jco-2013-aerts-1458-63

Results of a Multicenter Prospective Study on thePostoperative Treatment of Unilateral RetinoblastomaAfter Primary EnucleationIsabelle Aerts, Xavier Sastre-Garau, Alexia Savignoni, Livia Lumbroso-Le Rouic, Estelle Thebaud-Leculee,Didier Frappaz, Carole Coze, Caroline Thomas, Marion Gauthier-Villars, Christine Levy-Gabriel,Herve J. Brisse, Laurence Desjardins, and Francois Doz

Isabelle Aerts, Xavier Sastre-Garau,Alexia Savignoni, Livia Lumbroso-LeRouic, Marion Gauthier-Villars, ChristineLevy-Gabriel, Herve J. Brisse, LaurenceDesjardins, and Francois Doz, InstitutCurie; Francois Doz, University ParisDescartes, Sorbonne Paris Cite, Paris;Estelle Thebaud-Leculee, Centre OscarLambret, Lille; Didier Frappaz, InstitutdHemato-oncologie Pediatrique, Lyon;Carole Coze, Hopital de la Timone,Centre Hospitalier Universitaire (CHU)Marseille; and Caroline Thomas,CHU Nantes.

Published online ahead of print atwww.jco.org on March 4, 2013.

Supported by a grant from the Associa-tion Retinostop and by the local RotaryClubs in Saint-Cyr, la Cadire,le Castellet, and le Beausset.

Authors disclosures of potential con-flicts of interest and author contribu-tions are found at the end of thisarticle.

Corresponding author: Isabelle Aerts,MD, Department of Pediatric Oncology,Institut Curie-Hopital, 26 rue dUlm75248, Paris Cedex 05, France; e-mail:[email protected].

2013 by American Society of ClinicalOncology

0732-183X/13/3111-1458/$20.00

DOI: 10.1200/JCO.2012.42.3962

A B S T R A C T

PurposeThe objective of this prospective study was to assess overall survival and event-free survival inpatients with intraocular unilateral retinoblastoma (Reese-Ellsworth group V) treated by primaryenucleation with or without adjuvant therapy depending on histopathologic risk factors.

Patients and MethodsPatients (n 123) were divided into three groups on the basis of risk factors for extraocularrelapse and metastasis assessed on centralized histologic examination of enucleated eyes. Group1 (n 70) had minimal or no choroidal involvement and/or prelaminar or no optic nerve involvementand received no adjuvant therapy. Group 2 (n 52) had massive choroidal involvement and/or intra-or retrolaminar optic nerve involvement and/or anterior segment involvement and received fourcourses of adjuvant chemotherapy. Group 3 (n 1) had invasion of the surgical margin of the opticnerve and/or microscopic extrascleral involvement and received six courses of adjuvant chemo-therapy with intrathecal thiotepa, consolidation chemotherapy, and autologous stem-cell rescue.Genetic testing was also performed.

ResultsMedian follow-up for the 123 patients was 71 months. No disease progression, relapse, or distantmetastasis occurred during follow-up. No second malignancies occurred. This requires confirma-tion with longer follow-up. Secondary bilateralization occurred in two patients with identified RB1germline mutation. Adjuvant chemotherapy was well tolerated, with limited toxicity. Moleculartesting found constitutional RB1 gene mutations in only nine of 100 evaluated patients.

ConclusionThe survival rate of 100% was excellent, including 57% of patients who received no adjuvanttherapy, suggesting that chemotherapy could be de-escalated in some patients, especially thosewith massive choroidal involvement.

J Clin Oncol 31:1458-1463. 2013 by American Society of Clinical Oncology

INTRODUCTION

Retinoblastoma (RB), the most common primaryintraocular tumor in children, occurs unilaterally in60% of patients. In the vast majority of patients,unilateralRBisnonhereditary, andonly10%to15%of patients carry a germline mutation.1,2

Indeveloped countries, patientswithunilateralRB typically present at around 2 years of age withdisease that is still confined to the ocular globe buthas reached an advanced stage, so that conservativetreatment with chemotherapy and local adjuvanttherapy is often no longer an option.3 Inmost cases,after ruling out regional extension on clinical andradiologic findings, enucleation is performed. The

cure rate is higher than 95%4; however, some pa-tients are still at risk of secondary dissemination byorbital or metastatic relapse, which has a poor vi-tal prognosis.5-8

The risk is assessed on histologic examina-tion of the enucleated eye to determine the natureand extent of tumor invasion and adopt appropri-ate treatment strategies.5-12 Examination focuseson features related to the two main metastaticpathwaysthrough the ocular coats or via theoptic nerve (ON)as well as invasion of the an-terior segment.13

Assessment of the histologic features has led tothe development of risk categories, although there isnogeneral consensuson the risks associatedwith the

JOURNAL OF CLINICAL ONCOLOGY O R I G I N A L R E P O R T

VOLUME 31 NUMBER 11 APRIL 10 2013

1458 2013 by American Society of Clinical OncologyDownloaded from jco.ascopubs.org on September 12, 2015. For personal use only. No other uses without permission.

