this phase of −12.8 letters in the non-reimbursed group compared to −5.5 letters in the reimbursed group; p = 0.12). A numerically higher proportion of eyes from non-reimbursed areas lost ≥ 15 letters by the end of the extension phase compared with the reimbursed group (38.0% vs. 20.0%, p = 0.09). CSFT at the end of the extension phase was significantly lower in the reimbursed group than in the non-reimbursed group (196.0 ± 38.0 µm vs. 276.3 ± 137.1 µm, p = 0.004). Other disease outcomes also trended more favorably in reimbursed countries. Fibrosis at the last visit was observed in 29.4% in non-reimbursed regions compared to 18.2% in reimbursed regions (p = 0.18), and atrophy was seen in 17.6% versus 9.1%, respectively (p = 0.22).

Visual Trajectories by Treatment Intensity

To further explore whether treatment intensity influenced long-term outcomes, eyes were stratified by the total number of anti-VEGF injections received between the end of the original study (year 2) and end of the extension phase (year 6) (Fig. 2). Eyes were divided by the number of injections received between years 2 and 6 (> 13, 6–13, 1–5, and no treatment injection groups). BCVA changes showed that eyes receiving 6–13 injections (n = 20) had the most stable long-term outcomes, where BCVA declined modestly from 71.6 letters at the end of the original study (year 2) to 64.2 letters at the end of the extension phase (year 6). In contrast, eyes receiving 1–5 injections (n = 15) showed the steepest decline of 70.1 to 44.0 letters, while those with no injections (n = 22) declined from 72.7 to 59.3 letters. Eyes receiving > 13 injections (n = 15) also declined from 76.3 to 61.0, likely reflecting eyes with more active disease requiring frequent re-treatment.

Fig. 2.

Features predictive of visual outcomes. Among the baseline features and treatment exposure evaluated, better vision at year 2 and younger age were associated with achieving > 69 letters at year 6. PDT during the extension and thicker subfoveal choroid were associated with lower risk of ≥ 15-letter loss. BCVA best-corrected visual acuity

Features Associated with Visual Outcome at End of Extension Phase (Year 6)

Among clinical predictors, better mean BCVA at the end of the original study (year 2) was strongly associated with favorable long-term outcomes. Each additional letter at year 2 increased the odds of achieving good vision (> 69 letters) at year 6 by 22% (OR 1.22, 95% CI 1.10–1.42, p < 0.01). Conversely, older age was associated with poorer vision with each additional year, reducing the odds of achieving > 69 letters by 13% (OR 0.87, 95% CI 0.74–0.97, p = 0.03). While PDT administered at any time during the original PLANET study or during the extension phase (n = 11) was not associated with overall VA at year 6, it was associated with reduced odds of ≥ 15-letter loss (OR 0.09, 95% CI 0.01–0.72, p = 0.04). Thicker subfoveal choroid at the end of the original study (year 2) also predicted a lower risk of ≥ 15-letter loss (OR 0.99 per µm, 95% CI 0.98–1.00, p = 0.02). Other factors including gender, number of injections, and CSFT at year 2 were not significantly associated with long-term visual outcomes (Fig. 3).

Fig. 3.

Anatomical features at final visit and their association with vision. Among structural features assessed at the end of the extension phase, the presence of subretinal fibrosis and larger lesion size were associated with ≥ 15-letter loss. VA visual acuity, CSFT central subfield thickness

Among structural changes assessed at the end of the extension phase, the presence of subretinal fibrosis was strongly associated with worse outcomes. Fibrosis was linked to markedly lower vision at year 6 (β −22.97 letters, 95% CI −39.13 to −6.80, p < 0.01) and greatly increased odds of ≥ 15-letter loss (OR 26.6, 95% CI 3.89–285.1, p < 0.01). Atrophy showed a similar trend toward worse outcomes (β −16.96 letters, p = 0.08; OR 9.12 for ≥ 15-letter loss, p = 0.06) though not statistically significant. Larger GLD at the end of the study was also associated with higher risk of ≥ 15-letter loss (OR 1.56 per mm, 95% CI 1.03–2.51, p = 0.04). Other structural factors such as polypoidal lesion status and CSFT at the last visit were not significantly associated with vision at year 6 (Fig. 4). Examples of patients with good and poor outcome are shown in Fig. 5.

