Retinoblastoma represents the prototypic model for inherited cancers. The RB1 gene was the first tumor suppressor gene to be identified. It represents the most frequent primary eye cancer in children under 15 years old, habitually occurring in infancy, even in utero, but can be observed in older children or young adults. Many other retinal lesions may also simulate retinoblastoma. The two major presenting signs are leukocoria and strabismus, but other ocular or general signs may be observed. A highly malignant tumor, retinoblastoma can nowadays be cured. The heritable form, however, carries a high risk of second nonocular tumors. Treatment in the early stages of disease holds a good prognosis for survival and salvage of visual function. In very late stages, however, the prognosis for ocular function and even survival is jeopardized.
Retinoblastoma represents the prototypic model for inherited cancers (Gallie and Phillips, 1984; Conway et al., 2005). The Rb1 was the first tumor-suppressor gene to be identified, leading to the discovery of a whole new class of antioncogenes and greatly contributing to advances in the management of solid tumors in children.
Although a rare disease, retinoblastoma is the most frequent primary eye cancer in children under 15 years old (Mahoney et al., 1990). It may also be simulated by other lesions. Retinoblastoma is almost exclusively a disease of childhood, has even been detected in the foetus (Maat-Kievit et al., 1993) but can also occur in older children and, in exceptional cases, young adults (Shields et al., 1991). Diagnosis is often late, in spite of clear warning signs (Balmer and Munier, 2002a).
A malignant tumor now considered curable, some cases of retinoblastoma show ‘spontaneous’ regression. With the heritable form, however, there remains a major risk of second nonocular primary tumors, these even more lethal than the retinoblastoma itself (Abramson, 1999). Predisposition to malignant transformation may lead to a primary intracranial neuroblastic tumor, in particular pinealoblastoma (‘trilateral’ retinoblastoma) (Kivela, 1999).
Early treatment of the disease is straightforward, well tolerated, effective and at little cost, saving both life and vision. Undertaken at a late stage, however, treatment is heavy, mutilating and costly, with no certain outcome. This emphasizes the need for early diagnosis and the importance of awareness at all levels, both public and professional (Balmer and Munier, 2002c).
Retinoblastoma is a malignant tumor of the immature retina, the result of a double (‘two-hit’) oncogenic mutation occurring between the start of the third month postconception and the age of 4 years, this period representing the window of apparition and final maturity of retinoblasts. The cell of origin is most probably either a precursor cone photoreceptor cell or a multipotent retinoblast (Munier and Balmer, 2002). The disease may thus occur in utero and up to the age of 4 years, the average age at appearance of first signs being 7 months for bilateral cases and 24 months for unilateral cases (Balmer and Munier, 2002a).
Retinoblastoma may be nonheritable (60%) or heritable (40%), the latter by autosomal-dominant transmission with high penetrance (90%).
The incidence of retinoblastoma is given in the literature as 1 in 20 000 live births (Balmer and Munier, 2002b) with no gender or race predilection and no significant environmental or socio-economic factors. The age of the parents, paternal in particular, may play a role according to some authors (DerKinderen et al., 1990; Matsunaga et al., 1990). Although rare, retinoblastoma represents 80% of all primary ocular cancers in children up to 15 years old, 1% of all cancer-related deaths (Miller, 1969) and rates third highest of all intraocular tumors in all ages (Scat et al., 1996).
Leukocoria is the most common presenting sign in patients with retinoblastoma (60%) (Figure 1) but is a late sign, the survival rate high (88% at 5 years) but poor for globe salvage (Beaverson et al., 2001). The second major presenting sign is strabismus (20%), this usually the result of macular involvement and an early sign, the survival rate high and with a good chance of globe salvage (Beaverson et al., 2001). The remaining 20% present atypical signs, often inflammatory these usually late signs with a much poorer survival and globe salvage prognosis (Binder, 1974; Balmer et al., 1993).
