External Beam Radiation "Salvage" Therapy in Transgenic Murine Retinoblastoma

Objective  To determine the efficacy of low-dose "salvage" external beam radiationtherapy (EBRT) following failed subconjunctival carboplatin chemotherapy ina murine model of heritable retinoblastoma.

Methods  Eighty-four eyes from 8-week-old, simian virus 40, T-antigen–positivemice were treated with 6 serial subconjunctival carboplatin injections (100µg/25 µL). At 12 weeks of age, 64 eyes received EBRT for a totaldose of 480 (4.8 Gy), 1200 (12.0 Gy), 1560 (15.6 Gy), or 3000 (30.0 Gy) rad.Twenty eyes received no additional therapy following subconjunctival carboplatininjections. Ten eyes received a total dose EBRT of only 3000 rad. Eight eyesreceived subconjuctival injections of only an isotonic sodium chloride solution.Ten eyes served as untreated controls.

Main Outcome Measures  Eyes were enucleated at 20 weeks to assess the presence of tumor onhistopathological examination.

Results  Salvage therapy using low-dose EBRT was able to reestablish tumor controlin a dose-dependent manner. Increasing the EBRT dose to 3000 rad resultedin 100% tumor control. The dose-dependent curves were significantly differentbetween the treatment groups—EBRT alone vs salvage EBRT after receivingsubconjunctival carboplatin injections (P<.001).

Conclusion  Low-dose hyperfractionated salvage EBRT following failed primary subconjunctivalcarboplatin chemotherapy is efficacious in the treatment of retinoblastomain this animal model.

Clinical Relevance  Salvage EBRT using a reduced total radiation dose could be associatedwith a radiation-related treatment enhancement in pediatric retinoblastoma.

The early diagnosis and treatment of intraocular retinoblastoma havecontributed to a marked improvement in patient survival.^1,2 Treatmentstrategies for retinoblastoma have been evolving significantly over the pastseveral years. Current treatment modalities include enucleation, externalbeam radiation therapy (EBRT), scleral plaque radiotherapy, transpupillarythermoablation, cryotherapy, laser ablation, and chemotherapy.^3-6 Externalbeam radiation therapy has played a prominent role in the treatment of advancedstages of retinoblastoma, especially in children with a germline Rb mutation predisposing them to bilateral disease.^7-9 Althoughefficacious, ionizing radiation exposure has been associated with severe complicationssuch as facial deformities, cataract, radiation-induced retinopathy, radiation-inducedvasculopathy, optic neuropathy, and an enhanced risk of secondary malignanttumors.^10-19 Inrecent years, systemic chemotherapy has been increasingly used in the primarytreatment of moderate to large retinoblastoma tumors (Reese-Ellsworth stagesIII-V).^20-27 Chemotherapyhas been demonstrated to improve the rates of ocular salvage in selected patientsand has decreased the need for EBRT or postponed the use of EBRT in many cases.

Several clinical trials have documented the efficacy of systemic carboplatintherapy in the treatment of multiple neoplasms in children.^28,29 Althoughthis agent has played a pivotal role in virtually all chemotherapeutic regimensproposed for the treatment of retinoblastoma, owing to its low toxic effectprofile compared with other chemotherapeutic agents, concerns regarding significantmorbidity and potential mortality caused by drug-related toxicity still exist.^30-33 Asa result, the use of focal chemotherapy has been recently advocated. Thistreatment modality should include the benefits associated with chemoreductivetreatment and would conceivably spare patients the associated adverse toxiceffects and mutagenic potential of systemic delivery. Our laboratory has documentedthe efficacy of focal carboplatin chemotherapy, provided intravitreally andsubconjunctivally, in the treatment of a transgenic murine model of retinoblastoma.^34-37 Measurementof intraocular carboplatin concentration after peribulbar administration innontumor-bearing primates has shown dramatically higher concentrations ofcarboplatin in the vitreous and aqueous humor than in control animals receivingintravenous administration of the drug.³⁸ Abramsonet al³⁹ have reported significant efficacyin control of vitreous seeding and primary retinal tumor in a phase I/II clinicalstudy of subconjunctival carboplatin administration for intraocular retinoblastoma.