Fig. 4.

Visual outcomes from baseline to year 6 stratified by number of anti-VEGF injections received between exit of PLANET study and the 6-year follow-up visit. (A) Patients who received no injections during the follow-up period (n = 22). (B) Patients who received 1–5 injections during the follow-up period (n = 15). (C) Patients who received 6–13 injections during the follow-up period (n = 20). (D) Patients who received > 13 injections during the follow-up period (n = 15)

Fig. 5.

Representative examples of imaging for patients from commencement of the PLANET study to year 6. The visit and corresponding BCVA are indicated on the top and a summary of treatment on the bottom of each panel of images (A, B). (A) Imaging from a patient that had good outcomes at year 6. The indocyanine green angiography (ICGA) (top row) shows the presence of polypoidal lesions (PL) at baseline, which was completely regressed at month 12, but recurred at month 24, persisted until year 3, and was eventually completely closed at year 6. Disease activity as observed on OCT (bottom row) showed absence of sub- and intraretinal fluid at month 12. At month 24, recurrence of subretinal fluid (SRF) was observed, which persisted through year 3 following a treat-and-extend protocol. The patient was in the IAI with PDT arm in the PLANET study and received 14 IAI and one active PDT between baseline and month 24. After exiting PLANET, he received additional 20 IAI following the treat-and-extend protocol and two active PDT until year 6. (B) Imaging from a patient with poor outcomes at year 6. The ICGA (top row) showed the presence of PL at baseline which was completely regressed at month 12. At month 24, the original PL remained closed but a new PL appeared, with recurrence of SRF. The patient was in the IAI with sham PDT arm in the PLANET study and received 14 IAI and one sham PDT between baseline and month 24. After exiting PLANET, he received an additional six IAI and two active PDT on a pro re nata basis. He defaulted follow-up for financial reasons after year 3. At his year 6 visit, subretinal fibrosis was noted with poor vision. ICGA showed a mature network with no active PL

Adverse Events

One case of submacular hemorrhage with vitreous hemorrhage with vision dropping to one letter was reported. This occurred 3 months after the end of the 2-year PLANET study, and the patient was randomized into the aflibercept monotherapy arm. No other adverse events were reported.

Discussion

The long-term follow-up of the PLANET study provides important insights into the durability of aflibercept therapy in PCV. During the 2-year original study, patients achieved substantial improvements in both vision and anatomy, but mean BCVA returned to baseline at the end of the extension phase (year 6). While the mean gains were not maintained during the subsequent 4 years of clinical practice, a large proportion (45.8%) had ≥ 69 letters, with about 20% gaining 15 letters and 37% gaining 10 letters at year 6. At the exit of the PLANET study, PL were present in 70% of eyes according to reading center assessment. In comparison, only 30% of eyes had PL at the end of the extension phase (year 6), according to investigator grading.

The return to mean baseline vision may be partially explained by disease recurrence and potentially lower re-treatment intensity in some patients. Between the end of original study (year 2) and end of extension phase (year 6), additional anti-VEGF injections were received in 72% of the follow-up cohort in year 3, and 54% in subsequent years. However, the intensity of treatment was generally low, with patients receiving a mean of 10.2 injections in total, a pattern consistent with the chronic problem of undertreatment observed in real-world neovascular AMD practice [8, 13]. Importantly, treatment intensity was highly variable: some patients received no injections, while others received more than 40 anti-VEGF injections during this period, and participants from countries with reimbursed healthcare received substantially more re-treatment than their counterparts from countries where healthcare was self-funded [14, 15]. Interestingly, the best outcomes were seen in eyes treated with a moderate and sustained frequency (6–13 injections), where vision improved from 56 letters at baseline to 71 at the end of original study (year 2) and stabilized at 64 letters at the end of extension phase (year 6) (+10 letters from baseline). In contrast, presumably undertreated eyes (0–5 injections) experienced the poorest outcomes, while those receiving > 13 injections also declined, likely reflecting eyes with highly active disease requiring frequent re-treatment.