Copyright 2013 American Society of Clinical Oncology. All rights reserved.

various criteria.9-12,14 Microscopic evidence of invasion of orbital softtissues and residual tumor at the surgical margin of the ON are uni-formly accepted as high-risk factors.15,16 For many authors,5,17,18 butnot all,12 tumor involvement posterior to the lamina cribrosa is ahigh-risk feature.Conversely, tumor involvement anterior to the lam-ina cribrosa is widely considered a low-risk feature.15-18More contro-versial is the risk associated with tumor involving the ocular coatsand, in particular, isolated invasion of the choroid, since for someauthors, extensive choroidal involvement carries only a moderaterisk.9,10,12,19-21 When both choroidal and retrolaminar ON involve-mentarepresent, the risk is generally considered intermediate.9,12,17-19

Finally, invasion of the anterior segment has been reported to be amoderate- to high-risk feature.13,21

Adjuvant chemotherapy and radiotherapy are uniformly recom-mended forpatientswithhigh-riskhistologic features, suchas residualtumor at the surgical margin of the ON and/or microscopic orbitaldisease.9,12,15,17 Some patientsmay also benefit from amore intensivechemotherapeutic regimen with hematopoietic stem-cell rescue.6

However, there isnoconsensusonwhichotherpatients shouldreceiveadjuvant chemotherapy or on the most effective chemotherapeuticregimen.22Forpatientswith low-and intermediate-risk features, thereare no uniform protocols concerning the indication or regimen forpostoperative adjuvant chemotherapy, partially because of disparitiesin defining and quantifying the risk features. Practices therefore varywidely from one institution to another. The experience at the InstitutCurie does not support the use of adjuvant chemotherapy in patientsin the low-risk category (ie, those with prelaminar or noON involve-mentandminorchoroidal invasion)9but strongly suggests it isneededin intermediate-risk patients (ie, thosewith intra- or retrolaminarONinvolvement and/or massive choroidal involvement with or withoutintrascleral invasion).

Over the years, the changing regimens, choice of drugs, and dosereductions have primarily aimed at obtaining maximal efficacy byusing drugs that have proven effective in extraocular RB,7,8,16,21,23 andat minimizing toxic adverse effects, including the potential risk ofsecondmalignant neoplasms.

We report the results of a prospective study conducted by theFrench Society for Pediatric Oncology (formerly SFOP, nowSFCE). Patients with intraocular unilateral RB treated by primaryenucleation were selectively included in centers across France. Astandardized approach to selecting the regimen for adjuvant ther-apy based on histopathologic risk factors was used in all centers,with centralized reading to ensure homogeneous assessment. Pa-tients in the low-risk category received no adjuvant chemotherapy,those in the intermediate-risk group receivedmoderately intensivechemotherapy, and those in the high-risk group were to receive amore intensive regimen that included autologous hematopoieticstem-cell rescue and orbital irradiation.Overall survival and event-free survival were assessed.

PATIENTS AND METHODS

Thisprospective, nonrandomized studywas conducted in14 centers inFrancebetweenMay 2001 andNovember 2007. The study received ethics committeeapproval and authorization from the French health authorities. Written in-formedconsentwasobtained fromparents or legal guardians forparticipationin the study and for constitutional DNA testing.

Eligible patients were those with nonhereditary unilateral RB presentingas intraocular tumor requiring primary enucleation. The diagnosis of RB wasconfirmed on examination of the ocular fundus under general anesthesia. Allpatients were in group V according to the Reese-Ellsworth classification.24

Ocular and brain imaging were systematically performed at diagnosis torule out macroscopic extraocular invasion, by using ultrasound and eithercomputed tomography (CT) scan with contrast or gadolinium-enhancedmagnetic resonance imaging (MRI). In most cases, MRI was performed withstandard resolution. Later in the study, some patients underwent high-resolutionMRIobtainedwithorbit surface coils. A senior pediatric radiologistprospectively reviewed all data, and a retrospective radiologic review wasperformed at the end of the study, focusing on patients with retrolaminarONinvasion on pathologic examination.

