- © 2000 by American Society of Clinical Oncology
Chemoreduction and Local Ophthalmic Therapy for Intraocular Retinoblastoma
- Debra L. Friedman,
- Bruce Himelstein,
- Carol L. Shields,
- Jerry A. Shields,
- Michael Needle,
- David Miller,
- Greta R. Bunin and
- Anna T. Meadows
- From the Department of PediatricsDivision of Oncology, The Children’s Hospital of Philadelphia; University of Pennsylvania, Philadelphia PA; and Ocular Oncology Service, Wills Eye Hospital, Thomas Jefferson University, Philadelphia PA.
- Address reprint requests to Debra L. Friedman, Division of Hematology/Oncology, CH-29, Children’s Hospital and Regional Medical Center, 4800 Sand Point Way NE, Seattle, WA 98105; email dfried{at}chmc.org
Abstract
PURPOSE: To study the effectiveness of combined systemic chemotherapy and local ophthalmic therapy for retinoblastoma with the goal of avoiding enucleation and external-beam radiation therapy (EBRT).
PATIENTS AND METHODS: This was a prospective, nonrandomized, single-arm clinical trial. Seventy-five eyes were followed in 47 children. Patients were treated with a six-cycle protocol of vincristine, etoposide, and carboplatin. Most (83%) also received ophthalmic treatment (cryotherapy, laser photocoagulation, thermotherapy, or plaque radiation therapy) during and/or after the chemotherapy.
RESULTS: With a median follow-up of 13 months, event-free survival was 74%, with an event defined as enucleation and/or EBRT. Six children required EBRT in seven eyes (9%); five required enucleation of one eye (7%); five required a combination of EBRT and enucleation in six eyes (8%). Reese-Ellsworth groups 1, 2, and 3 eyes had excellent results, with avoidance of EBRT or enucleation in all 39. Treatment of groups 4 and 5 was less successful, with 33% of six eyes and 53% of 30 eyes, respectively, requiring EBRT and/or enucleation. Toxicities from chemotherapy were mild and included cytopenias (89%), fever and neutropenia (28%), infection (9%), and gastrointestinal symptoms, dehydration, and vincristine neurotoxicity (40%). No patients developed a second malignancy, metastatic disease, renal disease, or ototoxicity.
CONCLUSION: In retinoblastoma patients with Reese-Ellsworth eye groups 1, 2, or 3, systemic chemotherapy used with local ophthalmic therapies can eliminate the need for enucleation or EBRT without significant systemic toxicity. More effective therapy is required for Reese-Ellsworth eye groups 4 and 5.
RETINOBLASTOMA, THE most common intraocular malignancy of childhood is a relatively rare form of childhood cancer, with about 200 new cases occurring each year in the United States.1,2 With advances in therapy, survival has risen from 30% in the 1930s to nearly 95% in the 1990s.3 The therapies responsible for this marked improvement in survival, enucleation and external-beam radiation therapy (EBRT), are also associated with significant morbidity.3-5 Recently, advances in our understanding of retinoblastoma have led to trials of new treatment modalities aimed at decreasing morbidity and continuing excellent survival.
The genetics of retinoblastoma has enormous implications for therapy. Retinoblastoma is the paradigm for a genetically inherited cancer and provides the basis for the two-hit hypothesis of carcinogenesis.6 The disease results from loss or mutation of both copies of RB1, a tumor suppressor gene located at 13q14. Forty percent of children have the hereditary form of the disease, harboring a constitutional mutation of the RB1 gene. In the remaining 60%, the tumor results from two somatic mutations, each affecting one copy of the gene in a single retinoblast.7
Second malignancies following the hereditary form of retinoblastoma remain higher than those for any other pediatric malignancy.8 Bone and other sarcomas remain the most common second malignancies after retinoblastoma, with melanoma, leukemia, and lymphoma also reported in increased incidence.9-12 In a recent analysis of 1,604 survivors of retinoblastoma, the cumulative risk of a second cancer by 50 years of age was 51% for those with bilateral retinoblastoma and only 5% for those with unilateral disease. The latter statistic is likely representative of the small fraction of genetic cases with only a single eye affected.12 Choice of treatment modality must therefore include some analysis of genetic risk factors for second malignant neoplasm, as well as consideration of other associated morbidity.
