Ethnic, Racial, and Socioeconomic Disparities in Retinoblastoma | Pediatric Cancer | JAMA Pediatrics | JAMA Network
[Skip to Navigation]
Table 1.  Overview of the 830 Children in the Study Cohort
Overview of the 830 Children in the Study Cohort
Table 2.  Extent of Disease at Diagnosis, Occurrence of Enucleation, and 5-Year Relative Survival Rate According to Major Ethnic and Racial Groupsa
Extent of Disease at Diagnosis, Occurrence of Enucleation, and 5-Year Relative Survival Rate According to Major Ethnic and Racial Groupsa
Table 3.  Children With Retinoblastoma by Extent of Disease at Diagnosis, Ethnicity, and Selected SES Indicatorsa
Children With Retinoblastoma by Extent of Disease at Diagnosis, Ethnicity, and Selected SES Indicatorsa
Table 4.  Children With Retinoblastoma by Ocular Outcome, Ethnicity, and Selected SES Indicatorsa
Children With Retinoblastoma by Ocular Outcome, Ethnicity, and Selected SES Indicatorsa
Table 5.  Association of County Socioeconomic Deprivation Index With Extent of Disease and Ocular Outcomesa
Association of County Socioeconomic Deprivation Index With Extent of Disease and Ocular Outcomesa
Original Investigation
December 2015

Ethnic, Racial, and Socioeconomic Disparities in Retinoblastoma

Author Affiliations
  • 1Department of Pediatric Hematology/Oncology, Dana-Farber/Boston Children’s Cancer and Blood Disorders Center, Boston, Massachusetts
  • 2Department of Pediatrics, Harvard Medical School, Boston, Massachusetts
  • 3Department of Collective Health, Faculdade de Ciências Médicas da Santa Casa de São Paulo, São Paulo, Brazil
JAMA Pediatr. 2015;169(12):1096-1104. doi:10.1001/jamapediatrics.2015.2360

Importance  Most children with retinoblastoma in the United States are diagnosed as having a large intraocular tumor burden that requires intensive ocular-salvage treatment or enucleation.

Objective  To investigate the effect of socioeconomic status, race, and ethnicity on the extent of disease and the outcomes of retinoblastoma.

Design, Setting, and Participants  A population-based review of 18 Surveillance, Epidemiology, and End Results (SEER) registries. From January 1, 2000, through December 31, 2010, 830 cases of retinoblastoma were recorded for children aged 0 to 9 years. Data were collected and analyzed from January 1, 2000, through December 31, 2010, with the last follow-up on December 31, 2010.

Exposures  County-based socioeconomic variables analyzed included poverty level, educational attainment, language isolation, crowding, unemployment, and percentage of immigrants.

Main Outcomes and Measures  Extent of disease, ocular outcome, and children’s survival.

Results  Of the 830 individuals included, Hispanic children had a higher percentage of extraocular disease (86 of 261 [33.0%] vs 102 of 510 non-Hispanic children [20.0%]; odds ratio [OR], 1.97 [95% CI, 1.38-2.79]). The percentage of extraocular cases was also higher in counties with the following low socioeconomic status indicators: higher vs lower poverty status (115 of 413 [27.8%] vs 73 of 358 [20.4%]; OR, 1.51; 95% CI, 1.06-2.14); lower vs higher educational attainment (115 of 400 [28.7%] vs 73 of 371 [19.7%]; OR, 1.65; 95% CI, 1.16-2.34); higher vs lower levels of crowding (124 of 398 [31.2%] vs 64 of 373 [17.2%]; OR, 2.18; 95% CI, 1.53-3.13); higher vs lower unemployment (119 of 411 [28.9%] vs 69 of 360 [19.2%]; OR, 1.72; 95% CI, 1.21-2.45); higher vs lower language isolation (117 of 388 [30.2%] vs 71 of 383 [18.5%]; OR, 1.89; 95% CI, 1.34-2.70); and higher vs lower percentage of immigrants (109 of 386 [28.2%] vs 79 of 385 [20.5%]; OR, 1.52; 95% CI, 1.08-2.16). Higher rates of enucleation were associated with low educational attainment (265 of 401 [66.1%] vs 309 of 421 [73.4%]; OR, 1.42; 95% CI, 1.04-1.93), a higher level of crowding (316 of 416 [76.0%] vs 258 of 406 [63.5%]; OR, 1.81; 95% CI, 1.32-2.48), and Hispanic origin (202 of 271 [74.5%]; OR, 1.41; 95% CI, 1.01-1.98). Relative survival at 5 years was lower among black compared with non-Hispanic white children (92.7% vs 99.2%; P < .001).

