- Email: [email protected]
High-Risk Retinoblastoma Based on International Classification of Retinoblastoma: Analysis of 519 Enucleated Eyes
High-Risk Retinoblastoma Based on International Classification of Retinoblastoma: Analysis of 519 Enucleated Eyes Swathi Kaliki, MD,1,2,3 Carol L. Shields, MD,1 Duangnate Rojanaporn, MD,1,4 Saad Al-Dahmash, MD,1,5 John P. McLaughlin, BS,1 Jerry A. Shields, MD,1 Ralph C. Eagle, Jr., MD2 Purpose: To determine the correlation between the International Classification of Retinoblastoma (ICRB) and histopathologic high-risk retinoblastoma. Design: Retrospective study. Participants: A total of 519 patients. Intervention: Primary enucleation. Main Outcome Measures: High-risk retinoblastoma, metastasis, and death. Results: Of 519 primarily enucleated eyes, 87 (17%) were classified as group D and 432 (83%) were classified as group E on the basis of the ICRB. High-risk retinoblastoma was identified in 23% (117/519) of enucleated eyes, including 17% (15/87) group D and 24% (102/432) group E eyes. High-risk histopathologic features of retinoblastoma included anterior chamber involvement (5/15 [33%] group D eyes, 31/102 [30%] group E eyes), isolated massive posterior uveal invasion ⱖ3 mm (7/15 [47%] group D eyes, 22/102 [22%] group E eyes), isolated post-laminar optic nerve invasion (2/15 [13%] group D eyes, 46/102 [45%] group E eyes), and any combination of posterior uveal invasion and optic nerve involvement (7/15 [47%] group D eyes, 37/102 [36%] group E eyes). On logistic regression analysis, massive posterior uveal invasion ⱖ3 mm was more common in group D eyes (P ⫽ 0.0442), and post-laminar optic nerve invasion was more common in group E eyes (P ⫽ 0.0390). Of 117 patients with high-risk retinoblastoma, systemic adjuvant chemotherapy was administered in 83 patients (71%). Systemic metastasis developed in 0% (0/15) of those with high-risk group D retinoblastoma and 10% (10/102) of those with high-risk group E retinoblastoma over a mean follow-up period of 78 months (median, 62 months; range, 1– 419 months). There was no metastasis in any patient (n ⫽ 402) classified with non– high-risk retinoblastoma. Of the 10 patients who developed metastasis, 4 had received prior adjuvant chemotherapy and 6 had no prior adjuvant chemotherapy. There was no metastasis in high-risk patients treated with vincristine sulphate, etoposide phosphate, and carboplatin (VEC). Death from metastasis occurred in 4% of high-risk patients (5/117). Conclusions: On the basis of the ICRB, 17% of group D and 24% of group E eyes are at increased risk for metastatic disease. In this study, 8% of patients developed metastasis. There was no metastasis in any patient classified with non– high-risk retinoblastoma. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Ophthalmology 2013;120:997–1003 © 2013 by the American Academy of Ophthalmology.
Retinoblastoma is the most common primary malignant intraocular tumor in children.1 Several tumor classification systems have been used to describe retinoblastoma.1,2 Tumor classification is a powerful clinical tool for assessment of local and systemic treatment outcomes. The Reese–Ellsworth (RE) classification was successful in estimating treatment success (salvage of the affected eye) after primary external beam radiotherapy. Hernandez et al3 studied 34 eyes treated with external beam radiotherapy and reported tumor control in 79% of eyes in groups I and II and 20% of eyes in groups III and V. The International Classi© 2013 by the American Academy of Ophthalmology Published by Elsevier Inc.
fication of Retinoblastoma (ICRB) is a proven classification useful for prediction of the success of chemoreduction for retinoblastoma. Shields et al4 analyzed 249 eyes with retinoblastoma treated with intravenous chemoreduction and found that success was achieved in 100% of group A, 93% of group B, 90% of group C, and 47% of group D eyes. A later study identified treatment success with intra-arterial chemotherapy (IAC) in 100% of group C and D eyes, and 33% of group E eyes.5 Enucleation is the treatment of choice for advanced retinoblastoma in an eye with little or no potential for sight, or ISSN 0161-6420/13/$–see front matter http://dx.doi.org/10.1016/j.ophtha.2012.10.044
997
Ophthalmology Volume 120, Number 5, May 2013 if there is a concern for invasion of retinoblastoma into the optic nerve, choroid, or orbit.6 After enucleation, histopathologic factors predictive of metastatic disease can assist in identifying high-risk patients who require adjuvant systemic chemotherapy to prevent metastatic disease.7,8 Proper management of children with high-risk retinoblastoma using chemotherapy (vincristine sulphate, etoposide phosphate, and carboplatin [VEC] for 6 monthly cycles) can prevent metastasis, as noted by Kaliki et al.8 More recently, however, there has been a trend toward globe salvage therapy for eyes with markedly advanced tumors using IAC, particularly those eyes in groups D and E.5 Impressive local ocular control has been achieved with little systemic effect.5,9,10 The lack of systemic effect is beneficial for most children but might be detrimental for those whose eyes harbor high-risk retinoblastoma because the therapy likely would be insufficient to prevent metastatic disease. In this analysis, we specifically explore the likelihood of high-risk retinoblastoma, metastasis, and death based on the ICRB, RE, and American Joint Committee on Cancer (AJCC) classifications.
Materials and Methods This was a retrospective, nonrandomized, interventional case series. Institutional review board approval was obtained. The medical records of all patients with retinoblastoma managed with enucleation on the Ocular Oncology Service at Wills Eye Institute in Philadelphia between January 1, 1975, and December 15, 2011, were reviewed. Patients who underwent primary enucleation for treatment of retinoblastoma were included in this study. Those who underwent secondary enucleation after failure of other treatment modalities were excluded. The histopathologic features of the enucleated specimen were reviewed. High-risk histopathologic features were defined as the presence of 1 or more of the following features: tumor invasion into the anterior chamber, an area of massive posterior uveal invasion ⱖ3 mm in diameter, post-laminar optic nerve invasion, or a combination of any nonmassive posterior uveal invasion (⬍3 mm in diameter) with any degree of nonretrolaminar optic nerve invasion.11 Clinical records were reviewed for clinical and histopathologic findings. The ICRB and RE classification were recorded in all enucleated eyes. In eyes with high-risk histopathologic features, tumor was classified on the basis of AJCC classification;12 additional clinical data included patient age (weeks), greatest basal dimension of tumor in millimeters, and tumor thickness in millimeters. Tumor findings were documented by large fundus drawings, fundus photography with the RetCam camera (Massie Industries, Dublin, CA), fluorescein angiography, and ultrasonography. The histopathologic data recorded included growth pattern (exophytic, endophytic, combined exophytic-endophytic), tumor differentiation (well differentiated, moderately differentiated, poorly differentiated, undifferentiated), and status of anterior chamber, iris, ciliary body, choroid, optic nerve, sclera, and extrascleral structures. Optic nerve invasion was classified as pre-laminar, laminar, post-laminar, or to the site of transection. Depth and lateral extent of choroidal invasion (millimeters) and depth of optic nerve invasion (millimeters) were recorded when applicable. In the eyes with high-risk histopathologic features, the number of highrisk features was recorded. In patients receiving systemic adjuvant chemotherapy, the details of chemotherapy drugs and number of cycles of chemotherapy were recorded. After chemotherapy, metastatic evaluation
998
included history and physical examination, and computed tomography or magnetic resonance imaging (MRI) of the orbit and brain at 6-month intervals until age 5 years and yearly thereafter. Duration of follow-up and final systemic outcome (alive without metastasis, alive with metastasis, dead from metastasis, or dead from other causes) were recorded.
