SIOP-PODC recommendations for graduated-intensity treatment of retinoblastoma in developing countries


  • Conflict of interest: Nothing to declare.


Retinoblastoma remains incurable in many regions of the world. The major obstacles to cure are delayed diagnosis, poor treatment compliance, and lack of evidence-based recommendations for clinical management. Although enucleation is curative for intraocular disease, in developing countries retinoblastoma is often diagnosed after the disease has disseminated beyond the eye. A SIOP-PODC committee generated guidelines for the clinical management of retinoblastoma in developing countries and developed a classification system based on the resources available in those settings. Recommendations are provided for staging and treatment of unilateral and bilateral retinoblastoma and counseling of families for whom compliance is an issue. Pediatr Blood Cancer 2013; 60: 719–727. © 2013 Wiley Periodicals, Inc.


Retinoblastoma is highly curable in developed countries; however, most children with retinoblastoma in developing countries die as a result of late diagnosis and poor treatment compliance, which leads to extraocular dissemination and metastasis 1. Since it is estimated that about 2/3 of the population in the pediatric age live in developing countries, there may be more children dying of retinoblastoma than those surviving worldwide 2. In developed countries, patients with unilateral disease can be cured by enucleation of the affected eye, whereas those with bilateral disease undergo eye-conserving therapies in at least on one eye. Enucleation is a relatively simple surgical procedure that requires no special facility or equipment and is potentially available in centers with limited resources 3. However, even children who could be cured by enucleation, present with substantial challenges in developing countries. Their families do not always approve of enucleation 4, and the prevalence of patients presenting with invasion to critical eye structures (i.e., postlaminar optic nerve, sclera, or choroid) is high 5, so adjuvant therapy after enucleation is usually needed to prevent extraocular relapse.

Accurate risk assignment of patients with retinoblastoma requires standardized, expert evaluation of the pathologic features of the enucleated eye. This expertise is seldom available in less-developed countries. Thus, treatment decisions may be inaccurate in that setting. In addition, patients who present with overt extraocular disease are not candidates for enucleation, and the likelihood of cure is only realistic in those cases in which the tumor has disseminated to the orbit only. Children with metastatic retinoblastoma are seldom cured by conventional therapy 6.

Controversies are exist for conservative therapies. In developed countries, external-beam radiotherapy (EBRT) was the classic, initial effective conservative therapy for retinoblastoma. The use of radiation is associated with secondary malignancies, which are often fatal 7. Therefore, EBRT has been virtually replaced by chemotherapy combined with local therapies (e.g., cryotherapy, laser therapy, etc.) aimed at preventing secondary malignancies. 7 Unfortunately, local treatments require a sophisticated setting and intense use of qualified staff members who may not be available in many developing countries 8. In addition, localized approaches are less effective for treating more advanced disease. In such cases, radiotherapy is still needed, and enucleation may be ultimately required 9. Local therapies have not increased the risk of extraocular dissemination in developed countries when used by experienced teams 10, but the situation may be different in developing countries. Conservative therapy is simply not an option in many countries because advanced disease is present at diagnosis, the technology and/or agents are not available, or follow-up is inadequate.


The literature on managing extraocular retinoblastoma is limited, so the evidence is weak in comparison to information available on treating other pediatric malignancies. For this report we used the National Cancer Institute levels of evidence 11. We selected references from each setting and extrapolated those from other settings when no available reference was found. Thus, in cases with no published evidence, we assigned a level of evidence 4, based on recommendations of experts of this committee.

To develop these guidelines, an international group of retinoblastoma experts held teleconferences using the website. Because retinoblastoma is associated with unique clinical management challenges in developing countries, the committee's proposal was presented at the regional SIOP Africa and Asia meetings in 2012. Issues were further discussed with local healthcare professionals and representatives from parental groups in attendance; treating children whose families do not consent to therapy was a main focus of those discussions. This consensus document includes representative cooperative groups (such as AHOPCA from Central America, RetMex from Mexico, and GALOP from South America) and international organizations (e.g., INCTR) that lead retinoblastoma protocols or programs.

