We read “Reactive Retinal Astrocytic Tumors (So-called Vasoproliferative Tumors): Histopathologic, Immunohistochemical, and Genetic Studies of Four Cases” by Poole Perry and associates. They report histopathology following enucleation in 4 cases. They accurately demonstrated that the lesions were composed of glial cells (mostly astrocytes) with scarcity of blood vessels. Their observations confirm other reports demonstrating that end-stage retinal vasoproliferative tumors following failed conservative treatments are composed predominantly of glial cells. They showed a fundus photograph of 1 case in which enucleation was performed 16 months after cryotherapy and many intravitreal injections of bevacizumab. We suspect that these treatments, designed to reduce tumor vascularity, were responsible for the histopathologic findings of extensive gliosis.
Although the clinical features are well known, the pathogenesis, pathology, and terminology are still controversial. We continue to believe that retinal vasoproliferative tumor is not a true neoplasm, but rather represents florid retinal neovascularization, probably driven by vascular endothelial growth factors (VEGFs), that prompt the lesion to assume tumorous proportions. As in other retinal insults, neovascularization is gradually replaced by gliosis, resulting in what appears to be a glial neoplasm.
We have had extensive clinical experience with retinal vasoproliferative tumors and have published our experience with 334 tumors. Initially, they appear as a pink mass in the sensory retina, fed by a retinal artery and drained by a retinal vein, often with yellow intraretinal and subretinal lipoproteinaceous exudation. These features are entirely compatible with a lesion of retinal vascular origin and they are not compatible with gliosis.
We believe that retinal vasoproliferative tumor begins as retinal neovascularization, and this is further supported when one considers the entities that have been associated with secondary retinal vasoproliferative tumors, including intermediate uveitis, Coats disease, and familial exudative vitreoretinopathy, to name a few. These start as retinal vasculopathy without clinical evidence of gliosis. Eventually, gliosis can occur either by natural course or because of cryotherapy, laser, photodynamic therapy, irradiation, or anti-VEGF drugs, all therapies designed to induce vascular regression. Hence, by the time the more advanced cases come to enucleation, one would anticipate the glial component to prevail and the vascular component to be sparse, as observed by the authors.
Our final point concerns terminology. In their title, and in the text, the authors use the term “so-called vasoproliferative tumor.” To our knowledge, there is no histopathology available for early untreated cases, and from our extensive clinical experience, retinal vasoproliferative tumor is a reactive vascular mass without gliosis in the early stages and the astrocytic proliferation occurs as a late phenomenon. We, and others, have previously remarked that this lesion is most likely a reactive secondary retinal vascular mass and not a primary retinal or glial neoplasm.
For the above reasons, we believe that the term “reactive retinal astrocytic tumor,” proposed by the authors, represents the end stage of this condition and should not replace the more appropriate term “retinal vasoproliferative tumor.” The latter term is well established in the literature and is more accurate for the majority of tumors (99%) that are controlled by conservative methods and never reach the pathology laboratory. We thank our colleagues for their contribution regarding the postenucleation histopathology of advanced retinal vasoproliferative tumors.