Statistical Methods Data were summarized per the ICRB and RE classification for the categoric variables and compared between the stages using the Fisher exact test. The summaries for age, tumor base, and thickness were presented as mean, median, and range and compared between the stages by classification of retinoblastoma using independent-samples t test. Logistic regression analysis was performed to identify the high-risk features based on retinoblastoma classification. The factors found significant on univariate analysis at 0.10 level of significance were considered for multivariate analysis using the forward stepwise method. The factors significant at 5% level on multivariate analysis were reported. The association between high-risk retinoblastoma classification and prognosis was analyzed using the Fisher exact test. The factors predictive of each outcome (metastasis and death) were analyzed using the Cox proportional hazard model. Because there was no factor found significant on univariate analysis at the 0.10 level of significance, no multivariate analysis was performed. Kaplan– Meier analysis was performed to estimate the cumulative probability of metastasis and death at 1, 3, 5, 10, and 20 years in high-risk patients with retinoblastoma.
Results Of 639 eyes enucleated for retinoblastoma during this time period, 519 (81%) underwent primary enucleation and were included in this study. Secondary enucleation was performed in 120 eyes (19%), and these were excluded from the study. Of the 519 primarily enucleated eyes, 87 (17%) were classified as group D and 432 (83%) were classified as group E on the basis of the ICRB. There were no eyes in groups A, B, or C that underwent primary enucleation. According to RE classification, there were 2 (⬍1%) group IIa, 2 (⬍1%) group IIIa, 9 (2%) IIIb, 1 (⬍1%) group IVa, 4 (1%) group IVb, 112 (22%) group Va, and 389 (75%) group Vb eyes. Histopathologic study post-primary enucleation revealed high-risk retinoblastoma in 17% (15/87) group D and 24% (102/432) group E eyes by the ICRB. On the basis of the RE classification, high-risk features were identified in 20% (22/112) of group Va and 24% (95/389) of group Vb eyes (Table 1). Overall, 1 or more high-risk histopathologic features were present in 23% (117/519) of primarily enucleated eyes. On the basis of pathologic primary tumor, regional lymph node, metastasis classification of AJCC,12 there were 15% (17/117) pT1 with no optic nerve or choroidal invasion, 8% (9/117) pT2a with prelaminar/laminar optic nerve invasion or focal (⬍3 mm) choroidal invasion, 17% (20/117) pT2b with prelaminar/laminar optic nerve invasion and focal (⬍3 mm) choroidal invasion, 44% (52/117) pT3a with postlaminar optic nerve invasion or massive (ⱖ3 mm) choroidal invasion, 10% (12/117) pT3b with postlaminar optic nerve invasion and massive (ⱖ3 mm) choroidal invasion, 3% (4/117) pT4 with extraocular tumor extension, and 3% (3/117) pT4a with tumor invasion of optic nerve transection. By the ICRB, 13% (15/117) of group D eyes had high-risk features compared with 87% (102/117) of group E eyes. By RE classification, 19% (22/117) belonged to group Va and 81% (95/117) belonged to
Kaliki et al 䡠 High-Risk Retinoblastoma Table 1. Correlation between Retinoblastoma Classification and Histopathologic High-Risk Features in 519 Eyes with Retinoblastoma All Eyes with Retinoblastoma
Feature Histopathologic high-risk features (n ⫽ 519 eyes) Present Absent
n ⫽ 519 Eyes n (%)
International Classification of Retinoblastoma Group D n ⫽ 87 Eyes n (%)
Group E n ⫽ 432 Eyes n (%)
P Value*
Reese–Ellsworth Classification Va n ⫽ 112 Eyes n (%)
Vb n ⫽ 389 Eyes n (%)
0.2090 117 (23) 402 (77)
15 (17) 72 (83)
102 (24) 330 (76)
P Value* 0.3134
22 (20) 90 (80)
95 (24) 294 (76)
*Fisher exact test.
group Vb. There were no high-risk features in group II, III, or IV eyes. The demographic, clinical, and histopathologic features of eyes with high-risk retinoblastoma based on ICRB and RE classification are described in Table 2 (available at http://aaojournal. org). According to the ICRB classification, group E high-risk retinoblastoma eyes had a larger tumor base (16 vs. 22 mm; P⬍0.0001) and increased tumor thickness (9 vs. 13 mm; P ⫽ 0.0015) compared with group D. According to the RE classification, group Va high-risk retinoblastoma cases presented at an earlier age (65 vs. 139 weeks; P⬍0.0001) with exophytic growth pattern (68% vs. 18%; P⬍0.0001) and moderately differentiated retinoblastoma (14% vs. 1%; P ⫽ 0.0207) compared with group Vb. The histopathologic high-risk features based on ICRB and RE classification are listed in Tables 3 and 4, respectively (available at http://aaojournal.org). On the basis of multivariate analysis by logistic regression analysis according to ICRB, massive posterior
uveal invasion ⱖ3 mm was more common in group D eyes (50% vs. 22%; P ⫽ 0.0442), and post-laminar optic nerve invasion was more common in group E eyes (45% vs. 13%; P ⫽ 0.0390) (Fig 1). On the basis of RE classification, anterior chamber involvement was more common in group Vb (36% vs. 9%; P ⫽ 0.0247) compared with group Va. Of 117 patients with high-risk retinoblastoma, systemic adjuvant chemotherapy was administered in 83 (71%). Vincristine sulphate, etoposide phosphate, and carboplatin were administered in 60 patients; vincristine sulphate, cyclophosphamide, and doxorubicin hydrochloride (VCD) were administered in 10 patients; VCD and methotrexate were administered in 3 patients; and other regimens were administered in 10 patients. The mean number of cycles of adjuvant chemotherapy was 6 (median, 6 cycles; range, 4 – 6 cycles). Systemic metastasis developed in 8% of patients (10/117) with high-risk retinoblastoma. There were no metastatic events (0/15) in those with high-risk group D retinoblastoma, and
Figure 1. A 27-month-old boy presented with (A) group D retinoblastoma with solid tumor measuring 24 mm in the base and 12 mm in thickness and with prominent vitreous seeding. B, Gross examination confirmed clinical findings with an area of massive choroidal invasion. C, Grossly visible massive choroidal invasion (arrow). D, Histopathologic evidence of 9⫻1.2-mm massive choroidal invasion (hematoxylin– eosin stain, magnification 5⫻). A 12-month-old girl presented with (E) group E retinoblastoma with solid tumor measuring 24 mm in the base and 20 mm in thickness. There was total retinal detachment and extensive subretinal and vitreous seeding. F, Gross examination confirmed clinical findings with retinoblastoma filling the entire globe. G, Tumor in proximity to the widened optic nerve. H, On hematoxylin– eosin staining, there was histopathologic confirmation of 5 mm of postlaminar optic nerve invasion (hematoxylin– eosin stain, magnification 5⫻).
999
Ophthalmology Volume 120, Number 5, May 2013 Table 5. Association of International Classification of Retinoblastoma, Reese–Ellsworth Classification, and American Joint Committee on Cancer Classification with Metastasis and Death in 117 Patients with High-Risk Retinoblastoma Classification of Retinoblastoma
Patients with HighRisk Features n (%)
Metastasis n (%)
International Classification of Retinoblastoma Group D Group E RE classification Group Va Group Vb pTNM classification based on AJCC pT1† pT2a pT2b pT3a pT3b pT4* pT4a
n ⫽ 117 Patients n (%)
n ⫽ 10 Patients n (%)
15 (13) 102 (87)
0 (0) 10 (10)
0.357
0 (0) 5 (5)
1.000
21 (18) 96 (82)
2 (10) 8 (8)
1.000
1 (5) 4 (4)
1.000
17 (15) 9 (8) 20 (17) 52 (44) 12 (10) 4 (3) 3 (3)
1 (12) 1 (11) 0 (0) 7 (13) 1 (8) 0 (0) 0 (0)
1.000 1.000 0.208 0.107 1.000 1.000 1.000
1 (6) 0 (0) 0 (0) 3 (6) 1 (8) 0 (0) 0 (0)
0.557 1.000 0.590 0.654 0.424 1.000 1.000
P Value
Death n (%)
P Value
n ⫽ 5 Patients n (%)
AJCC ⫽ American Joint Committee on Cancer Classification; pTNM ⫽ pathologic primary tumor, regional lymph node, metastasis; RE ⫽ Reese– Ellsworth. *Extraocular tumor extension but not involving the optic nerve transection. † No optic nerve or choroidal invasion but with anterior chamber invasion.