The clinical presentation of retinoblastoma depends on the tumor growth pattern, duration, degree of vascularization and the presence of calcifications, vitreous seeding, retinal detachment or hemorrhage. Tumors may show an endophytic growth pattern (Figure 2), resulting from cell division within the internal retinal layers and tumor growth towards the vitreous. An exophytic growth pattern (Figure 3) is initiated in the external retinal layers, the tumor developing beneath the retina in the subretinal space and causing an overlying retinal detachment. A mixed growth represents the association of both endophytic and exophytic patterns. Exceptionally (2%), in its diffuse infiltrating (plaque-like) form, retinoblastoma may develop in a flat pattern on the surface of, or beneath, the retina, with no obvious mass and no calcifications, progressing slowly towards the anterior segment where late signs present in the form of pseudoinflammatory complications.
Endophytic retinoblastoma presents as one or several tumors, isolated or coalesced, of variable size, round or oval-shaped, yellowish-white in color (calcifications) or pink (vascularization), with a regular or sometimes turgescent and tortuous vascular network. Tumor localization is statistically age-dependent, the earliest tumors developing in the posterior pole, more peripheral tumors developing later (Abramson and Gombos, 1996). Endophytic tumors show a marked tendency to vitreous seeding (Figure 4).
Exophytic retinoblastoma presents with retinal detachment, this masking to a greater or lesser degree the underlying tumor masses.
Finally, retinoma or retinocytoma is a benign form of retinoblastoma, characterized by greyish, homogenous, more or less translucent masses, with calcifications and border pigmentation imitating an irradiated retinoblastoma (Gallie et al., 1982; Margo et al., 1983; Balmer et al., 1991; Singh et al., 2000). Importantly, this benign form of retinoblastoma is nevertheless susceptible to malignant transformation at a later stage.
The Reese–Ellsworth Classification, developed in 1963 (Reese and Ellsworth, 1963), remains the classification system most widely used. Although clearly outmoded, of the many systems proposed in recent years, none has yet found consensus within the medical community. The Reese–Ellsworth Classification is based on intraocular tumor staging and globe salvage prediction after external beam radiation, survival not being taken into account. The failure of external beam radiotherapy to effectively treat peripheral tumors and the difficulty of treating vitreous seeding at that time lead to these factors being determinant in staging the severity of the disease. Of the most recent staging proposals, the International Intraocular Retinoblastoma Classification (ABC Classification) constitutes that most commonly used and of greatest prognostic value (Murphree, 2005) (Table 1).
This classification stages intraocular tumors according to their prognosis after first-line chemotherapy and adjuvant focal therapy. It consists of five groups (A, B, C, D, E) in descending order of favourable prognosis, taking into account the size of the tumors, their proximity to the macula and the optic nerve, the degree of seeding and importance of retinal detachment and, finally, late complications (Figures 2, 3, 4, 5, 6 and 7).
The poorly cohesive high-index mitotic retinoblastoma cells have a great natural tendency to infiltrate the most readily accessible spaces, these being the vitreous, subretinal fluid, anterior segment and meningeal sheaths of the optic nerve. Retinoblastoma can thus develop in three directions:
Anteriorly via the vitreous or subretinal space to the anterior segment, reaching the ciliary body and anterior chamber and creating a pseudohypopyon.
Posteriorly, step by step towards the optic nerve and cephalo-rachidien liquid, usual gateway to endocranial dissemination.
Extraocular extension to the orbit via the choroid and sclera.
Past the barrier of the lamina cribrosa, infiltration may continue via the retrolaminal portion of the optic nerve to the subarachnoid space and further, transported by the cephalo-rachidien fluid, to the various cerebral structures. Metastatic dissemination may also occur via the vascular pathway (orbit, optic nerve) or lymphatic pathway (conjunctiva, eyelids). The tumor may also invade the orbit directly through the sclera or by following the orbital canals of the ciliary arteries and nerves.
Risk factors for metastatic disease
The risk of metastasis through optic nerve invasion depends on the stage of tumor penetration. The mortality rate is 50–85% (Uusitalo et al., 2001) if tumor cells have reached the surgical resection line, 13–69% if posterior to the lamina but not reaching the surgical resection, insignificant if anterior to the lamina cribrosa.
Massive choroidal invasion is generally considered the only other risk factor, the association of optic nerve and choroidal invasion carrying the greatest risk (Shields et al., 1993a; Khelfaoui et al., 1996; Chantada et al., 1999).
