Article Text


Retinoblastoma and retinal astrocytoma: unusual double tumour in one eye
  1. S M Imhof1,
  2. A C Moll1,
  3. P Van Der Valk2,
  4. A Y N Schouten-van Meeteren3
  1. 1Department Ophthalmology, University Hospital Vrije Universiteit, Amsterdam, Netherlands
  2. 2Department of Pathology, University Hospital Vrije Universiteit, Amsterdam, Netherlands
  3. 3Department of Pediatric Oncology, University Hospital Vrije Universiteit, Amsterdam, Netherlands
  1. Correspondence to: Saskia M Imhof, Department of Ophthalmology, University Hospital Vrije Universiteit, De Boelelaan 1117, NL-1081 HV Amsterdam, Netherlands; s.imhof{at}

Statistics from

In the recent literature there is controversy regarding the histopathological origin of retinoblastoma (RB) and retinal astrocytoma (RA). The common origin of both tumours from a multipotential stem cell has been studied in RB cell lines Y-79 and fresh RB materials with immunohistological techniques using GFAP (glial fibrillary acid protein), NSE (neuron specific enolase), photoreceptor cell markers (S-100, myelin basic protein), and synaptophysin.1–8 In immunohistopathological studies indications have been found for differentiation of RB into a neuronal and a partial glial pathway.1–8

To our knowledge no reports have been published on the simultaneous occurrence of both an RB and an RA in one eye. We present a case that demonstrates one eye of a 5 year old girl containing those two immunohistologically different tumours.

Case Report

In a 5 year old white girl a divergent strabismus of the right eye was discovered. No family history of RB or phacomatosis was present. Funduscopic examination under anaesthesia showed a central white mass. Nasally inferior in the peripheral retina a second large whitish mass with vitreous seeding on the top was discerned without abnormal vasculature.

On DNA investigation no mutation in the RB1 gene was found with SSCP (single strand conformational polymorphism) technique.

Enucleation was performed within 1 week after the diagnosis RB was established.

Histologically, the first, most anterior tumour was largely necrotic, but highly cellular (Fig 1A). Only few islands of vital cells with a central blood vessel were seen. The tumour cells had medium sized nuclei, with a very dense chromatin. The nuclear shape was oval to slightly irregular. Cytoplasm was scanty. There were many apoptotic tumour cells, and mitotic figures and occasional calcifications were present (Fig 1B).

Figure 1

Retinoblastoma. (A) Largely necrotic tumour with medium sized, dark stained cells, a typical picture for retinoblastoma (haematoxylin and eosin, ×80). (B) Cellular detail showing medium sized, hyperchromatic nuclei (haematoxylin and eosin, ×350).

The second and more posteriorly situated tumour was quite different. It showed areas with high cellularity, without necrosis. These high cellular areas were separated by areas of low cellularity, that appeared fibrous, or compact fascicular. Many calcifications were present (Fig 2A). The cells in the cellular areas had fairly small round nuclei (though slightly larger than the neuronal cells in the neighbouring retina), with a finely dispersed, somewhat granular chromatin, and occasional nucleoli. The cells in the compact fascicular areas were more spindle-shaped, but with a similar chromatin pattern. No mitotic or apoptotic figures were found (Fig 2B).

Figure 2

Retinal astrocytic tumour. (A) Glioma showing fibrillary areas and more hypercellular areas closely intermingled here; many calcifications are present (haematoxylin and eosin, ×80). (B) Nuclei are small and somewhat elongated. Their chromatin is somewhat finer, than in the first tumour (haematoxylin and eosin, ×350).

Immunohistochemically, the first, peripheral tumour stained positive for synaptophysin and negative for GFAP and neurofilament. The second, central tumour showed a negative synaptophysin, but showed strong staining for GFAP, especially in compact areas. The cells in the cellular areas only occasionally stained positive for GFAP.


Both the histology and the immunohistochemistry showed two entirely different tumours, one (the anterior tumour) clearly an RB, largely necrotic. In the second tumour, a glioma, with cellular areas resembling oligodendroglioma, the more compact “fibrous” areas resembled a pilocytic astrocytoma.

As the morphology of the two tumours clearly differed the occurrence of a “seedling” of RB, though theoretically possible, in itself is an unusual sequence of events and is not likely here. The differentiation would have increased the staining intensity for synaptophysin and may even have caused a positive neurofilament staining, but in our case both neuronal markers were negative. Also, the growth pattern of the second tumour with its intermingling of small cells (putative retinocytal) and glial components strongly suggest two different tumours.

However, that does not mean these two tumours cannot be related. We feel this case illustrates the discussion whether RB and RA are two totally different tumours, both originating in the retina, or a transitional form of simultaneous expression of neuronal and glial cells from a germline stem cell.1–8

The differentiation between RA and RB and the possible transitional form derived from the specific glial supporting cells (Mueller cell) were investigated on RB Y-79 cell lines and tumour material.1,3

The view that RBs are neuronal tumours, although sporadic cases may show immunohistochemical signs of partial astrocytic or Mueller’s cell differentiation, has been supported by several studies.2,4–6,8

In agreement with this finding, Gonzales-Fernandez et al report the in situ expression of photoreceptor and glial cell associated proteins in fresh RB tumour material.7

Their data suggest that RB has the histogenetic potential of the immature neural retinal epithelium which can give rise to both photoreceptor and Mueller cell lineages.

These studies suggest that indications for differentiation of RB into a neuronal and a partial glial pathway.

In our opinion it has been demonstrated by many studies that both retinal photosensory cells and Mueller glial cells are derived from a neuroretinal epithelium stem cell, with the capacity of different expression by immature retinal cells. Perhaps the occurrence of a double tumour, as our case, is therefore not so surprising. However, this case suggests, showing the clinical form of different expression of a germline retinal neuroectodermal stem cell, that differentiation of immature retinal cells could be a tightly regulated process.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.