10% metastatic events (10/102) in those with high-risk group E retinoblastoma over a mean follow-up period of 78 months (median, 62 months; range, 1– 419 months). There was no metastasis in any patient (n ⫽ 402 patients) classified with non– high-risk retinoblastoma. Systemic metastases based on ICRB, RE, and AJCC classifications are listed in Table 5. The mean time interval from diagnosis of high-risk retinoblastoma to metastasis was 13 months (median, 9 months; range, 1–51 months). The site of metastasis included bone (n ⫽ 4), brain (n ⫽ 3), and bone and brain (n ⫽ 1). Details of metastatic site were unavailable in 2 patients. Of the 10 patients who developed metastasis, 4 had received prior adjuvant chemotherapy with VCD (n ⫽ 3) and VCD and methotrexate (n ⫽ 1), whereas 6 patients had no prior adjuvant chemotherapy. No metastasis occurred in patients treated by VEC. Death due to metastasis occurred in 4% of patients (5/117) with high-risk retinoblastoma. Deaths based on ICRB, RE, and AJCC classifications are listed in Table 5. The mean time interval from detection of metastasis to death was 19 months (median, 24 months; range, 8 –24 months). On the basis of univariate analysis, there were no factors predictive of metastasis or death. The high-risk features in the patients with subsequent metastasis included postlaminar optic nerve invasion (n ⫽ 5), massive choroidal invasion (n ⫽ 2), combined postlaminar optic nerve invasion and massive choroidal
invasion (n ⫽ 1), anterior chamber invasion with a history of intraocular surgery (n⫽1), and anterior chamber invasion and nonmassive choroidal invasion (n ⫽ 1). Kaplan–Meier analysis of metastasis and death in patients with high-risk retinoblastoma was 7% and 2% at 1 year, 10% and 5% at 5 years, and 10% and 5% at 10 years, respectively (Table 6; Figs 2 and 3).
Discussion High-risk retinoblastoma detected on histopathologic examination of enucleated eyes can identify children at risk for
Table 6. Kaplan–Meier Analysis of Metastasis and Death in 117 Patients with High-Risk Retinoblastoma
1 yr 2 yrs 3 yrs 5 yrs 10 yrs 15 yrs 20 yrs
1000
Metastasis
Death
6.7⫾2.5 (7/92) 8.8⫾2.8 (9/83) 8.8⫾2.8 (9/78) 10.1⫾3.1 (10/61) 10.1⫾3.1 (10/23) 10.1⫾3.1 (10/12) 10.1⫾3.1 (10/3)
2.0⫾1.4 (2/96) 3.0⫾1.7 (3/87) 5.3⫾2.3 (5/80) 5.3⫾2.3 (5/63) 5.3⫾2.3 (5/24) 5.3⫾2.3 (5/13) 5.3⫾2.3 (5/4)
Figure 2. Kaplan–Meier analysis of metastasis in patients with high-risk retinoblastoma. Proportion of metastasis [# metastasis / # Left] ⫽ number of patients with metastasis/number of patients at risk for metastasis.
Kaliki et al 䡠 High-Risk Retinoblastoma
Figure 3. Kaplan–Meier analysis of death from metastasis in patients with high-risk retinoblastoma. Proportion of Death [# Died/ # Left] ⫽ number of deaths/number of patients at risk for death.
metastatic disease. High-risk features include tumor invasion into the anterior chamber, an area of posterior uveal invasion ⱖ3 mm, post-laminar optic nerve invasion, or a combination of any nonmassive posterior uveal invasion (⬍3 mm) with any degree of nonretrolaminar optic nerve invasion.11 Wilson and associates13 explored the relationship of ICRB classification with high-risk retinoblastoma in 67 eyes. They found high-risk histopathologic features in 15% (7/47) of group D and 50% (10/20) of group E eyes.13 On the basis of RE classification, high-risk retinoblastoma involved 16% (3/19) of group Va and 29% (14/48) of group Vb eyes.13 They found that group E eyes were more commonly associated with optic nerve (P ⫽ 0.026), choroid (P⬍0.001), ciliary body (P ⫽ 0.002), iris (P ⫽ 0.002), anterior chamber (P ⫽ 0.025), and scleral (P⬍0.001) invasion compared with group D eyes. The RE classification had no significant association with high-risk retinoblastoma.13 In our larger, more comprehensive study of 519 primarily enucleated eyes, high-risk histopathologic features were evident in 17% (15/87) of group D eyes and 24% (102/432) of group E eyes on the basis of the ICRB, and 20% (20/112) group Va and 24% (95/389) group Vb eyes on the basis of RE classification. Our study of 519 eyes was comparable to the previous study13 of 67 eyes with regard to the frequency of invasion for ICRB group D eyes. However, group E eyes manifested high-risk disease in 24% in our study compared with 50% in the previous study.13 This difference could be related to cohort size because our study had 432 group E eyes and the previous study14 had only 20 eyes. In addition, in contrast to the study by Wilson et al,13 massive posterior uveal invasion ⱖ3 mm was more common in ICRB group D eyes, post-laminar optic nerve invasion was more common in ICRB group E eyes, and anterior chamber involvement was more common in RE group Vb eyes. We believe that these dissimilarities could be attributed to a difference in cohort size.
