[The fundus autofluorescence of retinal astrocytic hamartomas in tuberous sclerosis complex]

Objective: To investigate the autofluorescence findings of retinal astrocytic hamartoma (RAH) in patients with tuberous sclerosis complex (TSC). Methods: It was a retrospective case series study. Twenty-three patients (35 eyes) who were referred to Department of Internal Medicine and Department of Ophthalmology, Peking Union Medical College Hospital between November 2012 and June 2018 with established TSC-associated RAH diagnosis were included. The findings of fundus autofluorescence, fundus photos and spectral-domain optical coherence tomography (SD-OCT) were retrospectively reviewed. RAH lesions were classified into three types based on the morphology shown in fundus photos. The fundus autofluorescence features of TSC-associated RAH were described. The Welch's test and Fisher's exact test were used for statistical analysis. Results: The patients were 8 males and 15 females aged (28±9) years old (range, 15-55 years). Seventy-two RAH lesions were examined, including 59 type 1 RAHs, 7 type 2 RAHs and 6 type 3 RAHs. According to fundus autofluorescence, type 1 RAHs could be further divided into reduced, speckled and background autofluorescence patterns, among which the hypoautofluorescence pattern accounted for the majority (69.5%, 41/59), while the speckled pattern was usually accompanied by outer retinal disorganization and discontinuation of photoreceptor outer segment as revealed by SD-OCT. No significant difference was revealed in tumor thickness for reduced, speckled and background autofluorescence patterns of type 1 RAHs [(490.2±97.9) vs. (589.2±221.6) vs. (463.0±76.2) μm respectively, F=1.426, P=0.283]. Among type 1 RAHs, the number of reduced autofluorescence pattern lesions found in perifoveal, peripapillary, inferonasal, inferotemporal, superonasal, superotemporal quadrants were 9, 4, 4, 7, 4, 13 respectively, while that of speckled autofluorescence pattern lesions were 3, 0, 3, 2, 3, 2 and background autoflurorescence pattern lesions 3, 0, 1, 1, 0, 0. No significant difference was revealed in location distribution (P=0.452) either. Type 2 RAHs featured numerous hyperautofluorescent spots or plaques, and calcification in type 2 RAHs varied in autofluorescence intensity. Type 3 RAHs, combining the features of type 1 and 2 RAHs, were characterized by central hyperautofluorescent spots and hypoautoflurescent rim, but the area of hyperautofluorescence was smaller than that of calcification as shown in fundus photos. Conclusions: In TSC, the fundus autofluorescence of RAHs varies from hypoautofluorescence to hyperautofluorescence patterns according to RAH types. The retinal involvement and calcification degree of TSC-associated RAHs could be reflected on the autofluorescence, which was beneficial to the full assessment. (Chin J Ophthalmol, 2020, 56: 211-216).