Retinoblastoma

Literature and regulatory information updated: 4 October 2026.

Heritable and non-heritable retinoblastoma: germline predisposition associated with RB1 variants, Knudson's hypothesis, mosaicism, 13q14 deletion syndrome, low-penetrance variants, RB1+/+ retinoblastoma with MYCN amplification, second primary tumours and family screening

0. Initial Reference Table

1. Introduction

Retinoblastoma is an embryonal retinal neoplasm caused by inactivation of the tumour suppressor gene RB1, located at 13q14. It affects approximately 1 in 20,000 live births and predominantly occurs in children younger than seven, with 90% diagnosed before age five; rarely, it may already be present prenatally [8]. It is the most common intraocular childhood cancer, accounts for 2.5–4% of all paediatric tumours and is the principal cause of enucleation in children [10]. GeneReviews estimates incidence at 1:15,000 to 1:20,000 live births, based on population registries from the Netherlands, Sweden and Finland [1]. Worldwide, 8,000 to 9,000 children are diagnosed annually, most living in low- or middle-income countries where mortality remains high [11, 27].

The history of retinoblastoma parallels that of cancer genetics. The historical tumour designation is glioma retinae [1]. In 1963, Lele and colleagues described a partial long-arm deletion of a group D chromosome in a patient with intellectual disability, growth retardation and retinoblastoma, an observation that would locate the predisposition locus on chromosome 13 [28]. In 1971, Alfred Knudson analysed 48 retinoblastoma cases and published data and, using Poisson statistics, proposed that the tumour results from two mutational events: in heritable disease, the first is transmitted through the germline and the second occurs in a somatic cell; in non-heritable disease, both occur somatically. Knudson calculated that each carrier develops an average of three tumours, explaining the coexistence of unaffected carriers, unilateral disease, and bilateral or multifocal disease [5]. Five years later, Knudson and colleagues directly linked the chromosomal deletion to retinoblastoma, consolidating the concept of cancer as a genetic disease of somatic cells [29]. In 1986, Friend and colleagues isolated a complementary DNA segment detecting a locus of at least 70 kb at 13q14, frequently deleted in retinoblastomas and osteosarcomas and lacking a transcript in these tumours: the first human tumour suppressor gene had been cloned [30].

The geneticist's role in retinoblastoma is defined and practical: informing the family, general practitioner and other clinicians about aetiology, recurrence risks for different relatives and second tumour risk; arranging genetic testing of the affected child and communicating results; extending testing to at-risk relatives; and arranging prenatal diagnosis when requested by the family [8]. Identification of the germline variant allows relatives who have not inherited it to be excluded from intensive ophthalmological surveillance with repeated examinations under anaesthesia, with documented reductions in costs and anaesthetic exposure [31].

2. Molecular Genetics and Pathophysiology

2.1 The RB1 gene and pRB protein

RB1 lies at 13q14.2 and encodes the ubiquitously expressed nuclear protein pRB, which regulates the transition from G1 to S phase of the cell cycle. In its hypophosphorylated state, pRB binds E2F-family transcription factors through its pocket domain, blocks the transactivation domain and recruits chromatin-modifying enzymes that actively repress E2F-responsive genes. Before S-phase entry, cyclin/cyclin-dependent kinase (CDK) complexes phosphorylate pRB, reducing its binding capacity and releasing associated proteins to permit DNA replication [1, 4]. Beyond cell cycle control, pRB contributes to constitutive heterochromatin stabilisation, apoptosis and cell differentiation [4, 10].

2.2 Cell of origin and tumour progression

Almost all retinoblastomas arise after biallelic RB1 inactivation in a developing retinal cone precursor. Xu and colleagues demonstrated that postmitotic human cone precursors are specifically sensitive to pRB depletion: suppression induces proliferation dependent on factors strongly expressed in maturing cones, such as MYCN and MDM2, as well as the related protein p107, SKP2 and reduced p27; transformed cells form tumours with typical retinoblastoma histology in orthotopic xenografts [6]. This cell-specific circuit explains the retinal tropism of a germline predisposition present throughout the body [11].

RB1 loss of function initiates retinoma, a benign growth-arrested lesion, and induces genomic instability, but is insufficient alone to cause retinoblastoma; this instability leads to additional changes in other genes that determine malignant progression [1, 10]. The RB1 variant spectrum in patients with retinoma and a family history of retinoblastoma does not differ from that in bilateral retinoblastoma [1].

2.3 First and second hits

In heritable retinoblastoma, the first hit is usually a point variant or small intragenic alteration (approximately 90–95%), and a deletion in 5–10%. The second hit is loss of heterozygosity (LOH) spanning an extensive 13q segment in approximately 70% of tumours; less frequently (approximately 30%), it is a second independent somatic variant, and in a few per cent, a methylation alteration [8]. Comparative genotyping of polymorphisms within and around RB1 in blood and tumour DNA demonstrates loss of the normal allele, with hemizygosity or duplication of the mutant allele (homozygosity), in 60–70% of enucleated-eye tumours [1]. RB1 promoter hypermethylation, silencing gene expression, occurs in approximately 15% of tumours in patients with sporadic unilateral retinoblastoma; this epigenetic pathway is not germline-transmissible [1]. The new mutation rate in RB1 is high, and 85% of new germline variants arise on the paternal allele [8].

Most pathogenic RB1 variants introduce a premature termination codon through single-base substitution, frameshift variants or out-of-frame exon skipping caused by splicing variants. Variants associated with low-penetrance retinoblastoma produce proteins retaining partial activity [1]. The mechanism is loss of function; RB1 variants are distributed throughout the gene without hotspots, except a small number of recurrent CGA>TGA variants [31, 32].

2.4 Genotype–phenotype correlation and penetrance

In most families with heritable retinoblastoma, all carriers develop multiple tumours in both eyes. However, the founder (the first affected family member) commonly has unilateral disease, either because null variants (frameshift or nonsense) have variable expressivity or because the founder is mosaic for the pathogenic allele. With few exceptions, RB1 null alleles have complete penetrance (above 99%) [1]. The Oxford Desk Reference estimates penetrance of inactivating variants at 90%, with approximately 30% presenting unilaterally and the remainder bilaterally [8].

Fewer than 10% of families have a "low-penetrance" phenotype with reduced expressivity (more unilateral disease) and reduced penetrance (25% or less). This phenotype is usually associated with missense variants, in-frame deletions or insertions, specific splicing variants, certain exon 1 insertion–deletion variants and promoter variants [1]. Individuals with partial loss-of-function variants have lower penetrance and more unilateral disease; mosaic individuals have reduced penetrance (approximately 60%) and expressivity [8]. A 2026 systematic review of 250 genetically confirmed low-penetrance retinoblastoma cases estimated penetrance at 50% without age adjustment and 64% with adjustment, more frequent paternal transmission (odds ratio [OR] 6.24), greater likelihood of unilateral disease (OR 9.3), later diagnosis (13 versus 6.5 months) and variant concentration in the N- and C-terminal pRB domains, proposing the term "intermediate-penetrance" retinoblastoma [33].

A third family group shows differential penetrance according to the parental origin of the pathogenic allele (parent-of-origin effect). In families segregating c.1981C>T (p.Arg661Trp), a carrier's probability of remaining unaffected is 90.3% with maternal inheritance and 32.5% with paternal inheritance. The proposed mechanism involves the differentially methylated CpG85 island in RB1 intron 2, favouring maternal allele expression: the maternal p.Arg661Trp allele retains sufficient suppressor activity, whereas the lower-expressed paternal allele behaves as a null allele [1, 34].

A practical estimate of family penetrance divides the number of affected eyes by the total number of eyes in obligate carriers. For example, a grandfather with unilateral retinoblastoma, an unaffected parent and a child with bilateral disease yield three affected eyes among six eyes in three carriers, corresponding to 50% penetrance in that family; the grandfather may, however, have been a germline mosaic [8].

2.5 Mosaicism

Mosaicism for the initiating RB1 variant is documented in approximately 10% of families with heritable retinoblastoma, either in the proband or a parent, and may be associated with unilateral or bilateral disease depending on variant type and the proportion of affected cells [8, 9]. In a study of 1,020 families using allele-specific PCR for 11 recurrent CGA>TGA variants, mosaicism was evident in 5.5% of bilateral probands and 3.8% of unilateral probands, and half of mosaic variants were detectable only with this highly sensitive technique [32]. Among 142 unaffected parental pairs, only one (0.7%) showed somatic mosaicism for the child's variant, indicating that mosaicism for an RB1 variant is highly likely to manifest as retinoblastoma [32]. Targeted deep sequencing in 124 unaffected, non-carrier parental couples of bilateral probands identified one mosaic parent, corresponding to a maximum parental mosaicism recurrence risk of 0.4%; sibling follow-up recorded one recurrence, corresponding to a maximum risk of 0.8% [35].

2.6 13q14 deletions and the role of MED4

Interstitial 13q deletions including RB1 may be submicroscopic or cytogenetically visible. In a series of 63 patients, sizes ranged from 4.2 kb to over 33.43 Mb, with non-recurrent breakpoints and signatures of non-homologous end joining [36]. Deletions larger than 1 Mb including MED4 are associated with milder retinoblastoma expression (unilateral disease or no tumour) [36]. Dehainault and colleagues demonstrated that RB1−/− retinoblastoma cells cannot survive without MED4, a Mediator complex subunit located centromeric to RB1; contiguous MED4 loss explains the low penetrance of large deletions encompassing both genes [37]. Neighbouring gene haploinsufficiency explains extraocular manifestations: NUFIP1 and PCDH8 loss was associated with psychomotor delay, MTLR1 with microcephaly and EDNRB with feeding difficulties [36].

2.7 RB1+/+ retinoblastoma with MYCN amplification

A second genetic abnormality capable of causing retinoblastoma is MYCN amplification (2p24.3), responsible for a small proportion of non-heritable retinoblastomas with normal RB1 [10]. Rushlow and colleagues analysed 1,068 non-familial unilateral retinoblastomas and found no RB1 variants in 29 (2.7%); 15 of these RB1+/+ tumours had high-level MYCN amplification (28 to 121 copies), whereas none of 93 primary RB1−/− tumours did. RB1+/+ MYCN^A^ tumours expressed functional pRB, had fewer copy number changes characteristic of retinoblastoma and showed aggressive histology; median diagnosis age in the 17 affected children was 4.5 months, versus 24 months in children with non-familial unilateral RB1−/− retinoblastoma [7]. GeneReviews estimates that approximately 1.5% of children with unilateral retinoblastoma and no family history have high-level MYCN amplification without inactivating RB1 variants; whether all these cases are non-heritable remains unestablished [1]. MYCN amplification is not exclusive to RB1+/+ tumours: in a German series of 139 enucleated unilateral retinoblastomas, 7.2% had MYCN gain or amplification, all with alteration of at least one RB1 allele, and more frequent secondary glaucoma and massive choroidal and scleral invasion [38]. Somatic tumour MYCN amplification should be distinguished from germline MYCN variants, whose loss of function causes Feingold syndrome type 1, a malformation syndrome without an oncological surveillance protocol [39].

2.8 Somatic RB1 in other tumours

Sporadic tumours such as urothelial bladder carcinoma, endometrial carcinoma, sarcomas, squamous cell lung carcinoma and lung adenocarcinoma frequently contain somatic RB1 variants absent from the germline. When these tumours occur in isolation without other findings suggesting heritable retinoblastoma, predisposition is not heritable, and an RB1 variant identified only in tumour tissue should not be interpreted as germline predisposition without constitutional confirmation [1]. Beyond the retina, RB1 loss contributes to other human cancer progression through loss of cell cycle control and genomic instability [10].