Orbital extension of retinoblastoma can lead to systemic dissemination via the blood vessels, lymphatic system or along the visual pathway to the brain. Orbital invasion is considered an important risk factor (Kopelman et al., 1987) but modern aggressive therapy may prolong survival.
Anterior chamber invasion does not increase the mortality rate, in spite of the risk of vascular dissemination.
Second nonocular primary tumors
The Rb1 mutation present in all cells of a patient with retinoblastoma predisposes them to developing other nonocular malignant tumors, albeit with a predilection for certain tissues.
Second primary tumors are the greatest cause of death in these patients, rather than the retinoblastoma itself (Abramson, 1999; Fletcher et al., 2004). The cumulative incidence of these additional tumors is around 1% per year, thus 50% after 50 years’ evolution (Abramson, 1999). Radiotherapy greatly increases the risk, particularly if administered before the age of 1 year (Abramson and Frank, 1998).
According to a study by Abramson et al. (2001), the most common sites of second nonocular tumors are, in order of frequency: soft head tissue (24%), cranium (18%), skin (15%), brain (8%) and others (25%).
In the early 20th century, a child suffering from retinoblastoma had little chance of survival. Progress in the management of this disease has lead to a steady increase in the survival rate, from 30% in the 1930s to 80% in the 1960s. In specialized centers today, almost 95% of these children can be saved thanks to modern therapeutic strategies: systematic enucleation, external beam radiotherapy, focal treatments, chemotherapy, stereotactic conformal radiotherapy.
Management of a child with retinoblastoma requires constant re-evaluation of the situation according to the response to treatment, with survival the primary aim, followed by globe salvage and the best possible visual function.
Treatment approaches changed notably in the 1990s with chemotherapy or, more exactly, chemoreduction, gradually replacing external beam radiation and enucleation as first-line treatment, shrinking the lesions until accessible to adjuvant focal tumor control (Figure 8). This new strategy has reduced the frequency of enucleation and spared the complications associated with primary external beam radiation, in particular the risk of second nonocular radiation-induced cancers, a risk particularly high during the first year of life.
Systemic chemotherapy is not, however, without its own hematologic side effects such as coagulation disorders or leukopenia, and as yet undetermined, possibly dangerous, long-term complications. Only strictly necessary agents should be administered and at a dosage that is truly effective, limiting the number of cycles to that needed to achieve the aim of treatment that is reduction of tumor volume to allow focal therapies. A permanent response is almost never achieved with chemotherapy alone.
According to Shields et al. (1997), primary chemotherapy comprising two or more agents will reduce tumor volume by over 50% within a few weeks (Figure 8), dry out a retinal detachment and permit consolidation with methods other than external beam radiotherapy.
The choice of agents, their combination and dosage, varies from one center to another. Most institutions utilize vincristine, carboplatin and an epipodophyllotoxin, either etopside or teniposide. Cyclosporine A is sometimes associated to prevent multidrug resistance (Chan et al., 2005). The number and frequency of cycles also varies.
Heat increases the permeability of the plasmatic membrane to antimitotics, thus reinforcing their cytoxic effect. Carboplatin is administered intravenously 1 or 2 h before the thermotherapy. Infrared radiation from a diode laser, this mounted on a surgical microscope, is focused on the tumor through a contact lens for a period of several minutes (5 to 30 according to the volume of the tumor). The aim is to deliver a temperature of 41–44°. A second session, this thermotherapy only, is carried out between 4 and 7 days later. The whole cycle is repeated two or three times at 28-day intervals until complete tumor control is achieved.
Thermotherapy, or hyperthermia, consists of utilizing the cytoxic effect of heat by raising the temperature of the tumor to over 45°, using infrared radiation from a diode laser (Shields et al., 1999). This method of treatment can be used alone for the treatment of small tumors, or in conjunction with chemotherapy or radiotherapy. If used in conjunction with these latter, a slightly lower temperature is sufficient (see above). Tumors under 3 mm in diameter and 2 mm in thickness usually respond well to thermotherapy alone.