On the basis of our results, both advanced group D (17%) and group E (24%) eyes are at increased risk for associated metastatic disease. The ICRB classification indicates that both groups carry the potential for high-risk retinoblastoma, and there is no statistically significant difference between the 2 groups (P ⫽ 0.2090) (Table 1). There are clinical features that predict histopathologic high-risk retinoblastoma. Shields and coworkers14,15 specifically evaluated clinical factors predictive of optic nerve invasion14 or choroidal invasion.15 Of 289 eyes, 84 (29%) demonstrated optic nerve invasion and 67 (23%) had choroidal invasion.14,15 Clinical factors predictive of optic nerve invasion included secondary glaucoma (P ⫽ 0.02), exophytic growth pattern (0.011), and tumor thickness ⱖ15 mm (P ⫽ 0.03).14 The main clinical risk feature predictive of choroidal invasion was iris neovascularization (P ⫽ 0.02).15 Chantada et al16 found high-risk retinoblastoma in 73 of 182 cases (40%), and clinical risk factors predicting high-risk retinoblastoma included glaucoma (P ⫽ 0.025) and buphthalmos (P ⫽ 0.00017). In contrast, Wilson and associates13 found no association between glaucoma and buphthalmos with highrisk retinoblastoma. Other clinical features predictive of high-risk retinoblastoma include older age at presentation (⬎2 years), longer duration lag period from symptoms to diagnosis (⬎3 months), hyphema, pseudohypopyon, staphyloma, and orbital cellulitis.17 Orbital MRI also can provide a correlation with high-risk retinoblastoma.18 –20 The accuracy, specificity, and sensitivity of MRI detection of high-risk retinoblastoma was 91%, 95%, and 60% for postlaminar invasion, respectively,19 and 68%, 80%, and 60% for choroidal invasion, respectively.18 The prediction and diagnosis of high-risk retinoblastoma are crucial because untreated high-risk retinoblastoma carries at least 24% risk for metastatic disease,7 and adjuvant systemic chemotherapy reduces the risk to 0% to 4%.7,8 In the present study, 18% (6/34) of patients with untreated high-risk retinoblastoma developed metastasis, and 5% (4/ 83) of patients with high-risk retinoblastoma who were treated with chemotherapy developed metastasis. Of the 402 patients with non– high-risk retinoblastoma, there were no metastatic events. Similar to a previous study by our group,8 there was no metastasis in high-risk patients treated with VEC. Metastasis from retinoblastoma generally occurs within 1 year of diagnosis of retinoblastoma.21–23 If there is no metastatic disease by 5 years after treatment of retinoblastoma, the patient is usually considered cured.21–23 In our study, 70% (7/10), 90% (9/10), and 100% (10/10) of patients displayed metastasis within 1 year, 2 years, and 5 years of diagnosis of retinoblastoma, respectively. The prognosis of metastatic retinoblastoma is poor, with death usually occurring within 6 months.21–23 In our study, all (n ⫽ 5) metastasis-related deaths occurred within 2 years of diagnosis of metastatic retinoblastoma. By Kaplan–Meier analysis, the probability of metastasis and death after 5 years of diagnosis of retinoblastoma remained stable at 10% and 5%, respectively. Infrequent number of events (metastasis and death) limited the statistical analysis of factors predictive of metastasis and death. However, all 10 patients developing metastasis had advanced retinoblastoma, with
1001
Ophthalmology Volume 120, Number 5, May 2013 100% (10/10) classified as group E according to the ICRB, 80% (8/10) classified as group Vb according to the RE classification, and 70% (7/10) classified as pT3a according to the AJCC classification. The clinical13–18 and MRI19,20 features can predict highrisk retinoblastoma, but ICRB and RE classifications provide limited correlation with high-risk retinoblastoma. The definitive diagnosis of high-risk retinoblastoma is achieved only with histopathologic inspection of the enucleated globe. It is clearly shown that systemic chemotherapy can inactivate invasive high-risk tumor and offer protection against systemic micrometastasis.7,8 However, the usefulness of IAC as a targeted drug directly to the eye and with minimal systemic effect might not offer protection against systemic disease in such cases. Even though excellent globe salvage is achieved with IAC for group D eyes, and moderate globe success is achieved for group E eyes,5 it should be realized that 17% of group D and 24% of group E eyes would be expected to harbor high-risk retinoblastoma and combining this information with the 24% risk of metastasis,7 one might anticipate that 4% of group D and 6% group E eyes treated without systemic chemotherapy might develop metastasis. The prediction of high-risk retinoblastoma based on the clinical, ultrasonography, and MRI features is pivotal in the selection of appropriate treatment in these cases. In conclusion, the ICRB can predict high-risk histopathologic features of retinoblastoma. In this study, histopathologic evidence of high-risk retinoblastoma was noted in 17% of group D eyes and 24% of group E eyes. This finding is useful for counseling families regarding the probable need of adjuvant chemotherapy after enucleation for retinoblastoma, and this information adds caution to the conservative management of advanced retinoblastoma with targeted therapies. Acknowledgment. Statistical analysis was provided by Rishita Nutheti, PhD, Hyderabad, India.
References 1. Shields JA, Shields CL. Retinoblastoma. In: Shields JA, Shields CL, eds. Intraocular Tumors: An Atlas and Textbook. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2008:293–318. 2. Ramasubramanian A, Shields CL. Staging and treatment strategies. In: Ramasubramanian A, Shields CL, eds. Retinoblastoma. New Delhi, India: Jaypee Brothers Medical Publishers; 2012:70 – 8. 3. Hernandez JC, Brady LW, Shields JA, et al. External beam radiation for retinoblastoma: results, patterns of failure, and proposal for treatment guidelines. Int J Radiat Oncol Biol Phys 1996;35:125–32. 4. Shields CL, Mashayekhi A, Au AK, et al. The International Classification of Retinoblastoma predicts chemoreduction success. Ophthalmology 2006;113:2276 – 80.
1002
5. Shields CL, Bianciotto CG, Jabbour P, et al. Intra-arterial chemotherapy for retinoblastoma: report no. 1, control of retinal tumors, subretinal seeds, and vitreous seeds. Arch Ophthalmol 2011;129:1399 – 406. 6. Shields CL, Shields JA. Retinoblastoma management: advances in enucleation, intravenous chemoreduction, and intra-arterial chemotherapy. Curr Opin Ophthalmol 2010;21:203–12. 7. Honavar SG, Singh AD, Shields CL, et al. Postenucleation adjuvant therapy in high-risk retinoblastoma. Arch Ophthalmol 2002;120:923–31. 8. Kaliki S, Shields CL, Shah SU, et al. Postenucleation adjuvant chemotherapy with vincristine, etoposide, and carboplatin for the treatment of high-risk retinoblastoma. Arch Ophthalmol 2011;129:1422–7. 9. Shields CL, Bianciotto CG, Jabbour P, et al. Intra-arterial chemotherapy for retinoblastoma: report no. 2, treatment complications. Arch Ophthalmol 2011;129:1407–15. 10. Gobin YP, Dunkel IJ, Marr BP, et al. Intra-arterial chemotherapy for the management of retinoblastoma: four-year experience. Arch Ophthalmol 2011;129:732–7. 11. Eagle RC Jr. High-risk features and tumor differentiation in retinoblastoma: a retrospective histopathologic study. Arch Pathol Lab Med 2009;133:1203–9. 12. Edge SB, Byrd DR, Compton CC, et al, eds. Retinoblastoma. In: AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2010:560 – 8. 13. Wilson MW, Qaddoumi I, Billups C, et al. A clinicopathological correlation of 67 eyes primarily enucleated for advanced intraocular retinoblastoma. Br J Ophthalmol 2011;95:553– 8. 14. Shields CL, Shields JA, Baez K, et al. Optic nerve invasion of retinoblastoma: metastatic potential and clinical risk factors. Cancer 1994;73:692– 8. 15. Shields CL, Shields JA, Baez KA, et al. Choroidal invasion of retinoblastoma: metastatic potential and clinical risk factors. Br J Ophthalmol 1993;77:544 – 8. 16. Chantada GL, Gonzalez A, Fandino A, et al. Some clinical findings at presentation can predict high-risk pathology features in unilateral retinoblastoma. J Pediatr Hematol Oncol 2009;31:325–9. 17. Kashyap S, Meel R, Puskker N, et al. Clinical predictors of high risk histopathology in retinoblastoma. Pediatr Blood Cancer 2012;58:356 – 61. 18. Chawla B, Sharma S, Sen S, et al. Correlation between clinical features, magnetic resonance imaging, and histopathologic findings in retinoblastoma: a prospective study. Ophthalmology 2012;119:850 – 6. 19. Brisse HJ, Guesmi M, Aerts I, et al. Relevance of CT and MRI in retinoblastoma for the diagnosis of postlaminar invasion with normal-size optic nerve: a retrospective study of 150 patients with histological comparison. Pediatr Radiol 2007;37: 649 –56. 20. de Graaf P, Moll AC, Imhof SM, et al. Retinoblastoma and optic nerve enhancement on MRI: not always extraocular tumour extension [letter]. Br J Ophthalmol 2006;90:800 –1. 21. MacKay CJ, Abramson DH, Ellsworth RM. Metastatic patterns of retinoblastoma. Arch Ophthalmol 1984;102:391– 6. 22. Kopelman JE, McLean IW, Rosenberg SH. Multivariate analysis of risk factors for metastasis in retinoblastoma treated by enucleation. Ophthalmology 1987;94:371–7. 23. Broaddus E, Topham A, Singh AD. Survival with retinoblastoma in the USA: 1975-2004. Br J Ophthalmol 2009;93:24 –7.