As presented by Hunger and Howard 12 for the management of acute lymphoblastic leukemia, we propose recommendations for three settings with different resource availability for the management of retinoblastoma (Table I). Countries of low income (setting 1) have the fewest available resources and minimal technology (e.g., low-dose chemotherapy and minimal imaging, ophthalmologic, and pathology services). Lower-middle income (setting 2) countries have some basic resources (e.g., moderate-dose chemotherapy, some ophthalmological therapies). Upper-middle income (setting 3) countries have more modern resources and technologies (e.g., MRI, high-dose chemotherapy, highly specialized ophthalmological, and pathology services), but the availability of these features may be limited to larger centers. Some retinoblastoma-specific issues must also be considered, such as the type of facility where treatment is delivered (eye hospital, pediatric hospital, or cancer center), as this variable also influences the final recommendation (Table II).

Table I. Proposed System for Categorizing Retinoblastoma Treatment Settings in Developing Countries
Resource availabilitySetting 1 (low-income)Setting 2 (lower–middle income)Setting 3 (upper–middle income)
  • CT, computed tomography; EBRT, external-beam radiation therapy; MRI, magnetic resonance imaging; RT, radiotherapy.

  • a

    Limited availability.

ImagingNot available or CT onlyCT and occasionally MRIMRI
Oncology treatmentLow-dose chemotherapy, Cobalt RTModerate-dose chemotherapy, Cobalt RT linear accelerator RTaModerate- and high-dose chemotherapy autologous hematopoietic stem cell rescue linear accelerator and 3D RT
Ophthalmologic treatmentMinimalCryotherapy, laser therapy, EBRTCryotherapy, laser-TTT, plaque oblation, EBRT (3D, IMRT), digital camera for funduscopy, Intra-arterial chemotherapya
Pathology assessmentMinimal, low specialty, diagnosis confirmationLow specialty, limited risk assessmentHigh-quality specialty, accurate risk assessment
Genetic testingNot availableNot availableLimited availability, usually low resolution
Criteria for reclassification of settingOphthalmologist with training in conservative therapyIncreased availability of localized therapies (laser-TTT, cryotherapy, and plaque ablation) 
 Availability of laser therapy, cryotherapy, and RT<5% mortality related to toxicity of intensive chemotherapy regimen 
 Safe pediatric anesthesiologySecond-line therapies for intraocular and extraocular relapse 
 <5% mortality rate related to toxicity of chemotherapyHighly specialized pathologic assessment capable of accurately stratifying patients to treatment 
 Pathology assessment capable of accurate disease staging  
Table II. Resources Typically Available at Various Centers in Developing Countries
ResourceEye hospitalChildren's hospitalCancer center
  • CT, computed tomography; ICU, intensive care unit; MRI, magnetic resonance imaging.

  • a

    Support services include psychosocial services, parental groups, etc.

Chemotherapy supportBelow standardStandard, pediatric ICUStandard, pediatric ICU may not be available
Local ophthalmic therapyUsually goodVariable, limited availabilityVariable, limited availability
RadiotherapyNot availableVariable, limited availabilityAvailable, but pediatric expertise may be limited
Palliative careNot availableUsually availableAvailable, but pediatric expertise may be limited
PathologyHighly specializedVariable, limited availabilityLow specialty
Support servicesaNot availableUsually availableAvailable, but pediatric expertise may be limited
On-site imaging facilitiesUltrasoundUltrasound, CT scans, MRIUltrasound, CT scans, MRI
Pediatric anesthesiaNot always availableAlways availableNot always available

In a given country, a center may have features of more than one setting. Because retinoblastoma is a rare malignancy, the care of patients with this disease should be centralized to referral centers that treat at least 5–10 patients with newly diagnosed disease per year 13. Care provided by multiple specialists, including ophthalmologists, pediatric oncologists, anesthesiologists, psychologists, nurses, social workers, pathologists, radiation oncologists, imaging specialists, and others, working as a multidisciplinary team is essential for achieving optimal results in retinoblastoma 14. The primary healthcare provider for children with retinoblastoma is typically an ophthalmologist. Therefore, a close relationship between the ophthalmologist and pediatric oncology team is essential.


