Retinoblastoma: Clinical Presentation and the Role of Neuroimaging
==================================================================

* Joan M. O'Brien

Retinoblastoma is the most common intraocular tumor of childhood, but it remains a rare disease. Retinoblastoma occurs in one of 15,000 to 20,000 live births (1). Approximately 200 new cases a year are diagnosed in the United States. The disease presents in infancy or early childhood, with the majority of cases diagnosed before the age of 4 years (2). The disease rarely occurs in older children or in adulthood. Unilateral disease occurs in approximately two thirds of all patients, whereas bilateral retinoblastoma comprises approximately one third of the total (2).

The most frequent presenting sign of retinoblastoma is leukocoria, a white pupillary reflex. Other presentations include strabismus (crossed or deviating eyes), decreased vision (particularly in bilateral cases), appearance of inflammation, retinal detachment, glaucoma, hypopyon (tumor cells anterior to the iris), or ocular pain (3).

The genetics of retinoblastoma follow Knudson's two-hit hypothesis. Patients who have unilateral, unifocal disease have mutations at the retinoblastoma locus in both alleles within a single retinal cell. This is an unlikely event; therefore these tumors are unifocal and unilateral. In contrast, patients who harbor an underlying germline mutation develop tumors that are characteristically multifocal and bilateral. These patients carry a mutant Rb gene in every cell of their bodies. Loss of heterozygosity tends to occur in the tumors of patients with germline mutations, and the normal allele is preferentially lost (4–6). These patients with germline mutation are predisposed to midline brain tumors (primitive neuroectodermal tumors [PNETs]) and to the development of second nonocular tumors (7).

Unfortunately, the patient's phenotype at presentation does not always correlate with the underlying genetic predisposition. Twelve percent of patients with unilateral disease have underlying germline mutations and are at risk to develop disease in the uninvolved eye (8). These patients are also prone to develop PNETs and second nonocular tumors, particularly sarcomas. When patients present with unilateral retinoblastoma, the tumor frequently fills 50% or more of the ocular volume; when the tumor is confluent, the ophthalmologist cannot determine whether it was unifocal or multifocal in origin.

Mosaicism exists in this disease as it does in other tumor suppressor syndromes such as neurofibromatosis. Mosaic carrier parents may not manifest the disease themselves, but may still transmit the mutation in a percentage of their gametes. Siblings of patients with retinoblastoma are therefore at risk to develop the disease and should be screened from birth (9).

Despite its low incidence, retinoblastoma has contributed greatly to our understanding of cancer. The retinoblastoma gene (Rb) was the first ocular disease gene to be cloned and also the first representative of a class of genes called tumor suppressors. These genes predispose individuals to neoplasia by their deletion. It is the function of the protein products of tumor suppressor genes to regulate cellular division. The retinoblastoma gene product exerts its regulatory role in the cell cycle at the G1 stop point. This stop allows repair of DNA to occur prior to replication. The under-phosphorylated form of the retinoblastoma gene product binds to numerous transcription factors and sequesters them, preventing transcription. Phosphorylation of the Rb gene is regulated by complex formation between cyclins and CDKs, an event that in turn is regulated by signals that extend from the outermost boundary of the cell. With the phosphorylation of the Rb gene product, transcription factors are synchronously available to exert their actions. An especially important binding partner of Rb is elongation factor (E2F). This transcription factor binds to consensus sequences across the genome to activate transcription and to allow the cell to proceed through the cell cycle. In the absence of a normal retinoblastoma gene product, cells divide without a regulated G1 checkpoint. This results in replication of DNA without repair and a predisposition to the development of multiple neoplasias over a lifetime (10).

Children with retinoblastoma develop ocular cancers from birth into their sixth year of life. They also develop midline primitive neuroectodermal brain tumors at a rate of 7% to 11% of germline cases. By the teen years, these patients develop sarcomas, particularly within the orbital radiation treatment field, as well as outside that field. Throughout their lives, retinoblastoma patients have a predisposition to many cancers with a 59% 35-year mortality for patients with underlying germline mutation (2).

Retinoblastoma in its diffuse form is particularly difficult to diagnose. As Brissé et al (page 499) describe in this issue of *AJNR,* diffuse retinoblastoma may mimic Coats disease, an exudative retinopathy that can produce massive yellow white lipid deposits within the retina. Although infrequent, misdiagnoses of patients with retinoblastoma can occur, sometimes leading to unnecessary surgical procedures that have the potential to disseminate neoplastic cells. We have never performed a fine-needle aspiration biopsy in a patient with retinoblastoma, and we believe strongly that there is never an indication to perform an intraocular procedure in these children. Intraocular procedures can result in mortality from tumor dissemination in this patient population, and even transcorneal approaches are not without risk (11).

We agree with the authors that, because biopsy is precluded, indirect approaches are required to make an accurate diagnosis of retinoblastoma. We believe that each imaging technique contributes different information to the management of this disease.