Kaliki et al 䡠 High-Risk Retinoblastoma
Footnotes and Financial Disclosures Originally received: May 7, 2012. Final revision: September 13, 2012. Accepted: October 31, 2012. Available online: February 8, 2013.
Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Manuscript no. 2012-658.
1
Ocular Oncology Service, Wills Eye Institute, Thomas Jefferson University, Philadelphia, Pennsylvania.
2
Department of Pathology, Wills Eye Institute, Thomas Jefferson University, Philadelphia, Pennsylvania.
3
Ocular Oncology Service, L.V. Prasad Eye Institute, Hyderabad, India.
4
Department of Ophthalmology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand.
5
Department of Ophthalmology, College of Medicine, King Saud University, Saudi Arabia.
Support was provided by Carlos G. Bianciotto Retinoblastoma Research Fund c/o Eye Tumor Research Foundation, Philadelphia, PA (C.L.S., J.A.S.); The Lucille Wiedman Fund for Pediatric Eye Cancer, Philadelphia, PA (C.L.S., J.A.S.); Lift for a Cure, Morrisdale, PA (C.L.S.); and the Noel T. and Sara L. Simmonds Endowment for Ophthalmic Pathology, Wills Eye Institute (R.C.E.). The funders had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; and in the preparation, review, or approval of the manuscript. Dr. Shields has 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. Correspondence: Carol L. Shields, MD, Ocular Oncology Service, Suite 1440, Wills Eye Institute, 840 Walnut Street, Philadelphia, PA 19107. E-mail: carol.shields@ shieldsoncology.com.
1003
Retinoblastoma is the most common primary malignant intraocular tumor in children.1 Several tumor classification systems have been used to describe retinoblastoma.1,2 Tumor classification is a powerful clinical tool for assessment of local and systemic treatment outcomes. The Reese–Ellsworth (RE) classification was successful in estimating treatment success (salvage of the affected eye) after primary external beam radiotherapy. Hernandez et al3 studied 34 eyes treated with external beam radiotherapy and reported tumor control in 79% of eyes in groups I and II and 20% of eyes in groups III and V. The International Classi© 2013 by the American Academy of Ophthalmology Published by Elsevier Inc.
fication of Retinoblastoma (ICRB) is a proven classification useful for prediction of the success of chemoreduction for retinoblastoma. Shields et al4 analyzed 249 eyes with retinoblastoma treated with intravenous chemoreduction and found that success was achieved in 100% of group A, 93% of group B, 90% of group C, and 47% of group D eyes. A later study identified treatment success with intra-arterial chemotherapy (IAC) in 100% of group C and D eyes, and 33% of group E eyes.5 Enucleation is the treatment of choice for advanced retinoblastoma in an eye with little or no potential for sight, or ISSN 0161-6420/13/$–see front matter http://dx.doi.org/10.1016/j.ophtha.2012.10.044
997
Ophthalmology Volume 120, Number 5, May 2013 if there is a concern for invasion of retinoblastoma into the optic nerve, choroid, or orbit.6 After enucleation, histopathologic factors predictive of metastatic disease can assist in identifying high-risk patients who require adjuvant systemic chemotherapy to prevent metastatic disease.7,8 Proper management of children with high-risk retinoblastoma using chemotherapy (vincristine sulphate, etoposide phosphate, and carboplatin [VEC] for 6 monthly cycles) can prevent metastasis, as noted by Kaliki et al.8 More recently, however, there has been a trend toward globe salvage therapy for eyes with markedly advanced tumors using IAC, particularly those eyes in groups D and E.5 Impressive local ocular control has been achieved with little systemic effect.5,9,10 The lack of systemic effect is beneficial for most children but might be detrimental for those whose eyes harbor high-risk retinoblastoma because the therapy likely would be insufficient to prevent metastatic disease. In this analysis, we specifically explore the likelihood of high-risk retinoblastoma, metastasis, and death based on the ICRB, RE, and American Joint Committee on Cancer (AJCC) classifications.
Materials and Methods This was a retrospective, nonrandomized, interventional case series. Institutional review board approval was obtained. The medical records of all patients with retinoblastoma managed with enucleation on the Ocular Oncology Service at Wills Eye Institute in Philadelphia between January 1, 1975, and December 15, 2011, were reviewed. Patients who underwent primary enucleation for treatment of retinoblastoma were included in this study. Those who underwent secondary enucleation after failure of other treatment modalities were excluded. The histopathologic features of the enucleated specimen were reviewed. High-risk histopathologic features were defined as the presence of 1 or more of the following features: tumor invasion into the anterior chamber, an area of massive posterior uveal invasion ⱖ3 mm in diameter, post-laminar optic nerve invasion, or a combination of any nonmassive posterior uveal invasion (⬍3 mm in diameter) with any degree of nonretrolaminar optic nerve invasion.11 Clinical records were reviewed for clinical and histopathologic findings. The ICRB and RE classification were recorded in all enucleated eyes. In eyes with high-risk histopathologic features, tumor was classified on the basis of AJCC classification;12 additional clinical data included patient age (weeks), greatest basal dimension of tumor in millimeters, and tumor thickness in millimeters. Tumor findings were documented by large fundus drawings, fundus photography with the RetCam camera (Massie Industries, Dublin, CA), fluorescein angiography, and ultrasonography. The histopathologic data recorded included growth pattern (exophytic, endophytic, combined exophytic-endophytic), tumor differentiation (well differentiated, moderately differentiated, poorly differentiated, undifferentiated), and status of anterior chamber, iris, ciliary body, choroid, optic nerve, sclera, and extrascleral structures. Optic nerve invasion was classified as pre-laminar, laminar, post-laminar, or to the site of transection. Depth and lateral extent of choroidal invasion (millimeters) and depth of optic nerve invasion (millimeters) were recorded when applicable. In the eyes with high-risk histopathologic features, the number of highrisk features was recorded. In patients receiving systemic adjuvant chemotherapy, the details of chemotherapy drugs and number of cycles of chemotherapy were recorded. After chemotherapy, metastatic evaluation
998
included history and physical examination, and computed tomography or magnetic resonance imaging (MRI) of the orbit and brain at 6-month intervals until age 5 years and yearly thereafter. Duration of follow-up and final systemic outcome (alive without metastasis, alive with metastasis, dead from metastasis, or dead from other causes) were recorded.
Statistical Methods Data were summarized per the ICRB and RE classification for the categoric variables and compared between the stages using the Fisher exact test. The summaries for age, tumor base, and thickness were presented as mean, median, and range and compared between the stages by classification of retinoblastoma using independent-samples t test. Logistic regression analysis was performed to identify the high-risk features based on retinoblastoma classification. The factors found significant on univariate analysis at 0.10 level of significance were considered for multivariate analysis using the forward stepwise method. The factors significant at 5% level on multivariate analysis were reported. The association between high-risk retinoblastoma classification and prognosis was analyzed using the Fisher exact test. The factors predictive of each outcome (metastasis and death) were analyzed using the Cox proportional hazard model. Because there was no factor found significant on univariate analysis at the 0.10 level of significance, no multivariate analysis was performed. Kaplan– Meier analysis was performed to estimate the cumulative probability of metastasis and death at 1, 3, 5, 10, and 20 years in high-risk patients with retinoblastoma.
