The development of anti-vascular endothelial growth factor (anti-VEGF) agents represented a dramatic breakthrough in the treatment of neovascular age-related macular degeneration (nAMD) and multiple phase III pivotal trials have established the efficacy of these agents. However, replicating these trial results in real world practice has been challenging and this has been documented in numerous real world studies [1,2,3,4,5,6,7,8].

The resulting efficacy gap between explanatory clinical trials and real world clinical practice provided strong impetus for a pragmatic treat & extend (T&E) paradigm. Most large T&E trails employed easy to replicate re-treatment criteria that could be successfully implemented in clinical practice [9,10,11,12,13,14]. A recent meta-analysis also demonstrated that T&E can achieve similar functional and anatomic outcomes compared to fixed or prn dosing regimens [15]. Consequently, there has been widespread adoption of T&E by clinicians globally [16].

In this issue of the EYE, Fu et al. present a large retrospective cohort study from Moorfields Eye hospital where they demonstrate that 68% of eyes could achieve at least one q12 dosing interval with aflibercept in treatment-naïve nAMD patients [17]. This study and others demonstrate that some lesions are more VEGF dependent than others.

It is imperative that clinicians continue to evolve and incorporate new knowledge to further refine and optimize this individualized approach to patient care. What are some important questions that need further evidence generation and evidence synthesis as it pertains to T&E in nAMD management?

  1. 1.

    Do we always need a loading phase? There are biological plausibility arguments to support a loading phase. Pivotal trials such as the VIEW trials with intravitreal aflibercept have demonstrated that intra-retinal fluid (IRF) reached its lowest level 1 week after the first injection, however sub-retinal fluid (SRF) could take up to 3 months to reach its trough [18,19,20]. Since IRF and SRF are important barometers to judge VEGF suppression in clinical practice, it seems logical to reduce fluid levels to the minimum possible level with 3 monthly loading doses. Additionally, visual acuity at month 3 appears to consistently be an important predictor of visual acuity outcomes at month 12 and month 24 [21, 22]. However, there are also valid arguments against a fixed loading dose in all nAMD cases. The CATT trial demonstrated that approximately 14% of patients in the PRN arm needed 3 or less injections in first year of treatment [23]. In addition, aggressive loading at the start does not appear to alter the underlying disease process in terms of long term anti-VEGF requirement for the patient. As such, individualizing the loading phase in future nAMD trials may provide further opportunity to optimize outcomes and reduce treatment burden.

  2. 2.

    What criteria should be used to define disease activity? Historically, treat and extend studies have defined active disease as presence of new macular hemorrhage, vison loss of > 5 ETDRS letters and/or presence of intraretinal fluid (IRF) or subretinal fluid (SRF) [11, 24,25,26]. However, what about other biomarkers such as new subretinal hyperreflective material (SHRM) or an enlarging serous pigment epithelial detachment? Evidence suggests that both of these are poor prognostic factors that can predict future fibrosis and exudation respectively [27,28,29]. Improvements in multi-modal imaging and quantitative analysis of biomarkers will be important for further refining disease activity criteria.

  3. 3.

    Differential impact of fluid compartments and how that might guide re-treatment intervals? A recent systematic review on the topic demonstrated either a positive or no negative effect of SRF at baseline and during the treatment cycle on visual acuity outcomes. However, IRF was consistently associated with worse VA outcomes [30]. The FLUID study demonstrated that tolerating up to 200 microns of subretinal fluid at the fovea resulted in non-inferior visual acuity outcomes compared to aggressively treating sub-retinal fluid [31]. Optimizing T&E algorithms in terms of fluid compartments needs further validation and evidence generation.

  4. 4.

    Can we further “individualize” the extension intervals for T&E? Vast majority of T&E trials extended stable patients at 2 week intervals. Applying a homogenous extension interval for all patients seems counter-intuitive to the rational behind an individualized care pathway that is the crux of a well designed T&E algorithm. What factors might dictate if stable patients could be extended not only at two week intervals, but potentially at three or four weeks?

    One important variable might be the anti-VEGF agent itself. Data suggests that different agents have different anti-VEGF suppression time [32]. As newer agents get regulatory approval, it will be important to generate evidence to assess whether extension intervals beyond the classic 2 week interval could be employed.

    Another variable might be the lesion type. Should type 3 (retinal angiomatous proliferation) lesions that are exquisitely sensitive to anti-VEGF therapy be treated with the same loading phase intensity and extension intervals as the more treatment resistant type 1 choroidal neovascular membranes [33]? Is tolerating stable SRF in the context of type 1 CNVM have different implications than tolerating stable SRF in type 3 CNMV? A recent retrospective study identified presence of SRF at baseline was the most significant independent negative predictor for visual outcomes for type 3 lesions [34].

  5. 5.

    How might technological advancements refine and improve the T&E paradigm? Disruptive technologies such as home OCT monitoring have the potential to dramatically modify many aspects of traditional T&E paradigms. Day to day remote monitoring of fluctuations in fluid status can potentially allow for increased precision in diagnosing early recurrence and allow for early identification of a patient’s ideal re-treatment interval, thus further optimizing T&E paradigms. There might also be further advantage of reducing fluid fluctuations by triggering earlier treatment before significant fluid develops. Multiple post-hoc analysis including those from the CATT and IVAN trials have demonstrated a negative association between large fluid fluctuations and visual acuity [23, 35]. Similarly, advances in Artificial Intelligence may allow more accurate prediction of a patient’s anti-VEGF “need” and may in-turn guide extension intervals.

  6. 6.

    How to incorporate new treatments for Geographic Atrophy (GA) within the T&E paradigm for anti-VEGF treatment for nAMD? One of the most important recent advancements recently in the field of AMD management has been a positive signal of a treatment affect for complement inhibition in reducing GA progression [36,37,38,39]. Given that GA is a common cause of progressive vision loss in nAMD patients [40], it will be imperative that any approved agent for GA can be incorporated into commonly used T&E algorithms for nAMD. This combination treatment will require innovative pragmatic trial design to assess how best to implement this in clinical practice. Which patients with nAMD undergoing anti-VEGF therapy should be candidates for combination treatment for GA atrophy? How best to combine intravitreal treatment for GA along with anti-VEGF treatment for nAMD in a T&E setting? How will this impact the treatment burden and patient compliance?

In summary, the development of pragmatic T&E paradigms for management of patients with nAMD has been a truly impactful step forward for patient care globally. However, as the art and science behind nAMD management continues to evolve, there will be an ever increasing need for robust, pragmatic randomized clinical trials that can move the field forward and further optimize treatment burden and patient outcomes.