Conclusions and Relevance  Significant disparities exist in the care and outcomes of children with retinoblastoma. A low socioeconomic status negatively affects disease extent and ocular outcomes, presumably by limiting access to primary and cancer-directed care. Hispanic children in particular have more advanced disease and higher rates of enucleation.


Retinoblastoma is a rare childhood cancer; approximately 300 cases are diagnosed yearly in the United States and 9000 are diagnosed yearly worldwide.1,2 Retinoblastoma presents in a heritable, often bilateral, form that accounts for 25% of the cases and is characterized by the presence of a germline RB1 mutation (RefSeq NM_000321.2) and in a nonheritable unilateral form.1 Treatment of retinoblastoma aims at cure, ocular salvage, and preservation of vision; thus, priorities must be established based on the laterality and the extent of disease.1,3

The presentation of retinoblastoma reflects the continuum of disease progression and the range of opportunities for diagnosis and ocular salvage. When diagnosed early, ocular salvage and preservation of vision are possible with relatively minimal therapy.4-7 The treatment of larger tumors is more complex; many patients require enucleation, and those with disease extension beyond the retina require chemotherapy and often orbital radiotherapy.8 Patients with advanced intraocular disease who are candidates for ocular salvage receive systemic or intra-arterial chemotherapy coupled with focal treatments, including radiotherapy, which is associated with an increased risk for second cancers.3,7,9-11 Although these complex therapies are effective, their consequences for the children and their families cannot be underestimated.12 Finally, a small proportion of patients present with extraocular disease, which is generally associated with poor survival.13,14

Thus, early diagnosis and referral and prompt initiation of therapy are key to successful outcomes; ocular survival, preservation of vision, and patient survival correlate directly with the time to diagnosis. Leukocoria, the most common presentation, can be detected readily with the red reflex test on a routine examination. The American Academy of Pediatrics recommends age-appropriate assessment of eye problems, including a symmetric red reflex, at all health supervision visits from birth to 5 years of age.15 Despite these practices that could help to detect retinoblastoma early, most US patients with retinoblastoma are still diagnosed as having overt leukocoria and a large intraocular tumor burden that requires intensive ocular-salvage treatments or enucleation. Thus, early diagnosis and successful outcomes for children with retinoblastoma are tightly related to the strength of the health care system; vulnerable, socioeconomically deprived populations are at risk for delayed diagnosis and adverse outcomes. Although disparities in access to care have been well evaluated and discussed for adult cancers,16 little is known about pediatric cancers; in this regard, retinoblastoma may offer a unique scenario in which disparities can be analyzed. In this study we examined the relationship among ethnicity, race, and socioeconomic factors and the extent of disease at diagnosis as a surrogate marker for delayed diagnosis and access to tertiary care and the outcomes in retinoblastoma.

Box Section Ref ID

At a Glance

  • Most studies investigating disparities in childhood cancer outcomes in high-income countries have failed to show significant differences based on socioeconomic or ethnic factors.

  • Hispanic children were more likely to present with extraocular disease (33.0% vs 20.0%) and were 41% more likely to undergo enucleation than non-Hispanic white children.

  • Low socioeconomic status affects disease extent and ocular outcomes, presumably by limiting access to primary and cancer-directed care.

  • Identifying and addressing such disparities is critical given the known morbidity and the long-term psychological, financial, and medical burden that these children and their families have to endure.

Selection of Study Population

The study population was selected from the Surveillance, Epidemiology, and End Results (SEER) database using the SEER*Stat program (version 8.1.2; Children aged 0 to 9 years who were diagnosed as having retinoblastoma from January 1, 2000, through December 31, 2010, were selected. Last follow-up was completed on December 31, 2010. This time frame allowed us to include children treated in the modern era for evaluation and to use all 18 population-based SEER registries (San Francisco and Oakland, California; Los Angeles, California; San Jose and Monterey, California; California, excluding the previous urban areas; Connecticut; metropolitan Detroit, Michigan; Hawaii; Iowa; New Mexico; Seattle and Puget Sound, Washington; Utah; metropolitan Atlanta, Georgia; rural Georgia; greater Georgia; Kentucky; Louisiana; New Jersey; and Native Alaska), which together represent 28% of the US population. One child was excluded, because the tumors were metachronous, with the first tumor detected before 2000. Another child was excluded owing to a concomitant diagnosis of retinoblastoma and glioma. We collected data from January 1, 2000, through December 31, 2010, and extracted the following variables for each individual: race, ethnicity, sex, age at diagnosis, and laterality. This study was conducted in compliance with the international regulations for the protection of human participants and was exempted from institutional review board review by the Dana-Farber/Harvard Cancer Center Office for Human Research Subjects.