3. Clinical Spectrum and Manifestations by Organ System

3.1 Terminology

Retinoblastoma terminology frequently overlaps and requires precision. Unifocal describes a single tumour in one eye; unilateral means only one eye is affected (often used synonymously with unifocal); multifocal means several independent tumours in the same eye; bilateral means tumours in both eyes; and trilateral means bilateral retinoblastoma associated with pineoblastoma, which may very rarely occur without evident ocular tumours [8]. Genetic or heritable retinoblastoma is defined by a germline RB1 variant present in all body cells, requiring only one somatic hit to initiate a tumour; it includes all trilateral, bilateral and multifocal cases and approximately 10% of unilateral unifocal cases. Non-genetic or non-heritable retinoblastoma occurs when both hits are confined to the tumour and includes most unilateral unifocal cases. "Sporadic" is used both for absence of family history and for non-genetic disease and should therefore be avoided or explicitly defined [8].

Table 1 summarises the differences between the two forms.

Table 1. Heritable and non-heritable retinoblastoma [1, 8, 10, 40]

3.2 Ocular manifestations

Approximately 60% of patients have unilateral retinoblastoma, diagnosed at a mean age of 24 months, and approximately 40% have bilateral disease, diagnosed at a mean of 15 months [1]. Unilateral tumours are usually unifocal, but some patients have multifocal tumours in one eye; intraocular seeding may mimic primary multifocal growth, and in most children with unilateral disease and no family history the tumour is too large to determine whether there is a single focus. In most children with bilateral disease, both eyes are affected at initial diagnosis, often with multiple tumours; some initially diagnosed with unilateral disease subsequently develop a contralateral tumour [1]. Probands generally present in three settings: negative family history with unilateral disease (60%), negative family history with bilateral disease (30%), or positive family history with unilateral or bilateral disease (approximately 10%); in the latter group, surveillance from birth often identifies tumours within the first month [1].

The most common presenting sign is a white pupillary reflex (leucocoria), followed by strabismus, which may accompany or precede leucocoria. Unusual presentations include glaucoma, orbital cellulitis, uveitis, hyphaema and vitreous haemorrhage, more frequent in older children [1]. In the 2017 global study of 4,351 patients from 153 countries, leucocoria prompted referral in 62.8%, strabismus in 10.2% and proptosis in 7.4% [12].

Retinoma is a benign retinal tumour with spontaneous growth arrest that may occur within retinal scars; calcified eyes with phthisis bulbi may result from spontaneous retinoblastoma regression through vascular occlusion [1]. Retinoma or a regressed tumour in a parent has the same genetic significance as retinoblastoma, so parents of an affected child should always be examined by an experienced ophthalmologist [8].

3.3 Extraocular extension and metastases

Undiagnosed retinoblastoma grows locally to fill the eye and extends into the optic nerve, orbit and brain; metastases predominantly involve bone marrow, bone and the central nervous system, and their occurrence relates to how long an active tumour remains in the eye [10]. Two high-risk metastatic profiles are described: a child aged three to five with advanced unilateral intraocular retinoblastoma and apparent symptoms for months or years without diagnosis or treatment; and a child with bilateral disease receiving repeated unsuccessful treatment over several years for intraocular recurrences in an attempt to preserve vision in one remaining eye, where parental refusal of second-eye enucleation may increase metastatic risk [10]. Heritable and non-heritable disease are not always distinguishable by tumour morphology, growth rate or extent, and there appears to be no intrinsic difference in metastatic risk [10].

3.4 Trilateral retinoblastoma

Pineoblastoma develops in retina-like pineal gland tissue. Pineoblastoma or a midline primitive neuroectodermal tumour associated with retinoblastoma is termed trilateral retinoblastoma; unlike intraocular retinoblastoma, which is usually curable, pineoblastoma is rare and often fatal [1, 10]. Approximately 5% of children with germline RB1 variants are at risk of trilateral retinoblastoma. In a meta-analysis of 174 patients, five-year survival after pineal trilateral retinoblastoma increased from 6% for diagnoses before 1995 to 44% from that year onwards, and survival in non-pineal cases from 0% to 57%, attributed to conventional and high-dose chemotherapy with stem cell rescue [40]. Pineoblastoma is not counted as a second primary tumour; it occurs in 2.4% of second tumour series [8, 41].

3.5 Second primary tumours

Survivors carrying a germline RB1 variant have increased risk of non-ocular neoplasms, termed second primary tumours or subsequent malignant neoplasms [8]. A review of 11 studies with 243 second tumours found osteosarcoma most common (37%), followed by melanoma (7.4%), soft tissue sarcomas (6.9%), brain tumours (4.5%), fibrosarcomas (3.3%), chondrosarcomas (3.3%) and other sarcomas (3.3%), with cumulative incidence of 19% by age 35 and relative risk of 15.4 in studies without selection bias [41]. The Oxford Desk Reference summarises the spectrum as osteosarcoma (relative risk 200–500; 37% of second tumours), particularly after radiotherapy; other sarcomas, including soft tissue sarcomas, fibrosarcomas and chondrosarcomas (14%); malignant melanoma (7%); and brain tumours (4.5%, excluding pineoblastoma) [8]. Tumours develop most frequently between the second and fourth decades, and excess retinoblastoma-related tumour risk is approximately 6–10% [8]. Soft tissue sarcomas are mainly leiomyosarcomas and rhabdomyosarcomas [1].

In a United States cohort of 1,601 survivors, new cancer risk in heritable disease (standardised incidence ratio [SIR] 19) greatly exceeded non-heritable disease (SIR 1.2), and radiotherapy increased risk 3.1-fold; cumulative new cancer incidence 50 years after diagnosis was 36% in heritable and 5.7% in non-heritable disease, with persistent excess sarcoma, melanoma, brain cancer and nasal cavity cancer [42]. In the Dutch registry, after median follow-up of 21.9 years, SIR was 20.4 in heritable and 1.86 in non-heritable disease; excess epithelial carcinomas (lung, bladder and breast) appeared after 30 years of follow-up, and absolute excess risk rose to 26.1 cases per 1,000 person-years after 40 years [13]. Second tumour incidence exceeds 50% among patients treated with external beam radiotherapy; heritable retinoblastoma survivors unexposed to high-dose radiotherapy have lower but still elevated risk [1].

Chemotherapy also contributes to risk. Among 906 heritable survivors, alkylating agents increased bone tumour risk (hazard ratio [HR] 1.60) and leiomyosarcoma risk (HR 2.67; HR 5.17 when administered before age one), and leiomyosarcoma was more frequent after alkylating agents plus radiotherapy than after radiotherapy alone (5.8% versus 1.6% by age 40) [43]. At the German referral centre, among 648 heritable patients, radiotherapy and systemic chemotherapy increased second tumour incidence 3.0-fold and 1.8-fold respectively; radiotherapy was associated with tumours in the periorbital field, and chemotherapy was the principal risk factor at other sites [44]. Genotype also modifies risk: second tumour incidence per 1,000 person-years was 8.4 in heterozygotes, 2.1 in mosaics, 10.3 in regular-penetrance variant carriers and 3.2 in incomplete-penetrance variant carriers [45]. Constitutional RB1 mutation increases secondary neoplasm risk primarily in the lungs, bladder, bones, soft tissues, skin and brain, particularly in irradiated children [10].

3.6 13q deletion syndrome

Partial deletion of chromosome 13's long arm produces a variable phenotype that can be divided into three groups according to the deleted segment. Common findings include prenatal-onset growth retardation; intellectual disability and microcephaly, with a tendency towards trigonocephaly and brain abnormalities in the holoprosencephaly spectrum; prominent nasal root, hypertelorism, ptosis, epicanthic folds, microphthalmia, colobomas and retinoblastoma (usually bilateral); prominent maxilla, micrognathia and prominent, oblique, low-set ears; short, webbed neck; small or absent thumbs, fifth-finger clinodactyly, fusion of metacarpals 4 and 5, talipes equinovarus and a short hallux; heart disease; hypospadias and cryptorchidism; and focal lumbar agenesis [28]. Occasional abnormalities include optic nerve and retinal dysplasia, iris hypoplasia, sensorineural deafness, oesophageal atresia with tracheo-oesophageal fistula, pyloric stenosis, imperforate anus, Hirschsprung disease, coeliac disease, renal abnormalities, neural tube defects, factor VII deficiency and VACTERL association [28].

Natural history depends on the deleted segment: proximal deletions not extending to q32 cause mild to moderate intellectual disability, variable minor anomalies and growth retardation; q14 involvement is associated with substantial retinoblastoma risk, intellectual disability, growth retardation and major malformations (microcephaly, central nervous system defects, distal limb abnormalities, ocular defects and gastrointestinal malformations); more distal deletions (q33–q34) cause severe intellectual disability, usually without growth retardation or gross structural malformations [28]. Although most patients with 13q14-involving deletions develop retinoblastoma, an estimated 13–20% remain unaffected [28]. In Mitter and colleagues' series, small deletions within 13q14 were associated with macrocephaly, tall stature, obesity and motor or language delay; medium deletions (13q12.3–q21.2) with characteristic facial dysmorphism, mild to moderate psychomotor delay, short stature and microcephaly; and large deletions (13q12–q31.2) with craniofacial dysmorphism, short stature, microcephaly, mild to severe psychomotor delay, hypotonia, constipation and feeding difficulties [36]. A substantial proportion of children with a deletion have no evident dysmorphism or developmental delay [8].

3.7 Retinoblastoma with MYCN amplification

RB1+/+ MYCN^A^ retinoblastoma presents as a large, rapidly growing unilateral tumour in very young infants (median 4.5 months), with undifferentiated, aggressive histology; no germline predisposition is known, although whether all cases are non-heritable remains unestablished [1, 7].

4. Clinical Recognition and Warning Signs

Retinoblastoma should be suspected in any child with leucocoria, strabismus, altered eye appearance or reduced visual acuity [1]. Early detection relies mainly on families: in a series of 1,831 patients, disease was first detected by family or friends in 80%, paediatricians in 8% and ophthalmologists in 10%, despite routine paediatric red-reflex screening. Presentation with leucocoria was associated with five-year patient survival above 86%, but five-year ocular survival was only 4% in unilateral and 29% in bilateral eyes; children with a family history monitored from birth were diagnosed earlier (8 months), at less advanced stages and with better ocular survival [46].

Heritable retinoblastoma should be suspected in any patient with retinoblastoma (unilateral unifocal, unilateral multifocal or bilateral), retinoma or a retinoblastoma family history [1]. Genetic warning signs include bilateral or multifocal disease, first-year diagnosis, retinoblastoma family history, retinoma or phthisis bulbi in a parent, pineoblastoma or a midline neuroectodermal tumour, and a personal or family history of osteosarcoma, soft tissue sarcoma or melanoma at a young age. Dysmorphism, developmental delay or congenital malformations in a child with retinoblastoma should prompt chromosomal testing directed at 13q14 [1, 8]. A large unilateral tumour in an infant only a few months old without a family history should suggest MYCN amplification [7].

History-taking should include a pedigree spanning at least three generations, documenting retinoblastoma and all other tumour types. Physical examination in older patients with heritable retinoblastoma should include a brief skin examination; parents of a child with retinoblastoma should always be examined by an experienced ophthalmologist for spontaneously regressed tumours [8]. In infants from heritable retinoblastoma families, early diagnosis depends on scheduled surveillance rather than signs, because tumours causing leucocoria are, by definition, advanced [19, 46].

