Defining High-risk Retinoblastoma

Key Points

Question

Is there uniformity in the definition of high-risk histopathologic features of retinoblastoma across retinoblastoma practices worldwide?

Findings

Among 27 oncologists and pathologists across 16 countries in 6 continents treating retinoblastoma who responded to a survey in 2021, there was little uniformity in the definition of high-risk retinoblastoma. Postlaminar optic nerve infiltration, involvement of optic nerve transection, and extrascleral tumor extension were the only features uniformly considered to be high risk for metastasis across all oncology practices.

Meaning

These results suggest a need to arrive at consensus definitions of high-risk histopathologic features of retinoblastoma and for prospective multicenter studies to avoid undertreatment or overtreatment of patients with retinoblastoma.

Abstract

Importance

High-risk histopathologic features of retinoblastoma are useful to assess the risk of systemic metastasis. In this era of globe salvage treatments for retinoblastoma, the definition of high-risk retinoblastoma is evolving.

Objective

To evaluate variations in the definition of high-risk histopathologic features for metastasis of retinoblastoma in different ocular oncology practices around the world.

Design, Setting, and Participants

An electronic web-based, nonvalidated 10-question survey was sent in December 2020 to 52 oncologists and pathologists treating retinoblastoma at referral retinoblastoma centers.

Intervention

Anonymized survey about the definition of high-risk histopathologic features for metastasis of retinoblastoma.

Main Outcomes and Measures

High-risk histopathologic features that determine further treatment with adjuvant systemic chemotherapy to prevent metastasis.

Results

Among the 52 survey recipients, the results are based on the responses from 27 individuals (52%) from 24 different retinoblastoma practices across 16 countries in 6 continents. The following were considered to be high-risk features: postlaminar optic nerve infiltration (27 [100%]), involvement of optic nerve transection (27 [100%]), extrascleral tissue infiltration (27 [100%]), massive (≥3 mm) choroidal invasion (25 [93%]), microscopic scleral infiltration (23 [85%]), ciliary body infiltration (20 [74%]), trabecular meshwork invasion (18 [67%]), iris infiltration (17 [63%]), anterior chamber seeds (14 [52%]), laminar optic nerve infiltration (13 [48%]), combination of prelaminar and laminar optic nerve infiltration and minor choroidal invasion (11 [41%]), minor (<3 mm) choroidal invasion (5 [19%]), and prelaminar optic nerve infiltration (2 [7%]). The other histopathologic features considered high risk included Schlemm canal invasion (4 [15%]) and severe anaplasia (1 [4%]). Four respondents (15%) said that the presence of more than 1 high-risk feature, especially a combination of massive peripapillary choroidal invasion and postlaminar optic nerve infiltration, should be considered very high risk for metastasis.

Conclusions and Relevance

Responses to this nonvalidated survey conducted in 2020-2021 showed little uniformity in the definition of high-risk retinoblastoma. Postlaminar optic nerve infiltration, involvement of optic nerve transection, and extrascleral tumor extension were the only features uniformly considered as high risk for metastasis across all oncology practices. These findings suggest that the relevance about their value in the current scenario with advanced disease being treated conservatively needs further evaluation; there is also a need to arrive at consensus definitions and conduct prospective multicenter studies to understand their relevance.

This survey study evaluates variations in the definition of high-risk histopathologic features for metastasis of retinoblastoma in different ocular oncology practices around the world.

Introduction

The management of retinoblastoma (RB) has evolved over the years toward globe salvage strategies. However, enucleation still remains the standard of care for advanced RB with clinical features associated with high-risk RB.^1,2,3,4 Identifying high-risk histopathologic features is crucial because the presence of high-risk RB puts approximately 24% of children with RB at risk for developing systemic metastasis.⁵ The presence of high-risk RB is an indication for adjuvant chemotherapy and radiotherapy for selected patients, and studies have found that adjuvant therapy may reduce the potential of systemic metastasis to 0% to 7%.^2,6,7,8,9,10 The Children’s Oncology Group (COG) clinical trial is the only prospective multicenter study with more than 300 patients and more than 2 years of complete follow-up based on high-risk histopathologic features and adjuvant chemotherapy.⁶

Although there are several publications in the literature about high-risk RB and the utility of adjuvant treatment in these cases, the definition of high-risk RB is quite variable.^{1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42} An earlier study by Chantada et al in 2008 highlighted the disparity in the definition of high-risk RB and treatment criteria.³⁷ There was a greater disparity in the definition of choroidal invasion, with 5 different definitions among the group. The COG standardized the definition of massive choroidal invasion (≥3 mm in maximum diameter), which is currently widely accepted.⁶ The disparity in the definition of high-risk RB and treatment criteria often interferes with comparison of outcomes across various treatment centers. In this electronic, anonymized, web-based survey, we intended to identify whether this disparity in the definition of high-risk RB and treatment criteria still exists in current practice across different RB centers around the world.

