The retinoblastoma protein

The RB1 gene encodes a 4.7 kilobase messenger RNA transcript that produces a protein of 110 kD (kilodaltons) and 928 amino acids (Fig. 128.2). The pRb protein is phosphorylated in a cell cycle-dependent manner and localizes to the nucleus.^56–58 The hypophosphorylated form predominates in quiescent and differentiating cells, whereas the hyperphosphorylated species accumulates in cycling cells as they enter DNA synthesis (S) phase (Fig. 128.3).^59–63 The hypophosphorylated form of pRb binds several viral oncoproteins, including SV40 large T antigen,⁶⁴ adenoviral E1a,⁶⁵ and human papillomavirus E7.⁶⁶ When bound to pRb, these oncoproteins stimulate cell division. Taken together, these findings provide evidence that hypophosphorylated pRb is important in negatively regulating the cell cycle, and that this inhibitory activity can be thwarted either by phosphorylation or viral oncoprotein binding. Further work has shown that the major cell cycle function of pRb is to inhibit the transition of cells out of gap 1 (G1) phase into S phase.⁶⁷ However, pRb may also have roles in other cell cycle phases.^68,69

Fig. 128.3 Role of the Rb protein in the cell cycle and apoptosis.

A major breakthrough in understanding how pRb regulates the cell cycle was the observation that pRb binds to members of the E2F transcription factor family (referred to here as E2F).^70–72 Further work has shown that pRb function is largely dependent on interactions with E2F.⁷³ E2F sites are found in the promoters of many genes that are important for cell cycle progression, and pRb represses transcription of these genes through its interaction with E2F.^74–77 Since E2F (but not pRb) has a DNA binding domain, the pRb-E2F association would explain how pRb is brought to specific DNA elements to exert its effect. Most E2Fs have a transactivation domain that stimulates expression of genes containing E2F binding sites in their promoters. pRb binds E2F within the transactivation domain,^78,79 thereby masking its activity. Since E2F activates genes involved in cell division,^74,80 inhibition of E2F provided a mechanistic explanation for how pRb inhibited cell division. However, the picture became more complicated with the recognition that pRb has intrinsic or “active” transcriptional repressor activity and is able to block the expression of genes when artificially brought to promoters by proteins other than E2F.⁸¹ These findings suggest a complex relationship between pRb and E2F. In some situations, pRb inhibits genes by simply masking the E2F transactivation domain, whereas in other situations E2F serves as a “courier” to deliver pRb to specific genes for active repression.

The importance of active repression by pRb was pointed out in several studies that showed that this activity is required for pRb to inhibit the G1-to-S phase transition of the cell cycle.^82,83 But how does pRb actively represses transcription? In a series of landmark papers, several groups showed that pRb binds to and recruits to promoters proteins that alter chromatin structure, such as histone deacetylases,^83–85 SWI-SNF ATPases,^86–88 DNA methyltransferases,⁸⁹ polycomb complexes,⁹⁰ and histone methylases.⁹¹ Alteration of local chromatin structure into a restricted conformation prevents access by the transcriptional machinery, thereby inhibiting expression, whereas dynamic reorganization of chromatin into an open configuration allows gene transcription. Depending on the nature of the chromatin-remodeling complex that pRb recruits, the cell cycle inhibition can be temporary, as occurs during the quiescent period between cell divisions, or permanent, as occurs during cell differentiation and senescence.⁹²

Studies of the tertiary (three-dimensional) structure of the pRb protein have provided insights into how the protein performs these complex functions (Fig. 128.2). The central region of the pRb protein contains the A box and B box, which are highly conserved from human to plants. These regions interact with each other along an extended interdomain interface to form the A–B pocket. The pocket is critical for the tumor suppressor function of pRb, and is disrupted by most germ line mutations in hereditary retinoblastoma patients and somatic mutations in tumors.^29,93 The pocket is required for binding to E2F, chromatin remodeling enzymes, viral oncoproteins, and other molecules. Many pRb-binding proteins contain an LxCxE (leucine – variable amino acid – cysteine – variable amino acid – glutamic acid) binding motif, which has been shown by crystallographic studies to be located in the B box.⁴ The B box does not assume an active confirmation unless bound to the A box, thereby providing an explanation for why both boxes of the pocket domain are required for pRb activity. Interestingly, E2F does not contain the LxCxE motif and binds pRb at a distinct site (Fig. 128.3). Recent crystallographic studies confirmed that the E2F binding site, located at the interface between the A and B boxes, is distinct from the LxCxE site.⁹⁴ These findings provide a structural explanation for how pRb simultaneously binds E2F and chromatin remodeling proteins, many of which interact with pRb through the LxCxE site.⁸⁸ The current molecular view is that pRb orchestrates the assembly of multi-protein complexes, which are then recruited to specific promoters by E2F, where they control access of the transcriptional machinery. Thus, pRb regulates in a dynamic and integrated manner the expression of specific genes involved in cell division, differentiation and apoptosis.⁹⁵

The carboxy-terminal end of the pRb protein is also critical for its function. This region is required, along with the pocket, for binding to E2F.⁹⁶ In addition, most of the phosphorylation sites that seem to be critical for regulating pRb activity are located in the carboxy-terminus.⁹⁷ In fact, recent work has shown that phosphorylation of pRb on the C-terminus initiates a novel intramolecular interaction that progressively strips pRb of activity as the cell moves through the cell cycle.⁹⁸ The carboxy-terminal region also contains binding sites for the oncoproteins c-abl and MDM2.^99,100 The tyrosine kinase activity of c-abl is blocked when it is complexed with pRb,⁹⁹ and this interaction appears to be important for Rb-mediated growth suppression.¹⁰¹ Besides directly blocking c-abl, the C-terminal region also appears to participate in the assembly of multimeric complexes containing pRb, E2F, c-abl and potentially other proteins.¹⁰² The importance of the pRb-MDM2 interaction is less clear. MDM2 interacts with the p53 tumor suppressor protein and opposes its proapoptotic activity by repressing p53 transcriptional activation and by mediating its degradation.^103,104 While initial results demonstrated that MDM2 blocks pRb function, more recent studies have shown that pRb can form a trimeric complex with MDM2 and p53 and thereby block the antiapoptotic activity of MDM2 by preventing the degradation of p53.¹⁰⁵

The function of the amino-terminal end of the pRb protein remains less clear. This region contains phosphorylation sites that may regulate pRb activity. Additionally, this region interacts with several proteins, including a replication-licensing factor (MCM7),¹⁰⁶ a novel G2/M cycle-regulated kinase,¹⁰⁷ and other proteins.¹⁰⁸ However, the function of these interactions has not been established. RB1 ± mice that develop pituitary tumors due to loss of Rb failed to be “rescued” by expression of a mutant form of pRb that lacked the amino-terminus, although the onset of tumors was delayed.¹⁰⁹ In other experiments, tumor suppression by pRb was actually enhanced when the amino-terminal region was removed.^110–112 Thus, the amino-terminus appears to contribute only weakly to the overall tumor suppressor activity of pRb.

Not only was the RB gene the first identified tumor suppressor gene, but the Rb pathway was the first, and still one of the most important, tumor suppressor pathways to be elucidated in human cancer.⁵⁵ As described above, the pRb protein is critical for regulation of the cell cycle, as well as senescence, differentiation and apoptosis, all of which are deregulated during cancer formation.^{67,92,113–120} While the pRb pathway is deregulated in virtually all cancers,⁵⁵ the RB1 gene is mutated in only a limited spectrum of cancers. In these other cancers, the pRb protein is inactivated by maintaining it in a hyperphosphorylated state. This occurs by deregulating the pRb pathway, which controls the phosphorylation state of pRb through kinases and kinase inhibitors.