Aside from its obvious adverse physiologic and psychologic effects, enucleation can be associated with chronic local effects, including discharge from the orbit, contraction of the socket, and extrusion of the implant.13 Both enucleation and high-dose EBRT (35 Gy) have been shown to result independently in significant compromise of orbital development and reduction in orbital volume.14
EBRT provides good local control in retinoblastoma, especially when used in conjunction with local therapy.15 However, it has significant local side effects, including xerophthalmia, cataract, retinopathy, and keratopathy. It can adversely affect orbital growth and often does not spare vision.4,13 It can multiply the already increased risk of second malignancies in retinoblastoma patients in a dose-dependent manner.12,16-18 For those with hereditary disease, the proportion of second cancers in patients receiving EBRT was 58% overall, compared with 27% in those not treated with radiation therapy.12 This effect may be age dependent, with the greatest risk in those retinoblastoma patients with hereditary disease treated under 1 year of age.7,19,20
Local treatment modalities, such as radioactive plaque, cryotherapy, thermotherapy, and laser therapy, have been used successfully to treat smaller tumors but are less effective against advanced disease.3,21-23 Although chemotherapy has not been successful in curing intraocular retinoblastoma, it has met with some success in treating extraocular disease.24-29 However, to avoid enucleation and EBRT, with their associated morbidity, and to use novel local ophthalmic therapies in more patients, research has been directed toward chemoreduction—using chemotherapy to reduce tumor volume and thereby increase the efficacy of local therapies.
The objective of this study was to use systemic chemotherapy and local ophthalmic treatment in children with retinoblastoma, thereby avoiding the long-term effects of enucleation and/or EBRT. Pilot data on chemoreduction therapy conducted at The Children’s Hospital of Philadelphia were promising.30,31 This study includes a larger number of patients, not previously reported, followed for a longer time and all treated consistently by both pediatric and ocular oncologists. Short- and long-term adverse effects of therapy are addressed and recommendations for future management are made.
PATIENTS AND METHODS
Newly diagnosed patients with intraocular retinoblastoma were identified at the Ocular Oncology Service at Wills Eye Hospital and at the Division of Oncology at The Children’s Hospital of Philadelphia. The study was a prospective, nonrandomized, single-arm clinical trial. It was approved by the institutional review board at The Children’s Hospital of Philadelphia. Children eligible for consideration were those with intraocular retinoblastoma that would ordinarily require treatment with enucleation and/or EBRT. Children with metastatic disease, prior treatment for retinoblastoma, or inadequate renal function, hepatic function, or hearing were excluded. Patients who could be treated by conservative local ophthalmic methods alone were also excluded.
Between September 1995 and June 1998, 59 children were considered for study. Twelve of the 59 patients were subsequently excluded. Eight had advanced unilateral disease treated with enucleation. These eight patients were all under 1 year of age; thus there was significant concern that they harbored a germline mutation. Children with unilateral disease who underwent enucleation at Wills Eye Hospital during this period and who were not considered for study had large unilateral, unifocal disease. Given the excellent results of enucleation in this group, the benefits of a trial of chemoreduction did not outweigh the potential risks. They were therefore not considered for this study. Three patients had diffuse bilateral disease previously treated with EBRT and/or enucleation; one refused chemotherapy. Nine children with unilateral disease were placed on study. Of the 38 children with bilateral disease, 10 eyes—all Reese-Ellsworth (RE) group 5B—required enucleation before commencing chemotherapy. In these patients, only one eye was followed for the purposes of the study. In all, 75 eyes were followed in 47 children.
Families eligible for participation in the study were counseled by one of the pediatric oncologists and signed informed-consent forms. Complete blood counts and blood chemistries, a urinalysis, and an audiologic test were obtained before chemotherapy. A computed tomography scan of the brain and orbits was also performed to evaluate the extent of intracranial disease. Bone marrow aspirates were obtained only if blood counts were abnormal, and lumbar punctures were performed if there were abnormal CNS findings on the computed tomography scan.
With respect to patients with bilateral disease, there were 18 (47%) female patients and 20 (53%) males with a mean age at diagnosis of 8 months and a median age of 5 months (range, 4 days to 33 months). Fourteen (37%) of the patients had a positive family history of retinoblastoma. One child was determined through amniocentesis to share with her mother an inversion on chromosome 13. (The family history was otherwise negative.) Another was found to have partial trisomy 13. For patients with unilateral disease, there were two females and seven males, with a mean age at diagnosis of 16 months and a median age of 7.5 months (range, 12 days to 6.6 years). One of these patients had a family history of retinoblastoma (Table 1).
Patient Characteristics
Patients received intravenous vincristine, etoposide, and carboplatin (VEC) in six cycles according to protocol CHP 582 (Table 2). Chemotherapy was coordinated with local ophthalmic treatment given during monthly or bimonthly ophthalmic examination under general anesthesia. The ocular oncologists determined the type and number of local ophthalmic treatments. Treatment methods included cryotherapy, argon laser photocoagulation, thermotherapy, and plaque radiation therapy.
CHP-582: Chemotherapy for Intraocular Retinoblastoma
Patients were followed both during and after chemotherapy by ophthalmic and physical examinations by the ocular and pediatric oncologists, respectively. Blood counts and chemistries were monitored closely during therapy. The primary end point of the study was to avoid enucleation or EBRT.