5. Diagnosis: Criteria, Tests and Algorithms

5.1 Clinical diagnosis and staging

Retinoblastoma diagnosis is established through retinal examination under full mydriasis by an ophthalmologist; confirmation and disease extent assessment use examination under anaesthesia with indirect ophthalmoscopy. Ocular imaging assists confirmation; histopathological examination is unnecessary, and biopsy is contraindicated because it may disseminate tumour outside the eye [1]. Intraocular and extraocular extent is assessed under anaesthesia and by ultrasound or magnetic resonance imaging (MRI), focusing on the tumour–optic nerve relationship; brain MRI assesses pineoblastoma. Computed tomography (CT) is not recommended because radiation increases second tumour risk [1]. For very large tumours with extraocular risk factors, or when examination of an enucleated eye shows optic nerve invasion or other risk factors, bone marrow aspiration and cerebrospinal fluid examination may be performed [1]. In children with a family history and rare presentations with strabismus or low vision, tumours may be small and detectable by optical coherence tomography [1].

Each affected eye is staged. The International Classification of Intraocular Retinoblastoma (ICRB), proposed by Linn Murphree in 2005, replaced Reese–Ellsworth classification and assigns eyes to groups A–E according to natural history and risk of eye loss after primary therapy; group A comprises small tumours away from the fovea and optic disc, while group E comprises extensive disease without a reasonable prospect of preservation [14]. The 8th edition of the AJCC Cancer Staging Manual adopted cTNMH, integrating intraocular extent (cT1 to cT3), extraocular extent (cT4), nodal and metastatic involvement and, as a new component, heritable category H [1, 15].

Table 2. AJCC heritable staging (H category) [1, 15]

5.2 Diagnosis of heritable retinoblastoma

Heritable retinoblastoma is diagnosed in a proband with retinoblastoma or retinoma and a retinoblastoma family history; because most patients lack a family history, identification of a heterozygous germline pathogenic or likely pathogenic RB1 variant is necessary. A variant of uncertain significance neither establishes nor excludes diagnosis [1]. Germline variant probability depends on family history and tumour presentation (Table 3).

Table 3. Germline RB1 variant probability in a proband [1]

In simplex bilateral cases, conventional molecular methods identify RB1 variants in 90–97%; the remainder may have translocations, deep intronic splicing variants or low-level mosaic variants [1]. The Oxford Desk Reference reports detection above 95% in bilateral or familial disease and approximately 8–10% in unilateral children [8]. Lohmann and colleagues identified constitutional variants in 17% (6/36) of isolated unilateral patients whose tumour variants could be characterised; one variant was associated with reduced expressivity [47].

5.3 Laboratory strategy

In bilateral, familial unilateral or multifocal unilateral disease, peripheral blood DNA undergoes RB1 sequencing and targeted deletion/duplication analysis. In unilateral unifocal disease with negative family history, the approach depends on tumour tissue availability: without tumour, blood is tested; with tumour, tumour DNA is tested first, then identified variants are sought in blood. If no tumour variants are identified, RB1 promoter CpG island methylation is analysed and, if normal, MYCN amplification is assessed [1]. Table 4 summarises each method's yield.

Table 4. Molecular tests in heritable retinoblastoma [1]

Tumour analysis should routinely use fresh tissue whenever enucleation is performed. A variant found in tumour but not blood cannot have been inherited from a parent, and siblings need neither screening nor testing; low-level mosaicism remains possible in the affected child, whose future offspring should therefore be tested [8]. When a pathogenic blood variant is identified, targeted deep sequencing of the proband and parents should be considered for mosaicism detection to inform recurrence risk [8, 35]. Cell-free DNA analysis of anterior chamber aqueous humour can identify biallelic RB1 variants without enucleation, but is not yet clinically available to exclude heritable predisposition [1].

5.4 Chromosomal testing

Approximately 3% of patients have an interstitial 13q14 deletion potentially undetected by conventional molecular techniques, so the Oxford Desk Reference recommends karyotyping with RB1 fluorescence in situ hybridisation (FISH) in all affected children [8]. GeneReviews estimates 13q14 deletions in 6–8% of retinoblastoma patients, cytogenetically visible deletions in up to 5% of unifocal and 7.5% of multifocal probands, and recommends considering CMA particularly with developmental delay or congenital anomalies; targeted deletion/duplication analysis detects deletions ranging from one exon to the entire gene but may not define breakpoints or adjacent gene loss [1]. Differing estimates reflect method resolution (conventional karyotyping versus arrays densely covering 13q14). Smith recommends chromosomal testing in all retinoblastoma patients [28]. If a parent carries a balanced translocation or another chromosome 13 rearrangement, siblings are at risk of inheriting an unbalanced rearrangement [1].

5.5 Linkage analysis

When the proband's variant remains unidentified, linkage analysis with polymorphic RB1 markers can exclude siblings lacking the risk haplotype from screening in informative families: comparison of genotypes in unaffected parents and affected and unaffected children gives a 1 in 4 probability of excluding a sibling. Tumour LOH defines the normal allele through the lost haplotype and permits identification of the risk allele [8]. However, Rushlow and colleagues documented an incorrect linkage conclusion in a two-generation family with a mosaic founder parent [32].

5.6 Histopathology after enucleation

Optic nerve invasion beyond the lamina cribrosa combined with massive choroidal invasion requires preventive or adjuvant chemotherapy [10]. In the prospective Children's Oncology Group (COG) ARET0332 study, with central review of 321 eyes enucleated for unilateral disease, institutional diagnosis discordance reached 23% among patients with high-risk histopathology; two-year event-free survival was 0.96 in high-risk patients receiving adjuvant chemotherapy and 0.99 in patients without these factors undergoing surveillance alone. Focal choroidal invasion below 3 mm with prelaminar or laminar optic nerve invasion had no recurrences and may not require adjuvant chemotherapy [48].

5.7 Genetic test limitations and report content

No single method detects all predisposition causes. Sequencing RB1 exons, exon–intron junctions and promoter does not itself detect deletions/duplications of one or more exons, requiring MLPA or copy number analysis; chromosomal microarray defines 13q14 deletion extent and contiguous gene loss, such as MED4, but misses balanced translocations and rearrangements without imbalance; deep intronic splicing-altering variants escape junction-only panels and may require RNA studies [1]. Low-level mosaicism is the principal blood false-negative cause: conventional sequencing may miss it (half the mosaic variants in Rushlow and colleagues' series were identified only through allele-specific PCR), requiring targeted deep sequencing or allele-specific PCR with report-declared sensitivity [1, 32, 35]. Promoter hypermethylation and MYCN amplification are assessed only in tumour tissue, and MYCN amplification is not a germline variant [1, 7]. Thus, a negative blood result in bilateral, multifocal or familial disease does not exclude predisposition: the child remains H1 and at-risk relatives undergo ophthalmological surveillance [1, 8]. A negative result in unilateral unifocal disease reduces risk to H0* only if the test is certified to detect low-level mosaicism, ideally after both tumour variants are identified [1].

The report should state the gene, the versioned reference transcript (for RB1, the MANE Select transcript is NM_000321.3 [3]), Human Genome Variation Society (HGVS) nomenclature at the coding DNA and protein levels, zygosity and allele fraction (needed to recognise mosaicism), classification according to the criteria of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP), with the criteria applied, the sample type (blood, tumour, aqueous humour) and technical limitations [49]. The variants NM_000321.3:c.1981C>T p.(Arg661Trp), of low penetrance and with a parent-of-origin effect, and NM_000321.3:c.958C>T p.(Arg320Ter), one of the recurrent CGA>TGA variants, are normalised in ClinVar and classified as pathogenic, with multiple submitters and no conflicts [50]. Part of the older literature described RB1 variants by position in GenBank genomic sequence L11910 (g. notation) or by exon and codon number; when a report is compared with those publications, both numbering systems should be given without mixing them [51]. A variant of uncertain significance does not become pathogenic through the occurrence of a tumour in a carrier, response to treatment or partial segregation in a family with reduced penetrance; segregation testing in this context is research for reclassification rather than predictive testing, and a variant of uncertain significance should not be used to exclude relatives from surveillance or as the basis for prenatal or preimplantation diagnosis [1, 49]. A variant found only in the tumour should not be interpreted as germline without confirmation in constitutional tissue [1].

6. Differential Diagnosis

Several inherited childhood eye disorders can clinically mimic retinoblastoma, as can ocular Toxocara canis infestation [1]. In retinal dysplasia, a bilateral congenital abnormality with retinal detachment at birth and frequent central nervous system malformations, the principal conditions to exclude are retinoblastoma, severe familial exudative vitreoretinopathy and, in premature infants, retinopathy of prematurity [52].

Table 5. Differential diagnosis of childhood leucocoria and retinal masses [1, 52]

7. Treatments and Risk Management Strategies

7.1 Principles

Treatment aims, in order of priority, to preserve life and then vision. Because treatment may be complex, multidisciplinary teams experienced in retinoblastoma should provide care, including ophthalmology, paediatric oncology, pathology and radiation oncology [1]. Choice depends on each eye's stage, number of foci, tumour location and size, vitreous seeding, potential for useful vision, extraocular extent and available resources [1]. Eye treatment options include enucleation, cryotherapy, laser, systemic or local chemotherapy (including intra-arterial chemotherapy) combined with or followed by laser or cryotherapy, episcleral plaque brachytherapy and, as a last resort, external beam radiotherapy [1]. Progress over the last two decades involved systemic chemotherapy with focal treatments in the late 1990s, allowing radiotherapy to be progressively abandoned, followed approximately ten years later by intra-arterial, intravitreal and intracameral in situ chemotherapy, increasing ocular salvage and reducing systemic chemotherapy exposure, always within what Munier and colleagues term the "metastatic grace period" [16].

None of the cytotoxic agents used intra-arterially or intravitreally has retinoblastoma as an approved label indication: injectable melphalan is FDA (Food and Drug Administration) approved for the palliative treatment of multiple myeloma, and topotecan (Hycamtin®) for ovarian cancer, small cell lung carcinoma and cervical carcinoma [53, 54]. Cyclophosphamide is the only one of these agents whose FDA label expressly includes retinoblastoma among susceptible malignancies [17]. Table 6 presents the recommended doses in retinoblastoma, as described in cooperative trials and specialist-centre series, and, in a separate column for comparison, the label dose in the authorised indication; these doses differ from the label in indication, route, age or dose, so their use is off-label; only intravenous cyclophosphamide has retinoblastoma within its label indication, and the GALOP regimen dose exceeds the initial course in the label [17, 55].

7.2 Focal treatments

Laser (photocoagulation or transpupillary thermotherapy), cryotherapy and episcleral plaque brachytherapy are used for accessible tumours, alone for small tumours or as consolidation after chemoreduction [1, 8]. Murphree and colleagues showed that systemic carboplatin (560 mg/m²) followed by 15 to 30 minutes of continuous diode laser hyperthermia (thermochemotherapy) controlled all Reese–Ellsworth group 1 and 2 eyes but failed in eyes with vitreous seeding; radiation retinopathy occurred in all eyes treated with iodine-125 plaques after prior chemotherapy in this study [56].