After enucleation,macroscopic and histologic examinations of the spec-imens were performed in each center, followed by prospective centralizedreadings by a single pathologist. The specimens, prepared as previously re-ported,25 were examined by light microscopy for evidence of risk factors

Table 1. Histologic Risk Criteria

Eye Structures/Involvement Criteria

Choroid and scleraNo choroidal involvement Possible presence of tumor cells under the pigment epithelium but not beyond Bruchs membraneMinimal superficial Presence of at most three microscopic clusters of tumor cells beyond Bruchs membrane with no invasion of the

deeper vascular layersMinimal deep Presence of a tumor cell cluster in contact with the sclera, with invasion of the vascular layers of the choroidExtensive deep (massive) Presence of tumor cell clusters on several samples in the form of extensive bands, with involvement of the entire

choroid including the vascular layers and tumor cells in contact with the scleraIntrascleral involvement Presence of tumor cells within the sclera extending to one third to two thirds or to the entire thickness of the scleraExtrascleral involvement Presence of tumor cells extending to the entire thickness of the sclera with microscopic orbital involvement

Optic nerveNo optic nerve involvement Absence of tumor cells anywhere on the optic nervePrelaminar Presence of tumor cells on the optic disc or head of the optic nerve, not extending into the lamina cribrosaIntralaminar Presence of tumor cells within the lamina cribrosaRetrolaminar Presence of tumor cells beyond the posterior margin of the lamina cribrosa but not extending to the resection line

of the optic nerveInvasion of surgical margin Presence of tumor cells at the surgical margin of the optic nerve

Anterior segmentAnterior segment involvement Presence of tumor cells in the ciliary body or iris

Retinoblastoma: Adjuvant Treatment After Enucleation

www.jco.org 2013 by American Society of Clinical Oncology 1459Downloaded from jco.ascopubs.org on September 12, 2015. For personal use only. No other uses without permission.


according to previously described criteria9 (Table 1). Patients were then as-signed to one of three groups: low risk, intermediate risk or high risk (ie, stageII of the international classification26; Table 2).

Patients in the intermediate- and high-risk groups underwent postoper-ative evaluation of distant metastasis, including lumbar puncture with exam-ination of the cytospin and bone marrow cytology. These patients alsounderwent hematologic, renal function, and hepatic function assessments, aswell as audiometry to determine their eligibility to receive one of two adjuvantchemotherapy regimens (Table 3).

Patient follow-up in all three groups to monitor for disease progressionafter treatment was essentially clinical, including monthly ophthalmologicexamination of the fundus oculi of the contralateral eye until age 18 months,examination once every 2 months until age 2 years, and once every 3 monthsthereafter. Patients in the intermediate- and high-risk groupsweremonitoredfor chemotherapy-related toxicity by using the National Cancer InstituteCommonTerminologyCriteria forAdverseEvents, version2.Ototoxicitywasassessed by using Brocks grading system.27 Long-term sequelae were alsomonitored, including annual audiometry, for 5 years. In addition, genetictesting was routinely proposed to screen patients for RB1 gene sequencealterations. Constitutional testing was performed as previously described.28

The sample sizewas estimatedon the basis of expected overall survival of96%, with 125 patients included over 5 years allowing calculation of the ratewith a precision of 2%. Overall and event-free survival were calculated byusing the Kaplan-Meier estimator and updated to June 2011.

RESULTS

A total of 123 patients (69 girls and 54 boys) were included, with amedian age of 23months (range, 1 to 124months) at enucleation. Allpatients underwent initial radiologic examination, including CT scan(59 patients), MRI scan (52 patients), or both (12 patients), whichconfirmed the intraocular nature of the disease in all patients. Themedian time to enucleation after onset of the initial signs ofRBwas 45days (range, 6 to 1,045 days) in the 117 patients for whom the infor-mation was available. The initial signs were leukocoria in 69 patients(56%), strabismus in 22 patients (18%), leukocoria and strabismusin16patients (13%), andother (13%).Themedian time fromdiagno-sis to enucleation was 7 days (range, 0 to 45 days) in 121 patients and 15 days in 118 of 121 patients.

Histopathologic examination found low-risk features for 70(57%) of 123 patients, intermediate-risk features for 52 patients(42%), andhigh-risk features foronepatient (1%).Histologic featuresfor the low-risk group are provided in Table 4 and for theintermediate-riskgroup inTable5.Postoperativeevaluationofdistantmetastasis was negative in all 53 patients with intermediate- or high-risk features.