Socioeconomic Factors

We obtained information on each child’s age at diagnosis, race, ethnicity, and state and county of residence directly from the SEER database. We also gathered area-based indicators of socioeconomic status (SES) based on the county of residence for each individual case; we chose 6 standard 2000 US Census variables, which reflected the levels of poverty, educational attainment, crowding, unemployment, immigration, and language isolation within each county. For each variable, we split the cohort into 2 groups based on the median value, which allowed for comparison of outcomes between more and less disadvantaged counties, as described in eTable 1 in the Supplement.

County Socioeconomic Deprivation Index

To analyze the joint effect of different socioeconomic factors on the outcomes of interest, we created a composite county socioeconomic deprivation index, which could be calculated for each county. An SES index is typically constructed by combining information on the levels of income, educational attainment, poverty, employment, and wealth and occupation.17 The variables included in the index were first standardized [x – (mean/SD)] based on the values found within the patient cohort. The index was then calculated based on 2 main, equally weighted domains. The first domain reflects the joint effects of purely economic factors and is calculated by obtaining the mean standardized variables for the level of poverty (P), low educational attainment (E), crowding (C), and unemployment (U). The other domain reflects factors representing potential cultural and language barriers and is calculated based on levels of immigration (I) and language isolation (L) within the counties. The formula used to construct the index is as follows:

[(Economic Index + Cultural and Language Index)/2] = {[(P + E + C + U)/4] + [(I + L)/2]}/2.

After the index value was calculated for each county represented in our cohort, we identified the median value. This process allowed us to compare the extent of disease and outcomes between the disadvantaged counties—those having deprivation index values that were greater than or equal to the median value—and the remaining counties.

Outcome Variables
Extent of Disease at Diagnosis

A tumor was considered to be intraocular if the value of SEER database historic stage A variable was localized, whereas regional or distant values indicated extraocular tumors (eTable 2 in the Supplement). Although central pathologic review was not possible, localized, regional, and distant values were interpreted as representing a continuum of disease progression that would allow us to identify children with more advanced disease for analysis purposes. Cases without specific SEER historic stage A values were excluded from analyses on extent of disease at diagnosis.

Occurrence of Enucleation (Ocular Outcome)

We used SEER surgical codes to identify children who underwent enucleation (eTable 3 in the Supplement). We excluded cases in which we could not discern whether surgery took place or the nature of the surgical intervention (n = 8). Children with bilateral disease who might have had 2 enucleations were counted once.

Statistical Analysis

Data were analyzed from January 1, 2000, through December 31, 2010. We used the χ2 test to verify the association between SES variables (including the individual patient’s race and ethnicity, isolated SES variables, and county socioeconomic deprivation index) and the extent of disease at diagnosis and enucleation. We calculated crude odds ratios (ORs) and corresponding 95% CIs. To evaluate confounding and effect modification, analysis of the SES effect was stratified by ethnicity. We used the Mantel-Haenszel test of homogeneity to compare the ORs of different strata and to identify effect modification.18 Given the small number of cases per registry, a multivariate analysis adjusting by registry site was not feasible. We used the SEER*Stat program with the Ederer II method and the US 1970-2007 expected survival table from the National Center for Health Statistics to analyze the differences in 5-year relative survival according to ethnicity, race, and SES variables.

Study Population

We identified 830 children with retinoblastoma (including 455 boys [54.8%] and 243 children with bilateral disease [29.3%]). Most of the children were white (606 [73.0%]), and almost one-third (262 [31.6%]) were Hispanic (Table 1).