The RB1 tumor suppressor pathway

Cell cycle progression normally occurs when pRb is inactivated by phosphorylation that is catalyzed by cyclin-dependent kinases (CDKs) in complex with their cyclin partners.^59,121,122 pRb contains 16 potential sites for CDK phosphorylation, and it oscillates between hypophosphorylated and hyperphosphorylated forms in cycling cells. At least three different cyclin/CDK complexes phosphorylate pRb during the cell cycle (Fig. 128.3). Cyclin D-ck4/6 phosphorylates pRb early in G1, cyclin E-CDK2 phosphorylates the protein near the end of G1, and cyclin A-CDK2 is thought to maintain phosphorylation of pRb during S phase.⁵⁵ Phosphorylation of specific sites appears to regulate distinct pRb functions, suggesting complex regulation of pRb by these phosphorylation events. For example, binding of E2F, LxCxE proteins, and c-abl are regulated by distinct sets of phosphorylation sites in the carboxy-terminus.^123,124

pRb is phosphorylated sequentially by different CDKs during the cell cycle. In fact, successive phosphorylation of pRb by cyclin D-CDK4/6 and cyclin E-CDK2 appears to be necessary to completely hyperphosphorylate pRb.¹²² Recently, a mechanistic explanation was suggested for how cyclin D-CDK4/6 and cyclin E-CDK2 may regulated distinct pRb functions.⁹⁸ Cyclin D-CDK4/6 appears to phosphorylate specific sites in the carboxy-terminal region of pRb, and this phosphorylation triggers an intramolecular interaction between the phosphorylated C-terminal region and a positively charged “lysine patch” encircling the LxCxE binding site in the B box of the pocket. This interaction displaces LxCxE proteins such as histone deacetylase from the pocket, thereby blocking the ability of pRb to arrest the cell cycle.^82,98 However, pRb can still bind to E2F in this partially phosphorylated state. Under hyperproliferative conditions, the intramolecular interaction between the carboxy-terminal region of pRb and the pocket also can recruit cyclin E/CDK2 to the pocket, where it phosphorylates serine-567, an otherwise inaccessible site buried within the pocket.¹⁰⁷ Ser-567 makes critical contacts between domain A and B,4 and this phosphorylation disrupts the A–B interface and disrupts pRb binding to E2F. The sensitive location of Ser-567 is further illustrated by the fact that it is the only CDK phosphoacceptor site in pRb that is a target of naturally occurring missense mutations in tumors.¹²⁵ The more complete inactivation of pRb, indicated by phosphorylation of serine-567, leads to release of E2F and increased apoptosis.¹¹⁸ Taken together, these findings suggest that the normal cell cycle may be regulated by partial phosphorylation of pRb that is catalyzed by cyclin D-CDK4/6, whereas the more complete phosphorylation that requires cyclin E-CDK2 may serve as a checkpoint for an abnormal hyperproliferative state that would trigger cell death.

Proteins called cyclin dependent kinase inhibitors (CDKIs), which inhibit the kinases that phosphorylate Rb (Fig. 128.3) represent another layer of complexity in the regulation of pRb. The p16INK4a protein is a CDKI that specifically inhibits CDK4, which catalyzes the early phosphorylation of pRb.¹²⁶ A tumor suppressor protein in its own right, p16INK4a is mutated or inactivated in many types of cancer, including cutaneous and uveal melanoma.^127,128 Loss of p16INK4a allows cyclin D-CDK4 to act unopposed in phosphorylating pRb, which results in constitutive functional inactivation of pRb. Since pRb and p16INK4a act in the same pathway, and mutation of either gene results in a similar deregulation of the cell cycle, both genes are rarely mutated in the same tumor.¹²⁹ Other CDKIs, such as p21 and p27 have more broad roles in regulating the cell cycle.¹³⁰

Reese–Ellsworth classification

The Reese–Ellsworth classification system, originally published in 1964, was a major advance in our collective understanding of retinoblastoma.^193,194 It is a group classification system and as such, only deals with organ-confined intraocular disease. In a staging system evaluating the whole child, a group classification would fit entirely within stage I disease. The Reese–Ellsworth classification was developed just as the indirect ophthalmoscope was being introduced into clinical practice. Anterior tumors, which can now be easily recognized and treated with cryotherapy or radioactive plaque, cause the eye to be classified in a more advanced group when using the Reese–Ellsworth classification. The classification does not take into account retinal detachment and subretinal tumor seeding. Also, vitreous seeding of any amount places the eye in group 5b (the last of 10 subgroups) with the poorest prognosis. Today, local vitreous seeding can often be successfully treated with brachytherapy.

International Classification for Intraocular Retinoblastoma

In 1989 Kingston and colleagues introduced the systemic chemotherapy combination currently in use in most centers (carboplatin, etoposide, vincristine) in combination with external beam radiotherapy as the primary treatment of retinoblastoma for Reese–Ellsworth group Vb eyes.^195,196 In January 1990, Murphree was the first to use chemotherapy plus local consolidation without external beam radiotherapy for all Reese–Ellsworth groups. Initially the early stage Reese–Ellsworth eyes in Los Angeles were treated with carboplatin alone combined with transpupillary diode laser hyperthermia (chemothermotherapy). Advanced intraocular disease was treated with CEV (carboplatin, etoposide, vincristine) systemic chemotherapy combined with sequential aggressive local therapy. These authors introduced the term “chemoreduction” to describe the concept of reducing the volume of the tumor with systemic chemotherapy followed by local consolidation in 1992.¹⁹⁷

Efforts at developing a revised classification that reflected the move away from external beam radiotherapy to primary chemotherapy for intraocular retinoblastoma began as early as 1994, when Murphree and Hungerford co-chaired a meeting of retinoblastoma specialists from around the world for a full day of discussions on the shape of a new classification system at the International Congress of Ophthalmology in Toronto. The system that grew out of those discussions took into account zone of disease similar to retinopathy of prematurity. However, the classification that this large group developed was too complicated and cumbersome to be useful. Eventually, the ABC classification system was described at the European Congress of Ophthalmology meeting in Istanbul in 2001. In May 2003, discussions at the Xth International Retinoblastoma Symposium led to a joint international effort in which a number of major retinoblastoma centers signed onto the effort. The case for a new group classification has recently been published.¹⁹⁵ The process of validating the International Classification is currently underway.

The International Classification System (Box 128.1) is based both on the natural history of retinoblastoma and on the likelihood of salvaging the eye when systemic chemotherapy is used as the primary treatment. Letters “A” through “E” instead of numbers were chosen to designate each classification group to avoid confusion with the Reese–Ellsworth system. The risk of loss of the eye due to retinoblastoma is graduated from “very low” for Group A to “very high” for Group E.

In this classification, the letter “A” is assigned to those eyes for which both the likelihood of curing the tumor and retaining excellent vision are both high. In intraocular group A eyes, the lesions are small and are away from critical visual structures (foveola and optic nerve). Groups A and B contain all eyes in which the tumor remains confined to the retina. In groups C and D eyes, the tumor has spread into the vitreous and subretinal space. In the case of group C eyes the spread is local. In the case of group D eyes the seeding is diffuse (Fig. 128.7). Group E eyes have been destroyed by the tumor and are rarely salvageable. In this system, the morbidity of the treatment increases from group A to group E, and the probability of salvaging the eye and useful vision decreases from A to E. Within limits, when chemotherapy is the primary treatment of retinoblastoma, absolute volume of the tumor is relatively less important than whether the tumor has become dispersed within the eye to either the vitreous or the subretinal fluid or both.

Fig. 128.7 Photographs depicting groups A to E; intraocular retinoblastoma as described in the new group classification (for details see text).

Retinoblastoma survival rates

The overall survival rate from retinoblastoma in developed countries results from a mixture of causes of death. In the first 4 years of life, most of the deaths will be from metastatic retinoblastoma.¹⁹⁸ Later deaths are increasingly likely to be the result of genetically predisposed second primary tumors such as osteosarcoma or fibrosarcoma. Thus, the overall survival rate will differ depending upon the time period examined. The survival rate in 1980 from one major retinoblastoma center in the USA was 92%.¹⁹⁸ The similar overall 5-year cumulative survival rate was 91% in SEER (Surveillance, Epidemiology, and End Results) data from 1974 to 1985.¹⁷¹

It is important to differentiate between survival from retinoblastoma alone or survival from both retinoblastoma and second primary neoplasms. For example, Abramson and colleagues reported that 86%, not 92%, of bilateral retinoblastoma patients survive 15 years.¹⁹⁸ By 5 years after diagnosis more children die of second malignant neoplasms than from retinoblastoma. In the Netherlands the presence of a National Registry shows the striking difference in overall survival between the subgroup of genetically predisposed patients and those who have no genetic predisposition.¹⁹⁹ A major controversy has arisen concerning the percentage of patients who will eventually succumb to second malignant neoplasms (SMNs). Abramson and colleagues had previously reported that as many as 90% of patients carrying the retinoblastoma predisposition allele develop a SMN within 30 years of diagnosis of the original retinoblastoma.²⁰⁰ In a more recent report of 1000 patients with nearly complete ascertainment from the New York and Boston group, 6% of bilateral patients had died of SMNs by 40 years after diagnosis, if they had not been treated with external beam radiation (EBR). In contrast, 35% died of SMNs if EBR had been used in their treatment.