7.3 Systemic chemotherapy

First-line intravenous chemotherapy usually combines vincristine, etoposide and carboplatin (VEC) in monthly cycles, followed by intensive sequential focal treatments during examinations under anaesthesia every two to three weeks; chemoreduction fails with diffuse vitreous or extensive subretinal seeding [8, 56]. In COG ARET0331, six cycles of vincristine and carboplatin without etoposide, with local treatment after the first cycle, produced two-year event-free survival of 65% in 25 group B eyes, with cumulative enucleation incidence of 15% and external beam radiotherapy incidence of 10%; all patients with treatment failure were salvaged [57]. The trial discussion notes that etoposide-associated secondary leukaemia was described mainly at cumulative doses above those used for retinoblastoma and with other regimens; in a cohort of 187 heritable patients treated with VEC without radiotherapy, receiving etoposide 10–12 mg/kg, only one leukaemia (promyelocytic) occurred [57]. The Oxford Desk Reference specifies systemic chemotherapy first-line, followed by local treatments (plaques, cryotherapy for accessible tumours or lens-sparing radiotherapy); large unilateral tumours are usually offered enucleation or chemotherapy delivered directly to the eye [8].

7.4 Intra-arterial chemotherapy

Intra-arterial chemotherapy (IAC) entails selective ophthalmic artery catheterisation through femoral access, under anaesthesia and anticoagulation, with cytotoxic infusion over approximately 30 minutes. In Abramson and colleagues' phase I/II study of ten children with Reese–Ellsworth group V eyes requiring enucleation, seven eyes were preserved, with minimal systemic toxicity (one grade 3 neutropenia case) and transient conjunctival and eyelid oedema in three patients [58]. The same group's four-year experience, involving 95 eyes and 289 injections, achieved successful catheterisation in 98.5% of procedures and two-year ocular event-free survival of 70% overall, 81.7% with primary IAC and 58.4% after intravenous chemotherapy or radiotherapy failure, without permanent extraocular complications [59]. Dose was determined by age and vascular anatomy [59].

In Brazil, a series of 357 eyes (300 patients) treated between 2010 and 2023 with 1,536 infusions (median four cycles per eye) used one to three agents per infusion: melphalan 3.0–7.5 mg, topotecan 0.3–2.0 mg and carboplatin 20–50 mg. Ocular survival was 90% at one year, 89% at two years and 86% at five years, without differences between primary, secondary or bridging indications; subretinal seeding increased enucleation risk, while infusion through the ophthalmic artery ostium in more than half the sessions was protective. The most frequent adverse effects were vascular, ischaemic or atrophic changes in the retina and choroid; no patient received external beam radiotherapy [60]. Other local effects include transient periocular erythema and eyelash madarosis, and access-limb vasospasm may require treatment interruption [61].

The 2026 Cochrane review of six studies concluded with moderate certainty, based on one randomised trial (143 eyes), that IAC probably increases globe preservation compared with intravenous chemotherapy (HR 2.01; 260 versus 454 eyes preserved per 1,000), without an overall survival difference (HR 0.97) and, with low certainty, no difference in tumour recurrence or grade 3–4 adverse effects; adding intravenous chemotherapy to IAC showed no additional ocular preservation benefit (RR 0.96) or survival benefit, with low-certainty evidence, and second tumours were not assessed [62]. A 2026 meta-analysis of 12 studies also favoured IAC for globe preservation (RR 1.33) and avoidance of enucleation (RR 1.69) [63]. The only randomised trial supporting the Cochrane review was Wen and colleagues' open-label multicentre study at six Chinese hospitals: 143 children with unilateral group D or E retinoblastoma without clinical high-risk factors were randomised to IAC (melphalan with carboplatin or topotecan, four cycles) or intravenous chemotherapy (vincristine, etoposide and carboplatin, six cycles); two-year progression-free globe salvage, the primary outcome, was 53% versus 27% (risk ratio 1.97; 95% CI 1.27–3.07), with less myelosuppression in the intra-arterial arm (51% versus 70%) but ophthalmic artery occlusion in 3% and stenosis in 18% of intra-arterially treated patients [64]. This finding applies to advanced unilateral disease; it was not tested in bilateral disease or patients with clinical high-risk factors, and most remaining IAC evidence is observational [62, 65].

In Portugal, the national intraocular tumour referral centre (Centro Hospitalar e Universitário de Coimbra) described 28 IAC-treated eyes between 2015 and 2020 (130 procedures; median four sessions per eye): 78.6% were preserved, overall survival was 100%, and three patients had permanent adverse effects (cataract, vitreous haemorrhage and choroidal ischaemia) [66]. This retrospective single-centre series describes national experience but cannot be directly compared with the randomised trial.

7.5 Intravitreal chemotherapy

Vitreous seeding is resistant to radiation and systemic chemotherapy. Munier and colleagues reintroduced intravitreal treatment using an injection technique with precautions against extraocular dissemination: in 23 heavily pretreated eyes, 122 melphalan injections (20–30 µg) every 7–10 days preserved 87% of globes, with two-year ocular survival of 84% and all patients alive without extraocular dissemination; peripheral "salt-and-pepper" retinopathy at the injection site occurred in 43% of eyes [67]. Seed eradication yields ocular preservation only if the retinal seed source receives concurrent treatment [67]. Intravitreal topotecan (30 µg in 0.15 mL every three weeks) produced complete regression of refractory or recurrent vitreous seeding in 17 of 17 eyes, without ocular or systemic complications, with 75% of evaluable eyes achieving acuity of 20/200 or better [68].

7.6 Enucleation

Enucleation is offered first-line for large unilateral tumours or when disease is difficult to control with other treatments [8]. The enucleated eye should undergo standardised histopathological examination to identify high-risk factors for metastasis (see 5.6) and, whenever possible, fresh tumour tissue should be collected for molecular testing [8, 48]. In the prospective Grupo de América Latina de Oncología Pediátrica (GALOP) study of 175 evaluable children with non-metastatic unilateral retinoblastoma, children with retrolaminar and/or scleral invasion received adjuvant chemotherapy consisting of four alternating cycles of cyclophosphamide (65 mg/kg/day, with mesna), idarubicin (10 mg/m²/day) and vincristine (0.05 mg/kg/day), and of carboplatin (500 mg/m²/day on days 1 and 2) with etoposide (100 mg/m²/day on days 1 to 3), doses per cycle in children, with no maximum dose defined in the source; three-year event-free survival was 97% and overall survival 98%, with substantial toxicity, including one death from sepsis [55].

7.7 Extraocular and metastatic disease

In COG ARET0321, locoregional disease (stages II–III) received four chemotherapy cycles followed by involved-field radiotherapy (45 Gy), while metastatic or trilateral disease (stages IVa–IVb) received four cycles followed, with at least a partial response, by one high-dose carboplatin, thiotepa and etoposide cycle with autologous haematopoietic stem cell rescue. One-year event-free survival was 88.1% in stages II–III, 82.6% in IVa and 28.3% in IVb or trilateral disease, with substantial toxicity and two treatment-related deaths [69]. Retinoblastoma extending beyond the eye retains an unfavourable prognosis, and trilateral retinoblastoma was almost invariably fatal for decades [10, 40].

7.8 Radiotherapy

External beam radiotherapy is now very rarely used because of associated second tumour risk; proton beam therapy may be used in selected circumstances [8]. Radiation, including radiographs and CT, should be avoided wherever possible in H1 carriers but may be used when indispensable for essential healthcare [1].

Table 6. Drugs used in retinoblastoma: routes, published doses and adverse effects

7.9 Ongoing trials

GeneReviews cites two phase I trials (NCT04428879 and NCT04156347) of episcleral topotecan for active, residual or recurrent intraocular retinoblastoma after first-line treatment [1]. In October 2026, NCT04428879 (Hospital for Sick Children, Toronto) was listed as completed, with 26 participants and primary completion in October 2024, but no registry-published results; NCT04156347 has had unknown status since 2023; the same episcleral topotecan implant (Chemoplaque) is now in a phase II expanded-access study for advanced or refractory eyes (NCT06679634; approximately 30 planned participants, primary safety outcome at 12 weeks, estimated primary completion 2029) [75, 76]. Oncolytic adenovirus VCN-01, designed to replicate in cells with dysfunctional RB1 signalling, showed activity in preclinical models and the first intravitreally treated patients, with local vitreous inflammation and no systemic complications [77]; the corresponding phase I study (NCT03284268, 13 participants) completed in 2024 without registry results, and EMA granted orphan designation for retinoblastoma treatment in November 2024 [78, 79]. Orphan designation is a development incentive, neither efficacy assessment nor authorisation. In October 2026, ClinicalTrials.gov included, among others, the COG phase II trial of intravitreal melphalan with standard chemotherapy for initial treatment of vitreous-seeded eyes (NCT05504291, primary completion planned for 2028) [80]; Institut Curie's RETINO 2018 protocol, including a randomised phase II comparison of IAC with melphalan plus topotecan versus melphalan alone (NCT04681417) [81]; and a prospective study of three adjuvant chemotherapy cycles after enucleation in high-risk patients (NCT05080010, registry not updated since 2021) [82]. None had registry results in October 2026, and no preliminary-result releases, accepted marketing applications or regulatory opinions for retinoblastoma drugs were identified in the preceding 12 months [75, 79]. Periodic whole-body MRI screening for second tumours in H1 carriers remains investigational [1].

7.10 Risk management and surveillance

Ophthalmological screening of at-risk relatives. The American Association of Ophthalmic Oncologists and Pathologists consensus defines any child with retinoblastoma in a parent, sibling or first- or second-degree relative as at risk, recommends dedicated ophthalmological screening for all children at above-population risk, stratifies examination frequency into high, intermediate and low risk based on genetic testing, and advises more frequent examinations, preferably under anaesthesia, in high-risk children [19]. The Oxford Desk Reference recommends examinations under anaesthesia from two to three weeks until age three for any child at increased risk, including siblings of an isolated unilateral case whose risk is only 1%, with decreasing frequency and approximately 14 anaesthetised examinations for most children [8]. Identified carriers should be examined from birth [1].

Table 7. Recommended surveillance [1]

H0* patients retain a low-level mosaicism risk and may develop a contralateral tumour; in unilateral retinoblastoma without a detectable blood variant, contralateral tumour risk is estimated at 3–5% [1, 8]. Successfully treated children require frequent examinations for early detection of new intraocular tumours, and routine ophthalmological examinations should continue until age seven [1, 10].

Second tumours. Patients and general practitioners should maintain a high level of awareness, with expedited referral for any suspicion and regular skin self-examination for melanoma, performed by the patient or partner or, with visual impairment, by the general practitioner or dermatologist [8]. Exposure to DNA-damaging agents (radiotherapy, tobacco and ultraviolet radiation) should be limited, and limiting chemotherapy exposure may plausibly also reduce risk [1]. Second tumours share RB1 inactivation, but treatment does not differ from equivalent neoplasms without heritable predisposition [1]. The American Society of Clinical Oncology classifies heritable retinoblastoma as a group 1 disorder, where genetic testing forms part of standard management of at-risk relatives [1], and ClinGen assigned strong paediatric actionability [20].

7.11 Pivotal trials and interpretation limits

Retinoblastoma treatment evidence mainly comprises single-arm phase II trials, prospective cooperative studies and centre series; a 2024 systematic review found only four randomised trials (174 participants, 188 eyes), all examining intravenous chemotherapy regimens, and Wen and colleagues' randomised trial was the first to compare administration routes [64, 65]. Table 7A summarises the studies informing the preceding treatment decisions.

