Retinoblastoma is the commonest malignant ocular Tumours of childhood and affects around 1 per 20,000 children. Survival rates in the USA approach 100 % though the rates are much lower in developing countries. The Tumours is derived from primitive retinal cells (retinoblasts) and usually presents in early childhood (90 % before the age of 5 years). Less than 10 % of children with retinoblastoma have a positive family history (where inheritance is autosomal dominant), but new mutations are frequent and approximately 40 % of retinoblastoma patients have a genetic predisposition. Retinoblastoma holds a unique place in human cancer genetics as the subject of Al Knudson’s pioneering work on the “two-hit model of Tumoursigenesis” and the paradigm of the Tumours suppressor gene.
Retinoblastoma typically presents as leukocoria (white eye, cat’s eye reflex) or strabismus. It is bilateral in about 30 % of cases, and these children have a younger age at diagnosis (mean 8 months) than those with unilateral Tumours (mean 25 months). Bilateral or multifocal Tumours occur in patients with germline mutations of the retinoblastoma (RB1) gene, but about 15 % of children with a single Tumours will have a germline mutation. The 40 % of all children with retinoblastoma who carry a germline retinoblastoma gene mutation are at risk for secondary Tumours, especially osteosarcoma and soft tissue sarcomas. This predisposition is exacerbated in those who receive radiotherapy. A few individuals (less than 2 %) with germline retinoblastoma gene mutations may develop a retinoma. These benign retinal lesions appear as focal translucencies with cottage-cheese-like calcification and underlying choroidal and retinal pigment epithelial disturbance. A retinoma is thought to result when a second RB1 mutation occurs in a retinoblast which is almost differentiated. About 2 % of all patients with retinoblastoma also have an intracranial lesion, usually in the pineal gland.
The association of pineal Tumours and retinoblastoma is often termed trilateral retinoblastoma and occurs in patients with germline retinoblastoma mutations.
The retinoblastoma mutation is non-penetrant in approximately 10 % of obligate carriers. Although there is no detectable paternal age effect on new mutations for retinoblastoma, most new hereditary mutations develop in paternal germ cells (Zhu et al. 1989). An excess of males among patients with bilateral sporadic disease has been noted, and Naumova and Sapienza (1994) suggested the involvement of genomic imprinting effects. Although an excess of retinoblastoma cases was reported in a Netherlands cohort of children conceived by assisted reproductive technologies, this finding requires replication (Moll et al. 2003).
About 60 % of retinoblastomas result from somatic (acquired) mutations inactivating both alleles of the retinoblastoma gene. Such patients develop single Tumours and there is no risk to their offspring. However, 15 % of patients with single Tumours will have a germline mutation. Risk estimates for the relatives of isolated cases of retinoblastoma are given in Table 2.1. The later estimates of Draper et al. (1992) are lower in some cases than earlier estimates. Genetic counseling for the relatives of patients with isolated unilateral retinoblastoma is complicated by the possibility for mosaicism of a germline RB1 mutation (Lohmann et al. 1997). Individuals with genetic retinoblastoma (germline mutation) are at increased risk for second Tumours, but children with nongenetic Tumours are not. The risk of second primary neoplasms reflects a genetic predisposition to non-retinoblastoma Tumours and the effects of treatment (e.g., radiation). Draper et al. (1986) estimated that the cumulative probability of a second primary neoplasm in genetic retinoblastoma patients at 18 years after diagnosis is 8.4 % for all Tumours and 6 % for osteosarcoma. The risks for osteosarcoma outside and within the radiation fields were 2.2 and 3.7 %, respectively, and thus patients with genetic retinoblastoma may be more sensitive to radiation-induced oncogenesis. The most common site of osteosarcoma outside the radiation field is in the femur, and genetic retinoblastoma patients are at a 200–500- fold increased risk of this complication. Soft tissue sarcomas also occur with increased frequency in patients with genetic retinoblastoma. Studies of non- ocular cancer in relatives of retinoblastoma patients have demonstrated that retinoblastoma mutation carriers are at increased risk of a variety of other cancers (overall relative risk 11.6) including cancer of the lung (relative risk 15), malignant melanoma, and bladder cancer (Sanders et al. 1988). Moll et al. (1996) estimated cumulative incidences of second primary Tumours in hereditary retinoblastoma of 4 and 18 % at ages 10 and 35 years, respectively. Eng et al. (1993) reported a cumulative probability of death from a second primary neoplasm of 26 % at 40 years after bilateral retinoblastoma. The most common second primary Tumours were bone and connective tissue neoplasms and malignant melanoma. Eng et al. (1993) also demonstrated that radiotherapy increased the risk of a second primary Tumours and suggested that all patients with inherited retinoblastoma should receive lifelong surveillance for second primary Tumours. In a follow-up study of UK hereditary retinoblastoma cases, Fletcher et al. (2004) found a cumulative cancer incidence and mortality of 69 and 48 %, respectively. In contrast to patients treated by radiotherapy, sarcomas accounted for only a minority of Tumours, and there was an increased mortality for lung cancer (standardized mortality ratio (SMR) = 7), bladder cancer (SMR = 26), and all other epithelial cancers combined (SMR = 3.3).