The overall survival rate in developing countries depends upon whether or not there was a delay in receiving medical attention. These delays are likely to be cultural, e.g., there is no tradition of seeking medical help until a process is far advanced. For example, in a report from Malaysia in 1980, only 6 of 20 patients survived retinoblastoma. All were initially seen with advanced disease and were older at initial diagnosis.²⁰¹ In comparison, in Great Britain where there is a tradition of early access to medical care, only 26 of 317 patients managed at Moorfield’s and St Bartholomew’s Hospitals in London between 1970 and 1984 developed orbital disease indicative of late diagnosis.²⁰² In Argentina, delayed diagnosis was related to lack of access to medical care and initial consultation with a pediatrician instead of an ophthalmologist.²⁰³ In the 1960s, Nigeria had an extremely high mortality rate because of late medical attention and lack of treatment facilities.²⁰⁴ In Japan 50 years ago, only 50% of patients with unilateral disease and 17% of patients with bilateral disease survived for 5 years.²⁰⁵

Factors affecting survival

Traditionally there have been histologic features in enucleated eyes that have seemed to correlate with increased risk of metastatic disease.²⁰⁶ Invasion by the tumor into the optic nerve posterior to the lamina cribrosa combined with massive invasion of the choroid is a combination that most centers would agree needs preventive or adjuvant chemotherapy (Fig. 128.8). Recent reports have questioned the value of retrolaminar involvement of the optic nerve and choroidal invasion in predicting risk to metastatic disease.²⁰⁷ Marback and colleagues have also questioned the effectiveness of invasion past the lamina cribrosa as a reliable predictor of increased risk of metastatic disease.¹⁸⁹

Fig. 128.8 (A) Histopathology showing invasion of the tumor into the optic nerve posterior to the lamina cribrosa. (B) Massive invasion of the choroid by tumor.

There appear to be no inherent differences in the risk for metastatic disease between patients that are genetically predisposed to retinoblastoma and those without genetic predisposition.²² Delay in diagnosis is a major factor adversely affecting survival of both heritable and nonheritable disease.²⁰⁸

Multiple episodes of recurrent disease in an eye is a bad prognostic sign for the survival of the eye and for the overall survival of the child. The more cell divisions a tumor cell goes through, the more likely it is that one of the progeny cells will achieve the enabling mutations that allow this clone of cells to survive outside the eye. With each visible recurrence inside the eye, more of those abilities may have been acquired through spontaneous mutations.

Factors affecting salvage of eye and vision

As suggested by Knudson in 1976,¹ early detection of signs of retinoblastoma with timely referral and treatment beginning when the tumors are small is perhaps the most important factor that increases the likelihood of salvage of the eye. A significant delay in initial diagnosis is associated with a decreased likelihood of retaining the eye. Haik and his colleagues evaluated the cause of the delay in diagnosis.²⁰⁹ According to these authors, the first delay in diagnosis of retinoblastoma occurs before the observation of the first symptom by the child’s parents or other observer. A second delay occurs between the observation of the first symptom and the visit to the primary care physician. These authors maintain that educating parents on signs to look for could possibly reduce this second source of delay. In addition, more than 50% of all primary care physicians made a referral to a pediatric ophthalmologist within 1 week, when there was no family history of disease, and 75% of the primary physicians referred the family to a pediatric ophthalmologist when there was a positive family history. However, 47% of the patients without a positive family history and 25% of patients with a positive family history had a delay averaging 4–5 months before the primary care physicians referred them to an ophthalmologist.²⁰⁹

With primary chemotherapy and focal consolidation, there is a direct correlation between prognosis and group classification. In Los Angeles Group A, eyes are virtually all preserved with excellent visual acuity. Group B eyes have nearly as good a prognosis (95%) but the visual outcome varies from 20/20 to 20/200 depending on whether or not the fovea is destroyed by the tumor. Group C eyes are salvaged at a rate approximating 70%. In contrast, group D eyes with diffuse dissemination and more advanced disease have no greater than a 50% chance of survival without the use of external beam radiotherapy. Only an occasional group E eye is salvaged (2%).

For patients with macular tumors, post-treatment vision is often 20/200 or worse, but can be better. In a series of 17 patients with large macular tumors, post-external beam radiotherapy treatment vision ranged from 5/200 to 20/50.²¹⁰ The authors pointed out that the ophthalmoscopic appearance and size of the tumors gave little reliable guidelines to visual outcome. In another series of 11 patients following treatment of macular retinoblastomas, two patients eventually recovered 20/20 vision despite having subretinal edema in the macula at the time of diagnosis.²¹¹ We have had experience with a patient who had a macular lesion completely filling the space within the vascular arcades. After treatment this tumor mass shrunk away from the disc and the fovea to a site 4 DD temporal to the fovea. Vision returned with excellent central fixation. In this case, and in some other cases, the tumor obviously had mushroomed over the surface of the retina and did not destroy the foveal cones.

Signs and symptoms

Leukocoria is the phenomenon created when light entering the eye is reflected back out through the pupil by the yellow or yellow-white tumor (Fig. 128.4). The observer sees this reflex or reflected light diffusely filling the pupil, giving rise to the term “leukocoria” or white pupil. Retinoblastomas in the posterior pole of the globe that have reached 3 mm in basal diameter are sufficiently large to generate a leukocoric reflex. In the USA more than half of all retinoblastomas are suspected or diagnosed after observation of a leukocoria, in the affected eye usually by a close family member.

In a retrospective review of 1265 patients with retinoblastoma from New York Hospital, leukocoria was the presenting sign in 56% of all cases.²¹² It is interesting and informative to note that in the New York series observation of leukocoria correlated with the presence of advanced (Reese–Ellsworth group Va or Vb) disease. Yet we know that with a large pupil diameter, even small tumors may create leukocoria. The red-reflex test, as used in a standard manner by pediatricians, has been documented to be positive in the presence of undilated pupils only when advanced or large retinoblastoma is present.²¹³

The likelihood of observing leukocoria relates directly to the diameter of the pupil at the time of observation. Family members often see leukocoria in the early evening or in dim light when the pupil dilates naturally. Attempts by the parents to take the child to “good light” in order to see the pupillary “glow” better, results in the opposite effect: it disappears because the pupil constricts. Because their observation is intermittent and due to the variations in pupil size, parents or family members often question their own observation. This tendency to be uncertain about the observation is reinforced if the pediatrician or primary care physician does not dilate the pupil, does not observe the leukocoria and assures the family that there is no abnormality.

Flash photography can easily document the presence of leukocoria. However, the camera must be either an inexpensive cameras that lacks the “eliminate-red eye” pre-flash or must have that feature turned off. The family often takes holiday or baby photographs in dim illumination when the child’s pupil is naturally dilated. The flash illuminates the interior of the eye, and the film records the reflected glow from the tumor (leukocoria) well before the pupil has a chance to constrict to the bright flash. Parents are urged to take and develop large numbers of flash photographs of their infants with cameras known to record “red-eye” in the first few weeks and months of life and become aware of the possibility of an abnormal pupil. If they observe such an abnormality, they should immediately take the photograph to their pediatrician demonstrating the leukokoria and insist on a referral to an ophthalmologist, preferably one who specializes in pediatrics. A number of parents in the last 4–5 years have searched the internet for “white pupil” and have found descriptions of the association between what they observe in their child and the possible presence of retinoblastoma. (Examples of websites where information is available for parents are: www.retinoblastoma.com and www.retinoblastoma.net, both accessed on December 9, 2011.)

Brightly lit rooms and high light intensity on the direct ophthalmoscope are two of the gremlins that actively conspire to prevent pediatricians and other primary care physicians from making an early diagnosis of retinoblastoma. In 80% of cases of retinoblastoma, failure to observe the clinical signs of this cancer can be minimized for pediatricians or primary care physicians by performing the red reflex examination in a darkened room with the light on the direct ophthalmoscope turned down to low.

The American Academy of Pediatrics (AAP) in May 2002 published a new policy statement regarding how their members should perform the red reflex test. This was subsequently updated in 2008.^213A (Their recommendations can be found on the AAP website at: www.aap.org, under Policy Statements; red reflex testing.) This policy statement is a tremendous step forward in the effort to assure early detection for all infants and toddlers with intraocular retinoblastoma.

Small tumors in the foveola or near the foveola can significantly reduce the visual acuity in that eye. Strabismus is the initial sign in one of every five patients with retinoblastoma and almost always is a result of either a tumor or of tumor-associated subretinal fluid in the macula.²¹² Decreased visual acuity caused by destruction or obscuration of the fovea is the immediate cause of the strabismus. If vision is decreased only because the tumor occludes the visual axis but does not destroy the fovea, there is a good possibility that the vision will improve after treatment. Before treatment begins it may not be clear whether or not the fovea has been destroyed by the tumor. Strabismus in the first 6 months of life always requires an immediate retinal examination to rule out retinoblastoma.

Other less common signs and symptoms include a red, painful eye with glaucoma, cloudy cornea, poor vision, and an aseptic orbital cellulitis (Fig. 128.9). In five cases of retinoblastoma presenting with the clinical picture of aseptic orbital cellulitis,²¹⁴ the imaging studies were misinterpreted as showing extraocular extension of the tumor into the optic nerve. We have had similar experiences in patients with retinoblastoma presenting as orbital cellulitis. Systemic treatment with 3 days of oral corticosteroids results in dramatic reduction of the swelling. In one case, pre-steroid and post-steroid CT scans were dramatically different. The edema in the perineural soft tissues responsible for the false-positive interpretation of tumor invasion into the nerve may disappear after steroid treatment.