The availability of services in developing countries guides the choice of treatment, so it is essential to identify the local facilities in which to either establish a new retinoblastoma program or optimize an existing one (Table II). Minimal requirements to for curative treatment of retinoblastoma include a dedicated ophthalmologist with adequate skill in eye enucleation in young children, the presence of a general pediatrician or pediatric oncologist and nursing staff familiar with chemotherapy administration and management of its side effect, supportive and palliative care, safe pediatric anesthesiology, availability of essential chemotherapy agents (cyclophosphamide, vincristine and doxorubicin 15, 16, or preferably carboplatin and etoposide), and a twinning initiative or availability of expert consultation for case referral or consult.

Children's hospitals are probably the best choice for developing retinoblastoma programs in most settings 1 and 2, because they are likely to have a pediatric oncology service and palliative care support. Few patients are candidates for eye-conserving therapy, so eye hospitals are usually an inadequate setting as the sole resource for treating children with retinoblastoma in these settings. Building a relationship between a children's hospital and an eye center may enhance treatment options for the few children who are candidates for eye salvage. In most instances of setting 3, children with retinoblastoma may be appropriately treated in any type of center, provided that adequate resources and personnel are available.

The availability of expert pathologists who have the skills to comprehensively assess eye tumors is usually lacking in children's hospitals in settings 1 and 2, and treatment tailored to risk factors associated with pathologic findings may be misleading 17. A recent report from the Children's Oncology Group showed that even in a high-income country, treating institutions misclassified the risk group assessment in 16% of cases 18. Training specialized pathologists would result in a more accurate use of adjuvant chemotherapy and it is, therefore, a priority in these settings 19. Radiotherapy may still play an important role in the treatment regimen of children with retinoblastoma in settings 1 and 2; however, its availability is usually limited to adult cancers, and cobalt is often the only radiation available. In some settings, radiotherapy is not available for children with retinoblastoma.

Although their value has not been unequivocally established, retinoblastoma awareness campaigns directed to the public and doctors may be important in settings 1 and 2, where metastatic disease is present in a high proportion of children at diagnosis 20. The impact of such campaigns in setting 3 may be lower, because most patients in that setting present with intraocular disease. In all settings, awareness campaigns should target familial cases, because in some middle-income countries, as many as 75% of familial retinoblastoma cases are not screened 21.


A disease-staging system for children with retinoblastoma is essential for the initial evaluation of the extent of extraretinal dissemination and prediction of survival. The International Retinoblastoma Staging System (IRSS) is easily applicable 22. Ideally, all patients should undergo a complete ophthalmological examination under anesthesia, including tonometry and slit-lamp examination, by an experienced ophthalmologist using indirect ophthalmoscopy. In programs where conservative therapy is undertaken with chemoreduction and localized therapy, a digital camera for documenting the fundoscopic findings may be helpful. Each eye should be assessed according to the International Classification of Intraocular Retinoblastoma 23 or others. At least a head and orbital contrast-enhanced computed tomography (CT) scan should be done, but magnetic resonance imaging (MRI) is preferable, if available, because MRI provides a more accurate imaging of the optic nerve extension.

Although the use of routine examinations of the CSF and bone marrow was questioned in publications from more developed countries 24, in developing countries with higher prevalence of extraocular disease, these procedures should be done more frequently 25, especially to distinguish between stage III (locoregional dissemination and still curable) and stage IV (metastatic dissemination and incurable when high dose therapies are not available) retinoblastoma. A lumbar puncture with examination of the cytocentrifugate and an extensive bone marrow evaluation, preferably including at least two sites, of both the aspirate for cytology and biopsies, should be done in all patients with stage 2 or more progressed disease 25. This procedure needs to be done under general anesthesia, which is not always available or recommended in children with advanced disease in low-income countries. Alternatively, a single bone marrow aspiration could be done, and if the results are positive, no other bone marrow study is needed. However, when a single aspiration fails to show malignant cells, a more exhaustive bone marrow evaluation should be done. This especially important when it is necessary to discriminate between stages III and IV, because children with stage IV disease benefit from more intensive chemotherapy with autologous stem cell rescue 26. On the other hand, in setting 2, where some children with stage III retinoblastoma are curable, it may be important to identify those with metastatic disease who would not be cured by lower-intensity therapy. The value of these examinations in cases of stage I disease, even for those with pathologic risk factors, is debatable.