For statistical analysis, an event was classified as either enucleation or EBRT. Event-free survival was calculated using the methods of Kaplan-Meier and differences between groups using log-rank tests.32,33
RESULTS
With a median follow-up time of 13 months (range, 21 days to 2.9 years) after the completion of chemotherapy, 57 eyes (76%) in 31 patients avoided both enucleation and EBRT after study. Four of these patients required additional systemic chemotherapy, consisting of carboplatin alone in three patients and VEC plus cyclosporine in the remaining patient. The decision to administer carboplatin alone was based on pilot data that showed platinum to be synergistic with heat.34 These patients received carboplatin that was delivered within 4 hours of thermotherapy in an attempt to optimize the potential for synergy. Fifty-one eyes (89%) in 29 patients received local therapy. Six children required EBRT in seven eyes (9%), five required enucleation of one eye (7%), and five required a combination of EBRT and enucleation in six eyes (8%). The treatment failure information is summarized in Table 3. Therapy failed in 18 eyes in 16 patients. In five eyes in six patients, with a median follow-up of 6 months, EBRT was effective salvage therapy. In two eyes in one patient, multiple episodes of tumor progression occurred that have been treated with subconjunctival carboplatin and systemic chemotherapy with VEC and cyclosporine at other centers. In five eyes in five patients, enucleation was required following failure of chemoreduction and local ophthalmic therapy. In six eyes in five patients, EBRT failed to provide long-term disease control, and with a median follow-up of 3 months after EBRT, enucleation was necessary. The 10 patients who required enucleation now remain disease-free, with a median follow-up of 12 months without further systemic chemotherapy. None of these patients was found to have disease involving the iris, ciliary body, or optic nerve at the time of enucleation.
Treatment Failures
Event-free survival (ie, free of EBRT or enucleation) for the study population was 74% (95% confidence interval, 60% to 84%) at the median follow-up time of 13 months (Fig 1). This was clearly associated with group of disease, with a P value of less than .0001. Patients with RE group 1, 2, or 3 eyes all had 100% event-free survival, whereas the event-free survival for patients for group 5 eyes was only 39% (95% confidence interval, 18% to 59%). Two of the six patients with group 4 eyes had an event by 13 months of follow-up (Fig 2).
Fig 1. Kaplan-Meier event-free survival.
Fig 2. Kaplan-Meier event-free survival by Reese-Ellsworth group (groups 1, 2, and 3, solid line; group 4, long dash; group 5, short dash).
Thirty-two (84%) of the patients with bilateral disease received local ophthalmic treatment while treated with chemotherapy, with a range of one to 14 treatments per patient. Thirty-three percent of the local treatments were cryotherapy; 63%, thermotherapy; 3%, laser photocoagulation; and the remaining 1%, plaque radiotherapy.
Twenty (53%) of these patients also received local ophthalmic treatment after chemotherapy. Most18 had already received local therapy while on study. During the follow-up, patients received as many as seven individual treatments, with 46% receiving cryotherapy; 35%, thermotherapy; 3%, laser photocoagulation; and 16%, plaque. Treatments continued for as long as 25 months after chemotherapy, with a mean of 145 days and a median of 31 days.
Of the nine patients with unilateral disease, seven received local ophthalmic treatment while on chemotherapy, with a range of one to seven treatments per patient. Local treatments were cryotherapy in four patients, transpupillary thermotherapy in six patients, and laser photocoagulation in one patient. No patients with unilateral disease were treated with plaque radiotherapy. Two patients required treatment with thermotherapy after completion of chemotherapy.
No treatment- or disease-related mortality, dissemination of disease outside the eye, or second primary neoplasms occurred. Toxicity was generally mild. Although cytopenias were seen in 83% of patients, only two required packed RBC transfusions for anemia and granulocyte colony growth factor for improved myeloid recovery. Fever and neutropenia were relatively common, with 21 episodes in 28% of patients, but documented bacterial infections were limited to six patients (9%), one of which had Escherichia coli sepsis. The other infections were central venous cath-eter related and due to Gram-positive organisms. Nonspecific gastrointestinal toxicity with feeding difficulties, dehydration, and/or vincristine neurotoxicity were common (40%), and four patients required some reduction in vincristine dose. There was no associated renal toxicity or ototoxicity.
DISCUSSION
Because of the morbidity and potential mortality (from second malignancies) associated with EBRT, pediatric and ocular oncologists have sought effective alternative methods for treating retinoblastoma. We have successfully used a coordinated approach to the care of children, with retinoblastoma with systemic chemotherapy administered by pediatric oncologists and local ophthalmic therapy administered by ocular oncologists. The physical proximity of The Children’s Hospital of Philadelphia and the Wills Eye Hospital contributed greatly to the ability to evaluate the most appropriate therapy efficiently for each child and to administer chemotherapy in close temporal relation to the ophthalmic therapies. Most patients were treated with both systemic and local therapy, which avoided EBRT and/or enucleation in 76% of eyes.