This focal approach has enabled researchers to concentrate on methodsof combining chemotherapy with other local treatment modalities.^40-42 Combinedmodality therapy, especially if acting synergistically, may allow a reductionin treatment dose, minimizing dose-dependent complications. Several investigatorshave evaluated the application of chemoreductive therapy with aggressive localtreatment to eliminate the need for EBRT.^40,43-46 Othershave evaluated focal plaque radiotherapy, as a primary and secondary treatment,in the hopes of decreasing orbital and central nervous system radiation exposure.^47,48 However, EBRT is often needed as"salvage" therapy after failure of systemic chemoreduction to eradicate vitreousor subretinal seeding in the management of large retinoblastoma tumors.^49-51 The aim of this studyis to evaluate the efficacy of low-dose salvage EBRT following failed subconjunctivalcarboplatin chemotherapy in a murine model of spontaneous heritable retinoblastoma.

This study protocol was approved by the University of Miami School ofMedicine Animal Care and Use Review Board, Miami, Fla. All experiments inthis study were conducted in accord with the Association for Research in Visionand Ophthalmology guidelines for the use of animals in ophthalmologic andvision research.

One hundred twelve eyes from simian virus 40, large T-antigen transgene-bearingmice were treated, as described below, beginning at 8 weeks of age. The transgenicmouse model used in this study has been characterized previously.^52-57 Briefly,a highly expressed murine oncogenic transgene drives bilateral retinal tumordevelopment by using a retinal-specific promoter sequence to direct the expressionof simian virus 40 T antigen, resulting in ocular tumor growth. Transgenicanimals were identified through polymerase chain reaction analysis of tailDNA. Samples positive and negative for the transgene were detected by visualizingethidium bromide–stained agarose gels. Transgene-positive animals developbilateral, heritable retinoblastoma that resembles human retinoblastoma includingsimilar histopathological, immunocytochemistry, and behavioral growth patterns.Pathological evidence of tumor is noted by the age of 4 weeks, a small intraoculartumor (corresponding to Reese-Ellsworth stage I) is present at 5 weeks, amoderate to large intraocular tumor (corresponding to Reese-Ellsworth stagesIII and IV) is noted at 10 weeks, and the tumor fills the globe by 16 weeks.Tumors in this animal model are typically small at 5 weeks, appearing withintraretinal involvement only (occupying <1% of retinal area and correspondinglyless ocular volume), and at 10 weeks the tumor is moderate in size (occupyingapproximately 20%-25% of the retinal area and 10%-25% of the ocular volume).

Eighty-four eyes from 8 week-old, simian virus 40, T-antigen–positivemice were treated with subconjunctival carboplatin (Paraplatin; Bristol-MyersSquibb, Princeton, NJ) injections (100 µg/25 µL) administeredtwice per week for a total of 6 injections. Animals were anesthetized witha combination of intraperitoneal ketamine hydrochloride and xylazine hydrochloride.Carboplatin injections were provided with a 33-gauge needle inserted intothe nasal and superior subconjunctival spaces. A microvolume delivery pumpwas used to ensure accurate and reproducible delivery of a 25-µL volume.Twenty eyes received no additional therapy following subconjunctival carboplatininjections. Ten eyes of litter-matched animals served as untreated controls.Eight eyes of litter-matched animals received 25-µL subconjunctivalinjections of balanced salt solution to provide a positive placebo control.Following each injection, all animals underwent external ophthalmologic andfundus examination to ensure no perforation of the globe had occurred.

Mouse eyes were treated, as per treatment groups described below, withfractionated EBRT using a 10-mV x-ray machine (CLINAC-2100; Varian MedicalSystems, Charlottesville, Va) (Figure 1).Unanesthetized animals were briefly immobilized in specially constructed cagesand shielded to minimize irradiation to midline nonocular structures. Treatmentports were confirmed and radiation was delivered at 324 rad/min (3.24 Gy/min)in a 7.0 × 7.0-cm field. All radiation treatments were delivered twiceper day in 120-rad (1.2-Gy) fractions from 6 to 8 hours apart. Ten eyes receivedonly EBRT at a total dose of 3000 rad (30.0 Gy) to provide a positive placebocontrol.

At 12 weeks of age, 64 eyes from the subconjunctival carboplatin-treatedmice received EBRT for a total dose of 480 (4.8 Gy) (n = 12), 1200 (12.0 Gy)(n = 10), 1560 (15.6 Gy) (n = 26), or 3000 (30.0 Gy) rad (n = 16).

At 20 weeks of age, all animals were killed with an overdose of combinedketamine and xylazine. Both eyes were enucleated and immediately immersionfixed in a 10% formalin solution. The eyes were sectioned serially and stainedwith hematoxylin-eosin. Light microscopic examination was performed on allhistopathological sections in a masked fashion. Eyes were graded positivefor tumor development if any histopathological evidence of tumor was present.Eyes were also evaluated for evidence of toxic effects to the cornea, lens,retina, or sclera.