Methods

A link to a 10-question survey (Survey Monkey Inc) was sent to 52 RB specialists (ocular oncologists and researchers, ocular pathologists, and medical oncologists) across 25 countries in 6 continents via email in December 2020. The survey recipients were included based on their publications or presentations on the topic of high-risk RB. Each participant was unaware of the fellow recipients and their responses. The LV Prasad Eye Institute institutional review board approved the circulation of this electronic web-based, nonvalidated survey. Informed written consent was received from all participants to participate in this survey. None of the participants were offered any compensation or incentives to participate.

The questions were objective, with response options of yes or no and a free-text box for additional comments for each question. The questions included whether the following factors were considered high-risk histopathologic features: anterior chamber (AC) tumor seeds or tumor infiltration of the iris, trabecular meshwork, ciliary body, choroid (minor [<3 mm] and/or major [≥3 mm]), optic nerve, sclera, and/or extrascleral tissue. The respondents were also asked to comment whether they considered any other histopathologic feature a high risk for metastasis. All responses completed before June 2021 were included for analysis. The name and the country of RB practice of the respondent were noted.

Results

Of the 52 recipients of the survey link worldwide, 27 specialists (52%) from 24 different RB practices across 16 countries in 6 continents responded to the survey (Figure 1). Of the 27 RB specialists who responded to the survey, 18 (67%) were ophthalmologists or researchers practicing ocular oncology, 2 (7%) were ocular pathologists, 3 (11%) were trained in both ocular oncology and pathology, and 4 (15%) were medical oncologists treating RB. Based on geographical distribution, the survey respondents included 8 (30%) from Asia, 8 (30%) from Europe, 5 (19%) from North America, 2 (7%) from Australia, 2 (7%) from South America, and 2 (7%) from Africa.

Figure 1. Survey Participants Based on Geographical Distribution.

There were variable definitions of high-risk RB between different practices, including variations within the same country (Figure 2). Ocular oncologists and medical oncologists (n = 24) were the primary clinical decision-makers regarding adjuvant chemotherapy based on histopathologic findings. The following were considered high-risk features: AC seeds (14 [52%]), iris infiltration (17 [63%]), ciliary body infiltration (20 [74%]), trabecular meshwork invasion (18 [67%]), minor choroidal invasion (5 [19%]), massive choroidal invasion (25 [93%]), prelaminar optic nerve infiltration (2 [7%]), laminar optic nerve infiltration (13 [48%]), postlaminar optic nerve infiltration (27 [100%]), involvement of optic nerve transection (27 [100%]), combination of prelaminar and laminar optic nerve infiltration and minor choroidal invasion (11 [41%]), microscopic scleral infiltration (23 [85%]), and extrascleral tissue infiltration (27 [100%]). Anterior segment (iris, ciliary body, and trabecular meshwork) invasion was considered a high-risk feature by 21 (78%) survey respondents.

Figure 2. Variation in the Definition of High-risk Retinoblastoma Based on Geographical Distribution.

There was variation in the definition of high-risk RB even in the same geographical location (Table 1) and between primary clinical decision-makers in different countries (Table 2). There was consensus in the definition of postlaminar optic nerve infiltration, infiltration of optic nerve transection, and extrascleral tumor infiltration as high-risk histopathologic features, while there was disparity in the definition of other histopathologic features as risk factors for metastasis. In addition to these features, some other histopathologic features and genetic alterations were considered risk factors for metastasis by a few respondents. The other histopathologic features considered high risk included Schlemm canal invasion (4 [15%]) and severe anaplasia (1 [4%]). One respondent (4%) also considered amplification of the N-Myc protein as an independent risk factor for metastasis in RB. Four respondents (15%; 2 of whom had been part of the COG prospective clinical trial on high-risk RB⁶) said that the presence of more than 1 high-risk feature, especially a combination of massive peripapillary choroidal invasion and postlaminar optic nerve infiltration, should be considered very high risk for metastasis.

Table 1. Definition of High-risk Histopathologic Features of Retinoblastoma Across Different Oncology Practices Around the World.

Continent	No.	Tissue invasion by tumor considered a high-risk feature by survey respondents, %
		AC seeds	Iris	CB	TM	MiCI	MaCI	PrLON	LON	PrLON or LON plus MiCI	PLON	ONT	Sclera	EST	Other
Africa	2	50	100	100	50	0	100	0	50	0	100	100	100	100	0
Asia	8	75	88	100	100	63	100	25	75	75	100	100	88	100	38
Australia	2	100	100	100	100	0	100	0	50	0	100	100	50	100	0
Europe	8	63	63	63	75	0	88	0	25	25	100	100	100	100	25
North America	5	0	20	60	20	0	100	0	60	40	100	100	60	100	20
South America	2	0	0	0	0	0	50	0	0	50	100	100	100	100	0

Abbreviations: AC, anterior chamber; CB, ciliary body; EST, extrascleral tissue; LON, laminar optic nerve; MaCI, massive (≥3 mm) choroidal invasion; MiCI, minor (<3 mm) choroidal invasion; ONT, optic nerve transection; PLON, postlaminar optic nerve; PrLON, prelaminar optic nerve; TM, trabecular meshwork.