A similar pattern was observed when comparing participants from reimbursed versus non-reimbursed healthcare systems. Although overall trajectories declined toward baseline in both groups, eyes from reimbursed regions achieved higher mean BCVA at year 6 and experienced fewer cases of ≥ 15-letter loss. This disparity was explained by greater treatment exposure: patients in reimbursed countries received significantly more injections (9.1 vs. 6.1, p < 0.01), leading to superior anatomical outcomes, including lower CSFT and numerically lower rates of fibrosis and atrophy. These findings highlight the influence of healthcare policy and access on long-term outcomes, emphasizing that consistent access to therapy enables sustained re-treatment and better disease control.

Predictor analyses reinforced the importance of early disease control. Better mean BCVA at year 2 strongly predicted favorable long-term outcomes, with each additional letter gained at year 2 increasing the odds of achieving ≥ 69 letters at year 6 by 22%. Conversely, older age reduced the likelihood of maintaining good vision. PDT was not associated with overall VA at year 6, although it was associated with a lower risk of ≥ 15-letter loss. However the number of patients who had received PDT in this cohort was small, and may not be generalizable to a wider cohort. Thicker subfoveal choroid at year 2 also predicted a lower risk of severe vision loss, suggesting that choroidal thickness may be considered as a biomarker of long-term outcomes.

We previously performed a similar follow-up exercise for the participants who completed the EVEREST-II study, in which participants in the original study were randomized to intravitreal ranibizumab with active versus sham PDT arms. In the EVEREST II 6-year extension, overall mean vision dropped below baseline, with only one-third of eyes maintaining ≥ 69 letters and about one-fifth declining to < 36 letters [8]. While combination therapy in EVEREST II achieved superior anatomical outcomes and slightly reduced injection burden, these differences amounted to only 1–2 injections over 4 years and are unlikely to be clinically meaningful [8]. In the present analysis, PLANET participants had mean VA return to baseline, and nearly half maintained ≥ 69 letters at year 6. The preservation of mean BCVA in the present study suggests that PDT may be positioned as a adjunctive strategy in eyes with persistent or recurrent activity, rather than a key component of treatment when IVA is used. Ultimately, the comparison between the two cohorts should be conducted with caution, as there are differences in baseline characteristics and treatment exposure within the original study and after their release (Table 3).

Table 3.

Comparison of the 6-year follow-up findings in the PLANET cohort and the EVEREST II follow-up cohort

	PLANET (n = 72)	EVEREST II monotherapy (n = 41)	EVEREST II combination (n = 49)
BCVA, mean (SD) letters
Baseline	58.1 (12.1)	65.6 (14.2)	63.0 (13.6)
Year 2	72.6 (11.5)	67.9 (14.3)	68.2 (14.1)
Year 6	57.8 (23.7)	58.5 (21.7)	56.8 (20.1)
CSFT, mean (SD) um
Baseline	361.7 (127.3)	406.7 (119.5)	457.1 (161.8)
Year 2	204.0 (46.2)	307.8 (107.2)	259.9 (86.5)
Year 6	261.2 (129.8)	332.0 (156.5)	276.5 (133.9)
Polypoidal lesion present, n (%)
Year 2	50 (69.4)	26 (65.0)	13 (26.5)
Year 6	21 (29.2)	22 (55.0)	22 (45.8)
Injection burden, mean
Year 2–6	10.2 ± 2.9	8.6 (3.8)	6.7 (2.0)
Any PDT	115.3% (11)	70%

Any PDT refers to receipt of photodynamic therapy at any time during the original PLANET study and subsequent follow-up period up to year 6

BCVA best-corrected visual acuity, CSFT central subfield thickness, PDT photodynamic therapy, SD standard deviation

A consistent observation between the two trials and extension studies was that additional treatment was required in up to three-quarters of the cohort, and roughly one-fifth of patients had poor vision (< 36 letters) in both follow-up cohorts. This highlights the fact that a subgroup of eyes are currently suboptimally managed, which is likely due to a combination of inadequate disease control by currently available strategies, recurrence, and undertreatment.