Overt extraocular retinoblastoma, regardless of the laterality, is classified as IRSS stages III or IV. Children with overt extraocular retinoblastoma usually present with severe pain caused by an orbital mass 27. They are frequently emaciated, and their quality of life significantly improves with chemotherapy and supportive measures, but intensive regimens are usually not tolerable.

Retinoblastoma is a highly chemosensitive tumor that responds well to many low-cost chemotherapeutic agents, so they should be offered to all children. Standard-dose chemotherapy with an intention of life prolongation should be given to children with stage IV disease in settings where treatment with high-dose chemotherapy and autologous stem cell rescue are not available. High dose chemotherapy followed by autologous stem cell rescue is the only effective therapy for patients with stage IV extraocular retinoblastoma 28–30; the cure rate may be as high as 70% if there is no CNS involvement, but it is still lower than 30% in those with CNS involvement 31.

Chemotherapy options include the combination of cyclophosphamide,which may be administered orally, and vincristine or carboplatin and etoposide (Tables III and IV), which seldom cause severe toxicity. Intrathecal chemotherapy may be considered when leptomeningeal dissemination is present but not when contraindicated by a CNS mass. The evidence supporting the use of intrathecal chemotherapy, however, is limited 32, 33. The use of radiotherapy after the orbital or CNS disease has shrunk in response to chemotherapy may also improve the quality of life of these children.

Table III. Treatment Recommendations and Evidence Levels Determined by Parental Compliance and Setting
Disease classificationSetting 1Setting 2Setting 3
Intraocular unilateral retinoblastomaUpfront enucleation or alternatively, preoperative chemotherapy followed by enucleation in cases of treatment refusal (Regimen 1 or 2; level 4)

Upfront enucleation followed by adjuvant chemotherapy according to risk (Regimen 1; level 3iiA) 70

Upfront enucleation followed by adjuvant chemotherapy according to risk (Regimens 3–5; level 3iiA) 53, 71
 Adjuvant chemotherapy to all patients (Regimen 1 or 2; level 4)  
Orbital retinoblastoma

Preoperative chemotherapy followed by secondary enucleation. Adjuvant chemotherapy and radiotherapy (Regimen 1 or 2; level 3iiA) 27

Preoperative chemotherapy followed by secondary enucleation. Adjuvant chemotherapy and radiotherapy (Regimens 1 and 6; level 3iiA) 70

Preoperative chemotherapy followed by secondary enucleation. Adjuvant chemotherapy and radiotherapy (Regimens 3–5; level 3iiA) 53, 72
Metastatic retinoblastoma (unilateral or bilateral)Palliative care and low dose chemotherapy (Regimens 2 and 6; level 4)Palliative care and low dose chemotherapy (Regimens 2 and 6; level 4)

Neoadjuvant chemotherapy, followed by local control and consolidation with high-dose chemotherapy and autologous hematopoietic stem cell rescue (in responding patients; level 3iiA) 26

Conservative therapy for unilateral diseaseNot recommended (level 4)Only for selected cases of groups A–C eyes, where chemoreduction and local therapy can be safely provided (Regimen 1; level 4)

Only for selected cases of group A–C eyes where chemoreduction and local therapy (Regimen 1; level 3iiDi) 61

or intra-arterial chemotherapy can be provided (level 3iiDiii) 65, 66
Conservative therapy for bilateral retinoblastomaEnucleation of group D and E eyes (level 4)

Enucleation of group D–E eyes (level 3iiA) 3

Enucleation of group E eyes (level 3iiA) 3

 Conservative therapy is usually not possible and need to be considered only in cases of groups A–C eyes where follow-up could be ascertained (Regimen 1) EBRT (where available) where treatment compliance is poor (level 4)

For group A–C eyes, chemoreduction and local therapy (Regimen 1) where available and adequate follow-up should be ascertained (level 3iiDi) 70