In the subgroup of 12 patients with retrolaminar involvement,retrospective review of radiologic data showed that in six patientsassessed by CT scan and in four patients assessed by conventionalMRI, imaging failed to detect ON abnormalities, whereas on patho-logic examination, ON invasionwasmeasured at between 0.7 and 2.5mm beyond the lamina cribrosa. In two patients assessed by high-resolutionMRI, abnormal retrolaminar enhancement was retrospec-tively identifiedmeasuring 1 and 1.7mmbeyond the lamina cribrosa,which is in good agreement with the pathologic findings (0.9 and 1.4mm, respectively).

The patient in the high-risk group had massive choroidal in-volvement, and tumoral invasionat the surgicalmarginof theONwassuspected.This patientunderwentpreoperative radiologic assessment

Table 2. Three-Level Risk Classification

Group Criteria

Low risk Minimal or no choroidal involvement and/or prelaminar involvement or absence of optic nerve involvementIntermediate risk Massive choroidal involvement and/or intra- and retrolaminar involvement and/or involvement of the anterior segmentHigh risk Invasion of the surgical margin of the optic nerve and/or microscopic extrascleral involvement

Corresponding to stage II of new international retinoblastoma staging system.26

Table 3. Chemotherapy Regimens, Intermediate- and High-Risk Groups

Drug Dose (per day) Cycles Days

Intermediate-risk group

Etoposide 100 mg/m2 (3.3 mg/kg) 1 and 3 1-5Carboplatin 160 mg/m2 (5.3 mg/kg) 1 and 3 1-5Vincristine 1.5 mg/m2 (0.05 mg/kg) 2 and 4 1 and 5Cyclophosphamide 300 mg/m2 (10 mg/kg) 2 and 4 1-5

High-risk groupEtoposide 100 mg/m2 1, 3, and 5 1-5Carboplatin 160 mg/m2 1, 3, and 5 1-5Thiotepa (intrathecal) 15 mg 1, 3, and 5 1Vincristine 1.5 mg/m2 2, 4, and 6 1Cyclophosphamide 1,000 mg/m2 2, 4, and 6 1-3CarboPEC

Etoposide 350 mg/m2 7 1-5Carboplatin 350 mg/m2 7 1-5Cyclophosphamide 1,600 mg/m2 7 2-5

Treated in outpatient setting, cycles 21 days apart, and no supportive careduring therapy.

Dose adjusted in patients 1 year old and/or 10 kg.Treated in outpatient setting only for cycles 1, 3, and 5; all cycles 21 days

apart. Supportive care during inpatient therapy: intravenous hydration, sodium2-sulfanylethanesulfonate with cyclophosphamide, granulocyte-colony stimu-lating factor.

Followed by autologous hematopoietic stem-cell rescue on day 8.

Table 4. Histologic Features in Low-Risk Group (n 70)

FeatureNo Optic Nerve

Involvement Prelaminar Total

No choroidal involvement 22 19 41Minimal choroidal involvement 17 12 29Total 39 31 70

NOTE. Among the 70 patients, 41 (59%) had no choroidal involvement and 29(41%) had minimal choroidal involvement; 39 patients (56%) had no opticnerve involvement and 31 (44%) had prelaminar involvement; and 12 patients(17%) had both choroidal and optic nerve involvement.

Aerts et al

1460 2013 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGYDownloaded from jco.ascopubs.org on September 12, 2015. For personal use only. No other uses without permission.


withCTscanalone,which failed to identify retrolaminar involvement.Examination of the specimen by the central pathologist confirmedthat tumorcellswere indeedpresentat the resection line,but theywereseenwithin the centralONvessels and notwithin the nerve fibers (Fig1 andAppendix Fig A1, online only). After consultationwith interna-tional experts, it was decided that the atypical high-risk histology didnot support a high risk of local relapse. Orbital radiotherapy, plannedin theprotocol,wasnotperformed.Thepatientnevertheless remainedin the high-risk group because of the risk of bone metastasis andreceived six cycles of chemotherapy as planned, with carboplatin,etoposide, and cyclophosphamide (CarboPEC) as consolidation.

In the intermediate-risk group, all 52 patients received fourcourses of chemotherapy (as planned), with a median interval of 19days (range, 11 to 88 days) between enucleation and initiation of thefirst cycle. In total, 208 courses of chemotherapy were administered,with a median total duration of 78 days (range, 66 to 95 days).

