The development of new CNV is the single most important cause of acute visual loss in patients with AS.12,17 Its common occurrence in young adults of working age (50 years of age or less) is even more relevant because its high incidence of recurrence, potential for scarring and permanent visual disability may damage a patient’s employment prospects, cause significant loss of personal revenue, disrupt the patient's family dynamics, and worsen his or her quality of life.12,13,18–20 The current study describes the long-term visual outcome, response to anti-VEGF treatment, and rate of disease recurrence over time in a group of patients with AS-CNV. Moreover, the risk of recurrence was calculated according to qualitative analysis of CNV lesions by OCT and OCTA, and the findings were compared against those of a group of patients who were long believed to exhibit a similar evolution and response to treatment (mCNV).
The results reported herein confirm that although AS-CNV did have a similar appearance on structural OCT to that observed in patients with mCNV (predominantly type 2 lesions) and although patients in the mCNV group had a similar rate of disease recurrence in a shorter period of time (4 years), the visual outcome at the end of the 6-year follow-up in the AS-CNV group was significantly worse than that in the mCNV group (1.095 vs. 0.6 logMAR). Moreover, they needed, on average, significantly more anti-VEGF intravitreal injections for the inactivation of the disease (3.5 ± 1.5 vs. 1.4 ± 0.2) and for the treatment of each recurrence episode. Finally, more patients had three or more episodes of recurrence during the entire length of follow-up in the AS-CNV group than in the mCNV group. The increased number of intravitreal injections in the AS-CNV group highlights the reactive nature of the CNV in such cases, and the significant challenge of the current therapeutic strategy (PRN) is avoiding future episodes of recurrence. Therefore, we hypothesized that a more proactive approach, such as a fixed interval regime or even a treat-and-extend regimen, could potentially lead to better visual and anatomical outcomes in the long term. We currently have at our disposal longer-acting anti-VEGF agents with higher molar concentrations, such as faricimab (6 mg) and aflibercept (8 mg), which can deliver longer exudate-free periods.21,22
According to the OCTA findings, patients with AS-CNV had a predominance of mixed-type morphology at the time of diagnosis, while patients in the control group had more interlacing and cogwheel morphologies. Our results also showed that patients with AS-CNV with mixed and cogwheel morphologies at presentation had an increased risk for disease recurrence during follow-up (OR 1.2 and 4.75, respectively). Although our sample size is representative of a relatively uncommon disease with a substantial follow-up length, the observed patterns failed to achieve significance. In the mCNV cohort, interlacing patterns were also associated with an increased risk of disease recurrence during the follow-up period (OR 2.0), but the difference was not significant.11
OCTA has demonstrated significant sensitivity and specificity for the early detection and characterization of CNV in numerous pathologies.23,24 However, the common occurrence of atrophic and fibrotic changes associated with AS-CNV progression makes the early detection of recurrences especially challenging, even when OCTA imaging is available.3,11 These changes could also be responsible for the apparent failure of anti-VEGF therapy in AS-CNV patients.3,11
In a retrospective study by Marchese et al., the authors described the 12-month follow-up of a group of 19 patients with AS-CNV.11 In addition to revealing the importance of qualitative studies of CNV lesions, the results of these studies are similar to the data reported herein. The final BCVA at the end of the 12-month follow-up in their group was very similar to that reported in our AS-CNV group after the first inactivation of the disease (0.42 ± 0.4 vs. 0.55 ± 0.5 logMAR);11 the same number of intravitreal injections were required to achieve this outcome (4 vs. 4.6),11 but with the difference of having a better BCVA at baseline compared with that reported in the present study.11 The marked difference between the final BCVA reported by Marchese et al. and that reported by us could be explained by our significantly longer follow-up time. This allowed the atrophic and fibrotic changes associated with disease recurrence to appear. Another important difference from our study is that Marchese et al. recognized two potential markers of neovascular activity that we did not account for: the presence of vascular branching plaque and a perilesional dark halo, which were observed in 63% and 58% of their samples, respectively.11 Another potential predictor marker for CNV activity that was unaccounted for in our baseline observations was the presence of densely packed capillary-like vessels called irregular vascular networks, as reported by Corbelli et al. and El Matri et al.3,25 At present, it is not clear whether the inclusion of these three markers during our baseline observations would have increased our ability to predict recurrence. Nevertheless, in the future, it may be possible to combine such biomarkers with the observed morphology by OCTA at baseline in a mathematical index/quotient to improve the individual power of these biomarkers to predict recurrence.26
Regarding CNV morphology determined by OCTA, Chapron et al. described the mixed type as the predominant form of CNV in their study, in concordance with our findings.5 Although Chapron et al. reported the interlacing pattern as more likely to be associated with disease activity and exudation,5 we believe that this does not necessarily translate to a higher risk of disease recurrence during follow-up. However, it is possible that the observation of this highly vascular morphology with straight fine vessels and no vascular loops at presentation, a morphology also observed by Falfoul et al. and Gal-Or et al. in their respective studies, could serve more as a predictor of the acute response to anti-VEGF treatment.7,8
Finally, in addition to its retrospective nature, our study has several limitations that we would like to acknowledge. There was a lack of treatment standardization across both groups, which suggested that the anti-VEGF agent was selected according to physician preference; all patients had individualized intervals between office visits during follow-up; general guidelines about what was considered a recurrence were followed; and the decision regarding the need for new anti-VEGF treatment was made by the physician on duty rather than by a blinded observer. Moreover, although each patient was instructed about signs and symptoms for early disease recurrence recognition, it is possible that some of the patients in the study group could have waited too long before seeking treatment. The ability of these patients to recognize such symptoms or slight variations in their BCVA could have also been impaired further due to recently developed scarring and atrophy throughout the follow-up. Such difficulty could also have mounted over time with each recurrence episode, negatively affecting the functional outcome of the study group. Likewise, the fact that we excluded patients with less than 12 months of follow-up possibly introduced unintended selection and survival bias. Patients with stable AS-CNV and good BCVA could have opted to not attend their corresponding follow-up visit, which pushed the mean toward a worse outcome. In contrast, patients who returned for their follow-up visit could have had more unstable disease and been more prone to recurrence and thus a negative outcome.