For group A–C eyes, chemoreduction, and local therapy (Regimen 1; level 3iiDi) 60, 61
  Chemoreduction followed by EBRT (Regimen 1) where localized therapy is not available (level 4)For group D eyes, enucleation if contralateral eye is group A–C (level 3iiDi) 60, 61
  EBRT when treatment compliance is a problem (level 4)Chemoreduction and EBRT when needed if bilateral group D eyes (level 3iiDi) or if contralateral eye is group E 60, 61
Table IV. Published Chemotherapy Regimens and Their Indications
Chemotherapy regimenaPossible use (level of evidence of the recommendation)AdvantagesDisadvantagesRefs.
  • a

    Dose modification may be necessary for children weighing less than 10 to 12 kg. Proposed dosages: Vincristine (0.05 mg/kg, IV), doxorubicin (1–2 mg/kg per dose), carboplatin (16–18 mg/kg per dose), etoposide (3.3–5 mg/kg per dose);

  • b

    Cyclophosphamide (20 mg/kg orally at night, 2 h after meals) may be given as palliative therapy (level 4);

  • c

    May be replaced by doxorubicin (30 mg/m2 on Day 1).

1) CarboplatinChemoreduction (setting 2 and 3) (level 3iiDi)Low mortality related to toxicityMay be insufficient for adjuvant therapy in children with high-risk disease54, 55, 61
 (500–560 mg/m2 on Day 1) + EtoposideAdjuvant therapy (setting 1 and 2) (level 3iiA)High availability  
 (100–150 mg/m2 on Days 1–2) + VincristineNeoadjuvant therapy (setting 1) (level 3iiA)Low cost  
 (1.5 mg/m2 on Day 1 (a) Ambulatory  
2) CyclophosphamidePalliative therapy (level 4)bLow toxicityMay be insufficient as adjuvant therapy15, 16
 (40 mg/kg on Day 1) + VincristineAdjuvant therapy (if carboplatin is not available) (level 3iiA)Low costLong-term gonadal and cardiac toxicity 
 (1.5 mg/m2 on Day 1) ± Doxorubicin Good antitumor activity  
 (30 mg/m2 on Day 1) Ambulatory  
3) CarboplatinChemoreduction for advanced cases (setting 3) (level 4)Good CNS penetrationHighly myelotoxic


 (500 mg/m2 on Days 1–2) + EtoposideAdjuvant therapy (setting 3) (level 3iiA)Probably more effective as adjuvant therapy (associated with Regimen 4) in patients with high-risk disease (compared to Regimen 1)Superiority over lower doses is not known 
 (100 mg/m2 on Days 1–3)Neoadjuvant therapy (setting 3) (level 3iiA) High cumulative doses of etoposide 
 Treatment of metastatic disease (setting 3) (level 3iiA)   
4) CyclophosphamideAdjuvant therapy (setting 3) (level 3iiA)Good CNS penetrationThe use of anthracyclines has not been proven to improve results


 (65 mg/kg on Day 1) + VincristineNeoadjuvant therapy (setting 3) (level 3iiA)Not cross-resistant to carboplatin-based drugsGonadal toxicity 
 (1.5 mg/m2 on Day 1) + IdarubicincTreatment of metastatic disease (setting 3) (level 3iiA)   
 (10 mg/m2 on Day 1)    
5) IfosfamideAdjuvant therapy (setting 3) (level 3iiA)Good CNS penetrationIfosfamide-induced nephrotoxicity in small children37, 56
 (1.8 g/m2 on Days 1–5) + EtoposideNeoadjuvant therapy (setting 3) (level 3iiiDiv)Not cross-resistant combinationInpatient administration 
 (100 mg/m2 on Days 1–5)Treatment of metastatic disease (setting 3) (level 4) Highly myelotoxic 
 ±Carboplatin (400 mg/m2 on Days 1 and 2)  Gonadotoxic 
Intrathecal chemotherapy (cytarabine or topotecan)Palliative treatment of leptomeningeal dissemination. (level 4)bEasy to administerEfficacy not proven15, 32, 33
 Possible role as further prevention of CNS relapses when low-dose adjuvant therapy is given (level 4)Limited toxicity  
  Low cost  

Children with stage III retinoblastoma may be curable with intensive therapy 34, which is available in some centers in setting 2 and in all centers in setting 3, but the results in setting 1 are poorer 35. Upfront surgery should not be attempted in children with stage III disease. Orbital exenteration is usually not recommended but may be necessary in those with poor response to neoadjuvant chemotherapy. These patients should be treated aggressively with a curative intent using carboplatin-based regimens and orbital radiotherapy (Table IV). However, a subgroup of children with stage III disease and massive enlargement of the optic nerve do poorly with this approach 36.