Table 1 Risk of retinoblastoma in relatives of a child with retinoblastoma and no family history
|Retinoblastoma in proband||Relationship to proband||Risk of carrying RB mutationa (%)||Risk of developing retinoblastoma Tumoursa (%)||Risk of developing retinoblastoma Tumoursb (%)|
|Bilateral||Sibling or dizygotic twin||5||2.7||2|
|Bilateral||Offspring of unaffected sibling||0.5||0.27|
|Unilateral||Sibling or dizygotic twin||0.8||0.4||1|
|Unilateral||Offspring of unaffected sibling||0.08||0.04|
The calculated risks from Musarella and Gallie (1987) take into account the retinoblastoma mutation rate and assume that 90 % of individuals with a germline mutation will develop a Tumours and 15 % of patients with unilateral retinoblastoma have a germline mutation bThe risks from Draper et al. (1992) for siblings relate to the first child, when there are further unaffected siblings the risk will be lower.
The retinoblastoma gene was initially assigned to chromosome 13 band q14 by reports of children with retinoblastoma and interstitial deletions of chromosome 13. About 3 % of children with retinoblastoma will have a cytogenetically visible chromosome 13 deletion or translocation.
Retinoblastoma cases with constitutional chromosome deletion may have associated mental retardation, and most have reduced serum levels of esterase D (the gene for which maps close to the retinoblastoma gene (RB1)). The RB1 gene spans 200 kb in 13q14 (Friend et al. 1986) and encodes a 110-kDa nuclear phosphoprotein with Tumours suppressor activity (Huang et al. 1988).
The identification of the retinoblastoma gene has enabled the presymptomatic identification of individuals with germline mutations by a variety of techniques. Intragenic restriction fragment length polymorphisms (RFLPs) are informative in about 95 % of families and usually allow accurate presymptomatic diagnosis of at-risk relatives in families with two or more affected members but are unhelpful in families with only a single case (Wiggs et al. 1988). Mutation detection rates of 80 % or more were reported by Lohmann et al. (1996) and Houdayer et al. (2004) using a variety of molecular genetic techniques. Germline deletions are not infrequent, and an appropriate deletion scanning approach (e.g., multiplex ligation-dependent probe amplification (MLPA)) should be performed as part of the mutation detection strategy. RB1 mutations are heterogeneous, and generally, no clear genotype-phenotype associations have been described except that promoter mutations and some intragenic mutations with residual protein activity may display reduced penetrance (Kratzke et al. 1994; Cowell et al. 1996). The identification of a germline mutation in a unilateral sporadic case distinguishes those individuals with simplex retinoblastoma and a new germline mutation from nonhereditary cases. When a germline mutation is identified, other relatives can be tested. When histopathological material (formalin fixed paraffin embedded) from a deceased patient is available, mutation analysis can be undertaken on the Tumours tissue and further investigations undertaken to determine if a characterized mutation is somatic or germline. Indeed, if Tumours tissue is available, then this will be analyzed first to define the two RB1 mutations. Constitutional DNA (e.g., blood) is then tested to determine if both are somatic or one is germline (as mosaicism is not uncommon, blood DNA testing must use methods sensitive enough to detect mosaic mutations). If two mutations are detected in the Tumours and neither is detected in blood, then siblings do not require surveillance though there will still be a small risk to the patient’s offspring (from mosaicism). A pathogenic mutation can be detected in blood DNA of ~95 % of patients with hereditary retinoblastoma. In patients who present with developmental delay and/or dysmorphic features, FISH or array CGH analysis to detect copy number variations is performed first.
Children with an RB1 mutation should undergo an eye examination every 3–4 weeks until age 1 year and then every 2–3 months until age 5 years followed by annual examination for the rest of their lives (Lohmann et al. 2011). There are no specific guidelines for the detection of non-ocular second Tumours, but complaints of bone pain or lumps should be investigated promptly.
In the absence of molecular genetic diagnosis, the parents and siblings of all children with retinoblastoma should undergo thorough ophthalmological assessment. Offspring and siblings of retinoblastoma patients should be followed up from birth, with a complete retinal examination. For example, examination should be performed at birth and monthly until 3 months of age without anesthesia, then under general anesthetic every 3 months until the age of 2 years, then every 4 months until the age of 3 years. Thereafter, examinations without general anesthesia can be performed every 6 months until the age of 5 years, and then every 12 months until the age of 11 years (although the frequency and duration of screening can be modified according to the results of DNA analysis). Parents of children with apparently sporadic retinoblastoma must be examined to exclude a regressed Tumours or retinoma because this would identify the child as having a germline RB1 gene mutation. Children with retinoblastoma will require careful follow-up to detect new Tumours or recurrence. New Tumours occurred in 11 % of children studied by Salmonsen et al. (1979). Traditionally, all children with retinoblastoma are followed up with regular retinal examinations under general anesthesia. However, the application of DNA techniques to identify RB1 gene mutations in Tumours and constitutional DNA enables follow-up to be restricted to those shown to have germline RB1 mutations. This not only enhances the management of at-risk relatives but is also a cost-effective strategy (Noorani et al. 1996). Survivors of bilateral retinoblastoma (and unilateral retinoblastoma in those with germline retinoblastoma gene mutations) are at high risk of osteosarcoma in adolescence and of the occurrence of other cancers.