Table 7A. Reference trials and prospective studies in retinoblastoma treatment

These studies cannot rank treatments through indirect comparisons: they differ in ICRB or Reese–Ellsworth group, laterality, age, previous treatment, analysis unit (patient or eye) and outcome definition, and "globe salvage" does not equate to useful vision preservation or patient survival. Centre-series percentages concern eyes selected for a technique rather than population rates. Wen's relative benefit requires consideration of residual absolute risk: even with IAC, 47% of children lost the globe or progressed within two years [64]. Negative or interrupted trials (ARET0331 and high-dose carboplatin in Roy's trial) are part of the evidence and show that intensification does not invariably improve outcomes [57, 83]. Finally, late toxicity relevant to germline carriers, namely second tumours after alkylating agents and radiation, was not an outcome in these trials; available data derive from survivor cohorts [43, 44, 62].

7.12 Comparative regulatory status: United States and European Union

No drug has a centrally authorised European retinoblastoma indication, and the only explicit United States label mention is cyclophosphamide among susceptible malignancies [17, 18]. Intra-arterial and intravitreal agents are therefore off-label, a status depending on three elements: the labelled disease indication, administration route and age; US topotecan and injectable melphalan labels state that paediatric safety and efficacy have not been established [53, 54]. Off-label use must be distinguished from evidence extrapolation: intravenous carboplatin in a published retinoblastoma regimen is off-label but supported by disease-specific trials; transferring doses from other routes, ages or tumours to intra-arterial or intravitreal administration would be unsupported extrapolation. Table 6 doses derive from cited studies rather than labels. In the EU, EMA performs scientific assessment and the European Commission grants authorisation; authorisation does not guarantee national marketing or funding, which depend in Portugal on INFARMED and hospital pathways, and in Brazil on ANVISA registration and incorporation into the Sistema Único de Saúde; CONITEC's oncology Diagnostic and Therapeutic Guidelines list, consulted in October 2026, did not include a retinoblastoma-specific guideline [84].

Table 7B. Regulatory status of retinoblastoma drugs (4 October 2026)

7.13 Reference networks and patient organisations

Retinoblastoma treatment should be concentrated in experienced centres. Relevant EU European Reference Networks include ERN PaedCan for paediatric cancer, offering retinoblastoma-specific training, and ERN-EYE for rare eye diseases; ERN GENTURIS covers heritable tumour predisposition syndromes generally [25, 26, 87]. Portugal's national intraocular tumour referral centre operates at Centro Hospitalar e Universitário de Coimbra and also receives children evacuated from Portuguese-speaking African countries [88, 89]. Patient organisations include Childhood Eye Cancer Trust (United Kingdom) and World Eye Cancer Hope, with family-directed information; Acreditar (Associação de Pais e Amigos de Crianças com Cancro) in Portugal; and GRAACC in Brazil, with a dedicated retinoblastoma service [1, 23, 24].

8. Reproductive Options and Family Counselling

8.1 Mode of transmission

Heritable retinoblastoma follows autosomal dominant inheritance. Most patients are simplex cases arising from a de novo variant; approximately 5% of simplex bilateral cases inherited the variant from an unaffected parent, who may be mosaic or carry a reduced-penetrance variant [1]. Alternatively, the initiating variant may arise in retinal cells or their embryonic precursors without germline involvement, underlying non-heritable unilateral unifocal retinoblastoma [8]. Mosaicism should always be considered when counselling newly identified families [8, 9].

8.2 Proband's parents

When a germline variant is identified in a proband without other affected relatives, parental molecular testing is recommended; if the variant is not known, the parents should be examined by an ophthalmologist experienced in retinoblastoma, retinoma and associated lesions [1]. If somatic mosaicism has been demonstrated in the proband (a postzygotic variant), parental carrier status can be excluded [1]. If the variant is detected in neither parent, a de novo variant or parental germline mosaicism should be considered; leucocyte DNA testing may miss somatic mosaicism, and highly sensitive methods such as high-coverage next-generation sequencing or allele-specific PCR may be used [1]. Parents who do not carry their child's variant should be given a low but non-negligible recurrence risk [8]. Family history may appear negative because of unrecognised retinoma, low-level mosaicism or reduced penetrance [1].

8.3 Empirical risks

When molecular testing is unavailable or uninformative, empirical risks based on tumour presentation and family history are used.

Table 8. Empirical retinoblastoma risks (pre-test) for siblings and offspring [1, 8]

Sibling risks in patients without family history apply when there are no unaffected siblings and reflect possible reduced-penetrance variants or parental germline mosaicism [1]. In low-penetrance retinoblastoma families, carrier tumour risk is below the 95% observed with null alleles [1].

8.4 Risks after genetic testing

The Oxford Desk Reference organises risk into four situations. In sporadic bilateral or multifocal retinoblastoma, a germline variant is presumed, with a 45% offspring risk; some patients are somatic or gonosomal mosaics with lower risk. In simplex unilateral retinoblastoma with a blood de novo variant, parental germline mosaicism remains possible, so siblings should be offered testing; offspring risk is 45%. In unilateral retinoblastoma without a blood variant, contralateral tumour risk is 3–5%, surveillance should continue, offspring risk is probably small, and if both tumour variants are absent from blood, postzygotic origin is presumed and siblings are not at risk. In sporadic bilateral or multifocal disease without a detected variant, offspring risk is substantial but may be below 45%, and offspring should undergo ophthalmological screening because genetic testing cannot be offered [8].

If tumour-identified variants are absent from proband leucocyte DNA, germline mosaicism probability for either is estimated at 1.2% and offspring transmission risk at 0.6%; offspring should be tested for both tumour variants [1, 31]. If one tumour variant is mosaic in leucocytes, germline involvement is uncertain and all offspring should be tested for that variant [1]. If a parent is heterozygous, sibling risk is 50%; if neither parent has the variant, sibling risk is below 5% but above population risk, and siblings may undergo testing, becoming H0 if they lack the familial variant [1]. Deep sequencing studies estimate maximum parental mosaicism recurrence risk at 0.4–0.8%, 266 to 533 times population risk [35].

8.5 Predictive testing and family strategy

Once the variant is identified, cascade testing should begin with parents and siblings and extend to other indicated relatives [8]. Predictive testing of at-risk newborns is appropriate given the burden of frequent screening under anaesthesia in early childhood; relatives shown not to be at risk through variant testing or linkage may be discharged from screening [8]. Retinoblastoma exemplifies a disorder where effective surveillance justifies childhood predictive testing to avoid unnecessary surveillance [90]. The strategy for assessing sibling and future pregnancy risk is to seek the variant in the affected child's blood and, if identified, test parents and siblings; without a variant, perform linkage analysis; examine the tumour, where LOH permits linkage-based exclusion and two variants absent from blood define a somatic de novo event, allowing siblings to be discharged from screening [8]. Genetic counselling, including reproductive options, should be offered to affected or at-risk young adults, ideally determining risk and discussing prenatal or preimplantation diagnosis before pregnancy [1].

8.6 Prenatal and preimplantation diagnosis

Prenatal diagnosis through chorionic villus sampling is possible in families with a known variant or informative linkage [8], and preimplantation genetic testing (PGT) is available once the familial variant is identified [1]. When a fetus is a carrier, obstetric ultrasound may detect medium-sized intraocular tumours; if present, preterm delivery may be considered to enable early treatment; even without visible tumours, delivery at 36 weeks may be recommended because approximately 30% of carrier newborns already have a small vision-threatening tumour [1]. In the Toronto series, newborns identified through amniocentesis and delivered electively at 36–38 weeks more frequently had cT1a eyes at first treatment (8 of 12 versus 1 of 8), greater treatment success (11 of 12 versus 3 of 8) and more eyes with acceptable vision (21 of 24 versus 8 of 16) than spontaneously born infants examined within the first week; null alleles caused earlier tumours than low-penetrance alleles [91]. When the variant is unknown, prenatal ultrasound or MRI may reveal a large retinoblastoma but lack sensitivity for small tumours [1]. Professional and family perspectives on prenatal diagnosis and PGT may differ, and the decision is personal [1].

9. Special Considerations: Population Diversity and Epidemiology

Retinoblastoma birth incidence is relatively uniform between populations, at 1:15,000 to 1:20,000 live births in European population-based series [1]; absolute case numbers therefore follow birth rates, concentrating disease burden in countries with more births, which also have lower incomes [92]. No substantial RB1 founder effects are described in the sources consulted, consistent with the high proportion of de novo variants and absence of gene hotspots [31].

The 2017 global study of 4,351 patients from 153 countries, estimated to represent over half of worldwide new cases that year, found 84.7% living in low- or middle-income countries. In high-income countries, median diagnosis age was 14.1 months and 98.5% had intraocular disease; in low-income countries, median age was 30.5 months, 49.1% had extraocular disease and 18.9% metastases. Low national income was associated with advanced disease even after age adjustment (OR 17.92 in low-income countries) and a lower proportion with family history, probably because many patients do not reach reproductive age [12]. At three-year follow-up of 4,064 children from 149 countries, survival was 99.5% in high-income, 91.2% in upper-middle-income, 80.3% in lower-middle-income and 57.3% in low-income countries; low-income residence (HR 16.67), cT4 stage (HR 8.98) and older diagnosis age up to three years (HR 1.38 per year) independently predicted poorer survival [22]. A meta-analysis of 38,130 patients from 80 regions (1980–2020) documented overall survival improvement from 79% to 88% and globe preservation from 22% to 44% over four decades, associated with income, Gini index, rurality, healthcare funding and education [93].

In Brazil, Instituto Nacional de Câncer population-based cancer registries (2000–2018) identified 675 patients, 91% younger than five; combined incidence was 7.02 per million at ages 0 to 4, ranging from 5.25 in the Central-West to 11.26 in the Northeast, and world-population-adjusted incidence was 2.23 per million at ages 0 to 14, suggesting underreporting except in the Northeast [94]. In hospital registries (2,269 cases from 86 institutions), 84% were treated publicly, 88% had unilateral disease, and diagnostic delay mainly reflected pre-hospital factors, with substantial missing data [95]. At a Ceará referral centre (105 patients, 2007–2021), mean diagnosis age was 1.7 years, 57.1% presented with leucocoria, 25.7% had bilateral disease and rural origin (57.4%) was associated with advanced stages [96]. In a São Paulo hospital cohort (465 patients, 1986–2003; 42% heritable), 77% were diagnosed at advanced stages, five-year survival was 88%, and heritable patients had a second tumour standardised mortality ratio of 86.1, with similar risk after radiotherapy plus chemotherapy or chemotherapy alone [97]. A Brazilian cohort of 159 patients without RB1 point variants showed a diverse spectrum of RB1 and adjacent-region deletions and duplications through MLPA [98].

Portugal's national intraocular tumour referral centre (Coimbra) characterised 28 children diagnosed between October 2015 and October 2020: 15 had heritable retinoblastoma (12 bilateral, 3 unilateral), mean diagnosis age was 13.6 months (9.6 months in heritable disease), only 14.3% had family history, pathogenic RB1 variants were identified in 46.4%, leucocoria prompted 71.4% of referrals, and most eyes presented at advanced stages; 14 children underwent enucleation and none died [88]. The authors estimated incidence at 4.04 per 100,000 live births (95% CI 1.59–6.49), approximately 1 in 25,000, consistent with European estimates; this estimate is based on cases reaching the referral centre over five years, with a wide confidence interval, rather than a population cancer registry [88]. During the same period, the centre received 15 evacuated children from Angola, Cape Verde, Guinea-Bissau and São Tomé and Príncipe, with median diagnosis age 20.9 months, extraocular disease in 33.3% and overall survival of 73.3% during follow-up, reproducing global-study inequality at single-centre scale [12, 89]. The sources consulted yielded no population-based incidence registries for most African and Asian countries or the Middle East, nor studies of low-penetrance variant frequency in non-European populations; available global data derive from treatment centres rather than population registries [12, 22].