A strength of this study is the long duration of follow-up, extending out to 6 years, which is rare for randomized clinical trials for PCV. The inclusion of both functional and anatomical endpoints allows for a comprehensive evaluation of long-term disease control. Furthermore, analysis by treatment intensity provides insights into the importance of consistent re-treatment and reflects real-world treatment variability.

However, several limitations should be acknowledged. First, attrition was substantial, with only a minority of the original PLANET cohort available for long-term follow-up, raising the possibility of selection bias. However, the baseline features between the 6-year follow-up participants were similar to the overall 2-year completers, suggesting that the results may be generalizable to the rest of the cohort. A similarly low recall rate was observed in our previous follow-up study of the EVEREST II participants [8]. Second, treatment after the end of the original study was not protocol-driven but physician- and patient-determined, introducing heterogeneity in re-treatment strategies. Third, PDT was rarely applied, limiting conclusions about its role in this cohort and making comparisons with EVEREST II imperfect. Regarding anatomical readouts at year 6, there was no central reading center involvement, and so the results cannot be directly compared with those from the reading center. Finally, cross-trial comparisons between PLANET and EVEREST II should be interpreted with caution, given differences in baseline characteristics, treatment regimens, and follow-up methodologies.

Conclusions

In summary, these findings highlight the fact that PCV is a chronic, relapsing disease that requires long-term management. Given that only a small percentage of patients required PDT during long-term follow-up, these data support the effectiveness of anti-VEGF monotherapy, with about half of patients maintaining > 69 Early Treatment Diabetic Retinopathy Study (ETDRS) letters at 6 years. PDT may be considered as an adjunct for anatomical stability and to reduce treatment burden, though its impact on visual outcomes appears neutral.

Author Contributions

Conception and design of analysis: Kelvin Yi Chong Teo, Chui ming Gemmy Cheung. Data collection and processing: Kelvin Yi Chong Teo, Young Hee Yoon, Shigeru Honda, Kyu Hyung Park, Hidenori Takahashi, Ryuichi Ideta, Tetsuju Sekiryu, Akira Hirata, Akira Obana, Ryusaburo Mori, Shih-Jen Chen, Caroline Chee, James G. Wong, Timothy Y Y Lai, Ian Wong, Nick Fung, Won Ki Lee, Hiroko Terasaki, Shih-Chou Chen, Takafumi Miyatani, Tomohiro Iida, Ryo Obata, Tobias Machewitz, Sergio Leal, Chui ming Gemmy Cheung. Writing: Kelvin Yi Chong Teo, Chui ming Gemmy Cheung. Data interpretation and analysis: Kelvin Yi Chong Teo, Tobias Machewitz, Sergio Leal, Chui ming Gemmy Cheung. Review and final approval: Kelvin Yi Chong Teo, Young Hee Yoon, Shigeru Honda, Kyu Hyung Park, Hidenori Takahashi, Ryuichi Ideta, Tetsuju Sekiryu, Akira Hirata, Akira Obana, Ryusaburo Mori, Shih-Jen Chen, Caroline Chee, James G. Wong, Timothy Y Y Lai, Ian Wong, Nick Fung, Won Ki Lee, Hiroko Terasaki, Shih-Chou Chen, Takafumi Miyatani, Tomohiro Iida, Ryo Obata, Tobias Machewitz, Sergio Leal, Chui ming Gemmy Cheung.

Funding

Sponsorship for this study was funded by the National Medical Research Council Singapore Open Fund Large Collaborative Grants: NMRC/LCG/004/2018, NMRC/LCG23 May/0032 and STaR 24-Jan-0005. No funding was received for the publication of this study.

Data Availability

All data analyzed during this study are included in this published article.