To determine the dose-response relationships among the different treatmentdoses, data were subjected to Probit statistical analysis. Probit is a standardmethod of dose response.⁵⁸ Probit regressionallows the assessment of the effects of drug concentration and treatment groupfor dose-response data in which the assumptions of standard linear regressionare not met. Total radiation dose was entered as a linear predictor in a maximumlikelihood Probit regression model. The model was assessed for goodness-of-fitto ensure that assumptions of the Probit model were not violated. The radiationenhancement ratio was calculated by dividing the tumor control dose for 50%of eyes (TCD₅₀) for 3000-rad EBRT alone by the TCD₅₀ for3000-rad salvage EBRT after subconjunctival carboplatin administration.

Histopathological examination at age 20 weeks revealed that all untreatedeyes (n = 10) and isotonic sodium chloride solution–treated eyes (n= 8) exhibited multiple large intraocular tumors (Figure 2). Tumor control was achieved in only 20% of the mice treatedwith 100 µg of subconjunctival carboplatin alone.

Salvage therapy using low-dose EBRT was able to reestablish tumor controlin a dose-dependent manner. Tumor control was not achieved in any of the eyestreated with a total dose of 480 rad after subconjunctival carboplatin therapy.Eradication of tumor was seen in 30% of eyes treated with subconjunctivalcarboplatin followed by EBRT for a total dose of 1200 rad (Figure 3). Average tumor size in this group treated with carboplatinand 1200-rad (12.0-Gy) EBRT was smaller than the average tumor size in thegroup treated with carboplatin alone. Sixty-five percent tumor control wasachieved with combined treatment of carboplatin and 1560-rad (15.6-Gy) totaldose EBRT (Figure 4) and completetumor eradication occurred in eyes treated with carboplatin and EBRT at atotal dose of 3000 (30.0 Gy) rad (Figure 5). In comparison, EBRT alone at 120-rad (1.2-Gy) fractions twiceper day failed to eradicate tumor in 70% of eyes treated with a total doseof 3000 rad (Figure 5). There wasno histopathological evidence of ocular toxic effects for any of the treatedeyes.

Probit regression analysis demonstrated that the dose-dependent curveswere significantly different between the treatment groups (EBRT alone vs salvageEBRT after subconjunctival carboplatin injections) (P<.001)(Figure 6). The regression coefficientfor EBRT salvage therapy was 2.80 with a 1.59 to 4.01 ninety-five percentconfidence interval. The radiation enhancement ratio was 2.39 with 100-µgsubconjunctival carboplatin injections.

Earlier detection and refinement of conservative methods of treatmentof retinoblastoma have led to fewer primary enucleations, high globe-conservationrates, and better prognosis for visual outcome.⁵⁹ Lifeexpectancy following treatment is excellent, but survivors with genetic predispositionface an increased risk of subsequent cancers.^3,12 Radiotherapyfurther increases the risk for a second primary tumor, with a cumulative incidenceof 35% during a 30-year follow-up in radiation-treated patients, comparedwith 5.8% in those not receiving EBRT.⁶⁰ Thishas led to the search for alternate treatment modalities.

Several groups have demonstrated the efficacy of initial systemic chemotherapyin reducing tumor size (chemoreduction).^24,61,62 Systemicchemotherapy also enhances the effectiveness of other treatment modalitiesand may reduce the possibility of metastatic dissemination.^63,64 Disadvantagesinclude serious toxic adverse effects, potential mutagenesis, and increasedrisk for development of other malignant neoplasms.^10,11,33,65 Furthermore,though effective in initial tumor reduction, this treatment modality aloneis unable to achieve complete tumor control and, therefore, often requiresadjuvant focal therapy. Chemoreduction combined with local therapies includinglaser hyperthermia and cryoablative therapy have been used with much successin recent years.^20,41,43,44 Despitethese advances in tumor control, recent studies indicate that even the combinationprotocols are often insufficient to eradicate vitreous or subretinal seeding,requiring salvage EBRT to establish more complete tumor control.^50,51 Thereasons for the failure of this treatment modality to eliminate vitreous andsubretinal seeding include insufficient penetration of the intraretinal andintravitreal spaces by the systemically administered chemotherapeutic drug,dose limitations, and rapid renal clearance of the drug.

