COMMENT & RESPONSEIn Reply Thank you very much for your interest in our recent publication. 1 We are happy to answer each of your questions in order.1. An injection was administered 8 weeks before optical coherence tomography angiography (OCTA) imaging and 8 weeks before date of death in the right eye. 2. Eight injections at about 8-week intervals were administered between OCTA imaging and date of death over 16 months. Before undergoing OCTA imaging, the patient had received 28 intravitreal injections over more than 3.5 years, indicating a chronic lesion with chronic changes. Substantial changes of the macular neovascularization structure in such a chronic complex are not likely. 3. We agree that connections between the deep capillary plexus (DCP) and the choriocapillaris are associated with advanced stages of disease. Choroidal contributions have been confirmed in surgical specimens 2 and in clinical imaging studies of treated eyes. 3 Thus we stated: "Our current findings do not exclude such a connection in general or in other cases." 1 We could have also said (but did not) that we did not see DCPoriginating vessels penetrate through basal laminar deposit (BLamD) and enter the sub-RPE-basal laminar (BL) space. As cited in the letter to the editor, Sacconi et al 4 described how DCP-originating flow signal extends past the RPE (and also BLamD) during periods of exudation. 4. Figure 1B of our article 1 shows a type 3 lesion with flow signal, positioned over calcified drusen that masquerade as a double layer sign. We chose to detail this lesion, and others in this case, in a separate manuscript. The type 3 lesion in Figure 1B spans a large area. As we state as a limitation, 1 our use of stepped sections may have missed connections between the retinal and choroidal circulations, and as mentioned above, may have also missed connections between the DCP and sub-RPE-BL space. Lack of serial sections is a trade-off we accepted in using epoxy resin embedding for high-resolution light microscopy covering a large area. However, in all analyzed tissue sections of the Figure 1B lesion, both BLamD and the Bruch membrane were intact. Furthermore, we did not see OCTA flow signal in the sub-RPE-BL space under this lesion, perhaps because we used volumetric projection artifact removal. This technique should reduce the chance of seeing spurious signal directly under type 3 neovascularization. This is especially important for accurate localization of flow in instances where the RPE-BL is not elevated by a hyporeflective structure, such as sub-RPE fluid or a calcific nodule. Sacconi et al, 4 who reported many treatment-naive eyes with flow signal in the sub-RPE-BL space, at the same time mentioned lack of pro-