10. Illustrative Clinical Case

Gargallo and colleagues described a girl conceived through in vitro fertilisation (maternal oocytes and anonymous donor sperm), admitted at six months with leucocoria and strabismus [61]. History and physical examination were normal except right fifth-finger clinodactyly. Examination under anaesthesia, orbital ultrasound and MRI revealed a 14 × 13 × 11 mm left intraocular mass in the inferotemporal sector with retinal detachment and no other retinal or brain lesions; the eye was ICRB group E. Intra-arterial melphalan was initiated but interrupted by left lower-limb vasospasm, followed by four vincristine, carboplatin and etoposide cycles; after partial response, disease progressed and the eye was enucleated [61].

Blood RB1 Sanger sequencing was normal, but MLPA showed low values within the normal range, raising suspicion of mosaic complete RB1 deletion. A SNP array confirmed a 35.7 Mb deletion from 13q12.13 to 13q21.2 in approximately 40% of cells (arr[hg19] 13q12.13q21.2(26,555,387–62,280,955)×1–2), and stimulated lymphocyte karyotyping identified two cell lines: 46,XX,del(13)(q12q21)[6]/46,XX[44]. RB1 probe FISH in oral mucosal cells detected the deletion in approximately 40%. Non-tumour retina had the deletion in approximately 50% of cells and tumour in all cells, without 13q LOH; a next-generation sequencing panel identified the second hit only in tumour, the nonsense RB1 variant c.958C>T (p.Arg320Ter), with 25% allele frequency, alongside 6p12.3pter gain and 6q25.3qter loss [61]. At 42 months, the right eye remained disease-free, growth was normal (weight and height at the 50th percentile, head circumference at the 90th), and there were no malformations or neurological developmental abnormalities. The karyotype figure is available at https://pmc.ncbi.nlm.nih.gov/articles/PMC8418035/figure/Fig2/ [61].

The case illustrates several principles. First, sequencing and MLPA may miss a mosaic complete RB1 deletion, and chromosomal testing (karyotyping, FISH or CMA) should form part of assessment in all retinoblastoma children, even without evident dysmorphism [8, 28]. Second, unilateral presentation agrees with reduced expressivity of deletions exceeding 1 Mb that include MED4 [36, 37]. Third, the child is classified as mosaic H1: contralateral-eye H1 surveillance and second tumour information are required, and future offspring should be tested, with transmission risk depending on germline involvement that cannot be determined from blood [1]. Because the deletion arose through postzygotic mosaicism, parents are not at increased risk and siblings need no surveillance [1]. Finally, IAC vascular toxicity may necessitate a treatment strategy change [60].

11. Prognosis and Long-Term Follow-up

Intraocular retinoblastoma is curable when diagnosed early: high-income-country survival exceeds 95% and reaches 99.5% at three years [11, 22]. Mean survival in developed countries reflects several death causes: in the first four years, most deaths result from metastatic retinoblastoma; later deaths mainly result from second primary tumours such as osteosarcoma or fibrosarcoma [10]. Diagnostic delay is the principal death-risk aggravating factor in both forms through high metastatic probability, and frequent intraocular recurrences in the same eye adversely affect survival prognosis [10]. Trilateral retinoblastoma was almost invariably fatal for decades, with improvement after 1995 associated with high-dose chemotherapy [10, 40].

Long-term mortality in heritable survivors is primarily determined by second tumours. In a cohort of 1,603 patients, 96 second tumour deaths occurred (relative risk 30), including 90 among 919 bilateral patients (relative risk 60), with significantly increased mortality from bone cancer, connective tissue cancer, melanoma and brain and meningeal neoplasms [99]. Radiotherapy as part of treatment increases second tumour mortality, and Moll and colleagues estimated cumulative second tumour incidence at 19% by age 35 [8, 41]. After more than 40 years of follow-up, epithelial carcinomas (lung, bladder and breast) become increasingly relevant [13].

Long-term follow-up should include ophthalmological surveillance of treated and contralateral eyes according to Table 7; visual rehabilitation, ocular prostheses and educational support for children with visual loss; clinical surveillance and continuing second tumour education, including skin self-examination and prompt symptom-triggered referral; avoidance of tobacco, excessive sunlight and non-essential ionising-radiation investigations; and reproductive genetic counselling during transition to adulthood [1, 8]. In unilateral survivors without germline variants (H0*), second tumour risk is not substantially increased and ophthalmological follow-up may be less intensive [1, 42]. Late systemic treatment effects, including leiomyosarcoma risk after alkylating agents before age one, justify surveillance stratification by received treatment [43, 44].

12. Summary and Final Considerations

Retinoblastoma is the model for two-hit theory and tumour suppressor genes. It results from biallelic RB1 inactivation in a retinal cone precursor and presents in two genetically distinct forms: heritable disease, with a germline variant, early bilateral or multifocal disease, pineoblastoma risk and high lifetime second tumour risk; and non-heritable disease, with two somatic hits and unilateral unifocal disease. Approximately 10–15% of unilateral cases are heritable, and mosaicism, present in approximately 10% of families, modifies laterality, penetrance and recurrence risks.

Tumour diagnosis is clinical, through fundus examination under anaesthesia and imaging without ionising radiation; predisposition diagnosis requires RB1 molecular testing in blood and, whenever possible, tumour, supplemented by 13q14 chromosomal testing and, without tumour RB1 variants, assessment of promoter hypermethylation and MYCN amplification. AJCC heritable category H guides patient and family surveillance. Low-penetrance variants, parent-of-origin effects and large deletions involving MED4 loss explain reduced expressivity in some families, and RB1+/+ retinoblastoma with MYCN amplification is a somatic entity of very young infants without demonstrated germline predisposition.

Treatment evolved from enucleation and external beam radiotherapy to systemic chemotherapy with focal consolidation, then intra-arterial and intravitreal chemotherapy, increasing ocular preservation without compromising survival, although evidence is mainly non-randomised. Avoiding radiation and limiting alkylating agents are central principles in germline carriers. Sibling and offspring screening through fundus examinations under anaesthesia from the first weeks of life, neonatal predictive testing, prenatal diagnosis with planned early delivery and PGT enable identification of small tumours treatable with vision preservation. The global survival disparity, from 99.5% to 57.3% according to national income, makes early leucocoria recognition the principal public health intervention.