ToxicityExpected hematotoxicity was observed in the intermediate-risk

group and in the high-risk patient. Only five serious adverse events

were reported in four patients in the intermediate-risk group: febrileneutropeniaandgrade4mucositis in twopatients, febrileneutropeniawith shock in one patient, vincristine-related sensory-motor neurop-athy and septicemia due to Staphylococcus in one patient, and septice-mia due to Pneumococcus in one patient. The high-risk patientexperienced three episodes of febrile neutropenia, including one withsepticemia due to Streptococcus. No deaths as a result of toxicity oc-curred. No late grade 3 or 4 ototoxicity was observed in the high-riskpatient or in the 36 intermediate-risk patients assessed. No late renaltoxicity was observed in the high-risk patient or in 18 intermediate-risk patients assessed.

Survival and Other OutcomesMedian follow-up time for the 123 patients was 71months. One

patient with no histologic risk factors was lost to follow-up 4 monthsafter diagnosis. For the remaining 122 patients, post enucleationfollow-up ranged from 25 to 120 months. Median follow-up in theintermediate-risk group was 56 months (range, 19 to 97 months); inthehigh-riskpatient, itwas68months.Nohomolateralprogressionorrelapse of RBwas observed, and no patients developed distantmetas-tasis or second malignancy during the follow-up period. Secondarybilateralizationwasobserved in twopatients:one in the low-riskgroup11 months after enucleation in whom the primary tumor was diag-nosed at 11months of age, and one in the intermediate-risk group 14months after enucleation in whom the primary tumor was diagnosedat 4 months of age.

Genetic TestsMolecular testingof theRB1genewasperformed in100 (81%)of

123 patients. ConstitutionalRB1mutations considered to be deleteri-ous were identified in only nine patients (9%). In two of the ninepatients,metachronic bilateralizationwas observed. No cases of bilat-eralization were observed among the patients in whom no constitu-tional RB1mutation was found.

Testing for somatic mutations was performed in 55 of 100 pa-tients. In all 55 patients, both mutations were found to be somaticevents, and constitutional testing was negative. The mutations weretherefore considered to be postzygotic, and monitoring of siblingscould be lifted in these patients.

Table 5. Histologic Features in Intermediate-Risk Group (n 52)

Involvement No Optic Nerve Involvement Prelaminar Intralaminar Retrolaminar Total

No choroidal 2 0 6 1 9Minimal superficial 1 0 3 2 6Minimal deep 5 6 2 2 15Extensive deep 6 5 2 7 20Scleral 0 0 2 0 2Total 14 11 15 12 52

NOTE. Among the 52 patients, nine (17%) had no choroidal involvement, six (12%) had minimal superficial involvement, 15 (29%) had minimal deep invasion,20 (38%) had extensive deep invasion, and two (4%) had scleral invasion. Fourteen (27%) of the 52 patients had no optic nerve involvement, 11 (21%) hadprelaminar involvement, 15 (29%) had intralaminar involvement, and 12 (23%) had retrolaminar involvement. Seven (13%) had both retrolaminar and extensivedeep choroidal involvement.Both patients also had anterior segment involvement.Patient also had anterior segment involvement.One patient also had anterior segment involvement.One patient also had anterior segment involvement.

Fig 1. Transverse section of optic nerve resection margin (detail). Tumor ispresent in the vessel (red arrow) but not in the surrounding nerve fibers.




CommentsThe goal of adjuvant therapy after enucleation in patients with

unilateral RB is to prevent both local and distant relapses. The usualapproach is to identify patients with histopathologic risk factors onexamination of enucleated eyes and to provide adjuvant therapywhere appropriate. Our results indicate that this approach is effective,since the overall survival rate, with amedian 71months of follow-up,was 100%, and event-free survival was also 100% in our series of 123patients. All patients underwent ophthalmologic follow-up, and sec-ondary bilateralization was observed in only two patients, both ofwhom had been diagnosed with RB before the age of 12 months andwere found to have constitutional RB1mutations.

Only a few prospective studies have been performed that usedhistologic risk factors as thebasis for assigningadjuvant therapy.12,15,29

Our study differs somewhat from other studies in that we definedthree separate risk categories. Inour series, 70patientswere assessed ashaving low-risk features and receivedno adjuvant therapy.All of thesepatients were alive with no disease dissemination at a median 51months of follow-up. These results confirm the findings in otherstudies29,30 that adjuvant therapy is not warranted in patients withminimal or no choroidal involvement and/or prelaminar or no ONinvolvement and no invasion of the anterior segment.