Fig. 128.9 (A) Aseptic orbital cellulitis in a patient with advanced intraocular retinoblastoma. (B) Corneal clouding and conjunctival chemosis in an eye with neovascular glaucoma secondary to advanced intraocular retinoblastoma.

Diseases simulating retinoblastoma (pseudoretinoblastoma)

In 500 consecutive referrals for retinoblastoma, simulating lesions were found in 212 of the 500 patients.²¹⁵ there were a total of 23 different conditions that referring pediatricians or ophthalmologists were unable to distinguish from retinoblastoma. The three most common causes of pseudoretinoblastoma were persistent hyperplastic primary vitreous (28%), Coats disease (16%), and presumed ocular toxocariasis (16%) (Fig. 128.10). Congenital cataract and retinopathy of prematurity were less frequently mistaken for possible retinoblastoma in this series than in other, earlier reports. Our experience in Los Angeles shows many fewer cases referred as possible retinoblastoma that turn out to be pseudoretinoblastoma. Other diagnoses that we have seen simulate retinoblastoma include Familial Exudative Vitreoretinopathy (FEVR), Norrie disease, and congenital retinal folds.

Fig. 128.10 (A) Unusual presentation of Coats disease in a 4-year-old simulating retinoblastoma. Foveal nodule is the result of chronic lipid from telangiectatic vessels in periphery seen here by fluorescein angiography (B).

In 1996, de Potter and colleagues²¹⁶ evaluated the role of MRI in differentiating solid intraocular tumors from intraocular lesions with primary retinal detachment. They reported that solid intraocular tumors appeared hyperintense on T1-weighted images and hypointense on T2-weighted images. Also, these lesions showed minimal to marked enhancement on contrast-enhanced T1-weighted images with fat suppression techniques. In secondary serous or exudative retinal detachment from intraocular lesions (Coats disease, persistent hyperplastic primary vitreous, phthisis bulbi, and retinopathy of prematurity), MRI showed hyperintensity of the subretinal space on both T1- and T2-weighted images and no enhancement in the subretinal space on contrast-enhanced sequences.

Unilateral nonheritable retinoblastoma

The ABC classification of intraocular retinoblastoma¹⁹⁵ helps clinicians decide when to treat unilateral retinoblastoma conservatively. Nearly all centers will attempt to salvage unilateral group A and B eyes. E eyes should be enucleated.

For group A disease, often detected in newborn members of a pedigree of retinoblastoma, photocoagulation and cryotherapy can be employed if these lesions are <3 mm in diameter and distant from the fovea and optic nerve. If they are too close to those vital structures, the classification would be group B.

The typical unilateral group B disease eye can be successfully treated with 3–4 cycles of two- or three-drug chemotherapy followed by focal consolidation. The exact chemotherapeutic regime varies from center to center with some experts preferring to exclude etoposide (due to its leukemogenic potential) while others treat with six cycles and three agents. Single-agent carboplatin with laser consolidation has also been described. In intraocular group B eyes with a solitary lesion outside the posterior pole, a radioactive plaque could be used as primary treatment. In those eyes, a plaque is preferred over EBR because of the time, expense, and number of anesthesias required to deliver radiation in a small child. Most experts do not believe that brachytherapy increases the risk of radiation induced tumors and risk of radiation retinopathy is small.

The case can be made for treating unilateral intraocular group C retinoblastoma, in which there is only local dissemination of tumor into the vitreous or subretinal space. Most oncology centers will attempt salvage in these cases especially if there is good visual potential. Systemic three-agent chemotherapy with carboplatin, vincristine, and etoposide, is the most common regimen administered over six cycles.

If the likelihood of salvaging group D eyes with diffusely disseminated intraocular tumor is ≤40%, a strong case can be made for recommending enucleation primarily and sparing the child the trauma of treatment to retain an eye that will likely be of no visual use. Shields and colleagues initially described their experience treating unilateral retinoblastoma with primary chemotherapy.²³⁰ The eyes that did well without needing either enucleation or external beam radiotherapy were eyes that would have been classified as intraocular groups A–C. Group D eyes did not do well. If the child and not the eye is taken into consideration in the decision-making process, strong consideration should be made for enucleation of unilateral group D eyes. If the chance of salvaging useful vision is not good, heroic treatment approaches are unwarranted. It makes little sense to subject a child with one completely normal eye to prolonged systemic chemotherapy and possibly external beam radiotherapy.

Bilateral retinoblastoma: symmetrical disease

Bilateral group A disease can follow the treatment strategy stated above-laser treatment and/or cryotherapy. For the very rare cases of bilateral group B disease, primary brachytherapy in each eye may be appropriate. For the more common bilateral group C disease, group C treatment protocol should be followed even if one eye appears to have the fovea destroyed by tumor. In patients with bilateral group D disease, many centers recommend 6 cycles of the three-drug chemotherapy consisting of carboplatin, etoposide, vincristine (CEV), combined with three cycles of with local subtenon carboplatin (20 mg in 2 mL) as the primary treatment of choice followed by local consolidation with laser or cryotherapy. Alternatively, systemic chemotherapy can be followed by low dose consolidative external beam radiation. The latter approach has been very successful in Los Angeles.

Advanced intraocular disease (groups D and E)

If the worse eye has features of group E disease, then enucleation is generally recommended. In some centers, if systemic chemotherapy is initiated for the contralateral eye, salvage of a group E is considered. There have been rare case reports of successful salvage of group E eyes.²³¹ If, in a group E eye the neuroimaging study suggests invasion of the tumor into the optic nerve and the other eye has group B or higher and requires systemic chemotherapy, there are two approaches. Many centers delay enucleation of the group E eye until after six cycles of chemotherapy. Others feel that this approach is flawed as it alters the histopathologic features of an eye potentially masking high risk features such as cut end disease or massive choroidal involvement. Those centers advocate immediate enucleation with adjuvant treatment based on histopathology.

When group D disease in one eye is associated with groups A–C disease in the second eye, the group D eye drives the selection of the treatment intensity if the choice is made to attempt to salvage the group D eye. It should not be automatically assumed that the eye with lesser grade tumor would ultimately be the better-seeing eye. Local treatment for groups A and B eyes can be delayed until the chemoreduction aimed at the fellow eye has its effect.

Terminology

There is still no firm agreement among pediatric oncologists on the precise descriptive terminology for systemic chemotherapy when treating intraocular retinoblastoma. Everyone agrees that systemic chemotherapy given in response to high-risk pathology is referred to as “adjuvant” chemotherapy in that it provides additional treatment after enucleation. When primary systemic chemotherapy came into general use as the initial or primary treatment modality for other childhood tumors, some oncologists recommended the use of the term “neoadjuvant” chemotherapy. Many oncologists argue, however, that the term “neoadjuvant” is confusing and the trend is away from its use. It appears that the most appropriate term might be “primary intravenous chemotherapy with focal consolidation.” At least this would assure that published articles about the treatment of retinoblastoma would routinely appear in PubMed searches by our colleagues in oncology.

Pre-1989 chemotherapy for extraocular disease

Before 1989, systemic intravenous chemotherapy was used rarely for the treatment of intraocular retinoblastoma. Most of the experience in that era was with using chemotherapy for extraocular retinoblastoma. Oncologists tested many different drugs. The usual agents employed at that time included cyclophosphamide (at what we know now was an inadequate dose), vincristine and doxorubicin. Pratt et al. described the St Jude experience with chemotherapy for extraocular retinoblastoma between 1962 and 1984.²³² A total of 11 of 114 children received chemotherapy for measurable extraocular disease that had been present at diagnosis (7/11 patients) or developed later (4/11 patients). The two patients with only orbital disease had complete response to chemotherapy and subsequent radiotherapy and are the only long-term survivors. Single-agent treatment was ineffective. Only cyclophosphamide and ifosfamide showed any value at all when given alone. Drugs that induced no response as a single agent included vincristine, doxorubicin, cisplatin, and VM-26.

Background of the currently used chemotherapy regimen

Kingston and colleagues at St Bartholomew’s Hospital in London were the first group systematically to explore the effect of the platinum group of drugs in this tumor.¹⁹⁶ In their 1987 report, they described 11 children with metastatic retinoblastoma treated between 1970 and 1984 with salvage chemotherapy. Three of the 11 had complete short-term remissions with a regimen of four drugs containing cisplatin. One additional child responded well to a combination of vincristine and cyclophosphamide. Based on these results they suggested that retinoblastoma appeared to be a chemosensitive tumor. In addition, these results led these authors to try a sandwich protocol of carboplatin, etoposide and vincristine with 40–44 Gy EBR as planned primary therapy in 14 group Vb eyes between March, 1989 and April, 1995.

