The term retinal vasoproliferative tumor (VPT, or what we call reactive retinal astrocytic tumor) was initially applied to the entity under discussion by J. Shields, and we can appreciate the Shieldses’ ongoing attachment to it. We do not wish, however, to be drawn into a sterile, medieval scholastic debate about terminology, but rather would like to keep the focus on understanding the underlying biology.

In their letter, the Shieldses make some statements about our paper that need clarification. As we pointed out, the smallest of our 4 tumors was 1.5 mm in elevation and the largest was 7.0 mm in elevation. There was no difference in the gliotic predominance among them. Our smallest lesion was a fortuitous finding and, to date, is the closest to an incipient lesion that has been described histopathologically. The 2 neuropathologists who evaluated the histopathology also thought our 4 tumors were pilocytic astrocytomas in an unusual location. Neither seriously considered a vascular tumor. Second, because the largest tumor had been subjected to numerous conservative therapies, the Shieldses attribute the gliosis in all of our lesions to the effects of local therapy. In fact, none of the others was ever directly treated, so that the gliosis was intrinsic to these reactive tumors. The main contribution of the Shieldses’ latest paper on 334 VPTs was to determine that 80% were primary and 20% secondary, attributable to an identifiable underlying ocular disease. The use of primary and secondary is curious in light of their belief, expressed in their letter, that all VPTs are reactive (ie, secondary), even if causation in the majority is cryptic or subclinical. Missing from their 334 cases was any description of an early lesion less than 2.8 mm in thickness (twice as elevated as our smallest), which would be of greatest value in addressing foundational lesional characteristics.

There are profound limitations to inferring pathogenesis from clinical observations that define a condition, no matter how detailed and meticulous they might be. For example, fluorescein angiography of small, early lesions of neoplastic astrocytic proliferations and retinoblastoma discloses a luxuriant microvasculature in each, which could lead to the misapprehension in the naïve observer that they are primarily vascular conditions. One previously reported astrocytic “neoplasm” was in retrospect actually a reactive lesion of the type beingdebated here. Thus, clinically descriptive terminology of a condition generally does not accurately reflect its pathogenesis and molecular biology. A compelling example of this limitation is offered by retinal hemangioblastoma, which for many years, based on clinical observations, was thought to be a retinal vascular condition. It was not until careful pathologic examination that the tumor was demonstrated to be a stromal cell tumor, and not a vascular tumor at all, identical to cerebellar hemangioblastoma. This was subsequently confirmed by molecular diagnostic studies. In early retinal hemangioblastoma, one cannot clinically visualize the stromal component of the tumor, which therefore gives the appearance of being primarily a vascular lesion.

We would urge the Shieldses to keep an open mind regarding their VPTs. We submit for their consideration a comparable and parallel situation from another tissue system. Reactive pyogenic granulomas commence unequivocally as vascular tumors that are ultimately overcome by progressive fibroblastic proliferation with collagenization. By contrast, nodular fasciitis is primarily a reactive proliferation of stromal myofibroblasts with a merely supportive vasculature. Pyogenic granuloma corresponds to the Shieldses’ conjecture regarding VPT, whereas fasciitis provides the better framework for our data. We hope that the Shieldses will be receptive to new data as more powerful biologic and genetic tools lead us toward a definitive resolution of our civilized controversy regarding biology and pathogenesis.