Table 2. Responses From the Primary Clinical Decision-Makers Regarding the Need for Adjuvant Therapy.

Continent	No.	Tissue invasion by tumor considered a high-risk feature by survey respondents, %
		AC seeds	Iris	CB	TM	MiCI	MaCI	PrLON	LON	PrLON or LON plus MiCI	PLON	ONT	Sclera	EST	Other
Africa	2	50	100	100	50	0	100	0	50	0	100	100	100	100	0
Asia	8	75	88	100	100	63	100	25	75	75	100	100	88	100	38
Australia	1	100	100	100	100	0	100	0	0	0	100	100	0	100	0
Europe	8	63	63	63	75	0	88	0	25	25	100	100	100	100	25
North America	3	0	33	33	33	0	100	0	66	66	100	100	66	100	33
South America	2	0	0	0	0	0	50	0	0	50	100	100	100	100	0

Abbreviations: AC, anterior chamber; CB, ciliary body; EST, extrascleral tissue; LON, laminar optic nerve; MaCI, massive (≥3 mm) choroidal invasion; MiCI, minor (<3 mm) choroidal invasion; ONT, optic nerve transection; PLON, postlaminar optic nerve; PrLON, prelaminar optic nerve; TM, trabecular meshwork.

Discussion

This study highlights the disparity in the definition of high-risk RB around the world. Because the decision to offer adjuvant chemotherapy and radiotherapy depends on the definition of high-risk RB at each individual treatment center, this disparity translates into a wide variation in practice in the management of high-risk RB, with highly variable outcomes. The results are thus not comparable, making meaningful conclusions from these multicenter studies challenging.

With the recent advances in RB treatment, there has been a decrease in enucleation rates, with eyes with more advanced disease receiving successful eye-conserving treatment. As a result, clinicians are observing AC seeds or Schlemm canal invasion in eyes with useful vision, a situation that was previously uncommon. Anterior chamber seeds could be primary in cases of anterior segment invasion, especially in patients with ciliary body infiltration resulting in contaminated aqueous or secondary to spillover of endophytic retinal tumor or vitreous seeds at the vitreous base or ciliary body area.⁴³ This infiltration is managed in selective cases with intracameral chemotherapy or plaque radiotherapy with acceptable outcomes, with no evidence of systemic metastasis during a short-term follow-up period.^43,44,45 In a study of 18 eyes with isolated AC seeds and no other high-risk features and without exposure to adjuvant chemotherapy, there was no incidence of metastasis or death over a period of 4 years.²² These limited reports suggest that isolated AC seeds, especially spillover seeds, might not be considered a high-risk feature or an indication for adjuvant treatment. In our survey, however, AC seeds were considered a high-risk feature by 52% of respondents.

Anterior segment (iris, ciliary body, and trabecular meshwork) invasion was considered a high-risk feature by 21 survey respondents (78%). To our knowledge, anterior segment invasion as an independent high-risk feature for metastasis has not been investigated and thus needs further studies. The significance of infiltration of the Schlemm canal as a high-risk feature also needs clarification. As the Schlemm canal drains into episcleral blood vessels via aqueous veins, it has the potential for hematogenous dissemination and may be considered a high-risk feature. These complex situations would require additional collaborative studies, further highlighting the need for uniformity in assessment and interpretation of high-risk features.

There has been much debate about the hematogenous route of the spread of RB compared with the spread via the optic nerve (central nervous system infiltration). It is generally accepted that central nervous system infiltration carries a worse prognosis, whereas the risk of hematogenous spread via the choroid is lower and can be managed, with lower mortality. This is further complicated because the definition of choroidal invasion as a high-risk feature in RB has been highly variable across RB centers, including the largest diameter (ie, >3 mm), more than 50% thickness of the choroid, more than 1 to 3 clusters of choroidal infiltration, or full-thickness choroidal infiltration.³⁷ The COG prospective study defined massive choroidal invasion as the largest diameter of choroidal infiltration (ie, ≥3 mm),⁶ which is currently widely accepted. Massive choroidal invasion was considered as a high-risk feature by 25 respondents (93%) in the current survey. In a study of 55 patients with isolated massive choroidal invasion (defined as the presence of >1 clump of tumor cells and/or invasion of more than half the thickness of choroid) who did not receive adjuvant chemotherapy, 3 (6%) developed extraocular tumor relapse during a mean follow-up of 3 years.³⁸ All 3 patients did well with intensive chemotherapy with or without external beam radiotherapy. Based on these results, the authors recommended that patients with isolated choroidal invasion do not receive adjuvant chemotherapy. This observation warrants further study in a larger cohort of patients. Most centers consider massive choroidal invasion an independent high-risk feature.