After hearing a report of their work at a meeting in March, 1990, Murphree et al. began the use of a carboplatin-based regimen for the primary treatment of all Reese–Ellsworth stage eyes. Initially in Los Angeles, carboplatin alone was combined with transpupillary hyperthermia for the treatment of smaller localized tumors. The results from thermochemotherapy were encouraging but better results and reduced time in the OR came with the use of triple agent chemotherapy combined with transpupillary laser or transscleral cryotherapy.²³³

By 1994, the Los Angeles three-drug protocol of primary chemotherapy with carboplatin, etoposide and vincristine (CEV) followed by focal consolidation was being used widely. By 1996, the patient outcomes from London, Los Angeles, Toronto, and Philadelphia were all published in the November issue of the Archives of Ophthalmology.^{196,233–235}

Primary systemic chemotherapy

Systemic chemotherapy has replaced external beam radiotherapy as the most commonly used primary treatment for intraocular retinoblastoma. The chemotherapy agents may vary slightly but most protocols consist of the systemic administration of carboplatin, etoposide, and vincristine (known as CEV or VEC in North American centers, and JOE in Britain), given over a 2-day period every 3 or 4 weeks (the administration of the drugs and the time required for the suppressed bone marrow to recover is considered one cycle). In Los Angeles, we use six cycles of CEV for intraocular group C and D disease and three cycles for intraocular group B disease (Fig. 128.13). There is some concern about etoposide, as it has been associated with the appearance of a secondary acute myelogenous leukemia (AML), in a small cohort of children treated for retinoblastoma (see below).^236,237 For this reason, some centers have dropped etoposide from the regimen for less advanced disease (groups A and B).

Fig. 128.13 Group B retinoblastoma, before (A) and after (B) primary chemotherapy.

Subtenon carboplatin

The addition of subtenon carboplatin to augment penetration of systemic carboplatin has been advocated by a number of groups over the past 10 years. Introduced and described as subconjunctival injections²³⁸ we modified the technique. Following a transconjunctival and trans-tenon incision, an olive tip cannulas is used to inject the chemotherapy in one of the quadrants of the eye between the rectus muscles (Fig. 128.14).

Fig. 128.14 (A) Small incision is made down to bare sclera. (B) An olive-tipped cannula is inserted and forced posteriorly. (C) The cannula is inserted to the hub to ensure proper positioning. (D) B-scan ultrasonography demonstrates a fluid cleft adjacent to sclera.

Following injection a number of complications have been described including, lid swelling, loss of orbital volume and secondary optic neuropathies. Severe orbital scarring can lead to strabismus and make subsequent enucleations challenging. This has led some centers to abandon the technique.

Complications of primary systemic chemotherapy

The chemotherapeutic agents currently used to treat intraocular retinoblastoma can cause many complications. Most complications are dose-related and are of short duration. As far as is known, these chemotherapeutic agents do not predispose children with heritable retinoblastoma to the same tumors for which they are genetically at risk, unlike the situation with external beam radiotherapy. The use of etoposide has been a source of significant debate. The drug is known to be leukemogenic and there has been concern that exposing patients with the RB1 mutation would increase their risk for secondary AML (sAML). While there have been a series of patients treated with systemic chemotherapy who later developed sAML, the number remain small; the majority of these patients were treated with high-dose multi-agent modalities well above the standard three-agent six-cycle regimen.^236,237 Most experts agree that if etoposide poses a risk for second tumors, it remains a small one. Perhaps a bigger concern is the potential risk of hearing loss following carboplatin use. Wilson and colleagues have reported an increase in carboplatin related ototoxicity at St Judes. Similar results have not been duplicated elsewhere.²³⁹ The risk of ototoxicity is significant when one considers that many retinoblastoma patients are visually disabled relying on other sensory input such as hearing. As with other intravenous forms of chemotherapy risks of bone marrow suppression, alopecia and central line infection exist; in rare cases these can severe and life-threatening.

Laser

In primary chemotherapy of intraocular retinoblastoma, focal consolidation is most often accomplished with transpupillary laser given directly to the residual tumor mass. Any wavelength laser from the green 532 nm argon, through yellow, orange, red, to the 810-nm diode infrared or the 1064-nm frequency-doubled YAG can theoretically deliver sufficient energy to reach photocoagulative temperatures of around 70°C in the tumor if it is absorbed. The energy of a laser is converted to heat only when it is absorbed. The shorter-wave length green 532-nm argon is more readily absorbed in the relatively nonpigment retinoblastoma, while the longer wave length 810-nm diode infrared laser is best absorbed by intact retinal pigment epithelium.

The technique we find useful with the argon 532 nm is essentially the same for both primary treatment of group A lesions and focal consolidation following primary chemotherapy in groups B–D. In general, focal consolidation begins concurrently with the beginning of the 2nd or 3rd cycle of systemic chemotherapy after the tumor volume has been reduced. The goal of the therapy is to completely cover each lesion with 30% overlap during at least three different sessions. We choose initial power setting of 250–300 mW, with a duration of 300–500 msec. The power and time settings are kept low to prevent tumor disruption and hemorrhage that may be associated with excessive energy delivery. The first burns are placed at the edge of the lesion with the spot half on and half off the tumor. The power and/or duration can be adjusted to achieve gentle whitening of the tumor. We do not recommend exceeding 500–600 mW and 700 msec with the 532 mm laser. Once the lesion is outlined, then the entire lesion including any type I regression-associated calcium is covered with overlapping rows of burns (Fig. 128.15). A small to moderate-sized lesion may require 200–400 burns for good coverage. The burns over the thicker areas of the tumor may be virtually invisible compared with those placed at the edge of the lesion. The power or duration should not be increased to compensate for the decreased “take” over the thicker parts of the lesion. Repeat the laser coverage at 2–4-week intervals during and/or after the administration of systemic chemotherapy until the entire lesion has been covered on at least three different occasions (Fig. 128.16).

Fig. 128.15 Technique for laser focal consolidation. 1. First burns are placed at the edge of the lesion with the spot half on and half off the tumor. 2. The lesion is then outlined with 30% overlap of the previous spot. 3. The lesion is then painted with 30% overlap.

Fig. 128.16 (A) Group B tumor prior to treatment with systemic chemotherapy and focal laser consolidation. (B) After treatment there is significant reduction in tumor volume.

Because the infrared 810-nm diode laser has a longer wavelength than the argon laser, it penetrates further and is absorbed mainly by the retinal pigment epithelium. It is useful primarily if retinal pigmented epithelium (RPE) is intact under the lesion to be treated. One major advantage of the infrared laser is its larger spot size allowing more rapid coverage of the lesion and offering less opportunity to deliver excessive concentrated energy that might cause bleeding or tumor disruption. The end point of energy application is, like that for the argon laser, a gentle whitening of a spot placed half on and half off the tumor. Because of the larger spot size, the power is generally set initially at between 400 and 600 mW for 500 msec. The power can be adjusted upward to 700–1000 mW if required but almost never above that setting. The appearance of punctuate hemorrhages within the treated area indicates maximum energy is being delivered.

Complications of focal laser consolidation include burns of the iris at the pupillary margin and focal lens opacities, both of which are very rare in experienced hands. Other complications that are associated with excessive energy delivered to the tumor include subhyaloid and vitreous hemorrhage. Theoretically, it is possible to mechanically disrupt the tumor and create vitreous seeding of the tumor by using excessive energy (power × time) levels but that complication too has been a very rare event in patient care if the above cautions are exercised. In approximately 1000 lesions in more than 300 eyes treated in Los Angeles, we have seen tumor disruption by the laser on only one tumor. In that case, early in the series, treatment was done before chemotherapy was given and the power out of the laser was increased to approximately 900 mW. Vitreous hemorrhage and tumor seeding resulted in loss of the eye.

We are also aware of one case in which repeated laser photocoagulation delivered to multiple recurrences of a lesion abutting the optic nerve was associated with the appearance of an extrascleral nodule of retinoblastoma in the orbit. It is likely that this case of extraocular disease was a complication of repeated laser applications adjacent to where the optic nerve exited the eye. This single case argues strongly for a limit to the number of treatment sessions with the same modality in the face of persistent tumor regrowth.

The most significant long-term complication of focal consolidation is due to decreased vision from RPE scar migration or “creep” in lesions near the foveola. Care must be taken when applying laser on the foveal side of a tumor near fixation. Lee and colleagues demonstrated an increase in the size of laser scars following red diode laser application.²⁷⁶ It is reasonable to consider close observation after sufficient primary chemotherapy of a small tumor located near the fovea until documented growth is seen. In some instances, central vision can be spared if regrowth does not occur. Tumors that exist in the maculopapillary bundle can be managed in this fashion, especially if the contralateral eye has been enucleated or has poor visual prognosis.

Focal laser treatment of the tumor and retina is associated with the locally increased production of growth factors. Rapid regrowth should signal resistance to heat and the emergence of a genetically altered, heat resistant tumor. Finally, it is not uncommon to see scleral excavation and thinning in areas of heavy, repeated laser photocoagulation.