13. References

1. Lohmann DR, Gallie BL. Retinoblastoma. 2000 Jul 18 [updated 2023 Sep 21]. In: Adam MP, Bick S, Mirzaa GM, Pagon RA, Wallace SE, Amemiya A, editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993–2026. PMID: 20301625. Available at: https://www.ncbi.nlm.nih.gov/books/NBK1452/
2. Orphanet. Retinoblastoma (ORPHA:790); Hereditary retinoblastoma (ORPHA:357027); Non-hereditary retinoblastoma (ORPHA:357034); Monosomy 13q14 syndrome (ORPHA:1587) [Internet]. Paris: INSERM [accessed 2026 Oct 4]. Available at: https://www.orpha.net/en/disease/detail/790
3. National Center for Biotechnology Information. RB1 RB transcriptional corepressor 1 [Homo sapiens], Gene ID 5925 — MANE Select transcript NM_000321.3 [Internet]. Bethesda (MD): NCBI Datasets [accessed 2026 Oct 4]. Available at: https://www.ncbi.nlm.nih.gov/gene/5925
4. UniProt Consortium. P06400 — RB_HUMAN, Retinoblastoma-associated protein [Internet] [accessed 2026 Oct 4]. Available at: https://www.uniprot.org/uniprotkb/P06400; NCBI Gene 5925: https://www.ncbi.nlm.nih.gov/gene/5925
5. Knudson AG Jr. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci U S A. 1971;68(4):820–823. doi:10.1073/pnas.68.4.820
6. Xu XL, Singh HP, Wang L, Qi DL, Poulos BK, Abramson DH, et al. Rb suppresses human cone-precursor-derived retinoblastoma tumours. Nature. 2014;514(7522):385–388. doi:10.1038/nature13813
7. Rushlow DE, Mol BM, Kennett JY, Yee S, Pajovic S, Thériault BL, et al. Characterisation of retinoblastomas without RB1 mutations: genomic, gene expression, and clinical studies. Lancet Oncol. 2013;14(4):327–334. doi:10.1016/S1470-2045(13)70045-7
8. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 614–617 (Retinoblastoma).
9. Sippel KC, Fraioli RE, Smith GD, Schalkoff ME, Sutherland J, Gallie BL, et al. Frequency of somatic and germ-line mosaicism in retinoblastoma: implications for genetic counseling. Am J Hum Genet. 1998;62(3):610–619. doi:10.1086/301766
10. Façanha AS, Vieira JCF, Nápravnik RGC. Retinoblastoma (RB). In: Ribeiro EM, coordinator. Síndromes Genéticas: Reconhecer para Tratar. Fortaleza: EdUnichristus; 2021. p. 173–175.
11. Dimaras H, Corson TW, Cobrinik D, White A, Zhao J, Munier FL, et al. Retinoblastoma. Nat Rev Dis Primers. 2015;1:15021. doi:10.1038/nrdp.2015.21
12. Global Retinoblastoma Study Group; Fabian ID, Abdallah E, Abdullahi SU, Abdulqader RA, Adamou Boubacar S, et al. Global retinoblastoma presentation and analysis by national income level. JAMA Oncol. 2020;6(5):685–695. doi:10.1001/jamaoncol.2019.6716
13. Marees T, Moll AC, Imhof SM, de Boer MR, Ringens PJ, van Leeuwen FE. Risk of second malignancies in survivors of retinoblastoma: more than 40 years of follow-up. J Natl Cancer Inst. 2008;100(24):1771–1779. doi:10.1093/jnci/djn394
14. Linn Murphree A. Intraocular retinoblastoma: the case for a new group classification. Ophthalmol Clin North Am. 2005;18(1):41–53. doi:10.1016/j.ohc.2004.11.003
15. Mallipatna A, Gallie BL, Chévez-Barrios P, Lumbroso-Le Rouic L, Chantada GL, Doz F, et al. Retinoblastoma. In: Amin MB, Edge SB, Greene FL, editors. AJCC Cancer Staging Manual. 8th ed. New York: Springer; 2017. p. 819–831.
16. Munier FL, Beck-Popovic M, Chantada GL, Cobrinik D, Kivelä TT, Lohmann D, et al. Conservative management of retinoblastoma: challenging orthodoxy without compromising the state of metastatic grace. "Alive, with good vision and no comorbidity". Prog Retin Eye Res. 2019;73:100764. doi:10.1016/j.preteyeres.2019.05.005
17. U.S. Food and Drug Administration. Cyclophosphamide for injection — prescribing information (indications and usage) [Internet]. openFDA Drug Label API [accessed 2026 Oct 4]. Available at: https://api.fda.gov/drug/label.json?search=indications_and_usage:retinoblastoma
18. European Medicines Agency. Hycamtin (topotecan) — EPAR, therapeutic indication and authorisation details [Internet]. Amsterdam: EMA; last updated 20 Jan 2026 [accessed 2026 Oct 4]. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/hycamtin
19. Skalet AH, Gombos DS, Gallie BL, Kim JW, Shields CL, Marr BP, et al. Screening children at risk for retinoblastoma: consensus report from the American Association of Ophthalmic Oncologists and Pathologists. Ophthalmology. 2018;125(3):453–458. doi:10.1016/j.ophtha.2017.09.001
20. Clinical Genome Resource (ClinGen) Actionability Working Group. Retinoblastoma (RB1), report AC001 — paediatric and adult contexts [Internet] [accessed 2026 Oct 4]. Available at: https://actionability.clinicalgenome.org/
21. Canadian Retinoblastoma Society. National Retinoblastoma Strategy Canadian Guidelines for Care / Stratégie thérapeutique du rétinoblastome guide clinique canadien. Can J Ophthalmol. 2009;44(Suppl 2):S1–88. doi:10.3129/i09-194
22. Global Retinoblastoma Study Group. The Global Retinoblastoma Outcome Study: a prospective, cluster-based analysis of 4064 patients from 149 countries. Lancet Glob Health. 2022;10(8):e1128–e1140. doi:10.1016/S2214-109X(22)00250-9
23. Acreditar — Associação de Pais e Amigos de Crianças com Cancro [Internet]. Lisbon [accessed 2026 Oct 4]. Available at: https://www.acreditar.org.pt/
24. GRAACC — Grupo de Apoio ao Adolescente e à Criança com Câncer. Retinoblastoma [Internet]. São Paulo [accessed 2026 Oct 4]. Available at: https://www.graacc.org.br/
25. European Reference Network for Paediatric Oncology (ERN PaedCan). Retinoblastoma webinar for paediatric oncologists [Internet] [accessed 2026 Oct 4]. Available at: https://paedcan.ern-net.eu/
26. European Reference Network for Rare Eye Diseases (ERN-EYE) [Internet] [accessed 2026 Oct 4]. Available at: https://www.ern-eye.eu/
27. Dimaras H, Kimani K, Dimba EA, Gronsdahl P, White A, Chan HS, et al. Retinoblastoma. Lancet. 2012;379(9824):1436–1446. doi:10.1016/S0140-6736(11)61137-9
28. Jones KL, Jones MC, del Campo M. Smith's Recognizable Patterns of Human Malformation. 8th ed. Philadelphia: Elsevier; 2022. p. 52–53 (Deletion 13q syndrome).
29. ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 2, 15 (Introduction to Cancer Genomics in Clinical Practice; Knudson AG, Meadows AT, Nichols WW, Hill R. Chromosomal deletion and retinoblastoma. N Engl J Med. 1976;295:1120–1123).
30. Friend SH, Bernards R, Rogelj S, Weinberg RA, Rapaport JM, Albert DM, et al. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature. 1986;323(6089):643–646. doi:10.1038/323643a0
31. Richter S, Vandezande K, Chen N, Zhang K, Sutherland J, Anderson J, et al. Sensitive and efficient detection of RB1 gene mutations enhances care for families with retinoblastoma. Am J Hum Genet. 2003;72(2):253–269. doi:10.1086/345651
32. Rushlow D, Piovesan B, Zhang K, Prigoda-Lee NL, Marchong MN, Clark RD, et al. Detection of mosaic RB1 mutations in families with retinoblastoma. Hum Mutat. 2009;30(5):842–851. doi:10.1002/humu.20940
33. Avnat E, Shapira G, Lustig Y, Citrin J, Rojanaporn D, Kaewkhaw R, et al. Revisiting low penetrance retinoblastoma: an integrated clinical, genetic, and bioinformatic analysis. Hum Mol Genet. 2026;35(6). doi:10.1093/hmg/ddag026
34. Eloy P, Dehainault C, Sefta M, Aerts I, Doz F, Cassoux N, et al. A parent-of-origin effect impacts the phenotype in low penetrance retinoblastoma families segregating the c.1981C>T/p.Arg661Trp mutation of RB1. PLoS Genet. 2016;12(2):e1005888. doi:10.1371/journal.pgen.1005888
35. Dehainault C, Golmard L, Millot GA, Charpin A, Laugé A, Tarabeux J, et al. Mosaicism and prenatal diagnosis options: insights from retinoblastoma. Eur J Hum Genet. 2017;25(3):381–383. doi:10.1038/ejhg.2016.174
36. Mitter D, Ullmann R, Muradyan A, Klein-Hitpass L, Kanber D, Ounap K, et al. Genotype-phenotype correlations in patients with retinoblastoma and interstitial 13q deletions. Eur J Hum Genet. 2011;19(9):947–958. doi:10.1038/ejhg.2011.58
37. Dehainault C, Garancher A, Castéra L, Cassoux N, Aerts I, Doz F, et al. The survival gene MED4 explains low penetrance retinoblastoma in patients with large RB1 deletion. Hum Mol Genet. 2014;23(19):5243–5250. doi:10.1093/hmg/ddu245
38. Papaioannou K, Kubica R, Fischhuber K, Beygo J, Reschke M, Ritter-Sovinz P, et al. MYCN amplification in RB1-inactivated retinoblastoma: association with high-risk features. Pediatr Blood Cancer. 2026;73(6):e70297. doi:10.1002/1545-5017.70297
39. Marcelis CL, Hol FA, Graham GE, Rieu PN, Kellermayer R, Meijer RP, et al. Genotype-phenotype correlations in MYCN-related Feingold syndrome. Hum Mutat. 2008;29(9):1125–1132. doi:10.1002/humu.20750
40. de Jong MC, Kors WA, de Graaf P, Castelijns JA, Kivelä T, Moll AC. Trilateral retinoblastoma: a systematic review and meta-analysis. Lancet Oncol. 2014;15(10):1157–1167. doi:10.1016/S1470-2045(14)70336-5
41. Moll AC, Imhof SM, Bouter LM, Tan KE. Second primary tumors in patients with retinoblastoma. A review of the literature. Ophthalmic Genet. 1997;18(1):27–34. doi:10.3109/13816819709057880
42. Kleinerman RA, Tucker MA, Tarone RE, Abramson DH, Seddon JM, Stovall M, et al. Risk of new cancers after radiotherapy in long-term survivors of retinoblastoma: an extended follow-up. J Clin Oncol. 2005;23(10):2272–2279. doi:10.1200/JCO.2005.05.054
43. Wong JR, Morton LM, Tucker MA, Abramson DH, Seddon JM, Sampson JN, et al. Risk of subsequent malignant neoplasms in long-term hereditary retinoblastoma survivors after chemotherapy and radiotherapy. J Clin Oncol. 2014;32(29):3284–3290. doi:10.1200/JCO.2013.54.7844
44. Temming P, Arendt M, Viehmann A, Eisele L, Le Guin CH, Schündeln MM, et al. Incidence of second cancers after radiotherapy and systemic chemotherapy in heritable retinoblastoma survivors: a report from the German reference center. Pediatr Blood Cancer. 2017;64(1):71–80. doi:10.1002/pbc.26193
45. Ketteler P, Hülsenbeck I, Frank M, Schmidt B, Jöckel KH, Lohmann DR. The impact of RB1 genotype on incidence of second tumours in heritable retinoblastoma. Eur J Cancer. 2020;133:47–55. doi:10.1016/j.ejca.2020.04.005
46. Abramson DH, Beaverson K, Sangani P, Vora RA, Lee TC, Hochberg HM, et al. Screening for retinoblastoma: presenting signs as prognosticators of patient and ocular survival. Pediatrics. 2003;112(6 Pt 1):1248–1255. doi:10.1542/peds.112.6.1248
47. Lohmann DR, Gerick M, Brandt B, Oelschläger U, Lorenz B, Passarge E, et al. Constitutional RB1-gene mutations in patients with isolated unilateral retinoblastoma. Am J Hum Genet. 1997;61(2):282–294. doi:10.1086/514845
48. Chévez-Barrios P, Eagle RC Jr, Krailo M, Piao J, Albert DM, Gao Y, et al. Study of unilateral retinoblastoma with and without histopathologic high-risk features and the role of adjuvant chemotherapy: a Children's Oncology Group study. J Clin Oncol. 2019;37(31):2883–2891. doi:10.1200/JCO.18.01808
49. Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405–424. doi:10.1038/gim.2015.30
50. National Center for Biotechnology Information. ClinVar: NM_000321.3(RB1):c.1981C>T (p.Arg661Trp), VCV000013087; NM_000321.3(RB1):c.958C>T (p.Arg320Ter), VCV000126824 [Internet]. Bethesda (MD): NCBI [accessed 2026 Oct 4]. Available at: https://www.ncbi.nlm.nih.gov/clinvar/variation/13087/
51. Alonso J, Moreno C, López A, Mendiola M, García-Miguel P, Abelairas J, et al. Five novel single nucleotide polymorphisms of the RB1 gene (g.5625T>C, g.70169T>G, g.76875A>T, g.78026delA, and g.150072T>C) in retinoblastoma patients. Hum Mutat. 2001;17(5):437. doi:10.1002/humu.1129. PMID: 11317369
52. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 300–301 (Retinal dysplasia).
53. U.S. Food and Drug Administration. Melphalan hydrochloride for injection — prescribing information (indications) [Internet]. openFDA Drug Label API [accessed 2026 Oct 4]. Available at: https://api.fda.gov/drug/label.json?search=openfda.generic_name:%22MELPHALAN%22