Brachytherapy or plaque radiotherapy
Radioactive plaques (Iodine-125, Ruthenium-106) (Shields et al., 1993b) can be used for the primary treatment of medium sized tumors (under 16 mm in diameter and 9 mm in thickness), but more often as a secondary measure after prior failed treatment or in the case of a relapse. Cobalt-60 plaques have been largely abandoned due to their high-energy gamma rays preventing all shielding and consequently giving off radiation in all directions.
Currently, the most widely used plaque is the Iodine-125. The low-energy gamma rays allow a unidirectional, shielded radiation field (Figure 9).
Ruthenium plaques may also be applied. These emit high-energy beta rays but with poor tissue penetrance, limiting their use to the treatment of small tumors under 6 mm in thickness.
Stereotactic conformal radiotherapy
When none of the above focal treatment modalities is applicable, external radiation therapy may be necessary. Ideally, radiotherapy should be applied only to the tumor itself, sparing the surrounding structures. This is now possible with stereotactic conformal radiation therapy for which new apparatus has been designed, a multileaf photon 6 MV collimator. A sophisticated system is required, comprising a custom-molded plastic mask, scanner for treatment planning, contact lens with suction cup to immobilize the globe and adapted to the mask, multiple conformal static fields. The radiation dose delivered is 50.4 Gy in 1.8 Gy fractions, representing 25–28 sessions.
Tumors in Groups B and C, with residual active tumoral activity in the foveal or peri/epi-papillary region after chemoreduction, are candidates for secondary stereotactic conformal radiotherapy.
Accelerated proton beam irradiation
Ionizing radiation for destroying tumor tissue is well known. Charged protons have a well-defined stopping point and reach maximum ionisation density at the end of their path, a point termed ‘Bragg Peak’. Modulation of the proton beam accelerator produces a succession of Bragg Peaks calculated to create a uniform ionisation plateau at the desired spot within the tumor, causing maximum damage where desired with minimal collateral tissue effect.
The radiotherapy equipment comprises a beam modulator and collimator, and a stereotactic couch fitted with a custom-molded plastic mask to immobilize the patient's head. Tumor localization is assured by suturing four tantalum rings to the sclera as reference points, marking the tumor edges (Zografos, 2002).
The precise indications for proton beam irradiation are in evolution but seem, at first sight, the same as those for stereotactic conformal radiotherapy, including orbital irradiation after enucleation with positive histopathology for tumor cells present at the surgical resection line of the optic nerve stump. More specific indications may develop with time, but proton beam radiotherapy requires the child's cooperation and is not suitable for children under 3 1/2 years.
Enucleation has become less common with earlier diagnosis and better alternative therapies, but remains the treatment of choice for advanced cases (massive involvement of the retina/vitreous, rubeosis iridis, secondary glaucoma) with no visual potential or with high risk of dissemination (massive choroidal or optic nerve invasion, anterior segment involvement, extraocular extension).
External beam irradiation is no longer considered a first-line conservative treatment due to its important long-term side effects. It remains indicated in advanced bilateral cases with or without vitreous seeding, or in the case of a relapse inaccessible to conservative focal treatment methods.
The habitual total dose administered is 40–50 Gy, delivered in fractions of 1.8 Gy through lateral or anterior fields, using a lens-sparing technique whenever possible.
Proton beam therapy may become common practice in future management of retinoblastoma, the advent of second-generation collimated Gantry machines allowing the irradiation of very small tumors.
Local delivery of chemotherapeutic agents (carboplatin) by periocular (subtenon) or intraocular injections has been experimented in animals (Murray et al., 1997; Van Quill et al., 2005) and used in selected retinoblastoma patients with some success. It is not, however, without risk and at the present time does in no way present an alternative to systemic chemotherapy. Further study is needed to determine fully the safety and applicability of these procedures.
Preimplantation genetic diagnosis, following an accurate mutation analysis procedure for retinoblastoma gene mutation sensitive at the single-cell level, is possible today for in vitro fertilization in order to achieve pregnancies without retinoblastoma (DerKinderen et al., 1990; Xu et al., 2004; Simpson et al., 2005).
Photodynamic therapy (PDT), based on the use of nontoxic photosensitizers activated with a visible nonionising laser light may represent an alternative nonmutagenic therapy. First experimental data indicate that PDT could efficiently slow down some retinoblastoma growth, with no acute toxicity (Aerts et al., 2005).