Extent of Disease at Diagnosis

Among the 771 children whose extent of disease could be determined, 188 (24.4%) presented with extraocular disease as defined in the study (Table 1). Hispanic children had a higher percentage of extraocular disease than non-Hispanic children (86 of 261 [33.0%] vs 102 of 510 [20.0%]; OR, 1.97; 95% CI, 1.38-2.79) (Table 2). Differences in the percentage of extraocular disease among racial groups were also significant (Table 2). For every socioeconomic factor under study, higher percentages of children had extraocular disease in counties with a more disadvantaged status, whether such status constituted a high poverty level (115 of 413 [27.8%] vs 73 of 358 [20.4%]; OR, 1.51; 95% CI, 1.06-2.14), a high proportion with an educational attainment of less than high school (115 of 400 [28.7%] vs 73 of 371 [19.7%]; OR, 1.65; 95% CI, 1.16-2.34), a high level of crowding (124 of 398 [31.2%] vs 64 of 373 [17.2%]; OR, 2.18; 95% CI, 1.53-3.13), a high level of language isolation (117 of 388 [30.2%] vs 71 of 383 [18.5%]; OR, 1.89; 95% CI, 1.34-2.70), a high percentage of unemployment (119 of 411 [28.9%] vs 69 of 360 [19.2%]; OR, 1.72; 95% CI, 1.21-2.45), or a high percentage of new immigrants (109 of 386 [28.2%] vs 79 of 385 [20.5%]; OR, 1.52; 95% CI, 1.08-2.16) (Table 3). Point estimates from stratified analysis showed that the effects of socioeconomic disadvantage on the risk of extraocular disease were higher among Hispanics for all indicators except immigration; however, homogeneity tests were not significant (Table 3). On the other hand, none of the county socioeconomic indicators was associated with a higher risk of extraocular disease among non-Hispanic children (white non-Hispanic, black, and others) (eTable 4 in the Supplement).

Ocular Outcome

Ocular outcomes could be determined for 822 children; 574 of these children (69.8%) underwent enucleation at some point during the course of treatment. The percentage of enucleation was higher among Hispanic compared with non-Hispanic children (202 of 271 [74.5%] vs 372 of 551 [67.5%]; OR, 1.41; 95% CI, 1.01-1.98) (Table 2). Black and Asian or Pacific Islander children also had a higher risk for enucleation compared with white, non-Hispanic children (92 of 124 [74.2%] vs 213 of 338 [63.0%]; OR, 1.69; 95% CI, 1.07-2.67), as did children of other races (66 of 86 [76.7%]; OR, 1.94; 95% CI, 1.10-3.53) (Table 2). Percentages of children undergoing enucleation were consistently higher in more disadvantaged counties for all the SES variables, although the risk was only significant for the following 3 indicators: a low level of educational attainment (OR, 1.42; 95% CI, 1.04-1.93), a high level of poverty (OR, 1.41; 95% CI, 1.04-1.93), and a high level of crowding (OR, 1.81; 95% CI, 1.32-2.48) (Table 4). Effect modification by ethnicity was only observed for language isolation, in which an increased risk for enucleation for those living in an area of high language isolation was greater for Hispanic (OR, 2.23; 95% CI, 1.19-4.16) than for non-Hispanic (OR, 0.99; 95% CI, 0.68-1.43) children (Table 4). No significant effect modification was observed for race or ethnicity for any of the SES indicators (eTable 5 in the Supplement).


The 5-year relative survival rate was 97.7%. Despite differences in percentages of extraocular disease and enucleation, we found no difference in the 5-year relative survival rates between Hispanic and non-Hispanic children (97.5% vs 97.9%). However, a lower 5-year relative survival rate was observed for black compared with white non-Hispanic children (92.7% vs 99.2%; P < .001). Although we found a lower relative survival rate for the more disadvantaged counties for all SES variables studied, the difference was statistically significant only for unemployment (99.7% and 95.6% for low and high levels of unemployment, respectively; P < .001) and language isolation (99.0% and 96.3% for low and high levels of isolation, respectively; P = .04) (eTable 6 in the Supplement).

Outcomes Based on the County Socioeconomic Deprivation Index

The county socioeconomic deprivation index was calculated for each child; the mean and median values for the index were 0 and −0.17, respectively. The highest value, which indicated a high level of socioeconomic disadvantage, was 1.51 and was found in Hudson County, New Jersey; the lowest value was −1.45 in Iowa County, Iowa. Results obtained from the use of this index were consistent with the findings for individual SES variables discussed above. Children who resided in more disadvantaged counties were at higher risk for presenting with extraocular disease and undergoing enucleation than those living in more advantaged counties (Table 4). Stratification by ethnicity (non-Hispanic and Hispanic) indicated that disparities existed within both populations; however, Hispanic children living in disadvantaged counties had a very high risk for undergoing enucleation, particularly for those with bilateral retinoblastoma (eFigure in the Supplement). Black children from more disadvantaged counties also had a higher risk for extraocular disease than their peers who resided in more advantaged counties (Table 5). Finally, children residing in more disadvantaged counties had a lower relative survival compared with those residing in more advantaged counties (96.5% vs 98.9%), although this difference was not statistically significant (eTable 6 in the Supplement).