Several groups have used the concept of chemoreduction and local ophthalmic therapy. Early trials have been promising but somewhat difficult to interpret fully. This is due to a combination of factors: small study samples, a mixture of chemotherapeutic regimens used even in a single report, and variation in what was reported (numbers of patients, tumors, or eyes) and the end point examined (response, survival, avoidance of other therapies).
Some studies have used chemotherapy and either EBRT or plaque together and achieved favorable results. As with our results, control was more problematic with higher-group patients.29,34-36 In a pilot study of 20 patients, Shields et al31 found that two cycles of VEC were sufficient to promote some regression in all 54 tumors studied in 31 eyes, with 89% showing evidence of calcification. However, it was later found that eyes with vitreous and subretinal seeds recurred at a rate of approximately 33%.37 Further follow-up revealed that six cycles of chemotherapy decreased the vitreous seed recurrence, but 75% of the RE group 5 patients required EBRT.38 Based on laboratory data in a rabbit model, thermotherapy combined with carboplatin treatment was assessed, with the rationale that heat is synergistic with platinum compounds. Murphree et al39 showed that this method was effective in treating 10 out of 10 tumors of RE groups 1 through 4. Hypothesizing that retinoblastomas overexpress the multidrug resistance protein P-glycoprotein, p170, Gallie et al have added cyclosporine to VEC and shown an actuarial relapse-free rate of 91% in 28 previously untreated eyes, with a median follow-up of 2.67 years.40 However, in the absence of a randomized trial with patients with otherwise similar prognostic factors who received the same ophthalmic therapy, the role of cyclosporine or any of the single agents remains unclear. Despite the use of cyclosporine, treatment failures occur, with these tumors overexpressing p170.40
Our data support the use of chemoreduction combined with local ophthalmic therapy in patients with eyes in RE groups 1 to 3. All patients with these lower-group tumors were successfully treated with minimal toxicity and with preservation of the globe. Although the numbers were small, this therapy may be less effective for group 4 patients, who have been in the unfavorable group with respect to vision and successful salvage treatment. However, if EBRT can be avoided in even a few children with more advanced disease, this approach appears warranted, given the significantly increased risk of second malignancy associated with EBRT in children with hereditary disease. The group 5 patients, with or without vitreous seeds, present the greatest obstacle to cure with combined chemotherapy and local therapy.38 As has been reported, it is also these patients who require plaque as secondary local treatment.23
Although this study reports on a relatively large number of patients treated in a consistent manner, it is limited by its size and by short follow-up time. Selection bias in the choice of patients could not be avoided, as these patients were evaluated by ocular oncology with respect to the ability to provide local therapy in the event the tumor volume could be reduced by chemotherapy. It is likely that additional patients with group 4 or 5 disease will continue to require intervention with EBRT and/or enucleation with longer follow-up. However, short-term toxicity is mild. No patients developed metastatic disease on therapy or sustained a second malignancy, although the follow-up period is still too short for the latter to be seen. Secondary acute leukemia has been reported following the use of etoposide with relatively short latency periods of 1 to 7 years.41,42 The dosage and schedule of administration of etoposide in this protocol have not been associated with higher incidences of secondary leukemia, and there is no rationale for believing that children with the hereditary form of retinoblastoma, those with bilateral disease, are more susceptible to the carcinogenic effects of etoposide.
We conclude that in appropriately selected patients with RE groups 1, 2, or 3, chemoreduction and local ophthalmic therapy are effective and carry little morbidity. However, other treatment options need to be developed for group 4 and 5 patients. The addition of subconjunctival chemotherapy, cyclosporine, and plaque or adjuvant low-dose radiation therapy would require a randomized trial with large numbers of patients to determine whether these approaches offer any advantages. An international, large, randomized trial is currently being planned with stratification by RE group, which has been shown repeatedly to be the most consistent prognostic factor. It will only be through longer follow-up with larger numbers of patients that this question can be adequately addressed.
Acknowledgments
Supported in part by the Eye Tumor Research Foundation, Philadelphia, PA (J.A.S., C.L.S.), the Award of Merit in Retina Research, Houston, TX (J.A.S.), the Macula Foundation (C.L.S.), and grant no. 3M01RR00240-32S9 from the National Institutes of Health (D.L.F.).
Footnotes
-
Presented in part at the Thirty-Fifth Annual Meeting of the American Society of Clinical Oncology, Atlanta, GA, and at the 1999 Joint Annual Meeting of the International Society of Paediatric Oncology and the American Society of Paediatric Hematology/Oncology, Montreal, Canada.
- Received April 12, 1999.
- Accepted July 23, 1999.
