Many studies have reported ocular manifestations in NF-1 patients, including choroidal abnormalities [9-16]. However, most of these cases are included in histopathologic studies or have abnormalities that are definitely observable with the fundoscope. Choroidal neurofibromatosis is thought to be a rare form of neurofibromatosis that affects the eyes [9,10,14-16]. The actual prevalence of this disorder may be higher than currently believed because patients with good vision tend not to undergo ocular examinations. Thus, without an ocular examination, which usually involves a conventional ophthalmoscopic examination or FA, patients with good vision are not tested for choroidal abnormalities. In some patients, the choroid neurofibromatosis may appear as yellowish nodules, but these may appear as bright patched areas upon fundoscopic examination in only weakly affected patients. FA shows irregular hyperfluorescence that is related to retinal pigmented epithelium (RPE) alteration, but it is not able to detect underlying choroidal abnormalities.

ICGA has overcome the limitations of FA in regard to imaging of the choroidal structure [17-20]. Several authors have previously reported the presence of choroidal neurofibromatosis with the use of ICGA. Yasunari et al. [21] have reported that they were able to detect choroidal abnormalities in all 17 of their patients with NF-1 using scanning laser ophthalmoscopy (SLO) and ICGA, and that the choroid was one of the structures most commonly affected by NF-1. They suggested that the hypofluorescent areas on the ICGA, that appeared to be obscure with SLO and FA in their less affected patients, were probably not caused by disturbance of the choriocapillaris circulation, but rather by the presence of refractile tissue or material in the choroid. They believed that such refractile material presumably blocks the fluorescence from the underlying choroidal vessels and produces a hypofluorescent appearance which disappears on the ICGA at the late phase due to the indocyanine-green dye being extravasated into the choroidal stroma. In contrast, Rescaldani et al. [10] have suggested in their study that the early hypofluorescent areas in the ICGA are due to the hypo- or non-perfused lobules of the choriocapillaris rather than the blocking of fluorescent light by the neurofibromatose choroidal nodules since these areas became smaller or disappeared in the late phase and because the deep choroidal vessels were visible within these areas. Additionally, one histological study has documented that Schwann cell growth with dysplasia of the smooth muscle cells of the arteriolar wall leads to the gradual narrowing of the vessel lumen [22]. According to Rescaldani et al. [10], chronic hypoperfusion of the choriocapillaris appears to cause lesions to the overlying RPE and cause its eventual atrophy (yellowish patches on ophthalmoscopy that correspond to the hyperfluorescent areas in the FA).

In our case 1, the fundus did not show yellowish nodules but rather bright patched areas around the posterior pole, similar to Yasunari's cases. These lesions appeared to be hyperfluorescent areas in the FA, which suggests RPE alterations are present. However, the ICGA showed diffuse hypofluorescence with visible deep choroidal vessels from the early phase to the late phase and was unrelated to the lesion observed by both fundoscopy and FA. This finding is believed to be due to the diffuse scattering of the neurofibromatose tissue in the choroid without remarkable prominence. In regard to case 2, the abnormal alterations were not clearly visible by either fundoscopy or FA. However, the ICGA showed multiple hypofluorescent areas that persisted into the late phase. It is thought that multiple neurofibromas occupying the choroid were detected by the ICGA, and that these neurofibromas were not yet producing RPE abnormalities. However, it is insufficient to conclude that these abnormalities presented with choroidal neurofibromas because various possibilities exist for the underlying cause. Choroidal involvement in NF-1 may occur more often than what is currently believed. Also, the underlying cause of this, whether it be due to the space-occupying neurofibroma itself or the vascular change associated with the neurofibromatosis, may lead to the hypoperfusion of the lesion that would later result in RPE alteration. We believe that these ICGA patterns that are seen in our cases may not be due to hypoperfusion or nonperfusion of the choroidal lobules, but due instead to blocked fluorescence by the choroid-occupying materials.

In conclusion, these two cases have shown that choroidal abnormalities occur more often than currently believed in patients with NF-1. Fundoscopic examination can identify grossly proliferative neurofibromatosis but only in the later stages, and FA can only describe actual RPE alterations, secondary to choriocapillaris nonperfusion. However, ICGA is a useful tool to detect the early changes that occur with choroid abnormalities. Although no abnormal finding was found by either fundoscopy or FA, ICGA seems to be useful for ocular examinations of NF-1 patients that have good vision.