The results of our study highlight the progress made in the treatment of retinoblastoma and the disparities that have emerged as opportunities for cure and ocular salvage have become available. Retinoblastoma, once a uniformly fatal disease, is now readily curable; clinical research and innovation are focused on how to improve ocular salvage with the least short- and long-term adverse effects.1,3 Despite these advances, our study demonstrates a higher risk for having more advanced disease and undergoing enucleation among specific subpopulations that can be characterized by race, ethnicity, and unfavorable socioeconomic factors, such as poverty, low educational attainment, crowding, language isolation, and being part of a new immigrant community. Thus, initiatives aimed at understanding those disparities and narrowing the outcome gaps should also be prioritized.

Although associations between socioeconomic factors and outcomes have been documented among cancers affecting the entire age spectrum,19-21 in many cases the underlying causes of such disparities are not fully understood, particularly for pediatric cancers.21 In acute lymphoblastic leukemia, poorer outcomes among black compared with white patients were noted in separate population-based22,23 and cooperative group24,25 studies, but not in a single-institution study,23 suggesting that an institution’s unique ability to provide relatively uniform care could erase the gaps caused by access and treatment disparities.21,26 Adherence to treatment has been found to be poorer among Hispanic children with acute lymphoblastic leukemia, a fact that increases the risk for relapse.27 Thus, treatment-related variables influenced by SES are the most likely factors to affect outcomes. However, many aspects of the patient’s long journey from having the first symptoms to receiving the diagnosis and starting treatment are not well understood and might play an important role for diseases, such as retinoblastoma, in which timely diagnosis is crucial. Measuring survival as the sole outcome on which disparities are addressed risks overlooking the granularity of a process that may affect vulnerable populations in many other ways, particularly in retinoblastoma.28 As shown in our study, although racial, ethnic, and socioeconomic factors do not have a significant effect on the survival of children with retinoblastoma, the extent of the disease (and thus the intensity of the treatment received) and the chances of eye salvage (and the ultimate effect on vision and quality of life) differ significantly.

The unique characteristics of the natural history of retinoblastoma, in which the disease stage clearly reflects the continuum of tumor growth until diagnosis and ocular salvage correlates directly with the intraocular disease burden, provided a unique opportunity to investigate the role of socioeconomic factors in the earlier stages along the sequence of presentation, diagnosis, and treatment. Unlike many other pediatric solid tumors, such as neuroblastoma, in which outcomes are influenced by underlying tumor-specific biological factors, no known genetic heterogeneity has been shown to correlate with outcomes in retinoblastoma.29 However, the genomics of retinoblastoma have not been evaluated consistently across racial and ethnic groups; this situation is particularly relevant in view of the differences seen in the incidence of retinoblastoma between countries and ethnic groups.30,31 An inverted correlation between the incidence of retinoblastoma and socioeconomic status in heterogeneous countries, such as Brazil and Mexico, has been documented32,33; thus, some of the ethnic disparities observed are driven by the biological features of the tumor.29 In comparison, genomic-wide interrogations have identified genetic variations as associated with susceptibility to childhood acute lymphoblastic leukemia, which could explain its increased incidence in North American Hispanic children.34,35

Differences in outcome by ethnic group traditionally attributed to the quality of health care may need to be redefined as genomic studies are conducted.36 When they interrogated genome-wide germline single-nucleotide polymorphisms in a cohort of children with acute lymphoblastic leukemia, Yang et al37 observed that the component of genomic variation that cosegregated with Native American ancestry was associated with a risk for relapse, even after adjustment for known prognostic factors. Whether the same phenomenon is applicable to embryonal tumors, such as retinoblastoma, is not known. The high prevalence of extraocular disease among Hispanic children, particularly among those living in counties with challenging socioeconomic conditions, suggests that socioeconomic factors might have made their largest impact even before any treatment started. Hispanic children are more than twice as likely as non-Hispanic white and black children to have no health care insurance.38 Despite improvements made in disparity reduction during the past decade, access to care remains the domain with the largest proportion of disparities for Hispanic children.39

Delivery of complex care, albeit important, might be not be as critical to prevent the loss of the eye and of vision as timely access to health supervision in the primary care setting, where early detection should occur. Given the fact that every child in the United States should undergo frequent health supervision visits, including ocular and visual examinations,15,40 our findings highlight the potential gaps between recommendations and the reality of access to primary care, which may be wider for the more vulnerable populations. Major efforts have been made to reduce the disparities in access to care for the pediatric socioeconomically deprived population and the ethnic and racial minorities39; however, as shown by our study, the complexity of pediatric cancer diagnosis and treatment requires dedicated analyses and interventions.

