Related Experiment Video
Updated: Mar 2, 2026
Establishment and Propagation of Human Retinoblastoma Tumors in Immune Deficient Mice
Published on: August 4, 2011
Vitreous Seeds in Retinoblastoma: Clinicopathologic Classification and Correlation
Alec L Amram1, Grecia Rico2, Jonathan W Kim3
1Department of Ophthalmology & Visual Sciences, University of Texas Medical Branch, Galveston, Texas.
Purpose:
A recent classification scheme for retinoblastoma vitreous seeds has shown promise in predicting treatment response. For the first time, we correlate this clinical classification scheme with its histopathologic features.
Design:
Retrospective review.
Participants:
Enucleated eyes received at the pathology department of the Retinoblastoma Center of Houston from 2010 to 2015.
Methods:
Macroscopic photographs of the enucleated eyes of patients with retinoblastoma were analyzed to select those with vitreous seeds. Cases with adequate material for clinicopathologic correlation were selected for further analysis, and clinical photographs were reviewed. Routine histopathologic slides were reviewed and compared with the clinical and macroscopic photographs. Seeds were classified as type 1 ("dust"), type 2 ("sphere"), or type 3 ("cloud"). To confirm the presence of macrophages, CD68 immunohistochemical staining was used. Synaptophysin was used to stain retinoblastoma cells.
Main Outcome Measures:
To correlate clinical vitreous seed type with histopathologic features.
Results:
A total of 14 eyes with adequate amounts of tumor seeds along with clinical and macroscopic photographic correlation were selected from a total of 138 eyes reviewed. Type 1 seeds consisted of individual viable tumor cells and scattered macrophages. Type 2 seeds consisted of 2 submorphologies: spheres with viable cells throughout and spheres with an outer rim of viable cells but necrotic cells centrally. Type 3 seeds were composed of more than 90% necrotic material admixed with few macrophages and viable cells at their outer rim. Untreated (8/14) and previously treated (6/14) eyes showed similar histopathologic features for each type of seeds. Treated eyes had more type 1 and 3 seeds.
Conclusions:
We provide the first histopathologic correlation of the clinical classification scheme for vitreous seeds in retinoblastoma. "Dust" is formed by scattered single cells alternating with macrophages. "Spheres" with translucent centers contain multiple layers of viable tumor cells that shed single cells and may be more clinically aggressive. "Cloud" seeds are mostly composed of necrotic material, explaining their lack of therapeutic response. Pretreated eyes showed tumor seeds morphologically similar to untreated eyes. Knowledge of the underlying histopathology of vitreous seed types is a fundamental component of classification and may aid in understanding clinical response to treatment.