Declarations

Conflict of interest

Dr. Kelvin Yi Chong Teo reports consultancy fees, honorarium, travel support and speaker fees from Bayer and Novartis outside the submitted work. Dr. Young Hee Yoon reports no relevant conflict of interest. Dr. Shigeru Honda reports support from Santen, Senju, Chugai, Alcon, HOYA, Novartis, and Bayer outside the submitted work. Dr. Kyu Hyung Park reports consultancy fees, honorarium, travel support and speaker fees from Bayer, Novartis, Alcon, Roche, Samsung Bioepis, Oculight, RetiMark outside the submitted work. Dr. Hidenori Takahashi reports consultancy fees, speaker honoraria, and travel support from Bayer and Novartis, and speaker honoraria from Santen, Senju, and Chugai, all outside the submitted work. He also reports financial support from SPS and Kyowa Kirin, and holds intellectual property and a founder position in DeepEyeVision. Dr. Ryuichi Ideta reports consultancy fees and speaker fees from Senju, Chugai, Santen, Novartis, and Kowa outside the submitted work. Dr Tetsuju Sekiryu reports support from Santen, Senju, Chugai, Alcon, HOYA, Novartis, and Bayer outside the submitted work. Dr. Akira Hirata reports consultancy fees and speaker fees from Novartis, Santen, Senju, and Kowa outside the submitted work. Dr. Akira Obana reports no relevant conflict of interest. Dr Ryusaburo Mori reports personal fees from Bayer Yakuhin Ltd., Chugai Pharmaceutical Co, Ltd., Novartis Pharma K.K, Senju Pharmaceutical Co, Ltd., Santen Pharmaceutical Co, Ltd, and Nippon Boehringer Ingelheim Co, Ltd. Dr. Shih Jen Chen reports consultation fees from Roche, Bayer, Novartis and Allergan outside the submitted work. Dr. Caroline Chee reports consultation fees from Roche, Speaker fees from Novartis outside the submitted work. Dr. James Wong reports consultancy fees and speaker honoraria from Bayer, Roche, and Apellis outside the submitted work. Dr. Timothy Lai reports consultancy fees, grants, travel support and speaker fees from Bayer, Novartis and Roche during the conduct of the study; consultancy fees, grants and speaker fees from 4DMT, Astellas, Bayer, Boehringer Ingelheim, Novartis, Ocular Therapeutix, Oculis, and Roche outside the submitted work. Dr. Ian Wong reports no relevant conflict of interest. Dr Nicholas Fung reports travel support from Alcon outside the submitted work. Dr. Won Ki Lee reports consultancy fees, travel support and speaker fees from Bayer, Novartis, Roche outside the submitted work. Dr. Hiroko Terasaki reports speaker fees from Alcon, Bayer, Santen, and Sanofi outside the submitted work. Dr. Shih-Chou Chen reports consultation fees from Roche, Bayer, Novartis and Allergan outside the submitted work. Dr Takafumi Miyatani reports no conflict of interest. Dr Tomohiro Iida reports support from Santen, Senju, Chugai, Novartis, Kyowa Kirin, Topcon and Bayer outside the submitted work. Dr Ryo Obata reports speaker fees from Bayer, Chugai, Novartis, Santen, and Senju outside the submitted work. Tobias Machewitz and Dr. Sergio Leal are employees of Bayer AG. Dr. Chui Ming Gemmy Cheung reports consultancy fees, honorarium, travel support and speaker fees from 4DMT, Astellas, Bayer, Novartis, Roche, Boehringer Ingelheim, Zeiss and Topcon outside the submitted work. Dr. Chui Ming Gemmy Cheung is an Editorial Board member of Ophthalmology and Therapy. Dr. Chui Ming Gemmy Cheung was not involved in selection of peer reviewers for the manuscript nor any of the subsequent editorial decisions.

Ethics/Ethical Approval

The study was conducted in accordance with the tenets of the Declaration of Helsinki and with approval of the respective institutional review boards of the participating centers (SingHealth IRB: R1597/96/2018). Written informed consent was obtained from each participant.