Cryotherapy

Destruction of the tumor by cryotherapy results from disruption of cellular membranes following the freeze–thaw cycle. It can also have a local vaso-occlusive effect on the tumor and nearby retina/choroid. Cryotherapy is useful and can be used successfully in tumors up to 3.5 mm in diameter and 2.0 mm in thickness; more than one treatment is generally recommended.^277–279 The difficulty in using cryotherapy as local treatment for posterior pole tumors is that a surgical procedure is required to open the conjunctiva so that accurate placement of the probe can be obtained. In addition, because the probe tip cannot be visualized while the freezing is taking place, it is possible to freeze the optic nerve. An important consideration is that cryotherapy routinely destroys a great deal of normal retina surrounding the lesion, thereby increasing the visual deficit and the resulting chorioretinal scar.

The probe of the cryotherapy unit is used to localize the tumor and to elevate the tumor on the tip of the probe. Once the probe is directly beneath the tumor, freezing is begun, and the ice ball is maintained until it encompasses the entire tumor mass. Then the ice ball is allowed to thaw, and this freeze–thaw cycle is repeated for a total of three applications.

Complications of cryotherapy include vitreous hemorrhage, subretinal fluid, and retinal holes. Retinal detachment can result from a combination of atrophic retina and vitreous traction. Cryotherapy results in strong adhesions at the margins of scars.^117,280,281 Extensive cryotherapy can cause atrophy of the sclera, with formation of a pseudocoloboma of the sclera. We have observed the creation of retinal breaks and rhegmatogenous retinal detachment when the calcium portions of type I regression is included in the ice ball. The presence of retinal detachment in the region of proposed cryotherapy is a relative contraindication. Cryotherapy should not be combined with the administration of subtenon carboplatin because toxic levels of the drug may accumulate inside the eyes.

External beam radiotherapy (teletherapy)

External beam radiotherapy (EBR) was the treatment of choice as primary therapy of intraocular retinoblastoma for much of the 20th century. Verhoeff and Reese pioneered the method in the early 1900s. Retinoblastoma is undoubtedly a radiosensitive tumor. Tumor control rates and ocular salvage rates are relatively high.^282–284 However, with decades of use, one late effect of the treatment in heritable retinoblastoma, second malignant neoplasms (SMNs), caused particular alarm in retinoblastoma centers. As a result most centers now use EBR as a salvage technique if primary chemotherapy has failed or cases where extraocular spread has developed (cut end disease or eyes inadvertently biopsied).

Measures to concentrate the radiation dose in the vitreous and/or retina are of particular importance in order to avoid induction of a second malignancy or cataract formation and permit normal development of the facial bones and brain. Recent advances in treatment planning and delivery permit considerably better sparing of the normal tissues adjacent to the globe than in the past. Sedation of the patient is necessary to ensure immobilization of the child’s head during the daily delivery of radiotherapy. Propofol produces short-term, deep sedation. During the sedation, monitoring parameters include respiratory rate, blood pressure, heart rate, and oxygen saturation.

A custom-made head immobilization device is also required to reproduce the position of the patient’s head throughout the course of 15–25 daily treatments. A CT scan of the orbit using a slice separation of 1–3 mm, performed with the patient in the head immobilization device, is utilized for designing the size, shape, and angles of the beam(s) that will optimally irradiate the retina while minimizing radiation dose to the surrounding structures.

There are many different delivery protocols, as well as several different machines. Many centers use between 35 and 45 Gy given in a modified lateral approach. In some protocols, 15 Gy are given from an anterior approach. The remaining 30 Gy are delivered to the eye and anterior orbit with a lateral approach. The total dose of 45 Gy for primary therapy is frequently given as daily doses of 2 Gy or as alternate-day doses of 4 Gy. A wedge is used to block the posterior surface of the lens. Because of this, the anterior retina (especially the anterior nasal retina) frequently receives only approximately 30% of the total dose. The isodose curves show that, posterior to the equator, virtually 100% of the dose is delivered. This includes the optic nerve to about 10 mm. However, the isodose curves fall off rapidly anterior to the equator, precisely the site at which recurrence is often seen after primary radiotherapy. The approach of giving 33% of the dose through an anterior approach with a lens shield obviates some of the problems that occur when the minimal dose is given to the anterior retina through the lateral approach.

In a report from the Essen group on the prevention of tumor recurrence, primary teletherapy was most effective at high doses, with 49% of patients who received 40 Gy having recurrent disease compared with only 22% of those who received a dose of 50 Gy.²⁸³ A highly accurate alignment of the beam tended to decrease the incidence of recurrent tumors by half: 22% versus 48% recurrent tumors with conventional alignment made to the lateral orbital wall.

Two different techniques for EBR have been compared.²⁸⁴ In this study, lens-sparing electron beam approach was used between 1979 and 1984, and the modified lateral beam technique was used from 1984 to 1987. The latter approach was clearly superior for eyes with RE groups I–III disease, as measured by freedom from recurrent tumors and decreased frequency of enucleation. No apparent difference between methods was noted in groups IV and V retinoblastoma.

A problem that has always been difficult to manage is the eye movement that occurs even under general anesthesia. Schipper^285–287 in Utrecht has been a pioneer in devising a method to fixate both the eye and the head for the duration of the treatment. Variations on the Utrecht method have been developed at St Bartholomew’s Hospital in London by Harnett et al.^288,289 Both groups use either a 6- or 8-mV linear accelerator with a lateral D-shaped field of 26–32 mm. The D-shaped field is contoured to include the entire retinal surface, the vitreous, and 10 mm of the anterior optic nerve. The eye is fixed to the beam-defining collimator by using a low-vacuum contact lens, which is held in place by a magnetic millimeter scale. Because the eye is fixed in the isocenter of the accelerator, rotation of the gantry directs the beam. In the series reported by Schipper et al.,²⁸⁷ 18 of 54 radiated eyes developed clinically detectable cataract, but only five of the 18 eyes so affected required cataract surgery. Lens opacity developed only in lenses in which a posterior portion of >1 mm had to be included in the treatment field. The minimum cataractogenic dose in that series was 8 Gy to the lens. Hungerford et al. noted that, by using the Utrecht device (the contact lens) for fixating the eye, the lens and anterior segment can be effectively spared while the whole retina is treated with a full dose of radiation.²⁹⁰

Intensity modulated radiation therapy

Intensity modulated radiation therapy (IMRT) is an alternate method of delivering radiotherapy to the eye. It is effective in reducing the dose to normal tissues and is available at most large radiation treatment centers. IMRT uses computer controlled tungsten blades selectively to block areas of the treatment volume for a prescribed fraction of each beam’s treatment time to make a radiation field with varying intensities. Figure 128.17 is a three-dimensional reconstruction of a CT scan done on a retinoblastoma patient with his head positioned using a custom-molded headrest and a custom dental impression and custom bolus material over his retinoblastoma-containing right globe. The green arrows represent the central rays of the eight incident IMRT radiation beams converging on the patient’s right retina. Figure 128.18 compares the radiation dose distributions produced by irradiating the entire globe using the 8-beam IMRT arrangement (A) and a single 20 million volt electron beam (B). The IMRT substantially reduces the dose to the frontal lobe posterior to the orbit. Figure 128.19 compares the radiation dose distributions produced by irradiating the retina using the 8-beam IMRT arrangement (A) and a single lateral 6 million volt photon beam. Both methods limit the radiation dose to the lens of the eye. The lateral photon beam, however, gives a substantially higher dose to the sphenoid sinus and contralateral orbit.

Fig. 128.17 Three-dimensional reconstruction of a CT scan. The green arrows represent the central rays of the eight incident intensity-modulated radiation therapy beams converging on the patient’s right retina.

Fig. 128.18 Radiation dose distributions produced by irradiating the entire globe using the 8-beam intensity-modulated radiation therapy (IMRT) arrangement (A) and a single 20 million volt electron beam (B). The IMRT substantially reduces the dose to the frontal lobe posterior to the orbit.

Fig. 128.19 Radiation dose distributions produced by irradiating the retina using (A) the 8-beam intensity-modulated radiation therapy arrangement and (B) a single lateral 6 million volt photon beam. Both methods limit the radiation dose to the lens of the eye.

Proton beam radiotherapy

This form of teletherapy uses charged particles with a very sharp Bragg peak curve. This results in dosimetric plans that are much tighter with less scatter to critical adjacent structures including the orbital bones, CNS and pineal region. Recent studies have demonstrated the dosimetric superiority of protons over more traditional EBR and even IMR.²⁹¹ The challenge remains the high cost of such equipment and the relatively few centers capable of administering fractionated proton beam treatments under anesthesia to the pediatric population.