When a child is referred to us with retinoblastoma, we first examine the child, while awake, in the office. This allows us to take a careful family history, looking for near relatives with retinoblastoma, tumor predisposition syndromes, or eye loss. We also attempt to narrow the differential diagnosis by exploring the child's medical history (episodes of pica are correlated with toxocara canis exposure, low birth weight with retinopathy of prematurity, etc.). We evaluate the child, while awake, to determine the visual potential of each eye. We then schedule an examination under anesthesia. At that examination, we perform indirect ophthalmoscopy, fundus photography, and A- and B-scan ocular sonography. We check corneal dimensions and the axial length of the eye, because persistent hyperplastic primary vitreous is associated with a foreshortened globe and is important in the differential diagnosis of retinoblastoma. We check intraocular pressure and evaluate the anterior segment of the eye. Neovascular glaucoma can be diagnosed by these means, and children with this diagnosis are unlikely to retain vision over the long term. We draw blood for toxocara titers and other uveitis laboratory analyses if intraocular inflammation is high in the differential diagnosis. We look for hyperechogenic flecks on B-scan sonography, as these represent intrinsic tumor calcification and make a diagnosis of retinoblastoma very likely. We occasionally see fine flecks of calcium by use of sonography, which cannot be depicted by CT scanning. We schedule CT under the same anesthesia to confirm the presence of intraocular calcification and to exclude extraocular disease. We use MR imaging selectively to exclude the presence of orbital or optic nerve disease at diagnosis and to follow patients with germline retinoblastoma for the development of PNETs.

If midline PNETs present symptomatically, they respond poorly to therapy. We attempt to determine this diagnosis prior to symptoms with routine MR neuroimaging on a 6-month basis for children at risk. We also routinely perform MR studies in children who have histopathologic risk factors for local disease recurrence on an enucleation specimen. Such risk factors include invasion of the optic nerve posterior to the lamina cribrosa (where the meninges insert on the optic nerve) or massive choroidal invasion. If we see evidence of invasion of the optic nerve posterior to the lamina cribrosa, we also offer the patient 6 months of adjuvant chemotherapy with carboplatin, vincristine, and etoposide. We believe that patients who manifest tumor at the cut-end of the optic nerve are at great risk for systemic relapse, and we offer these patients aggressive chemotherapeutic regimens, including bone marrow transplantation, as well as orbital or whole-brain radiation therapy (4). These patients are followed up very closely with sequential MR neuroimaging to diagnose orbital or intracranial relapse.

In recent years, treating physicians have moved away from radiation as the primary treatment for retinoblastoma toward chemoreduction with laser hyperthermia and cryotherapy. These latter therapies seek to avoid problems associated with radiation in children, especially the development of midface hypoplasia, cataracts, and the much-increased rate of sarcomas within the radiation field. These newer treatment regimens, however, require increased vigilance on the part of the ophthalmologist and the radiologist, because tumors require aggressive local control in conjunction with chemotherapy to prevent relapse. Because chemotherapy combined with local treatment has the advantage of not being associated with cataract formation, the ophthalmologist is able to survey the retina serially by indirect ophthalmoscopy. This diagnostic approach demonstrates greater sensitivity and specificity than any imaging technique can provide (12). Despite the trend toward chemotherapy combined with local treatment, many children with retinoblastoma still require radiation therapy. These children develop cataracts, and visualization of the retina becomes limited. In these children, neuroimaging and sonography are the only methods available to diagnose disease relapse. We use a combination of A- and B-scan sonography, thin-section CT scanning, and MR imaging using 3D fast spin-echo sequences to follow intraocular disease in patients with cataracts.

We congratulate Brissé et al on their excellent discussion of the diffuse form of retinoblastoma. This is an entity that can be easily misdiagnosed. Inappropriate intraocular procedures can result in disease dissemination. Retinoblastoma should always be considered in the differential diagnosis of any intraocular pathologic process in a child (Table). Frequently children with retinoblastoma have histories of ocular or periocular trauma, and retinoblastoma can present with hyphema, hypopyon, or glaucoma. The children with less characteristic presentations are those who are most frequently misdiagnosed. No child who presents to an ophthalmologist with a limited view to the back of the eye should undergo an intraocular procedure without appropriate preoperative imaging. We believe that sonography, CT, and MR imaging have complementary and important roles in the management of childhood ocular disease. Because retinoblastoma has such a high mortality once neoplastic cells are no longer confined to the eye, the ophthalmologist and radiologist should always keep this disease in mind when considering the differential diagnosis of childhood eye disease (11).