Although timely access to care and the resulting effect on ocular stage is a main contributor to ocular outcome, disparities in treatment received may also exist. Treatment for ocular salvage requires sophisticated multidisciplinary care and the development of individualized treatment plans that include frequent visits for examination under anesthesia and administration of focal treatments.12 A high adherence level is key to avoid treatment delays; vulnerable populations with limited support who face language, insurance, and financial barriers are at the greater risk for failure. As treatments for retinoblastoma become increasingly complex, disparities may worsen.

Finally, we must assume that the disparities described herein exist for all children with cancer; a careful, disease-specific evaluation of their impact should be performed to assess the magnitude of the problem further. In addition, our findings need to be interpreted in the context of the limitations inherent to the use of the SEER registry and census data and to extrapolating area-based indicators to investigate social determinants of health in this specific population. Thus, prospective studies should be conducted; the composite socioeconomic deprivation index generated herein may provide a tool for an extended evaluation across the disease spectrum.


The unique features of retinoblastoma have provided a scenario for the analysis of the impact of racial, ethnic, and socioeconomic factors and have led to the identification of populations at risk for disparate outcomes in childhood cancer. Identifying and addressing such disparities is critical, given the known morbidity and the long-term psychological, financial, and medical burden that these children and their families have to endure.

Back to top
Article Information

Corresponding Author: Carlos Rodriguez-Galindo, MD, Department of Pediatric Hematology/Oncology, Dana-Farber/Boston Children’s Cancer and Blood Disorders Center, 450 Brookline Ave, Room D3-313, Boston, MA 02215 (

Accepted for Publication: July 8, 2015.

Published Online: October 5, 2015. doi:10.1001/jamapediatrics.2015.2360.

Author Contributions: Drs Ribeiro and Rodriguez-Galindo jointly directed this work. Dr Rodriguez-Galindo had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: Truong, Friedrich, Ribeiro, Rodriguez-Galindo.

Drafting of the manuscript: Truong, Ribeiro, Rodriguez-Galindo.

Critical revision of the manuscript for important intellectual content: Green, Friedrich, Ribeiro, Rodriguez-Galindo.

Statistical analysis: Truong, Friedrich , Ribeiro, Rodriguez-Galindo.

Obtained funding: Rodriguez-Galindo.

Administrative, technical, or material support: Rodriguez-Galindo.

Study supervision: Green, Friedrich, Ribeiro, Rodriguez-Galindo.

Conflict of Interest Disclosures: None reported.

Previous Presentation: This paper was presented at the 49th Annual Meeting of the American Society of Clinical Oncology; June 2, 2013; Chicago, Illinois.