Results Of 639 eyes enucleated for retinoblastoma during this time period, 519 (81%) underwent primary enucleation and were included in this study. Secondary enucleation was performed in 120 eyes (19%), and these were excluded from the study. Of the 519 primarily enucleated eyes, 87 (17%) were classified as group D and 432 (83%) were classified as group E on the basis of the ICRB. There were no eyes in groups A, B, or C that underwent primary enucleation. According to RE classification, there were 2 (⬍1%) group IIa, 2 (⬍1%) group IIIa, 9 (2%) IIIb, 1 (⬍1%) group IVa, 4 (1%) group IVb, 112 (22%) group Va, and 389 (75%) group Vb eyes. Histopathologic study post-primary enucleation revealed high-risk retinoblastoma in 17% (15/87) group D and 24% (102/432) group E eyes by the ICRB. On the basis of the RE classification, high-risk features were identified in 20% (22/112) of group Va and 24% (95/389) of group Vb eyes (Table 1). Overall, 1 or more high-risk histopathologic features were present in 23% (117/519) of primarily enucleated eyes. On the basis of pathologic primary tumor, regional lymph node, metastasis classification of AJCC,12 there were 15% (17/117) pT1 with no optic nerve or choroidal invasion, 8% (9/117) pT2a with prelaminar/laminar optic nerve invasion or focal (⬍3 mm) choroidal invasion, 17% (20/117) pT2b with prelaminar/laminar optic nerve invasion and focal (⬍3 mm) choroidal invasion, 44% (52/117) pT3a with postlaminar optic nerve invasion or massive (ⱖ3 mm) choroidal invasion, 10% (12/117) pT3b with postlaminar optic nerve invasion and massive (ⱖ3 mm) choroidal invasion, 3% (4/117) pT4 with extraocular tumor extension, and 3% (3/117) pT4a with tumor invasion of optic nerve transection. By the ICRB, 13% (15/117) of group D eyes had high-risk features compared with 87% (102/117) of group E eyes. By RE classification, 19% (22/117) belonged to group Va and 81% (95/117) belonged to
Kaliki et al 䡠 High-Risk Retinoblastoma Table 1. Correlation between Retinoblastoma Classification and Histopathologic High-Risk Features in 519 Eyes with Retinoblastoma All Eyes with Retinoblastoma
Feature Histopathologic high-risk features (n ⫽ 519 eyes) Present Absent
n ⫽ 519 Eyes n (%)
International Classification of Retinoblastoma Group D n ⫽ 87 Eyes n (%)
Group E n ⫽ 432 Eyes n (%)
P Value*
Reese–Ellsworth Classification Va n ⫽ 112 Eyes n (%)
Vb n ⫽ 389 Eyes n (%)
0.2090 117 (23) 402 (77)
15 (17) 72 (83)
102 (24) 330 (76)
P Value* 0.3134
22 (20) 90 (80)
95 (24) 294 (76)
*Fisher exact test.
group Vb. There were no high-risk features in group II, III, or IV eyes. The demographic, clinical, and histopathologic features of eyes with high-risk retinoblastoma based on ICRB and RE classification are described in Table 2 (available at http://aaojournal. org). According to the ICRB classification, group E high-risk retinoblastoma eyes had a larger tumor base (16 vs. 22 mm; P⬍0.0001) and increased tumor thickness (9 vs. 13 mm; P ⫽ 0.0015) compared with group D. According to the RE classification, group Va high-risk retinoblastoma cases presented at an earlier age (65 vs. 139 weeks; P⬍0.0001) with exophytic growth pattern (68% vs. 18%; P⬍0.0001) and moderately differentiated retinoblastoma (14% vs. 1%; P ⫽ 0.0207) compared with group Vb. The histopathologic high-risk features based on ICRB and RE classification are listed in Tables 3 and 4, respectively (available at http://aaojournal.org). On the basis of multivariate analysis by logistic regression analysis according to ICRB, massive posterior
uveal invasion ⱖ3 mm was more common in group D eyes (50% vs. 22%; P ⫽ 0.0442), and post-laminar optic nerve invasion was more common in group E eyes (45% vs. 13%; P ⫽ 0.0390) (Fig 1). On the basis of RE classification, anterior chamber involvement was more common in group Vb (36% vs. 9%; P ⫽ 0.0247) compared with group Va. Of 117 patients with high-risk retinoblastoma, systemic adjuvant chemotherapy was administered in 83 (71%). Vincristine sulphate, etoposide phosphate, and carboplatin were administered in 60 patients; vincristine sulphate, cyclophosphamide, and doxorubicin hydrochloride (VCD) were administered in 10 patients; VCD and methotrexate were administered in 3 patients; and other regimens were administered in 10 patients. The mean number of cycles of adjuvant chemotherapy was 6 (median, 6 cycles; range, 4 – 6 cycles). Systemic metastasis developed in 8% of patients (10/117) with high-risk retinoblastoma. There were no metastatic events (0/15) in those with high-risk group D retinoblastoma, and
Figure 1. A 27-month-old boy presented with (A) group D retinoblastoma with solid tumor measuring 24 mm in the base and 12 mm in thickness and with prominent vitreous seeding. B, Gross examination confirmed clinical findings with an area of massive choroidal invasion. C, Grossly visible massive choroidal invasion (arrow). D, Histopathologic evidence of 9⫻1.2-mm massive choroidal invasion (hematoxylin– eosin stain, magnification 5⫻). A 12-month-old girl presented with (E) group E retinoblastoma with solid tumor measuring 24 mm in the base and 20 mm in thickness. There was total retinal detachment and extensive subretinal and vitreous seeding. F, Gross examination confirmed clinical findings with retinoblastoma filling the entire globe. G, Tumor in proximity to the widened optic nerve. H, On hematoxylin– eosin staining, there was histopathologic confirmation of 5 mm of postlaminar optic nerve invasion (hematoxylin– eosin stain, magnification 5⫻).
999
Ophthalmology Volume 120, Number 5, May 2013 Table 5. Association of International Classification of Retinoblastoma, Reese–Ellsworth Classification, and American Joint Committee on Cancer Classification with Metastasis and Death in 117 Patients with High-Risk Retinoblastoma Classification of Retinoblastoma
Patients with HighRisk Features n (%)
Metastasis n (%)
International Classification of Retinoblastoma Group D Group E RE classification Group Va Group Vb pTNM classification based on AJCC pT1† pT2a pT2b pT3a pT3b pT4* pT4a
n ⫽ 117 Patients n (%)
n ⫽ 10 Patients n (%)
15 (13) 102 (87)
0 (0) 10 (10)
0.357
0 (0) 5 (5)
1.000
21 (18) 96 (82)
2 (10) 8 (8)
1.000
1 (5) 4 (4)
1.000
17 (15) 9 (8) 20 (17) 52 (44) 12 (10) 4 (3) 3 (3)
1 (12) 1 (11) 0 (0) 7 (13) 1 (8) 0 (0) 0 (0)
1.000 1.000 0.208 0.107 1.000 1.000 1.000
1 (6) 0 (0) 0 (0) 3 (6) 1 (8) 0 (0) 0 (0)
0.557 1.000 0.590 0.654 0.424 1.000 1.000
P Value
Death n (%)
P Value
n ⫽ 5 Patients n (%)
AJCC ⫽ American Joint Committee on Cancer Classification; pTNM ⫽ pathologic primary tumor, regional lymph node, metastasis; RE ⫽ Reese– Ellsworth. *Extraocular tumor extension but not involving the optic nerve transection. † No optic nerve or choroidal invasion but with anterior chamber invasion.