Our series also included 52 patients with intermediate-risk fea-tures who received four cycles of adjuvant chemotherapy.No relapsesor metastasis was observed with a median follow-up of 56months. It is of particular note that in 10 of the 15 patients withminimal deep choroidal involvement and in 12 of the 20 patientswith extensive deep choroidal involvement, invasion of the ocularcoat was not associated with either ON or anterior segment in-volvement. Thus, in our series, a total of 22 patients (42%) hadisolated deep (or massive) choroidal involvement. In a survey ofthe literature, this particular subgroup has clearly been the focus ofmuch discussion concerning the need for adjuvant chemotherapy,given that the relapse rate in patients in this subgroup who receiveno adjuvant therapy after enucleation is reportedly less than5%.12,22,29 In addition, Chantada et al30 reported that, in a recentstudy, only three (2.6%) of 114 patients with isolated choroidalinvasion developed relapse with no adjuvant therapy after enucle-ation. It is therefore conceivable that only a small proportion of the22 patients in our series benefited from treatment, although theothers were exposed to avoidable toxicities. However, in our retro-spective study,9 massive choroidal involvement was an identifiedrisk factor: the debate is between overtreating some patients byproviding adjuvant chemotherapy but without observing extraoc-ular relapse versus not providing adjuvant chemotherapy and tak-ing the risk of extraocular relapse, a life-threatening condition thatrequires highly toxic treatment. Of course, our study, like all oth-ers, was not designed to compare adjuvant therapy versus noadjuvant therapy within the intermediate-risk group or with re-gard to individual risk features taken in isolation: indeed, thenumber of patients required for such a study makes it infeasible.

Only onepatient in our series hadhigh-risk features and receivedsix cycles of adjuvant chemotherapy followed by high-dose chemo-therapy and hematopoietic stem-cell rescue. Because of the particularnature of the ON involvement, the patient did not receive orbitalradiotherapy as planned in the protocol, but no local disease progres-sion, relapse, or metastasis was observed at 68 months of follow-up.However,theefficacyofadjuvanttherapyinvolvinganintensivechem-

otherapeutic regimen clearly cannot be assessed on the basis of asingle patient.

The main role of imaging in patients with RB is to rule outextraocular extension. Extrascleral/orbital or regional lymph nodeextension remain rare in developed countries, and themajor issue forradiologists is therefore to detect ON involvement. Our study is inagreement with our previous results regarding the accuracy of preop-erative imaging methods.31 Early-stage retrolaminar invasion (withnormal size ON) was not detected in the six patients assessed by CTscan alone or in the four patients assessed by conventional MRI.Conversely, high-resolution MRI, used in two patients, identifiedsubtle retrolaminar enhancement, which correlated well withpathologic findings. The most recent imaging guidelines stronglyrecommend the use of high-resolutionMRI.32 The accuracy of high-resolution MRI to differentiate intra- versus early retrolaminar inva-sion is not yet established, and pathology still remains the goldstandard. However, by using high-resolutionMRI, it may be possibleto differentiate patients with early-stage retrolaminar invasion whocan be treated by primary enucleation frompatientswithmore exten-sive retrolaminar invasion inwhom front-line enucleationmight leadto microscopic residual tumor. Further studies are needed; however,theuseofhigh-resolutionMRI for local stagingat the timeofdiagnosiscould improve the prospect of complete tumor resection and informdecisionsontherapeuticstrategies, includingneoadjuvantchemother-apy for either conservative treatment or locally advanced tumors.33

In a subsequent study by the French Society for PediatricCancer,RB-SFCE 2009, which is currently enrolling patients with intraocularunilateral RB treated by primary enucleation, we propose to use sys-tematic high-resolution MRI for preoperative staging and a similarstandardized approach to assigning adjuvant therapy on the basis ofhistologic risk factors. The three-level risk classification in this studyhasbeenslightlymodified.Patientswith intralaminar involvementareincluded in the low-risk group, according to the recent pathologyguidelines from the International Retinoblastoma Staging WorkingGroup.27 In the intermediate-risk group, we propose to de-escalateadjuvant chemotherapy in patients with isolatedmassive choroidalinvasion and to provide these patients with only two cyclesof vincristine-carboplatin.