Upfront enucleation is the treatment of choice for children with intraocular unilateral retinoblastoma. In developed countries, fewer than 1% of these children present with buphthalmia 37, and fewer than 20% present with significant risk factors upon pathology examination 38. Thus, in more than 95% of cases, enucleation results in complete removal of the tumor, and fewer than 5% have microscopically residual disease after enucleation.

In many countries classified as setting 1, as many as two thirds of children present with enlarged eyeballs, many of whom have microscopic extraocular dissemination 35, 39. Enlarged eyes may be difficult to enucleate and are at high risk of rupture 40, which would seed the tumor in the orbit, thereby theoretically increasing the risk of death and necessitating intensive chemotherapy and orbital radiotherapy. In addition, the tumor may be left behind in the resection margin of the optic nerve. Theoretically, pre-enucleation chemotherapy should reduce tumoral volume in severely buphthalmic eyes, thereby reducing the risk of eye rupture and tumoral residue at the optic nerve margin 40. This especially important in settings where no radiotherapy is available since children with this condition need it for tumor control 41.

Extraocular extension of disease may be difficult to assess when only low-resolution CT scans are available; gross invasion to the optic nerve or extrascleral invasion may be missed. Thus, clinical–pathological correlations may be important for initial management 42. In centers where pathology is poor or not available, older age at presentation, longer lag time from the onset of symptoms to diagnosis, presence of hyphema, pseudohypopyon, staphyloma, massive bupthalmia, and history of orbital cellulitis may provide a valuable indication for considering adjuvant chemotherapy in such cases after enucleation 42. In these children, preoperative chemotherapy to shrink the tumor may facilitate enucleation easier without tumor residue 40, 43. In these instances, these children should be considered at higher risk for extraocular relapse and adjuvant chemotherapy should always be used, regardless of the pathologic findings upon examination of the enucleated eye 44, 45. Enucleation should not be performed later than 2 or 3 chemotherapy cycles, because chemotherapy resistance may ensue, and the child may die of disseminated disease 46. Even when tumor response to neoadjuvant chemotherapy is spectacular, enucleation is still required.

The choice of chemotherapy regimen depends on the local availability of chemotherapy drugs and the supportive care facilities. Carboplatin-based regimens should be the first choice 47, but if this drug is not available, a regimen including cyclophosphamide and vincristine, with the possible addition of doxorubicin 16, may be an alternative (Table IV).

In cases in which parents consent to upfront enucleation and expert surgery and pathologic assessment are available, enucleation of the affected eye should be performed as soon as extraocular disease has been ruled out. Adjuvant therapy should be instituted after pathologic examination of the enucleated eye per international standards 48. Adjuvant chemotherapy is necessary for children with postlaminar optic nerve involvement 49, 50, with or without tumor in the resection margin, or any degree of scleral involvement 51. Its use in children with other risk factors and lower risk of extraocular relapse should be balanced with the potential toxicity of chemotherapy and the availability of second-line therapy in a given setting 52.

Children with isolated choroidal or anterior segment invasion have a low risk of relapse (<5%); thus, when treating these patients with adjuvant chemotherapy three issues should be considered: (1) The risk of toxicity-related death during a neutropenic episode or other toxic event may outweigh the benefit of adjuvant chemotherapy in children with low-risk disease, (2) high-quality pathology assessments are essential, and (3) high-dose chemotherapy and stem cell rescue must be available for treatment of relapse. Before withdrawing chemotherapy in children with isolated massive choroidal invasion, an experienced ocular pathologist following international standards should provide a full examination of the enucleated eyeball 48 to ensure that scleral or postlaminar optic nerve invasion, which would require adjuvant therapy, are not present. Because advanced pathologic assessments may not be available in many centers, using adjuvant chemotherapy to treat all children may be a safer approach. More intensive regimens may yield better results in children with high-risk disease 51, 53, but they may also be associated with increased risk of toxicity-related death in settings with limited resources. Thus, the benefit of preventing extraocular relapse may be outweighed by the risk of toxic death. Therefore, more intensive regimens should be used only in setting 3 and preferably under prospective protocols.