Dimaras  H, Kimani  K, Dimba  EAO,  et al.  Retinoblastoma.  Lancet. 2012;379(9824):1436-1446.PubMedGoogle ScholarCrossref
Chantada  GL, Qaddoumi  I, Canturk  S,  et al.  Strategies to manage retinoblastoma in developing countries.  Pediatr Blood Cancer. 2011;56(3):341-348.PubMedGoogle ScholarCrossref
Rodriguez-Galindo  C, Chantada  GL, Haik  BG, Wilson  MW.  Treatment of retinoblastoma: current status and future perspectives.  Curr Treat Options Neurol. 2007;9(4):294-307.PubMedGoogle ScholarCrossref
Rodriguez-Galindo  C, Wilson  MW, Haik  BG,  et al.  Treatment of intraocular retinoblastoma with vincristine and carboplatin.  J Clin Oncol. 2003;21(10):2019-2025.PubMedGoogle ScholarCrossref
Shields  CL, Mashayekhi  A, Cater  J, Shelil  A, Meadows  AT, Shields  JA.  Chemoreduction for retinoblastoma: analysis of tumor control and risks for recurrence in 457 tumors.  Am J Ophthalmol. 2004;138(3):329-337.PubMedGoogle ScholarCrossref
Abramson  DH, Francis  JH, Dunkel  IJ, Marr  BP, Brodie  SE, Gobin  YP.  Ophthalmic artery chemosurgery for retinoblastoma prevents new intraocular tumors.  Ophthalmology. 2013;120(3):560-565.PubMedGoogle ScholarCrossref
Abramson  DH, Marr  BP, Brodie  SE, Dunkel  I, Palioura  S, Gobin  YP.  Ophthalmic artery chemosurgery for less advanced intraocular retinoblastoma: five year review.  PLoS One. 2012;7(4):e34120. doi:10.1371/journal.pone.0034120.PubMedGoogle ScholarCrossref
Kaliki  S, Shields  CL, Shah  SU, Eagle  RC  Jr, Shields  JA, Leahey  A.  Postenucleation adjuvant chemotherapy with vincristine, etoposide, and carboplatin for the treatment of high-risk retinoblastoma.  Arch Ophthalmol. 2011;129(11):1422-1427.PubMedGoogle ScholarCrossref
Marr  BP, Brodie  SE, Dunkel  IJ, Gobin  YP, Abramson  DH.  Three-drug intra-arterial chemotherapy using simultaneous carboplatin, topotecan and melphalan for intraocular retinoblastoma: preliminary results.  Br J Ophthalmol. 2012;96(10):1300-1303.PubMedGoogle ScholarCrossref
Kleinerman  RA, Tucker  MA, Tarone  RE,  et al.  Risk of new cancers after radiotherapy in long-term survivors of retinoblastoma: an extended follow-up.  J Clin Oncol. 2005;23(10):2272-2279.PubMedGoogle ScholarCrossref
Shields  CL, Kaliki  S, Al-Dahmash  S,  et al.  Management of advanced retinoblastoma with intravenous chemotherapy then intra-arterial chemotherapy as alternative to enucleation.  Retina. 2013;33(10):2103-2109.PubMedGoogle ScholarCrossref
Wilson  MW, Haik  BG, Rodriguez-Galindo  C.  Socioeconomic impact of modern multidisciplinary management of retinoblastoma.  Pediatrics. 2006;118(2):e331-e336. doi:10.1542/peds.2006-0226.PubMedGoogle ScholarCrossref
Dunkel  IJ, Khakoo  Y, Kernan  NA,  et al.  Intensive multimodality therapy for patients with stage 4a metastatic retinoblastoma.  Pediatr Blood Cancer. 2010;55(1):55-59.PubMedGoogle Scholar
Dunkel  IJ, Chan  HSL, Jubran  R,  et al.  High-dose chemotherapy with autologous hematopoietic stem cell rescue for stage 4B retinoblastoma.  Pediatr Blood Cancer. 2010;55(1):149-152.PubMedGoogle Scholar
Cynthia Baker, Graham A Barden  III, Brown  OW,  et al; Geoffrey R Simon; Committee on Practice and Ambulatory Medicine; Bright Futures Periodicity Schedule Workgroup.  2014 recommendations for pediatric preventive health care.  Pediatrics. 2014;133(3):568-570.PubMedGoogle ScholarCrossref
Kogevinas  M, Pearce  N, Susser  M, Boffetta  P, eds.  Social Inequalities and Cancer. Lyon, France: International Agency for Research on Cancer; 1997. IARC Scientific publication 138.
Yu  M, Tatalovich  Z, Gibson  JT, Cronin  KA.  Using a composite index of socioeconomic status to investigate health disparities while protecting the confidentiality of cancer registry data.  Cancer Causes Control. 2014;25(1):81-92.PubMedGoogle ScholarCrossref
Kuritz  SJ, Landis  JR, Koch  GG.  A general overview of Mantel-Haenszel methods: applications and recent developments.  Annu Rev Public Health. 1988;9:123-160.PubMedGoogle ScholarCrossref
Hines  R, Markossian  T, Johnson  A, Dong  F, Bayakly  R.  Geographic residency status and census tract socioeconomic status as determinants of colorectal cancer outcomes.  Am J Public Health. 2014;104(3):e63-e71. doi:10.2105/AJPH.2013.301572.PubMedGoogle ScholarCrossref
Singh  G, Williams  S, Siahpush  M, Mulhollen  A.  Socioeconomic, rural-urban, and racial inequalities in US cancer mortality: part I—all cancers and lung cancer and part II—colorectal, prostate, breast, and cervical cancers.  J Cancer Epidemiol. 2011;2011:107497. doi:10.1155/2011/107497. PubMedGoogle Scholar