54. U.S. Food and Drug Administration. HYCAMTIN (topotecan) for injection and capsules — prescribing information [Internet]. openFDA Drug Label API [accessed 2026 Oct 4]. Available at: https://api.fda.gov/drug/label.json?search=openfda.brand_name:%22HYCAMTIN%22
55. Pérez V, Sampor C, Rey G, Parareda-Salles A, Kopp K, Dabezies AP, et al. Treatment of nonmetastatic unilateral retinoblastoma in children. JAMA Ophthalmol. 2018;136(7):747–752. doi:10.1001/jamaophthalmol.2018.1501
56. Murphree AL, Villablanca JG, Deegan WF 3rd, Sato JK, Malogolowkin M, Fisher A, et al. Chemotherapy plus local treatment in the management of intraocular retinoblastoma. Arch Ophthalmol. 1996;114(11):1348–1356. doi:10.1001/archopht.1996.01100140548005
57. Friedman DL, Krailo M, Villaluna D, Gombos D, Langholz B, Jubran R, et al. Systemic neoadjuvant chemotherapy for Group B intraocular retinoblastoma (ARET0331): a report from the Children's Oncology Group. Pediatr Blood Cancer. 2017;64(7):e26394. doi:10.1002/pbc.26394
58. Abramson DH, Dunkel IJ, Brodie SE, Kim JW, Gobin YP. A phase I/II study of direct intraarterial (ophthalmic artery) chemotherapy with melphalan for intraocular retinoblastoma: initial results. Ophthalmology. 2008;115(8):1398–1404. doi:10.1016/j.ophtha.2007.12.014
59. Gobin YP, Dunkel IJ, Marr BP, Brodie SE, Abramson DH. Intra-arterial chemotherapy for the management of retinoblastoma: four-year experience. Arch Ophthalmol. 2011;129(6):732–737. doi:10.1001/archophthalmol.2011.5
60. Teixeira LF, Macedo CRPD, Fonseca JRF, Morales B, Mangeon MK, Miranda BA, et al. Intra-arterial chemotherapy for retinoblastoma, outcomes analysis in 357 eyes: thirteen years of experience in a referral center in Brazil. Ophthalmol Retina. 2025;9(8):798–806. doi:10.1016/j.oret.2025.02.013
61. Gargallo P, Oltra S, Balaguer J, Barranco H, Yáñez Y, Segura V, et al. Retinoblastoma and mosaic 13q deletion: a case report. Int J Retina Vitreous. 2021;7(1):50. doi:10.1186/s40942-021-00321-9
62. Rokohl AC, Lohmann N, Reinking N, Skoetz N, Heindl LM. Intravenous chemotherapy versus intra-arterial chemotherapy for retinoblastoma. Cochrane Database Syst Rev. 2026;2(2):CD013695. doi:10.1002/14651858.CD013695.pub2
63. Chen KY, Chan HC, Chan CM. Comparative effectiveness and safety of intra-arterial chemotherapy and intravenous chemotherapy for retinoblastoma: a systematic review and meta-analysis. Surv Ophthalmol. 2026;71(3):878–891. doi:10.1016/j.survophthal.2025.12.009
64. Wen X, Fan J, Jin M, Jiang H, Li J, Han M, et al. Intravenous versus super-selected intra-arterial chemotherapy in children with advanced unilateral retinoblastoma: an open-label, multicentre, randomised trial. Lancet Child Adolesc Health. 2023;7(9):613–620. doi:10.1016/S2352-4642(23)00141-4
65. Lavasidis G, Papaioannou K, Anagnostou N, Ketteler P, Bechrakis NE, Ntzani E. Evidence in focus: the sparse landscape of randomized trials on retinoblastoma treatment. Ocul Oncol Pathol. 2024;10(1):53–62. doi:10.1159/000536410
66. Castela G, Providência J, Monteiro M, Oliveiros B, Silva S, Brito M, et al. Effectiveness of intra-arterial chemotherapy for the treatment of intraocular retinoblastoma: relevance of a multidisciplinary setting. Clin Ophthalmol. 2023;17:487–496. doi:10.2147/OPTH.S398488
67. Munier FL, Gaillard MC, Balmer A, Soliman S, Podilsky G, Moulin AP, et al. Intravitreal chemotherapy for vitreous disease in retinoblastoma revisited: from prohibition to conditional indications. Br J Ophthalmol. 2012;96(8):1078–1083. doi:10.1136/bjophthalmol-2011-301450
68. Rao R, Honavar SG, Sharma V, Reddy VAP. Intravitreal topotecan in the management of refractory and recurrent vitreous seeds in retinoblastoma. Br J Ophthalmol. 2018;102(4):490–495. doi:10.1136/bjophthalmol-2017-310641
69. Dunkel IJ, Piao J, Chantada GL, Banerjee A, Abouelnaga S, Buchsbaum JC, et al. Intensive multimodality therapy for extraocular retinoblastoma: a Children's Oncology Group trial (ARET0321). J Clin Oncol. 2022;40(33):3839–3847. doi:10.1200/JCO.21.02337
70. European Medicines Agency. Hycamtin (topotecan): EPAR — product information (summary of product characteristics) [Internet]. Amsterdam: EMA [accessed 2026 Oct 26]. Available at: https://www.ema.europa.eu/en/documents/product-information/hycamtin-epar-product-information_en.pdf
71. U.S. Food and Drug Administration. Topotecan injection (NDA 200582) — prescribing information, dosage and administration [Internet]. openFDA Drug Label API; revision of 2021-02-01 [accessed 2026 Oct 26]. Available at: https://api.fda.gov/drug/label.json?search=openfda.generic_name:%22TOPOTECAN%22+AND+openfda.route:%22INTRAVENOUS%22
72. U.S. Food and Drug Administration. Carboplatin injection — prescribing information (boxed warning) [Internet]. openFDA Drug Label API [accessed 2026 Oct 4]. Available at: https://api.fda.gov/drug/label.json?search=openfda.generic_name:%22CARBOPLATIN%22
73. U.S. Food and Drug Administration. Etoposide injection (ANDA 074529) — prescribing information, dosage and administration [Internet]. openFDA Drug Label API; revision of 2026-05-20 [accessed 2026 Oct 26]. Available at: https://api.fda.gov/drug/label.json?search=openfda.generic_name:%22ETOPOSIDE%22
74. U.S. Food and Drug Administration. Vincristine sulfate injection — prescribing information (warnings) [Internet]. openFDA Drug Label API [accessed 2026 Oct 4]. Available at: https://api.fda.gov/drug/label.json?search=openfda.generic_name:%22VINCRISTINE%22
75. The Hospital for Sick Children. Topotecan episcleral plaque for treatment of retinoblastoma. ClinicalTrials.gov identifier: NCT04428879 [Internet]; updated 13 Dec 2024 [accessed 2026 Oct 4]. Available at: https://clinicaltrials.gov/study/NCT04428879
76. Targeted Therapy Technologies, LLC. Retinoblastoma phase II expanded access clinical trial (Chemoplaque/episcleral topotecan). ClinicalTrials.gov identifier: NCT06679634 [Internet]; updated 22 May 2026 [accessed 2026 Oct 4]. Available at: https://clinicaltrials.gov/study/NCT06679634
77. Pascual-Pasto G, Bazan-Peregrino M, Olaciregui NG, Restrepo-Perdomo CA, Mato-Berciano A, Ottaviani D, et al. Therapeutic targeting of the RB1 pathway in retinoblastoma with the oncolytic adenovirus VCN-01. Sci Transl Med. 2019;11(476):eaat9321. doi:10.1126/scitranslmed.aat9321
78. Fundació Sant Joan de Déu. Evaluate safety and the oncolytic adenovirus VCN-01 activity in patients with refractory retinoblastoma. ClinicalTrials.gov identifier: NCT03284268 [Internet]; updated 27 Aug 2024 [accessed 2026 Oct 4]. Available at: https://clinicaltrials.gov/study/NCT03284268
79. European Medicines Agency. EU/3/24/2982 — orphan designation for the treatment of retinoblastoma (genetically modified human adenovirus encoding human PH20 hyaluronidase) [Internet]. Amsterdam: EMA; 2024 [accessed 2026 Oct 4]. Available at: https://www.ema.europa.eu/en/medicines/human/orphan-designations/eu-3-24-2982
80. Children's Oncology Group. Intravitreal melphalan for intraocular retinoblastoma. ClinicalTrials.gov identifier: NCT05504291 [Internet] [accessed 2026 Oct 4]. Available at: https://clinicaltrials.gov/study/NCT05504291
81. Institut Curie. Ocular conservative treatment for retinoblastoma: efficacy of the new management strategies and visual outcome — RETINO 2018. ClinicalTrials.gov identifier: NCT04681417 [Internet] [accessed 2026 Oct 4]. Available at: https://clinicaltrials.gov/study/NCT04681417
82. Sun Yat-sen University. Three cycles of adjuvant chemotherapy for the patients with high-risk retinoblastoma after enucleation: prospective study. ClinicalTrials.gov identifier: NCT05080010 [Internet] [accessed 2026 Oct 4]. Available at: https://clinicaltrials.gov/study/NCT05080010
83. Roy PS, Muhammed S, Singh U, Gowravajhala S, Jain R, Trehan A, et al. A single-blinded, randomized controlled trial of standard versus higher dose carboplatin-based intravenous chemotherapy for group D and E retinoblastoma. Pediatr Blood Cancer. 2023;70(8):e30444. doi:10.1002/pbc.30444
84. Comissão Nacional de Incorporação de Tecnologias no SUS (CONITEC). Diretrizes Diagnósticas e Terapêuticas (DDT) em Oncologia — lista geral [Internet]. Brasília: Ministério da Saúde; modified 14 Sep 2026 [accessed 2026 Oct 4]. Available at: https://www.gov.br/conitec/pt-br/assuntos/avaliacao-de-tecnologias-em-saude/protocolos-clinicos-e-diretrizes-terapeuticas/ddt
85. U.S. Food and Drug Administration. Drugs@FDA: NDA 014691 and 020207 (Alkeran), 020671 (Hycamtin), 019880 (Paraplatin), 018768 (VePesid), 014103 (Oncovin), 012141/012142 (Cytoxan) — initial approval dates [Internet]. openFDA [accessed 2026 Oct 4]. Available at: https://api.fda.gov/drug/drugsfda.json
86. European Medicines Agency. Phelinun (melphalan) — EPAR, therapeutic indication [Internet]. Amsterdam: EMA; last updated 18 Jun 2025 [accessed 2026 Oct 4]. Available at: https://www.ema.europa.eu/en/medicines/human/EPAR/phelinun
87. European Reference Network on Genetic Tumour Risk Syndromes (ERN GENTURIS) [Internet] [accessed 2026 Oct 4]. Available at: https://www.genturis.eu/
88. Castela G, Providência J, Monteiro M, Silva S, Brito M, Sá J, et al. Characterization of the Portuguese population diagnosed with retinoblastoma. Sci Rep. 2022;12(1):4378. doi:10.1038/s41598-022-08326-6
89. Castela G, Providência J, Monteiro M, Silva S, Brito M, Murta JN, et al. Treatment of advanced retinoblastoma in children evacuated from low-income countries: experience from a national referral center in Portugal. Clin Ophthalmol. 2021;15:4765–4773. doi:10.2147/OPTH.S343919
90. Firth HV, Hurst JA. Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 38 (Testing for genetic status).
91. Soliman SE, Dimaras H, Khetan V, Gardiner JA, Chan HS, Héon E, et al. Prenatal versus postnatal screening for familial retinoblastoma. Ophthalmology. 2016;123(12):2610–2617. doi:10.1016/j.ophtha.2016.08.027
92. Kivelä T. The epidemiological challenge of the most frequent eye cancer: retinoblastoma, an issue of birth and death. Br J Ophthalmol. 2009;93(9):1129–1131. doi:10.1136/bjo.2008.150292
93. Wong ES, Choy RW, Zhang Y, Chu WK, Chen LJ, Pang CP, et al. Global retinoblastoma survival and globe preservation: a systematic review and meta-analysis of associations with socioeconomic and health-care factors. Lancet Glob Health. 2022;10(3):e380–e389. doi:10.1016/S2214-109X(21)00555-6
94. Barbosa AC, de Magalhães-Barbosa MC, Moreira JPL, Colombini GNUI, Prata-Barbosa A. Incidence of retinoblastoma in children and adolescents in Brazil: a population-based study. Front Pediatr. 2022;10:1048792. doi:10.3389/fped.2022.1048792
95. Barbosa AC, Magalhães-Barbosa MC, Colombini GNUI, Prata-Barbosa A. Clinical, epidemiological, and care profile of hospitalized patients with retinoblastoma in Brazil. Arq Bras Oftalmol. 2024;88(1):e20230073. doi:10.5935/0004-2749.2023-0073
96. Pereira SM, Lima RV, Bezerra Filho COA, Santos CM, Muniz MCR, Ferreira DMR, et al. Retinoblastoma in Ceara: an epidemiological study in a Brazilian pediatric oncology referral center. Arq Bras Oftalmol. 2024;88(2):e20230265. doi:10.5935/0004-2749.2023-0265
97. Ribeiro KB, Veiga LHS, Carvalho Filho NP, Morton LM, Kleinerman RA, Antoneli CBG. Overall survival and cause-specific mortality in a hospital-based cohort of retinoblastoma patients in São Paulo, Brazil. Int J Cancer. 2025;156(1):69–78. doi:10.1002/ijc.35127
98. Mendonça V, Pereira Sena P, Evangelista dos Santos AC, Rodrigues Bonvicino C, Ashton-Prolla P, Epelman S, et al. Diverse mutational spectrum in the 13q14 chromosomal region in a Brazilian cohort of retinoblastoma. Exp Eye Res. 2022;224:109211. doi:10.1016/j.exer.2022.109211
99. Eng C, Li FP, Abramson DH, Ellsworth RM, Wong FL, Goldman MB, et al. Mortality from second tumors among long-term survivors of retinoblastoma. J Natl Cancer Inst. 1993;85(14):1121–1128. doi:10.1093/jnci/85.14.1121