Conservative therapy is usually not a priority in setting 1, where most children die of extraocular retinoblastoma. Enucleation would cure a high proportion of children with bilateral retinoblastoma, so it is important that patients with intraocular disease not be exposed to treatments with conservative intent in a setting that has no facilities or experience in localized therapy. Chemoreduction followed by focal therapy to avoid EBRT, the standard conservative treatment in developed countries 54, 55, may not be feasible in developing countries, because most children there present with advanced disease requiring EBRT or enucleation. This treatment is particularly dangerous in settings with a high rate of abandonment of follow-up, because partially treated tumors may reactivate and disseminate 56, 57.

Telemedicine facilities are helpful during follow-up 58. Late relapses tend to occur 59, even in those whose disease was treated appropriately. Thus, patients lost to follow-up may be at a higher risk of mortality 56. Chemoreduction may be advantageous in cases in which pathologic assessment of the contralateral enucleated eye shows risk factors indicating the need for adjuvant chemotherapy. As a general rule, conservative therapy of Group D eyes should not be considered routinely in centers with limited resources in setting 2, because their preservation rate is low, especially if EBRT is not available. In most centers in setting 3 and some in setting 2, state-of-the-art conservative treatments are possible 60, 61. Centers of excellence have been created in many countries that are capable of providing state-of-the-art conservative treatment by adequately trained, experienced teams.

The only benefit of chemoreduction is avoiding or delaying EBRT, which is associated with 6–17% increased risk of mortality caused by radiation-induced second tumors during adulthood in developed countries 62, 63. To justify the use of chemoreduction over EBRT, toxic mortality associated with a regimen such as standard CEV (Regimen 1, Table IV) should be less than 1–2%. Avoidance of EBRT not only decreases the risk of secondary malignancies but also results in better cosmesis and lower prevalence of ocular side effects. However, in terms of ocular salvage, no benefit has been proven. Patients treated with EBRT need less-intensive follow-up and are likely to be cured with one 6-week course of radiotherapy, whereas children treated with chemoreduction and local therapy usually need a more intensive, longer follow-up to consolidate tumor response and treat later relapses. However, surgery for repairing radiation-induced cataracts that occur in almost all patients within a few years of irradiation should be available 64. Therefore, the availability of a high-quality EBRT facility is a priority in this scenario, especially in setting 2. Training of EBRT personnel is also a challenge in developing countries. In setting 3, most resources for localized therapy are available, but more cases with advanced disease are likely to be seen, so second-line therapy should be available.

Intra-arterial chemotherapy is widely used in developed countries 65 and has become gradually available in some developing countries (setting 3) 66, 67. This modality may be important for treating eyes with advanced disease or as secondary treatment, but it should be used with caution as initial treatment because of the higher prevalence of eyes with pathologic risk factors in this setting. Intra-arterial chemotherapy is usually not recommended for initial treatment of most cases of unilateral disease, which are best managed by enucleation in developing countries. Adjuvant therapy for enucleated eyes in cases of bilateral retinoblastoma should follow the same guidelines as those for cases of unilateral disease.


Treatment compliance is a substantial problem that occurs in many developing countries 68. This is critically important in cases of intraocular retinoblastoma requiring enucleation, because some patients could be cured by this simple surgical procedure. If left untreated, retinoblastoma is uniformly fatal. As many as one third of the patients in a series from Indonesia abandoned therapy temporarily 4; but this may occur in as many as 75% of cases in other settings 69. Some patients returned for medical care but only after the globe became grossly proptotic and metastatic disease occurred, at which time it was too late; fewer than 20% were cured 4. As many as 18% of patients never came back; they probably died of disease 4.

