Clinicopathologic Correlation of Progressive Fibrovascular ProliferationAssociated With Occult Choroidal Neovascularization in Age-Related MacularDegeneration

The development of fibrous tissue in patients with occult choroidalneovascularization (CNV) associated with age-related macular degeneration(AMD) is one of the major determinants of vision loss. Usually this processis slow with the development of disciform scarring that exceeds 50% of thelesion occurring in only 20% of eyes by 12 months.¹ Thetypical time course is between 30 and 100 months of follow-up.² Interestingly,in the absence of subretinal blood or classic CNV, no eyes with occult CNVdeveloped more than 50% disciform scarring in a 9- to 12-month follow-up.³

In the present case, we describe clinicopathologic findings in a patientwith AMD who developed progressive subretinal fibrovascular proliferationassociated with occult CNV.

A 76-year-old man had a 3-month history of vision loss and a visualacuity of 20/60 OD as well as turbid subretinal fluid in that eye (Figure 1A). The fellow eye demonstrated avisual acuity of 20/30 and the retinal examination findings in this eye showeddrusen and pigmentary changes of the retinal pigment epithelium (RPE). Onstereofluorescein angiography, leakage of undetermined origin at the levelof the RPE consistent with a 4-disc area of occult CNV was identified in theright eye (Figure 1B and C). Theright eye underwent a rapid and progressive fibrosis over the ensuing 6 months(Figure 1D-F) with visual acuitydeclining to 20/400. The patient underwent vitrectomy with removal of thesubretinal neovascular fibrotic scar.

The surgical specimen measured 4 × 3.5 × 1.2 mm. Microscopicanalysis of the neovascular complex demonstrated a well-demarcated lesionwith a 2-component fibrous scar (Figure 2A). The overall thickness of the lesion was approximately 1200 µmwith the subretinal and intra-Bruch's membrane components measuring 1100 and100 µm, respectively. Fibroblastlike cells were evident throughout theextensive matrix of collagen. Many microvascular channels were prominent inthe region external to the RPE and within the outer subretinal area. The innermostregion of fibrosis was relatively acellular and contained few vascular structures.

DAPI staining for cellular nuclei demonstrated striking hypercellularityin the subRPE component of the scar (Figure2B). Many cells in this area were positive for CD68⁺,a macrophage marker (Figure 3A).The neovascular structures stained positively for CD34⁺, a markerfor microvascular endothelium and hematopoetic stem cells (Figure 4A), and for von Willebrand factor (vWF) (Figure 4B), a marker for differentiated endothelial cells. Numerousindividual cells identified as CD34⁺ cells were also located throughoutthe scar. Within the complex, multiple individual isolated cells were CD34⁺ but vWF negative, indicating possible immature endothelial cells (Figure 4A and B). Tissue stains for vascularendothelial growth factor (VEGF) and pigment epithelium–derived factorwere negative but strongly positive for platelet-derived growth factor (PDGF)B within neovascular endothelial cells (Figure3B). The VEGF from the subretinal fluid was below the detectablelevels by enyme-linked immunosorbent assay (Quantikine Human VEGF Immunoassay;R&D Systems, Inc, Minneapolis, Minn).

This patient demonstrated rapid fibrosis of an occult fibrous neovascularmembrane. The histopathology of the present lesion contained microvascularchannels, a prominent fibrotic response, hypercellularity of the tissue comprisingboth inflammatory and endothelial cells, and expression of PDFG B within vascularchannels.

A hypercellular response was most prominent in that portion of the scaradjacent to the RPE. Part of the cellular responses appeared to be a prominentinflammatory component consisting of CD68⁺ macrophages. This findingsuggests that these cells may participate in the evolution of fibrous neovascularmembranes and support the inflammatory paradigm of CNV in AMD. The originof vascular elements contributing to the subretinal fibrovascular membranein AMD remains uncertain. In the present case, endothelial cells located withinthe microvascular channels were strongly positive for CD34⁺ andvWF. Many individual cells isolated within the fibrous matrix of this membranewere CD34⁺ but vWF negative. In addition, many endothelial markerpositive cells were identified as isolated, individual cells rather than embeddedin a neovascular channel. These findings, in this case, suggest that a cellularrather than a vascular invasion may have played a role in the pathogenesisof the neovascular structures. In addition, while vWF staining is specificto endothelial cells, CD34⁺ also serves as a marker for hematopoeticstem cells. Recently, work with stem cells has highlighted the capacity ofthese cells to differentiate into vascular elements both in vitro and in vivounder the influence of VEGF and PDGF.⁴ Whetherthis process occurs in the pathogenesis of CNV remains speculative.

Consistent with a previous report on cytokine production in CNV undergoinginvolution,⁵ this case exhibited a lackof VEGF expression both by immunohistochemistry and quantitative enzyme-linkedimmunosorbent assay. However, prominent PDGF staining was detected withinthe vascular channels. In summary, this surgically excised fibrovascular membranedemonstrates additional findings that add to our knowledge of the molecularbiology of neovascular AMD.

None of the authors has a proprietary interest in any test or productdescribed within this article.

Corresponding author and reprints: Karl Csaky, MD, PhD, Bldg 10,Room 10N119, NEI/NIH, 9000 Rockville Pike, Bethesda, MD 20895-1857 (e-mail: kcsaky@helix.nih.gov).