View this table:
[Table1](http://www.ajnr.org/content/22/3/427/T1)

Table: Different diagnosis of leukocoria

## References

1.  Tamboli A, Podgor MJ, Horm JW. **The incidence of retinoblastoma in the United States: 1974 through 1985.** Arch Ophthalmol 1990;108:128-132
    
    [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1001/archopht.1990.01070030134045&link_type=DOI) 
    
    [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=2288550&link_type=MED&atom=%2Fajnr%2F22%2F3%2F427.atom) 
    
    [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=A1990CH30700039&link_type=ISI) 

2.  Abramson DH, Ellsworth RM, Grumbach N, et al. **Retinoblastoma: survival, age at detection, and comparison 1914–1958, 1958–1983.** J Pediatr Ophthalmol Strabismus 1985;22:246-250
    
    [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=4078667&link_type=MED&atom=%2Fajnr%2F22%2F3%2F427.atom) 

3.  Shields JA, Augsburger JJ. **Current approaches to the diagnosis and management of retinoblastoma.** Surv Ophthalmol 1981;25:347-372
    
    [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1016/0039-6257(81)90072-2&link_type=DOI) 
    
    [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=7017985&link_type=MED&atom=%2Fajnr%2F22%2F3%2F427.atom) 
    
    [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=A1981LT58800001&link_type=ISI) 

4.  Yandell DW, Campbell TA, Dayton SH, et al. **Identification of oncogenic point mutations in the human retinoblastoma gene and application to genetic counseling.** N Engl J Med 1989;321:1689-1695
    
    [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1056/NEJM198912213212501&link_type=DOI) 
    
    [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=2594029&link_type=MED&atom=%2Fajnr%2F22%2F3%2F427.atom) 
    
    [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=A1989CE29400001&link_type=ISI) 

5.  Higgins MJ, Hansen MF, Cavenee WK, et al. **Molecular detection of chromosomal translocations that disrupt the putative retinoblastoma susceptibility locus.** Mol Cell Biol 1989;9:1-5
    
    [Abstract/FREE Full Text](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MzoibWNiIjtzOjU6InJlc2lkIjtzOjU6IjkvMS8xIjtzOjQ6ImF0b20iO3M6MTk6Ii9ham5yLzIyLzMvNDI3LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 

6.  Wiggs J, Nordenskjold M, Yandell D, et al. **Prediction of the risk of hereditary retinoblastoma, using DNA polymorphisms within the retinoblastoma gene.** N Engl J Med 1988;318:151-157
    
    [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=2892131&link_type=MED&atom=%2Fajnr%2F22%2F3%2F427.atom) 
    
    [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=A1988L714100005&link_type=ISI) 

7.  Abramson DH, Ellsworth RM, Kitchin D, et al. **Second nonocular tumors in retinoblastoma survivors: are they radiation induced?** Ophthalmology 1984;91:1351-1355
    
    [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=6595610&link_type=MED&atom=%2Fajnr%2F22%2F3%2F427.atom) 
    
    [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=A1984TS48600014&link_type=ISI) 

8.  Dryja TP. **Assessment of risk in hereditary retinoblastoma.** Albert DM, Jakobiec FA, Robinson NL, eds. *Principles and Practice of Ophthalmology* Philadelphia:W.B. Saunders Co. 1994;3270-3279
    
    

9.  Hasenyager-Smith J, Murray TG, Fulton L, et al. **Siblings of retinoblastoma patients: Are we underestimating their risk? A brief report.** Am J Ophthalmol 2000;129:396-398
    
    [CrossRef](http://www.ajnr.org/lookup/external-ref?access_num=10.1016/S0002-9394(99)00385-2&link_type=DOI) 
    
    [PubMed](http://www.ajnr.org/lookup/external-ref?access_num=10704567&link_type=MED&atom=%2Fajnr%2F22%2F3%2F427.atom) 
    
    [Web of Science](http://www.ajnr.org/lookup/external-ref?access_num=000085807500029&link_type=ISI) 

10. O'Brien JM. **Retinoblastoma: a current review.** Jameson JL, ed. *Principles of Molecular Medicine* Totowa, NJ: Humana Pres Inc.; 1998: 955-962
    
    

11. Uusitalo MS, Van Quill KR, Scott IV, et al. ** Evaluation of chemoprophylaxis in patients with unilateral retinoblastoma with high-risk features on histopathologic examination.** Arch Ophthalmol 2001; 119:41–48 
    
    

12. Chan HSL, DeBoer G, Thiessen JJ, et al. **Cyclosporin-modulated chemotherapy with focal therapy as a new approach that avoids radiation of retinoblastoma.** Clin Cancer Res 1996;2:1499-1508
    
    [Abstract](http://www.ajnr.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6MTA6ImNsaW5jYW5yZXMiO3M6NToicmVzaWQiO3M6ODoiMi85LzE0OTkiO3M6NDoiYXRvbSI7czoxOToiL2FqbnIvMjIvMy80MjcuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 

*   Copyright © American Society of Neuroradiology