Bhatia  S.  Disparities in cancer outcomes: lessons learned from children with cancer.  Pediatr Blood Cancer. 2011;56(6):994-1002.PubMedGoogle ScholarCrossref
Kadan-Lottick  NS, Ness  KK, Bhatia  S, Gurney  JG.  Survival variability by race and ethnicity in childhood acute lymphoblastic leukemia.  JAMA. 2003;290(15):2008-2014.PubMedGoogle ScholarCrossref
Pui  CH, Pei  D, Pappo  AS,  et al.  Treatment outcomes in black and white children with cancer: results from the SEER database and St Jude Children’s Research Hospital, 1992 through 2007.  J Clin Oncol. 2012;30(16):2005-2012.PubMedGoogle ScholarCrossref
Bhatia  S, Sather  HN, Heerema  NA, Trigg  ME, Gaynon  PS, Robison  LL.  Racial and ethnic differences in survival of children with acute lymphoblastic leukemia.  Blood. 2002;100(6):1957-1964.PubMedGoogle ScholarCrossref
Pollock  BH, DeBaun  MR, Camitta  BM,  et al.  Racial differences in the survival of childhood B-precursor acute lymphoblastic leukemia: a Pediatric Oncology Group Study.  J Clin Oncol. 2000;18(4):813-823.PubMedGoogle Scholar
Caposole  MZ, Miller  K, Kim  J-N, Steward  NA, Bauer  TL.  Elimination of socioeconomic and racial disparities related to lung cancer: closing the gap at a high volume community cancer center.  Surg Oncol. 2014;23(2):46-52.PubMedGoogle ScholarCrossref
Bhatia  S, Landier  W, Shangguan  M,  et al.  Nonadherence to oral mercaptopurine and risk of relapse in Hispanic and non-Hispanic white children with acute lymphoblastic leukemia: a report from the children’s oncology group.  J Clin Oncol. 2012;30(17):2094-2101.PubMedGoogle ScholarCrossref
Cheung  R.  Impact of socioeconomic disparities on cause-specific survival of retinoblastoma.  Mol Clin Oncol. 2013;1(3):535-540.PubMedGoogle Scholar
Zhang  J, Benavente  CA, McEvoy  J,  et al.  A novel retinoblastoma therapy from genomic and epigenetic analyses.  Nature. 2012;481(7381):329-334.PubMedGoogle Scholar
Stiller  CA, Parkin  DM.  Geographic and ethnic variations in the incidence of childhood cancer.  Br Med Bull. 1996;52(4):682-703.PubMedGoogle ScholarCrossref
Valery  PC, Moore  SP, Meiklejohn  J, Bray  F.  International variations in childhood cancer in indigenous populations: a systematic review.  Lancet Oncol. 2014;15(2):e90-e103. doi:10.1016/S1470-2045(13)70553-9.PubMedGoogle ScholarCrossref
de Camargo  B, de Oliveira Ferreira  JM, de Souza Reis  R, Ferman  S, de Oliveira Santos  M, Pombo-de-Oliveira  MS.  Socioeconomic status and the incidence of non-central nervous system childhood embryonic tumours in Brazil.  BMC Cancer. 2011;11:160. doi:10.1186/1471-2407-11-160.PubMedGoogle ScholarCrossref
Fajardo-Gutiérrez  A, Juárez-Ocaña  S, González-Miranda  G,  et al.  Incidence of cancer in children residing in ten jurisdictions of the Mexican Republic: importance of the cancer registry (a population-based study).  BMC Cancer. 2007;7:68. doi:10.1186/1471-2407-7-68.PubMedGoogle ScholarCrossref
Xu  H, Cheng  C, Devidas  M,  et al.  ARID5B genetic polymorphisms contribute to racial disparities in the incidence and treatment outcome of childhood acute lymphoblastic leukemia.  J Clin Oncol. 2012;30(7):751-757.PubMedGoogle ScholarCrossref
Chow  EJ, Puumala  SE, Mueller  BA,  et al.  Childhood cancer in relation to parental race and ethnicity: a 5-state pooled analysis.  Cancer. 2010;116(12):3045-3053.PubMedGoogle ScholarCrossref
Lim  JY-S, Bhatia  S, Robison  LL, Yang  JJ.  Genomics of racial and ethnic disparities in childhood acute lymphoblastic leukemia.  Cancer. 2014;120(7):955-962.PubMedGoogle ScholarCrossref
Yang  JJ, Cheng  C, Devidas  M,  et al.  Ancestry and pharmacogenomics of relapse in acute lymphoblastic leukemia.  Nat Genet. 2011;43(3):237-241.PubMedGoogle ScholarCrossref
Bloom  B, Jones  LI, Freeman  G.  Summary health statistics for US Children: National Health Interview Survey, 2012.  Vital Health Stat 10. 2013;(258):1-81.Google Scholar
Dougherty  D, Chen  X, Gray  DT, Simon  AE.  Child and adolescent health care quality and disparities: are we making progress?  Acad Pediatr. 2014;14(2):137-148.PubMedGoogle ScholarCrossref
Committee on Practice and Ambulatory Medicine, Section on Ophthalmology. American Association of Certified Orthoptists; American Association for Pediatric Ophthalmology and Strabismus; American Academy of Ophthalmology.  Eye examination in infants, children, and young adults by pediatricians.  Pediatrics. 2003;111(4, pt 1):902-907.PubMedGoogle ScholarCrossref