Furthermore, the workup for metastasis following enucleationwas negative in all 53 patients in the intermediate- and high-riskgroups.Wepropose, as others have,34,35 to select patients who requirescreening and to perform the work-up only in stage II patients.26

This prospective study confirms that unilateral RB is rarely he-reditary and that constitutional as well as somatic genetic studiesusefully contribute to the assessment of familial recurrence of RB andto the follow-up of siblings.3

Inconclusion,our studyshowedthat riskassignmenton thebasisof histopathologic features is an effective approach to selecting appro-priate adjuvant therapy since all patients were alive and event-free at amedian 71 months of follow-up. Clearly, any effective strategy oftreatmentadaptationbasedonrisk factors reliesona thoroughknowl-edge ofwhich regimens are best adapted to the various risk features. Aprospective study has recently been started to further refine our stan-dardizedapproachandtoconfirmtheexcellent resultsobtained in thisstudy, which indicate that adjuvant chemotherapy is safe and effectiveoverall, with no secondary cancer, and could be reduced in somepatients in the intermediate-risk group, particularly those with iso-lated choroidal involvement.

Aerts et al

1462 2013 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGYDownloaded from jco.ascopubs.org on September 12, 2015. For personal use only. No other uses without permission.


AUTHORS DISCLOSURES OF POTENTIAL CONFLICTSOF INTEREST

The author(s) indicated no potential conflicts of interest.

AUTHOR CONTRIBUTIONS

Conception and design: Xavier Sastre-Garau, Didier Frappaz, CaroleCoze, Laurence Desjardins, Francois Doz

Provision of study materials or patients: Isabelle Aerts, XavierSastre-Garau, Livia Lumbroso-Le Rouic, Francois DozCollection and assembly of data: Isabelle Aerts, Xavier Sastre-Garau,Livia Lumbroso-Le Rouic, Estelle Thebaud-Leculee, Didier Frappaz,Carole Coze, Caroline Thomas, Marion Gauthier-Villars, ChristineLevy-Gabriel, Laurence Desjardins, Francois DozData analysis and interpretation: Alexia Savignoni, Carole Coze, HerveJ. Brisse, Francois DozManuscript writing: All authorsFinal approval of manuscript: All authors

REFERENCES

1. Blanquet V, Turleau C, Gross-Morand MS, etal: Spectrum of germline mutations in the RB1 gene:A study of 232 patients with hereditary and nonhe-reditary retinoblastoma. Hum Mol Genet 4:383-388,1995

2. Lohmann DR: RB1 gene mutations in retino-blastoma. Hum Mutat 14:283-288, 1999

3. Aerts I, Lumbroso-Le Rouic L, Gauthier-VillarsM, et al: Retinoblastoma. Orphanet J Rare Dis 1:31,2006

4. Howarth C, Meyer D, Hustu HO, et al: Stage-related combined modality treatment of retinoblas-toma: Results of a prospective study. Cancer 45:851-858, 1980

5. Doz F, Khelfaoui F, Mosseri V, et al: The roleof chemotherapy in orbital involvement of retino-blastoma: The experience of a single institution with33 patients. Cancer 74:722-732, 1994

6. Namouni F, Doz F, Tanguy ML, et al: High-dose chemotherapy with carboplatin, etoposide andcyclophosphamide followed by a haematopoieticstem cell rescue in patients with high-risk retinoblas-toma: A SFOP and SFGM study. Eur J Cancer33:2368-2375, 1997

7. Doz F, Neuenschwander S, Plantaz D, et al:Etoposide and carboplatin in extraocular retinoblas-toma: A study by the Societe Francaise dOncologiePediatrique. J Clin Oncol 13:902-909, 1995

8. Chantada G, Fandino A, Casak S, et al: Treat-ment of overt extraocular retinoblastoma. Med Pe-diatr Oncol 40:158-161, 2003

9. Khelfaoui F, Validire P, Auperin A, et al: His-topathologic risk factors in retinoblastoma: A retro-spective study of 172 patients treated in a singleinstitution. Cancer 77:1206-1213, 1996

10. Kopelman JE, McLean IW, Rosenberg SH:Multivariate analysis of risk factors for metastasis inretinoblastoma treated by enucleation. Ophthalmol-ogy 94:371-377, 1987

11. Messmer EP, Heinrich T, Hopping W, et al:Risk factors for metastases in patients with retino-blastoma. Ophthalmology 98:136-141, 1991

12. Schvartzman E, Chantada G, Fandino A, et al:Results of a stage-based protocol for the treatmentof retinoblastoma. J Clin Oncol 14:1532-1536, 1996

13. Haik BG, Dunleavy SA, Cooke C, et al: Reti-noblastoma with anterior chamber extension. Oph-thalmology 94:367-370, 1985