Brachytherapy

The use of local radioactivity in the treatment of retinoblastoma was originally pioneered by Moore and Scott in 1929. Interstitial implants were used initially, and modifications of this technique were reported later. Ophthalmic applications were designed using radium and, later, cobalt 60. In 1977, Rosengren and Tengroth reported on the reasonably successful results of such treatment in 20 patients.²⁹²

Primary brachytherapy may be the treatment of choice in isolated group B intraocular retinoblastoma located at or anterior to the equator. The use of radioactive plaques in the primary and secondary management of retinoblastoma has increased as the danger of external beam therapy in genetically predisposed patients has become recognized. The results of plaque therapy in 50 selected patients have been reported.²⁹³ Of 51 eyes treated, vitreous seeding was evident in 49. Only two of 18 eyes were salvaged with radioactive plaque after everything else had failed. Thirty-three patients with small tumors did well, with most retaining vision. The difficulty with plaque therapy for posterior pole tumors is the likelihood of permanent visual loss associated with the massive doses of radiation to the retina and other vital structures. However, radioactive plaques have distinct advantages over EBR in cases in which the area of treatment is limited to the tumor and a small marginal rim.

Iodine125 (125I) isotope became a common radiation source used in brachytherapy when cobalt was abandoned decades ago. The seeds are secured in a gold carrier. Gold prevents radiation from penetrating the substance of the plaque and shields normal bone and tissue from most of the radiation. Luxton and colleagues worked out much of the dosimetry planning for the 125I isotope.²⁹⁴ Dosimetry planning is carried out with the help of sophisticated software, such as that available from Bebig Corporation.²⁹⁵ The gold carrier for the iodine was later modified with deeper wells for the seeds creating a conformal treatment plan (Fig. 128.20).²⁹⁶ In primary brachytherapy, the calculated dose to the apex of the tumor as measured by the peak height on ultrasonography is generally in the range of 40–45 Gy.

Fig. 128.20 (A) Gold carrier for the iodine was modified with deeper wells for the seeds creating a conformal treatment. The seeds can be loaded specifically depending on location of tumor. (B) The ruthenium plaque itself contains the radiation source. Therefore the possibility of differentially loading radiation seeds in the plaque to conform to the shape of the tumor is avoided with ruthenium.

In addition to 125I, the isotope ruthenium106 (a β-emitter), is commonly used in Europe. The advantage of ruthenium is that the half-life is much longer than iodine so that a single plaque may be reused for up to 1 year. There are two major disadvantages. As a β-emitter the scleral dose of a ruthenium plaque is higher than a similar plan with 125I; as a result, retinoblastoma lesions higher than 5 mm cannot be treated easily. Also, in ruthenium plaques, the plaque itself contains the radiation sources. Therefore, the possibility of differentially loading radiation seeds in the plaque to conform to the shape of the tumor is not possible. However, ruthenium plaques are available in numerous shapes and sizes to address this issue.

The technique of localizing the plaque when treating retinoblastoma differs from the localization technique for choroidal melanoma, primarily because of absence of pigment in the intraocular tumor. Transillumination techniques used with melanomas are not helpful in localizing retinoblastoma. The surgical approach is otherwise identical. The computerized treatment plan should be available to the surgeon. With the Bebig Corp. Software, the radiation physicist can tell the surgeon the position in clock-hours for the center of the anterior edge of the plaque, the location of the eyelets, and how many millimeters those points are located posterior to the limbus (Fig. 128.21). After the 12, 3, 6, and 9 o’clock locations are marked at the limbus, a 360° peritomy is performed, and traction is placed on appropriate rectus muscles by 4–0 silk sutures passed behind the muscles without needles.

Fig. 128.21 Dosimetry planning is carried out with the help of sophisticated software, such as that available from Bebig Corporation. The position in clock-hours for the center of the anterior edge of the plaque, the location of the eyelets, and how many millimeters those points are located posterior to the limbus are shown.

A generous margin of at least 2 mm is incorporated into the treatment plan. In the computer-generated model, the exact meridian and the distance in millimeters to the center of the anterior edge of the plaque are available to the surgeon. Once the tumor is localized, a “cold” plaque (not containing radioactive seeds) is secured with temporary nonabsorbable sutures and its location verified by scleral depression along the edges. Once the correct position of the plaque is verified, the “cold” plaque is replaced by the “hot” plaque containing the radioactive seeds. The conjunctiva is closed with interrupted absorbable sutures. A sheet of th-inch lead is fashioned as a shield to be worn during the postoperative course for iodine plaques. Because ruthenium is a beta emitter, the lead is not necessary.

In California, state law allows for outpatient brachytherapy, however, in many states the law prohibits individuals with implanted radioactive sources from leaving the hospital because of the concern of radioactivity in the environment. Removal of the plaque is scheduled at the completion of the required hours of the brachytherapy. The time the plaque is in place is determined by the activity of the radioactive seeds and the total dose to be delivered.

Care must be taken when combining external beam radiotherapy and brachytherapy sources. This may result in serious complications that can result in destruction of the eye.²⁹⁷ Use of 125I brachytherapy for consolidation immediately following primary systemic chemotherapy is associated with an extremely high risk of aggressive radiation retinopathy that can destroy the eye if traditional doses for isolated brachytherapy are used. We experienced significant rapid radiation retinopathy in 6/6 eyes when 40 Gy was given to the apex of the regressed tumor immediately following completion of CEV chemotherapy. A dose reduction from 40–45 Gy to 20–25 Gy would be appropriate. Perhaps a better alternative is to reserve brachytherapy for treatment of recurrent focal disease several months following the completion of chemotherapy.

Surgical technique

Before prepping for surgery, the tumor should always be visualized through a dilated pupil with the indirect ophthalmoscope. Prior to performing a limbal peritomy, we dry the conjunctival surface thoroughly and mark 4–5 mm of limbal conjunctiva with a marking pen. Once the peritomy has been made, the purple ink marks the conjunctival edge allowing a precise, quick closure at the end of the procedure and reduces the possibility of conjunctival inclusion cysts in the postoperative period. A curved Stephens scissors is used to spread Tenon’s widely in each quadrant. Vortex veins exit the sclera about 16–18 mm posterior to the limbus so there is little risk of inducing bleeding with deep spreading at this point we irrigate 2 cc of local anesthesia into the retrobulbar space.

The rectus muscles are then isolated and imbricated in the usual manner with exception of the lateral rectus muscle which we imbricate 8 mm posterior to its insertion. This leaves a stump of muscle, which when grasped at its insertion with a wide Adair clamp, allows for control of the eye during the severing of the optic nerve. It is not safe to pass large needles through the rectus insertions for securing the eye. The inferior oblique is easily obtained where it runs in the lower lid while the inferior rectus or lateral rectus is on a Jamison hook and the eye is turned up and in. Interruption of the inferior oblique with hot-temp cautery usually controls postoperative bleeding.

Prior to cutting the optic nerve, we irrigate the remaining 6–8 cc of a 1 : 1 mixture of 1% lidocaine with epinephrine and 0.5% bupivacaine with epinephrine into the retrobulbar space using a blunt 20-gauge irrigating cannula passed along the sclera. The introduction of this much fluid into the posterior orbit causes proptosis and temporary high intraocular pressure. After 8–10 minutes, the intraocular pressure has returned to normal. The introduction of the local anesthesia deep into the orbit has two beneficial effects: enhancing control of hemostasis and providing 3–4 hours of postoperative pain control. Allowing a full 10 minutes to pass between irrigation of the local anesthesia and cutting the nerve effectively controls bleeding from the severed ophthalmic artery.

Obtaining the longest section of optic nerve is important and can be facilitated using straight mayo or Metzenbaum scissors passed along the nasal wall of the orbit. The optic nerve follows a temporal to nasal route as it plunges to the orbital apex. The tips of the scissors should be felt to plunge through posterior Tenon’s capsule. It is counterintuitive to keep the tips of the scissors pressed nasally but with this technique optic nerve lengths of 12–15 mm can be obtained routinely. A snare, which crushes the optic nerve, should be used in enucleation of eyes with retinoblastoma only with the approval of the ocular pathologist. Likewise, clamping the optic nerve is unwise and should be avoided for similar reasons. Immediately after removing the eye we use a test-tube filled with a frozen-slush saline solution to tamponade the orbital apex. The ice-filled test-tube is placed into the orbital apex and held firmly for 10 minutes (Fig. 128.22). The combination of infusing the orbit with epinephrine-enhanced local anesthetic and using the ice-filled test tube as a tamponade results in little swelling or bruising, eliminates the need for a pressure patch and allows us to remove the patch the morning following outpatient surgery.

Fig. 128.22 (A) Test-tube with slushed ice is held in the orbital apex after cutting the optic nerve to aid in hemostasis. (B) Medpor implant with pre-drilled holes for muscle attachment. Marking the holes with a pen improves visualization. (C) Bisected enucleated globe filled with retinoblastoma. Tumor from one calotte can be taken for DNA diagnosis studies. (D) Hemostats are used to grasp the edge of the wound to open the posterior orbital cavity. This helps greatly with visualization during closure.