The best treatment is, and will always be, early treatment. This signifies early diagnosis. In cases of positive family history all members of the family should be screened for retinoblastoma and all newborns at birth. Pregnancies within a family at risk require special monitoring with antenatal ultrasonography, genetic testing and, if in doubt, a 3 to 4-week preterm delivery (Figure 10).
In children, and until proved otherwise, the risk of retinoblastoma should always be forefront in the presence of leukocoria, strabismus or any other unexplained sign. The family, gynecologist, pediatrician or primary care physician are the first likely to be confronted with these signs, well before the multidisciplinary specialized team. This emphasizes the importance of awareness at all levels, both public and professional.
Abramson DH . (1989). Acta Ophthalmol Suppl 194: 3–63.
Abramson DH . (1999). Ophthalmic Genet 20: 193–204.
Abramson DH, Frank CM . (1998). Ophthalmology 105: 573–580.
Abramson DH, Gombos DS . (1996). Retina 16: 232–239.
Abramson DH, Melson MR, Dunkel IJ, Frank CM . (2001). Ophthalmology 108: 1868–1876.
Aerts I, Leuraud P, Laville I, Blais J, Maillard P, Desjardins L et al. (2005) In International Retinoblastoma Symposium. International Society of Genetic Eye Disease and International Retinoblastoma Symposium: Whistler, Canada.
Balmer A, Gailloud C, Munier F, Uffer S, Guex-Crosier Y . (1993). Ophthalmic Paediatr Genet 14: 33–38.
Balmer A, Munier F . (2002a) In: Zografos L (ed). Tumeurs Intraoculaires. Société Française d'Ophtalmologie et Masson: Paris, pp 485–533.
Balmer A, Munier F . (2002b) In: Zografos L (ed). Tumeurs Intraoculaires. Société Française d'Ophtalmologie et Masson: Paris, pp 481–485.
Balmer A, Munier F . (2002c) In: Zografos L (ed). Tumeurs Intraoculaires. Société Française d'Ophtalmologie et Masson: Paris, pp 587–617.
Balmer A, Munier F, Gailloud C . (1991). Ophthalmic Paediatr Genet 12: 131–137.
Beaverson K, Abramson DH, Lee TC, Hochberg HM, Kirszrot J, Sangani P et al. (2001) In: Keunen JEE, Imhof SM, de Keizer RJW, Moll AC (eds). Xth International Congress of Ocular Oncology. Amsterdam, Conference Proceedings, published by the eds 202.
Binder PS . (1974). Am J Ophthalmol 77: 674–679.
Chan HS, Gallie BL, Munier FL, Beck Popovic M . (2005). Ophthalmol Clin North Am 18: 55–63, viii.
Chantada GL, de Davila MT, Fandino A, Manzitti J, Raslawski E, Casak S et al. (1999). Ophthalmic Genet 20: 133–140.
Chevez-Barrios P, Chintagumpala M, Mieler W, Paysse E, Boniuk M, Kozinetz C et al. (2005). J Clin Oncol 23: 7927–7935.
Conway RM, Wheeler SM, Murray TG, Jockovich ME, O'Brien JM . (2005). Ophthalmol Clin North Am 18: 25–39 vii.
DerKinderen DJ, Koten JW, Tan KE, Beemer FA, Van Romunde LK, Den Otter W . (1990). Am J Ophthalmol 110: 605–609.
Fletcher O, Easton D, Anderson K, Gilham C, Jay M, Peto J . (2004). J Natl Cancer Inst 96: 357–363.
Gallie BL, Phillips RA . (1984). Ophthalmology 91: 666–672.
Gallie BL, Phillips RA, Ellsworth RM, Abramson DH . (1982). Ophthalmology 89: 1393–1399.
Girard B, Le Hoang P, D'Hermies F, Quere MA, Rousselie F . (1989). J Fr Ophtalmol 12: 369–381.
Huang HJ, Yee JK, Shew JY, Chen PL, Bookstein R, Friedmann T et al. (1988). Science 242: 1563–1566.
Khelfaoui F, Validire P, Auperin A, Quintana E, Michon J, Pacquement H et al. (1996). Cancer 77: 1206–1213.