10% metastatic events (10/102) in those with high-risk group E retinoblastoma over a mean follow-up period of 78 months (median, 62 months; range, 1– 419 months). There was no metastasis in any patient (n ⫽ 402 patients) classified with non– high-risk retinoblastoma. Systemic metastases based on ICRB, RE, and AJCC classifications are listed in Table 5. The mean time interval from diagnosis of high-risk retinoblastoma to metastasis was 13 months (median, 9 months; range, 1–51 months). The site of metastasis included bone (n ⫽ 4), brain (n ⫽ 3), and bone and brain (n ⫽ 1). Details of metastatic site were unavailable in 2 patients. Of the 10 patients who developed metastasis, 4 had received prior adjuvant chemotherapy with VCD (n ⫽ 3) and VCD and methotrexate (n ⫽ 1), whereas 6 patients had no prior adjuvant chemotherapy. No metastasis occurred in patients treated by VEC. Death due to metastasis occurred in 4% of patients (5/117) with high-risk retinoblastoma. Deaths based on ICRB, RE, and AJCC classifications are listed in Table 5. The mean time interval from detection of metastasis to death was 19 months (median, 24 months; range, 8 –24 months). On the basis of univariate analysis, there were no factors predictive of metastasis or death. The high-risk features in the patients with subsequent metastasis included postlaminar optic nerve invasion (n ⫽ 5), massive choroidal invasion (n ⫽ 2), combined postlaminar optic nerve invasion and massive choroidal
invasion (n ⫽ 1), anterior chamber invasion with a history of intraocular surgery (n⫽1), and anterior chamber invasion and nonmassive choroidal invasion (n ⫽ 1). Kaplan–Meier analysis of metastasis and death in patients with high-risk retinoblastoma was 7% and 2% at 1 year, 10% and 5% at 5 years, and 10% and 5% at 10 years, respectively (Table 6; Figs 2 and 3).
Discussion High-risk retinoblastoma detected on histopathologic examination of enucleated eyes can identify children at risk for
Table 6. Kaplan–Meier Analysis of Metastasis and Death in 117 Patients with High-Risk Retinoblastoma
1 yr 2 yrs 3 yrs 5 yrs 10 yrs 15 yrs 20 yrs
1000
Metastasis
Death
6.7⫾2.5 (7/92) 8.8⫾2.8 (9/83) 8.8⫾2.8 (9/78) 10.1⫾3.1 (10/61) 10.1⫾3.1 (10/23) 10.1⫾3.1 (10/12) 10.1⫾3.1 (10/3)
2.0⫾1.4 (2/96) 3.0⫾1.7 (3/87) 5.3⫾2.3 (5/80) 5.3⫾2.3 (5/63) 5.3⫾2.3 (5/24) 5.3⫾2.3 (5/13) 5.3⫾2.3 (5/4)
Figure 2. Kaplan–Meier analysis of metastasis in patients with high-risk retinoblastoma. Proportion of metastasis [# metastasis / # Left] ⫽ number of patients with metastasis/number of patients at risk for metastasis.
Kaliki et al 䡠 High-Risk Retinoblastoma
Figure 3. Kaplan–Meier analysis of death from metastasis in patients with high-risk retinoblastoma. Proportion of Death [# Died/ # Left] ⫽ number of deaths/number of patients at risk for death.
metastatic disease. High-risk features include tumor invasion into the anterior chamber, an area of posterior uveal invasion ⱖ3 mm, post-laminar optic nerve invasion, or a combination of any nonmassive posterior uveal invasion (⬍3 mm) with any degree of nonretrolaminar optic nerve invasion.11 Wilson and associates13 explored the relationship of ICRB classification with high-risk retinoblastoma in 67 eyes. They found high-risk histopathologic features in 15% (7/47) of group D and 50% (10/20) of group E eyes.13 On the basis of RE classification, high-risk retinoblastoma involved 16% (3/19) of group Va and 29% (14/48) of group Vb eyes.13 They found that group E eyes were more commonly associated with optic nerve (P ⫽ 0.026), choroid (P⬍0.001), ciliary body (P ⫽ 0.002), iris (P ⫽ 0.002), anterior chamber (P ⫽ 0.025), and scleral (P⬍0.001) invasion compared with group D eyes. The RE classification had no significant association with high-risk retinoblastoma.13 In our larger, more comprehensive study of 519 primarily enucleated eyes, high-risk histopathologic features were evident in 17% (15/87) of group D eyes and 24% (102/432) of group E eyes on the basis of the ICRB, and 20% (20/112) group Va and 24% (95/389) group Vb eyes on the basis of RE classification. Our study of 519 eyes was comparable to the previous study13 of 67 eyes with regard to the frequency of invasion for ICRB group D eyes. However, group E eyes manifested high-risk disease in 24% in our study compared with 50% in the previous study.13 This difference could be related to cohort size because our study had 432 group E eyes and the previous study14 had only 20 eyes. In addition, in contrast to the study by Wilson et al,13 massive posterior uveal invasion ⱖ3 mm was more common in ICRB group D eyes, post-laminar optic nerve invasion was more common in ICRB group E eyes, and anterior chamber involvement was more common in RE group Vb eyes. We believe that these dissimilarities could be attributed to a difference in cohort size.
On the basis of our results, both advanced group D (17%) and group E (24%) eyes are at increased risk for associated metastatic disease. The ICRB classification indicates that both groups carry the potential for high-risk retinoblastoma, and there is no statistically significant difference between the 2 groups (P ⫽ 0.2090) (Table 1). There are clinical features that predict histopathologic high-risk retinoblastoma. Shields and coworkers14,15 specifically evaluated clinical factors predictive of optic nerve invasion14 or choroidal invasion.15 Of 289 eyes, 84 (29%) demonstrated optic nerve invasion and 67 (23%) had choroidal invasion.14,15 Clinical factors predictive of optic nerve invasion included secondary glaucoma (P ⫽ 0.02), exophytic growth pattern (0.011), and tumor thickness ⱖ15 mm (P ⫽ 0.03).14 The main clinical risk feature predictive of choroidal invasion was iris neovascularization (P ⫽ 0.02).15 Chantada et al16 found high-risk retinoblastoma in 73 of 182 cases (40%), and clinical risk factors predicting high-risk retinoblastoma included glaucoma (P ⫽ 0.025) and buphthalmos (P ⫽ 0.00017). In contrast, Wilson and associates13 found no association between glaucoma and buphthalmos with highrisk retinoblastoma. Other clinical features predictive of high-risk retinoblastoma include older age at presentation (⬎2 years), longer duration lag period from symptoms to diagnosis (⬎3 months), hyphema, pseudohypopyon, staphyloma, and orbital cellulitis.17 Orbital MRI also can provide a correlation with high-risk retinoblastoma.18 –20 The accuracy, specificity, and sensitivity of MRI detection of high-risk retinoblastoma was 91%, 95%, and 60% for postlaminar invasion, respectively,19 and 68%, 80%, and 60% for choroidal invasion, respectively.18 The prediction and diagnosis of high-risk retinoblastoma are crucial because untreated high-risk retinoblastoma carries at least 24% risk for metastatic disease,7 and adjuvant systemic chemotherapy reduces the risk to 0% to 4%.7,8 In the present study, 18% (6/34) of patients with untreated high-risk retinoblastoma developed metastasis, and 5% (4/ 83) of patients with high-risk retinoblastoma who were treated with chemotherapy developed metastasis. Of the 402 patients with non– high-risk retinoblastoma, there were no metastatic events. Similar to a previous study by our group,8 there was no metastasis in high-risk patients treated with VEC. Metastasis from retinoblastoma generally occurs within 1 year of diagnosis of retinoblastoma.21–23 If there is no metastatic disease by 5 years after treatment of retinoblastoma, the patient is usually considered cured.21–23 In our study, 70% (7/10), 90% (9/10), and 100% (10/10) of patients displayed metastasis within 1 year, 2 years, and 5 years of diagnosis of retinoblastoma, respectively. The prognosis of metastatic retinoblastoma is poor, with death usually occurring within 6 months.21–23 In our study, all (n ⫽ 5) metastasis-related deaths occurred within 2 years of diagnosis of metastatic retinoblastoma. By Kaplan–Meier analysis, the probability of metastasis and death after 5 years of diagnosis of retinoblastoma remained stable at 10% and 5%, respectively. Infrequent number of events (metastasis and death) limited the statistical analysis of factors predictive of metastasis and death. However, all 10 patients developing metastasis had advanced retinoblastoma, with
1001
Ophthalmology Volume 120, Number 5, May 2013 100% (10/10) classified as group E according to the ICRB, 80% (8/10) classified as group Vb according to the RE classification, and 70% (7/10) classified as pT3a according to the AJCC classification. The clinical13–18 and MRI19,20 features can predict highrisk retinoblastoma, but ICRB and RE classifications provide limited correlation with high-risk retinoblastoma. The definitive diagnosis of high-risk retinoblastoma is achieved only with histopathologic inspection of the enucleated globe. It is clearly shown that systemic chemotherapy can inactivate invasive high-risk tumor and offer protection against systemic micrometastasis.7,8 However, the usefulness of IAC as a targeted drug directly to the eye and with minimal systemic effect might not offer protection against systemic disease in such cases. Even though excellent globe salvage is achieved with IAC for group D eyes, and moderate globe success is achieved for group E eyes,5 it should be realized that 17% of group D and 24% of group E eyes would be expected to harbor high-risk retinoblastoma and combining this information with the 24% risk of metastasis,7 one might anticipate that 4% of group D and 6% group E eyes treated without systemic chemotherapy might develop metastasis. The prediction of high-risk retinoblastoma based on the clinical, ultrasonography, and MRI features is pivotal in the selection of appropriate treatment in these cases. In conclusion, the ICRB can predict high-risk histopathologic features of retinoblastoma. In this study, histopathologic evidence of high-risk retinoblastoma was noted in 17% of group D eyes and 24% of group E eyes. This finding is useful for counseling families regarding the probable need of adjuvant chemotherapy after enucleation for retinoblastoma, and this information adds caution to the conservative management of advanced retinoblastoma with targeted therapies. Acknowledgment. Statistical analysis was provided by Rishita Nutheti, PhD, Hyderabad, India.