Our experience supports the recommendations of most ocularists is that the implant be as large and as anterior as possible. Thin prostheses allow the best movement. The long-term complications reported in association with pegging the hydroxyapatite implant have convinced us to abandon its use. We currently use an integrated pre-drilled, nonwrapped conical modification of the Medpor implant described in 1994, by Karesh and Dresner.²⁹⁸ Medpor orbital implants are available in three sizes: a 16-mm cone with an 18-mm volume; an 18-mm cone with a 20-mm volume, and a 20-mm cone with a 22-mm volume. We use a 16-mm cone only for infants and very young children. An 18-mm conical implant works well for most enucleations in children 6 months and older. The 20-mm conical implant is reserved for children older than 2.5 years. The material is well tolerated, resists infection, is nonantigenic, and promotes rapid tissue ingrowth because of it porosity. We prefer the conical shape which provides an additional volume posteriorly. The smooth, porous anterior surface of this particular implant is designed to minimize the possibility of late exposure. Many centers prefer hydroxyapatite. The only significant difference with this implant is that is requires a wrap to cover the surface. Sclera is used by many centers but others have abandoned this material in preference for the prewrapped or polymer coated variety.²⁹⁹

Both Medpor and coated hydroxyapatite implants have predrilled suture tunnels which allow for direct attachment of the rectus muscles. The preformed suture tunnels can be difficult to see if not highlighted with a marking pen (Fig. 128.22). Bending the needle to increase its curvature and then passing the hub through the tunnel first, makes the process quick and easy. Recently, we have been carefully dissecting extraocular fat and connective tissue from the outside of the removed eye and reinserting that fat over the surface of the implant after the rectus muscles have been attached. This tissue “cushion” prevents suture Knots from being the source of implant exposure. We have not seen an expose implant since introducing the implant “cushion.”

Tumor harvesting

Tumor harvesting requires good communication between the surgeon and pathologist and a consent signed by the parents for use of excess tumor tissue. We have performed the following procedure successfully for more than 25 years. Immediately after enucleation, the globe is placed on a sterile Mayo tray along with a small pair of tissue forceps, a curved Stevens scissors, a single-edge 2-inch razor blade, a caliper, and a millimeter rule. Two small Petri dishes and two specimen jars with formalin should be available in the operating room.

The enucleated specimen is examined carefully by the surgeon, ocular pathologist, or both looking for evidence of extraocular tumor. If no extraocular tumor is seen on gross examination and on preoperative neuroimaging, the retrobulbar connective tissue and adherent fat can be bluntly dissected from the attached optic nerve stump, placed on a saline soaked gauze pad, and returned to the surgical field for later use as filler between Tenon’s and the anterior implant surface as described above. After the connective tissue has been dissected free, the globe and the attached optic nerve can be photographed. The length of optic nerve obtained is measured with the caliper or millimeter rule. The nerve is then cut leaving a 1-mm stump on the back of the globe. The severed nerve is placed alone in one of the containers of fixative. The globe is stabilized with the fingers and bisected through the cornea with a sterile, single-edge 2-inch razor blade so that half the tumor mass and the entire optic nerve stump remain in one calotte. Photographs of the bisected globe are then taken (Fig. 128.22). The calotte containing the optic nerve stump and half the tumor mass is placed in the second container of formalin without being disturbed. The tumor material in the remaining calotte can be placed into a small Petri dish to be hand-collected as soon after harvesting as possible for tumor karyotyping and/or DNA studies as provided in the consent. Viable tumor is relatively firm and gray-white and can be lifted with forceps. Necrotic tumor is liquid, often contains calcium grains, and is usually yellow-white.

If most of the tumor material is inadvertently left in the calotte to be submitted for pathologic study, forceps and scissors can be used to remove a sufficient volume of that portion of the tumor farthest from the choroid. The viable tumor specimen should reach the laboratory as soon as possible. If the specimen is to be mailed to a distant laboratory, specific instructions should be obtained from that laboratory.

Once the surgeon has completed opening the globe and handling the tumor specimen, his/her surgical gloves should be changed before returning to the operating table for the closure. Instruments used to handle the tumor or to open the globe should never be returned to the operative field.

Surgical closure

We have found it very useful to grasp the edge of the opened conjunctiva, which is now easily identified by the purple ink from the marking pen, with five or six hemostats. This elevates the edge of the wound and opens the posterior orbital cavity for easy visualization (Fig. 128.22). After the implant has been inserted and the muscles attached to the pre-drilled tracts as described above closure can begin. During closure of the Tenon’s fascia layer it is helpful if Tenon’s is hydrated with copious saline irrigation. We use a lubricated 5–0 Vicryl suture on a P-3 needle to close Tenon’s fascia with 3–5 horizontal mattress sutures passed 3–4 mm posterior to the conjunctival edge. Alternatively, this step can be closed with a pursestring suture. Before or during Tenon closure, the fat and connective tissue that was previously harvested from the enucleated eye, is placed on top of the implant and spread evenly. A second row of reinforcing vertical interrupted sutures should be placed to provide strength to the closure. This double suturing of Tenon’s fascia provides superior strength to the wound and significantly decreases the risk of exposure and extrusion. The wound should be meticulously inspected to ensure that there are no holes or gaps in the Tenon’s capsule. The conjunctiva is then closed with running 6–0 plain gut or Vicryl suture.

At the end of the procedure, a generous amount of steroid–antibiotic ointment is placed over the closure site, and a hard plastic conformer is inserted between the lids. If the lids close completely over the conformer immediately after insertion, then a larger conformer is necessary. Ideally, the lids should close only halfway over a correctly fitting conformer. Once the surgical edema has subsided, the lids will close without difficulty. In Los Angeles, the patch remains in place for only 1 day, and a true pressure patch is not needed. We routinely discharge the child on the day of surgery and remove the patch the next day when the child returns to the office. Performing the procedure on an outpatient basis is acceptable if postoperative observation for 1–2 hours is possible and if the child can return the following day for dressing removal. Systemic prophylactic antibiotics are not routinely necessary but some centers advocate their use. If subsequent chemotherapy is necessary, at least 1 week should pass before beginning treatment.

Postoperative care following enucleation

The immediate and short-term complications of enucleation for retinoblastoma include inadvertently opening the eye during the process of cutting the optic nerve, bruising or cutting the nerve to the levator muscle, cutting across tumor in the optic nerve because of obtaining a short section of optic nerve, and postoperative nausea and vomiting. Long-term complications include extrusion of the implant (often due to inadequate surgical closure), conjunctival cysts from inadequate surgical closure, and an insufficiently large implant giving poor cosmetic outcome.

Orbital growth following enucleation will be normal if a large orbital implant is placed in the orbit at the time of surgery and no orbital radiation from an external source is given.³⁰⁰ However, failure to replace the eye with an orbital implant and failure to maintain the presence of an ocular prosthesis will result in orbital growth deficiency.³⁰¹ Presumably, the severe bony hypoplasia associated with EBR would not be a cosmetic problem in these cases. Imhoff and colleagues demonstrated that the growth of irradiated orbits was significantly impaired (P<0.001) when compared with nonirradiated orbits and that secondary enucleation did not add to the bony growth retardation triggered by external beam radiotherapy. Not surprisingly, these authors observe the growth-impairing effect of EBR to be most profound when the child is irradiated before 6 months of age (P<0.01).³⁰² When reports assume that both enucleation and radiation contribute to retardation of orbital growth and do not differentiate between the two, confusion is the outcome.³⁰³ Orbital growth studies in rabbits show a decelerated increase in orbital mass following enucleation that was mitigated by an expandable but not static orbital implant.³⁰⁴ Based on the findings of Fountain et al. growth in enucleated human orbits may be normal if a large but nonexpandable orbital implant is used.³⁰⁰

Follow-up after enucleation varies from center to center. We do not currently order MRI scans more than once yearly except when heritable disease is diagnosed before 12 months of age. In that case, we do MRI every 6 months until age 3, looking for possible midline PNET. The generally recognized risk period for extraocular spread after successful treatment or enucleation is 12–18 months.

Retrolaminar optic nerve involvement

If a pathologic evaluation demonstrates optic nerve involvement beyond the lamina cribrosa and massive choroidal invasion, 6 months of adjuvant chemotherapy is recommended by some centers but not all. Chantada and colleagues question the need for this adjuvant therapy.²⁰⁷ The initial series reporting the risk of retrolaminar optic nerve involvement came from New York in 1989.³⁰⁵ It is instructive that currently in New York, this group is not treating retrolaminar retinoblastoma with adjuvant chemotherapy.

The chemotherapy protocol commonly used for adjuvant therapy is the combination of carboplatin, etoposide, and vincristine given at 3-week intervals for six cycles. The Children’s Oncology Group has now completed accrual in a trial assessing the risks and benefits associated with administering adjuvant chemotherapy in eye with so called high risk histopathologic features (see below).³⁰⁶