Kivela T . (1999). J Clin Oncol 17: 1829–1837.
Kopelman JE, McLean IW, Rosenberg SH . (1987). Ophthalmology 94: 371–377.
Maat-Kievit JA, Oepkes D, Hartwig NG, Vermeij-Keers C, van Kamp IL, van de Kamp JJ . (1993). Prenat Diagn 13: 377–384.
Mahoney MC, Burnett WS, Majerovics A, Tanenbaum H . (1990). Ophthalmology 97: 1143–1147.
Margo CE, Hidayat A, Kopelman J, Zimmerman LE . (1983). Arch Ophthalmol 101: 1519–1531.
Matsunaga E, Minoda K, Sasaki MS . (1990). Hum Genet 84: 155–158.
Messmer EP, Heinrich T, Höpping W, de Sutter E, Havers W, Sauerwein W . (1991). Ophthalmology 98: 136–141.
Miller RW . (1969). J Pediatr 75: 685–689.
Morgan G . (1971). Br J Ophthalmol 55: 600.
Munier F, Balmer A . (2002) In: Zografos L (ed). Tumeurs Intraoculaires. Société Française d'Ophtalmologie et Masson: Paris, pp 466–487.
Murphree AL . (2005). Ophthalmol Clin North Am 18: 41–53, viii.
Murray TG, Cicciarelli N, O'Brien JM, Hernandez E, Mueller RL, Smith BJ et al. (1997). Arch Ophthalmol 115: 1286–1290.
Reese AB, Ellsworth R . (1963). Trans Am Acad Ophthalmol Otolaryngol 67: 164–172.
Scat Y, Liotet S, Carre F . (1996). J Fr Ophtalmol 19: 83–88.
Shields CL, Santos MC, Diniz W, Gunduz K, Mercado G, Cater JR et al. (1999). Arch Ophthalmol 117: 885–893.
Shields CL, Shields JA, Baez KA, Cater J, De Potter PV . (1993a). Br J Ophthalmol 77: 544–548.
Shields CL, Shields JA, De Potter P, Minelli S, Hernandez C, Brady LW et al. (1993b). Ophthalmology 100: 216–224.
Shields CL, Shields JA, DePotter P, Himelstein BP, Meadows AT . (1997). J Pediatr Ophthalmol Strabismus 34: 165–169.
Shields CL, Shields JA, Shah P . (1991). Ophthalmology 98: 395–399.
Simpson JL, Carson SA, Cisneros P . (2005). J Natl Cancer Inst Monogr, 87–90.
Singh AD, Santos CM, Shields CL, Shields JA, Eagle Jr RC . (2000). Arch Ophthalmol 118: 199–205.
Uusitalo MS, Van Quill KR, Scott IU, Matthay KK, Murray TG, O'Brien JM . (2001). Arch Ophthalmol 119: 41–48.
Van Quill KR, Dioguardi PK, Tong CT, Gilbert JA, Aaberg Jr TM, Grossniklaus HE et al. (2005). Ophthalmology 112: 1151–1158.
Xu HJ . (2003). Ophthalmol Clin North Am 16: 621–629.
Xu K, Rosenwaks Z, Beaverson K, Cholst I, Veeck L, Abramson DH . (2004). Am J Ophthalmol 137: 18–23.
Zografos L . (2002) In: Tumeurs Intraoculaires. SFO: Paris, pp 21–64.
About this article
Cite this article
Balmer, A., Zografos, L. & Munier, F. Diagnosis and current management of retinoblastoma. Oncogene 25, 5341–5349 (2006). https://doi.org/10.1038/sj.onc.1209622
- tumor suppressor gene
- diagnosis; management
miR-498 promotes cell proliferation and inhibits cell apoptosis in retinoblastoma by directly targeting CCPG1
Child's Nervous System (2018)
Journal of Experimental & Clinical Cancer Research (2016)
CRISPR/Cas9 mediated knockout of rb1 and rbl1 leads to rapid and penetrant retinoblastoma development in Xenopus tropicalis
Scientific Reports (2016)
Cancer and Metastasis Reviews (2015)
Tumor Biology (2015)