References 1. Shields JA, Shields CL. Retinoblastoma. In: Shields JA, Shields CL, eds. Intraocular Tumors: An Atlas and Textbook. 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2008:293–318. 2. Ramasubramanian A, Shields CL. Staging and treatment strategies. In: Ramasubramanian A, Shields CL, eds. Retinoblastoma. New Delhi, India: Jaypee Brothers Medical Publishers; 2012:70 – 8. 3. Hernandez JC, Brady LW, Shields JA, et al. External beam radiation for retinoblastoma: results, patterns of failure, and proposal for treatment guidelines. Int J Radiat Oncol Biol Phys 1996;35:125–32. 4. Shields CL, Mashayekhi A, Au AK, et al. The International Classification of Retinoblastoma predicts chemoreduction success. Ophthalmology 2006;113:2276 – 80.
1002
5. Shields CL, Bianciotto CG, Jabbour P, et al. Intra-arterial chemotherapy for retinoblastoma: report no. 1, control of retinal tumors, subretinal seeds, and vitreous seeds. Arch Ophthalmol 2011;129:1399 – 406. 6. Shields CL, Shields JA. Retinoblastoma management: advances in enucleation, intravenous chemoreduction, and intra-arterial chemotherapy. Curr Opin Ophthalmol 2010;21:203–12. 7. Honavar SG, Singh AD, Shields CL, et al. Postenucleation adjuvant therapy in high-risk retinoblastoma. Arch Ophthalmol 2002;120:923–31. 8. Kaliki S, Shields CL, Shah SU, et al. Postenucleation adjuvant chemotherapy with vincristine, etoposide, and carboplatin for the treatment of high-risk retinoblastoma. Arch Ophthalmol 2011;129:1422–7. 9. Shields CL, Bianciotto CG, Jabbour P, et al. Intra-arterial chemotherapy for retinoblastoma: report no. 2, treatment complications. Arch Ophthalmol 2011;129:1407–15. 10. Gobin YP, Dunkel IJ, Marr BP, et al. Intra-arterial chemotherapy for the management of retinoblastoma: four-year experience. Arch Ophthalmol 2011;129:732–7. 11. Eagle RC Jr. High-risk features and tumor differentiation in retinoblastoma: a retrospective histopathologic study. Arch Pathol Lab Med 2009;133:1203–9. 12. Edge SB, Byrd DR, Compton CC, et al, eds. Retinoblastoma. In: AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2010:560 – 8. 13. Wilson MW, Qaddoumi I, Billups C, et al. A clinicopathological correlation of 67 eyes primarily enucleated for advanced intraocular retinoblastoma. Br J Ophthalmol 2011;95:553– 8. 14. Shields CL, Shields JA, Baez K, et al. Optic nerve invasion of retinoblastoma: metastatic potential and clinical risk factors. Cancer 1994;73:692– 8. 15. Shields CL, Shields JA, Baez KA, et al. Choroidal invasion of retinoblastoma: metastatic potential and clinical risk factors. Br J Ophthalmol 1993;77:544 – 8. 16. Chantada GL, Gonzalez A, Fandino A, et al. Some clinical findings at presentation can predict high-risk pathology features in unilateral retinoblastoma. J Pediatr Hematol Oncol 2009;31:325–9. 17. Kashyap S, Meel R, Puskker N, et al. Clinical predictors of high risk histopathology in retinoblastoma. Pediatr Blood Cancer 2012;58:356 – 61. 18. Chawla B, Sharma S, Sen S, et al. Correlation between clinical features, magnetic resonance imaging, and histopathologic findings in retinoblastoma: a prospective study. Ophthalmology 2012;119:850 – 6. 19. Brisse HJ, Guesmi M, Aerts I, et al. Relevance of CT and MRI in retinoblastoma for the diagnosis of postlaminar invasion with normal-size optic nerve: a retrospective study of 150 patients with histological comparison. Pediatr Radiol 2007;37: 649 –56. 20. de Graaf P, Moll AC, Imhof SM, et al. Retinoblastoma and optic nerve enhancement on MRI: not always extraocular tumour extension [letter]. Br J Ophthalmol 2006;90:800 –1. 21. MacKay CJ, Abramson DH, Ellsworth RM. Metastatic patterns of retinoblastoma. Arch Ophthalmol 1984;102:391– 6. 22. Kopelman JE, McLean IW, Rosenberg SH. Multivariate analysis of risk factors for metastasis in retinoblastoma treated by enucleation. Ophthalmology 1987;94:371–7. 23. Broaddus E, Topham A, Singh AD. Survival with retinoblastoma in the USA: 1975-2004. Br J Ophthalmol 2009;93:24 –7.
Kaliki et al 䡠 High-Risk Retinoblastoma
Footnotes and Financial Disclosures Originally received: May 7, 2012. Final revision: September 13, 2012. Accepted: October 31, 2012. Available online: February 8, 2013.
Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article. Manuscript no. 2012-658.
1
Ocular Oncology Service, Wills Eye Institute, Thomas Jefferson University, Philadelphia, Pennsylvania.
2
Department of Pathology, Wills Eye Institute, Thomas Jefferson University, Philadelphia, Pennsylvania.
3
Ocular Oncology Service, L.V. Prasad Eye Institute, Hyderabad, India.
4
Department of Ophthalmology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand.
5
Department of Ophthalmology, College of Medicine, King Saud University, Saudi Arabia.
Support was provided by Carlos G. Bianciotto Retinoblastoma Research Fund c/o Eye Tumor Research Foundation, Philadelphia, PA (C.L.S., J.A.S.); The Lucille Wiedman Fund for Pediatric Eye Cancer, Philadelphia, PA (C.L.S., J.A.S.); Lift for a Cure, Morrisdale, PA (C.L.S.); and the Noel T. and Sara L. Simmonds Endowment for Ophthalmic Pathology, Wills Eye Institute (R.C.E.). The funders had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; and in the preparation, review, or approval of the manuscript. Dr. Shields has 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. Correspondence: Carol L. Shields, MD, Ocular Oncology Service, Suite 1440, Wills Eye Institute, 840 Walnut Street, Philadelphia, PA 19107. E-mail: carol.shields@ shieldsoncology.com.
1003