14. Uusitalo MS, Van Quill KR, Scott IU, et al:Evaluation of chemoprophylaxis in patients withunilateral retinoblastoma with high-risk features onhistopathologic examination. Arch Ophthalmol 119:41-48, 2001

15. Zelter M, Damel A, Gonzalez G, et al: Aprospective study on the treatment of retinoblas-toma in 72 patients. Cancer 68:1685-1690, 1991

16. Freeman CR, Esseltine DL, Whitehead VM, etal: Retinoblastoma: The case for radiotherapy andfor adjuvant chemotherapy. Cancer 46:1913-1918,1980

17. Shields CL, Shields JA, Baez K, et al: Opticnerve invasion of retinoblastoma: Metastatic poten-tial and clinical risk factors. Cancer 73:692-698, 1994

18. Hungerford J: Factors influencing metastasisin retinoblastoma. Br J Ophthalmol 77:541, 1993

19. Shields CL, Shields JA, Baez KA, et al:Choroidal invasion of retinoblastoma: Metastaticpotential and clinical risk factors. Br J Ophthalmol77:544-548, 1993

20. Redler LD, Ellsworth RM: Prognostic impor-tance of choroidal invasion in retinoblastoma. ArchOphthalmol 90:294-296, 1973

21. Rubin CM, Robison LL, Cameron JD, et al:Intraocular retinoblastoma group V: An analysis ofprognostic factors. J Clin Oncol 3:680-685, 1985

22. Chantada GL, Dunkel IJ, de Davila MT, et al:Retinoblastoma patients with high risk ocular path-ological features: Who needs adjuvant therapy? Br JOphthalmol 88:1069-1073, 2004

23. Pratt CB, Fontanesi J, Chenaille P, et al:Chemotherapy for extraocular retinoblastoma. Pedi-atr Hematol Oncol 11:301-309, 1994

24. Reese AB, Ellsworth RM: The evaluation andcurrent concept of retinoblastoma therapy. TransAm Acad Ophthalmol Otolaryngol 67:164-172, 1963

25. Sastre X, Chantada GL, Doz F, et al: Proceed-ings of the consensus meetings from the Interna-

tional Retinoblastoma Staging Working Group onthe pathology guidelines for the examination ofenucleated eyes and evaluation of prognostic riskfactors in retinoblastoma. Arch Pathol Lab Med133:1199-1202, 2009

26. Chantada G, Doz F, Antoneli CB, et al: Aproposal for an international retinoblastoma stagingsystem. Pediatr Blood Cancer 47:801-805, 2006

27. Brock PR, Bellman SC, Yeomans EC, et al:Cisplatin ototoxicity in children: A practical gradingsystem. Med Pediatr Oncol 19:295-300, 1991

28. Houdayer C, Gauthier-Villars M, Lauge A, et al:Comprehensive screening for constitutional RB1mutations by DHPLC and QMPSF. Hum Mutat 23:193-202, 2004

29. Chantada G, Fandino A, Davila MT, et al:Results of a prospective study for the treatment ofretinoblastoma. Cancer 100:834-842, 2004

30. Chantada GL, Fandino A, Guitter MR, et al:Results of a prospective study for the treatment ofunilateral retinoblastoma. Pediatr Blood Cancer 55:60-66, 2010

31. Brisse HJ, Guesmi M, Aerts I, et al: Relevanceof CT and MRI in retinoblastoma for the diagnosis ofpostlaminar invasion with normal-size optic nerve: Aretrospective study of 150 patients with histologicalcomparison. Pediatr Radiol 37:649-656, 2007

32. de Graaf P, Goricke S, Rodjan F, et al: Guide-lines for imaging retinoblastoma: Imaging principlesand MRI standardization. Pediatr Radiol 42:2-14,2012

33. Bellaton E, Bertozzi AI, Behar C, et al: Neoad-juvant chemotherapy for extensive unilateral retino-blastoma. Br J Ophthalmol 87:327-329, 2003

34. Azar D, Donaldson C, Dalla-Pozza L: Question-ing the need for routine bone marrow aspiration andlumbar puncture in patients with retinoblastoma.Clin Experiment Ophthalmol 31:57-60, 2003

35. Zacharoulis S, Abramson DH, Dunkel IJ: Moreaggressive bone marrow screening in retinoblas-toma patients is not indicated: The Memorial Sloan-Kettering Cancer Center experience. Pediatr BloodCancer 46:56-61, 2006




jco-2013-aerts-1458-63

Documents