Local delivery of carboplatin therapy may be more beneficial than intravenousdelivery because it is associated with increased drug penetration of ocularstructures and tumor control in animal models. Previous studies in our laboratoryhave demonstrated the ability of carboplatin administered intravitreally,subconjunctivally, and iontophoretically to completely eliminate tumor inthe treatment of murine retinoblastoma.^34-36,42,49 Peribulbarcarboplatin administration in nontumor-bearing primates produces dramaticallyhigher concentrations in the vitreous and aqueous humor compared with intravenousadministration of the drug.³⁸ Recent experimentsin our laboratory on the pharmokinetics of local vs systemic carboplatin administrationin the rabbit eye show that intraretinal and intravitreal drug concentrationsfollowing subconjunctival injection or iontophoretic provision are significantlyhigher than those achieved with systemic administration (B.C.H., T.G.M., E.H.,and Monika Viogt, MD, Peter Milne, PhD, Martina Kralinger, MD, and Jean-MarieParel, PhD, unpublished data, January 2001 ). One clinical trial of subconjunctivalcarboplatin administration for intraocular retinoblastoma has also shown itto be efficacious in children.³⁹ Direct subconjunctivaldelivery of carboplatin chemotherapy may be associated with decreased toxiceffects and considered safe when administered at optimum concentrations anddose schedules.³⁴

Use of local chemotherapeutic drug delivery as part of a combined modalityprotocol further limits toxic effects. Platinum compounds may have a radiosensitizingeffect on hypoxic tumors, such as retinoblastoma, which could contribute tothe potential efficacy of combined modality therapy with carboplatin and low-doseirradiation.⁴⁹ Our current study shows thatlow-dose hyperfractionated salvage EBRT following failed primary carboplatinchemotherapy is efficacious in the treatment of retinoblastoma in this transgenicanimal model. The total dose provided with EBRT may be reduced by 58%, whilemaintaining local tumor control.

In this murine retinoblastoma model, the TCD₅₀ has recentlybeen found to be 3370 rad (33.7 Gy) for those animals administered radiationtherapy in 120-rad (1.2-Gy) fractions twice per day.⁶⁶ Inour study, the TCD₅₀ for those animals administered salvage EBRTafter failed carboplatin therapy decreased to 1410 rad (14.1 Gy). The radiationenhancement ratio, which describes the potential enhancement of combined therapyof carboplatin and radiation in this tumor model, is calculated to be 2.39.The TCD₅₀ for EBRT used in the calculation was taken from the Haydenet al⁶⁶ study because of its larger samplesize (n = 42). In this prior study, the mice were killed at 16 weeks insteadof 20 weeks. Therefore, the radiation enhancement ratio may actually be underestimatedas earlier age at death in the EBRT-only group may mask subsequent tumor growth.

The degree of enhancement described by the radiation enhancement ratiosuggests the ability to decrease the dose of both carboplatin and radiationtherapy without any compromise in tumor control. This approach may allow fora decrease in the morbidity associated with radiation therapy or chemotherapy.If used as a primary modality for the treatment of retinoblastoma, the potentiationbetween carboplatin and radiation therapy may result in a more complete initialresponse, leading to fewer recurrences. If used as secondary treatment afterfailure of chemotherapy, EBRT may be used at lower doses than previously believednecessary to achieve tumor control, minimizing the adverse effects of radiationtherapy and the long-term secondary cancer risks of this treatment. Becausecarboplatin is a known radiation sensitizer, however, it could increase normaltissue complications or increase the risk of secondary cancers in combinationwith EBRT.

Several considerations remain before applying the information gainedin this study to childhood retinoblastoma. First, the transgenic model maydiffer from human retinoblastoma in its response to treatment. For example,radiation-related complications, including optic neuropathy and retinal vasculopathy,are rarely seen in the mouse eye. Second, although subconjunctival carboplatintherapy avoids problems of systemic toxic effects and showed no signs of toxiceffects in mice, it is difficult to predict the effect of carboplatin therapyon the final visual acuity in humans. The potential for long-term ocular toxiceffects from combination local carboplatin therapy and EBRT is unknown.

External beam radiation therapy using a reduced total radiation doseas salvage therapy could be associated with a radiation-related treatmentenhancement, permitting tumor control, while minimizing treatment-relatedtoxic effects in pediatric retinoblastoma.

Corresponding author: Timothy G. Murray, MD, Bascom Palmer Eye Institute,Department of Ophthalmology, University of Miami, PO Box 016880, Miami, FL33101 (e-mail: tmurray@med.miami.edu).

Submitted for publication April 25, 2003; final revision received August24, 2003; accepted September 10, 2003.

This study was supported by the Macula Society, Beachwood, Ohio; Fightfor Sight and Research to Prevent Blindness Inc, New York, NY; Miami and theAmerican Cancer Society Florida Division, Tampa; and the Knights Templar EyeFoundation, Chicago, Ill.

Carboplatin was generously donated by Bristol-Meyers Squibb, Princeton.