Centers where compliance is a substantial problem should establish a comprehensive program to approach these families. Poor compliance to therapy is associated with poor socioeconomic conditions, especially in settings where anticancer treatment is not free of charge 68. In other situations, cultural and/or religious reasons may limit the acceptance of enucleation. In large countries with geographic barriers, patients are usually admitted to hospitals in large cities, and their mothers are usually the sole caregiver. In some settings, mothers need the approval of the extended family for procedures such as enucleation. Actions to improve compliance should be actively pursued. Successful experiences have been reported in Central America 70, where families of children with high-risk disease are approached by a multidisciplinary team and given special support 70. In that setting, the rate of treatment refusal decreased from 21% of patients in 2000–2003 to 11% in 2004–2008 after an intervention program. Aspects of the care that likely led to decreased refusal rate included twinning with established centers of excellence via internet-based consultations, donations of equipment, and sponsoring the Central American ophthalmologists to attend international conferences as well as providing training locally 71. This program also provided prosthesis and other centers use a provisional prosthesis in the operating room after enucleation, so that the families tolerate the procedure better. Parental groups play a key role providing emotional support, contact with survivors who are living normal lives, and financial support are important to preventing treatment refusal.

Therefore, in addition to these general guidelines that are applicable to all settings, specific recommendations for the management of patients whose families refuse enucleation in each setting include:

Settings 1 and 2: Especially in centers where expert pathology examination is not available, these patients may treated with pre-enucleation chemotherapy followed by adjuvant therapy (Table III). Despite the fact that pre-enucleation chemotherapy may obscure the pathology of the enucleated eye, the reported results in setting 2 favor pre-enucleation chemotherapy in patients with poor compliance 4, 44. Many families that do not consent to initial enucleation do consent to chemotherapy 72. This provides time to approach the families, in a multidisciplinary fashion, to reconsider their decision. In these situations, the clinician must balance the risks and benefits of this approach. The risks include chemotherapy-related toxicity, including death, in children who do not clearly benefit from it. Chemotherapy also may alter the pathology of the enucleated eye in such a way that compromises the accuracy of the assessment of risk of extraocular relapse 44. These risks must be weighed against the fact that if no chemotherapy is given, the child will die in more than 80% of the cases 4. Chemotherapy in this situation may improve the results of children from families who temporarily refuse enucleation. Therefore, clinicians should either not administer any therapy while approaching the families about consenting to enucleation or administer chemotherapy in the meantime. No prospective studies have compared these approaches and there may be regional variations in the efficacy of this approach in each setting. If chemotherapy is chosen, it should be done as a last resort, to prevent imminent drop out. Tumor response is almost always seen with chemotherapy; however, this does not imply that the diseased eye can become salvageable. No focal therapy should be administered to the eye, and enucleation should be done as soon as the family consents to it.

A recent retrospective series from China 44 and preliminary data from a prospective multicenter study from Central America 41 have shown a survival rate greater than 80% in children with retinoblastoma whose families were at high risk of treatment abandonment, when the child was given pre-enucleation chemotherapy. The survival results from the Chinese series were inferior to those seen in children whose families consented to enucleation upfront 44. The authors were concerned that pre-enucleation chemotherapy increases mortality, because the pathologic features of the enucleated eye may be interpreted inaccurately after chemotherapy. Thus, children may not receive adequate post-surgical therapy 44. Although this may explain the poorer survival results, compared to those of children who received adequate therapy, all of the patients who suffered extraocular relapse had undergone a very late enucleation, often after more than six chemotherapy cycles. This delay led to extraocular dissemination before enucleation, which was evident in four of five relapsed cases. Delayed enucleation was caused by a lack of treatment compliance 45.

Setting 3: Refusal of enucleation is less common in this setting 1 and pre-enucleation chemotherapy has been used less frequently 72. In this setting as well as in selected cases of setting 2, the necessary legal framework in support of the Children's Bill of Rights are usually available, so parental refusal of treatment for their child should be addressed via the legal support systems when other supporting alternatives have failed.


We thank the Fund for Ophthalmic Knowledge for financial support, the St Jude International Outreach Program for facilitating the use of their web-conferencing facility, and the local organizing committee of the 2012 SIOP Africa and Asia for scheduling these discussions during their program. We also thank the following members of the SIOP-PODC graduated-intensity retinoblastoma guidelines working group for their valuable participation in developing these guidelines: Sidnei Epelman MD, John Hungerford MD, Laurence Desjardins MD, Clare Stannard MD, PhD, and Carlos Rodriguez Galindo MD. We thank Angela Mc Arthur for editorial assistance.