Central nervous system involvement with RB carries a very poor prognosis and occurs owing to optic nerve infiltration. In this survey, the combination of prelaminar and laminar optic nerve infiltration and minor choroidal invasion was considered a high-risk feature by 41% of respondents. In the COG study of 15 patients with a combination of prelaminar and laminar optic nerve infiltration and minor choroidal invasion, none developed extraocular tumor relapse at 2 years of follow-up, with the conclusion that adjuvant chemotherapy was not indicated in this group of patients.⁶ The risk of extraocular tumor relapse was highest for children with a 1.5-mm postlaminar optic nerve extension and massive peripapillary choroidal invasion despite adjuvant chemotherapy. These findings suggest that this combination is a very high-risk feature and may need a more aggressive approach.

Isolated postlaminar optic nerve invasion, involvement of optic nerve transection, and microscopic extrascleral tumor infiltration were considered high-risk features by all 27 participants. In a study of 187 children with RB and high-risk features of postlaminar optic nerve invasion and/or scleral infiltration, adjuvant chemotherapy improved the 3-year event-free survival to 97%.¹⁴ Infiltration of optic nerve transection is a poor prognostic factor and requires a combination of adjuvant chemotherapy and radiotherapy. In a study of 650 enucleated eyes with RB, the survival rate was 54% among patients with orbital invasion, 86% among those with scleral invasion, and 85% among those with extrascleral tumor infiltration.¹³ The survival rate was also low among patients with optic nerve (postlaminar and cut end) infiltration. However, the authors do not mention whether the poor prognosis for these patients was with or without any adjuvant treatment. Based on available evidence, the involvement of optic nerve transection and microscopic extrascleral tumor infiltration should be considered microscopic residual disease rather than a histopathologic risk factor and should warrant aggressive treatment.

In our survey, 1 respondent also defined high-risk RB by the presence of severe anaplasia. In a study of 266 enucleated eyes, Mendoza et al³² showed that severe anaplasia was associated with increased risk of metastasis even in the absence of defined high-risk features. Anaplasia may or may not correlate with the other high-risk features described in all patients. The term severe anaplasia needs a better pathologic definition as to what percentage of undifferentiated tumor cells are required to be considered anaplastic. The significance of high-grade anaplasia as an independent high-risk feature needs more evidence, further highlighting the need for collaborative studies.

One respondent to our survey also considered N-Myc amplification an independent risk factor for metastasis in RB. Amplification of N-Myc is a known poor prognostic indicator for neuroblastoma associated with poor histologic characteristics.^46,47,48,49 Although both neuroblastoma and RB are tumors of neuroectodermal origin, Lillington et al⁴⁹ found no definitive association between N-Myc amplification and poor outcomes in RB. Xu et al⁵⁰ have shown an association between chromosome 6p gain and aggressive histopathologic features. The association between N-Myc amplification and other genetic alterations (chromosome 6p gain) and high-risk histopathologic features needs further research.

Limitations

This study has some limitations, including that the survey was not validated and the respondents had different backgrounds and opportunities to treat, follow up, and study their patients, which is also a unique and positive quality of the survey. However, the conclusions for the interpretation of high-risk histopathologic features may be associated with the type of histopathologic studies, the access to specialized ocular pathologists, and the retrospective vs prospective clinicopathologic studies available for each respondent.

Conclusions

Based on this international multicenter survey, there continues to be a disparity in the definition of high-risk RB in current practice around the world. Although isolated postlaminar optic nerve invasion, involvement of optic nerve transection, and microscopic extrascleral tumor infiltration were uniformly considered high-risk features, there was disparity with regard to other features, such as anterior segment invasion, laminar optic nerve, and choroidal and scleral infiltration. However, the definition of most of these histopathologic features as higher-risk factors for metastasis is based on the experience of initially enucleated eyes and, to some extent, on secondarily enucleated eyes after systemic chemotherapy. There is no information about their value in the current scenario with advanced disease treated conservatively with intra-arterial chemotherapy. Also, in this era of advanced rescue treatments and with a low relapse rate for patients with high-risk histopathologic features, the decision of adjuvant treatment is more complicated and should be weighed against compliance with follow-up, risks of systemic chemotherapy, and availability of rescue treatments at the center. These findings suggest a need to arrive at consensus definitions and to conduct prospective multicenter studies with respect to risk stratification (low, intermediate, and high) based on histopathologic features and genetic alterations to plan appropriate treatment strategies so the treatment outcomes for prevention of metastasis at different centers can be compared accurately.