Kaplan-Meier estimates of patients free of new tumor after 6 cycles of chemoreduction for retinoblastoma.
Shields CL, Shelil A, Cater J, Meadows AT, Shields JA. Development of New Retinoblastomas After 6 Cycles of Chemoreduction for Retinoblastoma in 162 Eyes of 106 Consecutive Patients. Arch Ophthalmol. 2003;121(11):1571-1576. doi:10.1001/archopht.121.11.1571
Copyright 2003 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2003
To evaluate the occurrence of new retinoblastomas in patients treated with 6 cycles of chemoreduction.
Prospective nonrandomized single-center case series.
Ocular Oncology Service at Wills Eye Hospital of Thomas Jefferson University, Philadelphia, Pa, in conjunction with the Division of Oncology at The Children's Hospital of Philadelphia.
A total of 162 eyes of 106 patients with retinoblastoma treated with 6 cycles of chemoreduction between January 1, 1995, and May 31, 2002.
All patients received intravenous chemoreduction with vincristine sulfate, etoposide, and carboplatin, combined with focal treatment (cryotherapy or thermotherapy) to each retinal tumor.
Main Outcome Measure
Development of new intraretinal retinoblastoma during or after treatment with chemoreduction. Recurrent subretinal tumor seeds or vitreous seeds were excluded from this analysis, and only primary new intraretinal tumors were included.
Of 28 patients with unilateral retinoblastoma, new intraretinal tumor development was found during or after chemoreduction in 2 (9%) of the 23 patients with sporadic disease and 4 (80%) of the 5 patients with familial disease. The new tumor was located in the macula in none, between the macula and equator in 7 (54%), and between the equator and ora serrata in 6 (46%). Of the 78 patients with bilateral retinoblastoma, new tumor development was found during or after chemoreduction in 11 (19%) of the 57 patients with sporadic disease and 8 (38%) of the 21 patients with familial disease. The new tumor was macula in 2 (4%), between the macula and equator in 30 (55%), and between the equator and ora serrata in 23 (42%). Overall, according to Kaplan-Meier analysis, new tumor development occurred in 23% of patients by 1-year follow-up and 24% by 5-year follow-up. By multivariate analysis, the most important risk factors for the development of new tumors was younger age at presentation(median age, 2 months with new tumor vs 9 months without new tumor) and family history of retinoblastoma (12 [48%] of patients with new tumor vs 14 [17%] without new tumor).
Children with retinoblastoma treated with chemoreduction should be followed for new intraretinal tumor development, as it peaks at a mean interval of 5 months after initiation of chemoreduction and affects 24% of patients by 5 years of follow-up. New tumors are most commonly found in those who develop disease as young infants and those with a family history of retinoblastoma.
DURING THE PAST decade, chemoreduction has emerged as the most important conservative therapy in the management of retinoblastoma.1- 6 With this method, chemotherapy is delivered intravenously to the affected child to reduce the size of each retinoblastoma. Rarely does the chemotherapy alone completely control retinoblastoma; hence, the method is specifically termed chemoreduction. This is followed by focal consolidation with the use of cryotherapy or thermotherapy to each tumor to permanently devitalize the retinoblastoma.7,8 Chemoreduction has been most successful for tumors without associated subretinal fluid, subretinal seeds, and vitreous seeds. Overall, globe salvage by 5 years after chemoreduction for retinoblastoma is 85% for eyes classified as Reese-Ellsworth groups I to IV and 47% for eyes in Reese-Ellsworth group V.9
The most haunting problem with chemoreduction for retinoblastoma is recurrence of tumor seeds in the subretinal space or vitreous cavity months or years after completion of therapy. For children who have retinoblastoma and seeds in the subretinal space, recurrence of at least 1 seed is noted in 62% of eyes by 5 years of follow-up.10 For those who have retinoblastoma and vitreous seeds, recurrence of vitreous seeds is found in 50% of eyes by 5 years of follow-up.10 These findings underscore the importance of long-term, cautious, and meticulous observation of children with retinoblastoma.
Another important concern with chemoreduction for retinoblastoma is the development of new intraretinal tumors during or after therapy. These are different from recurrent seeds in that they occur within the retina as an independent new focus of tumor, often with slightly dilated retinal feeder vessels. Recurrent seeds, on the other hand, occur either under the retina(subretinal) or over the retina (vitreous) as detached tiny viable portions of one of the existing tumors and typically without feeding retinal vessels. Occasionally, subretinal seeds closely resemble new intraretinal tumors, but can be differentiated by the fact that seeds are generally visualized from the initial examination, whereas new tumors are not visualized at the initial examination and the retina, albeit quite thin, is draped overlying the subretinal seeds. In this report, we analyze the incidence and features of new retinal tumors in 162 eyes treated with chemoreduction for retinoblastoma.
All new patients with intraocular retinoblastoma who were treated initially with chemoreduction (approved by the Institutional Review Board of The Children's Hospital of Philadelphia, No. 582) on the Ocular Oncology Service, Wills Eye Hospital, Thomas Jefferson University, Philadelphia, Pa, in conjunction with the Division of Oncology at The Children's Hospital of Philadelphia were identified. The eligibility criteria for treatment with chemoreduction6 were the presence of intraocular retinoblastoma in which either eye would ordinarily require enucleation or external beam radiotherapy for cure of the disease on the basis of published indications.1,2,11,12 Patients whose tumor(s) could be properly controlled with conservative methods alone(cryotherapy, laser photocoagulation, thermotherapy, chemothermotherapy, and/or plaque radiotherapy) were not eligible for inclusion in the chemoreduction protocol. Exclusion criteria for treatment with chemoreduction included biomicroscopic evidence of iris neovascularization, neovascular glaucoma, and tumor invasion into the anterior chamber, iris, optic nerve, choroid, or extraocular tissues as documented by clinical, ultrasonographic, and neuroimaging modalities. Exclusion criteria from a systemic standpoint were evidence of systemic metastasis, previous chemotherapy, previous treatment for retinoblastoma, and inadequate organ function of the kidney, liver, or ear. The chemotherapeutic agents included intravenous vincristine sulfate (0.05 mg/kg), etoposide (5 mg/kg), and carboplatin(18.6 mg/kg). All 3 were given on day 0, and just etoposide was given on day 1. The duration of treatment was planned for 6 monthly cycles. The potential risks and benefits of the chemoreduction protocol were discussed with the patient's family and informed consent was signed.
Ocular oncologic follow-up was provided at examination with the patient under anesthesia every 1 to 2 months after initiation of chemoreduction until complete control of the disease was achieved. Thereafter, examinations were provided every 2 to 4 months as needed. At each examination, the status of the individual retinal tumors, vitreous seeds, subretinal seeds, and subretinal fluid was noted. Adjuvant treatment to the regressed retinal tumors by means of thermotherapy or cryotherapy was provided while the patient was on the chemoreduction protocol. The patients were not randomized to treatment.
All data were collected in a prospective fashion. Photographic documentation and ultrasonography of the retinal tumors were performed. For each patient, age at diagnosis, race, sex, and hereditary pattern (sporadic or familial) were determined. The eye was assessed for laterality of involvement (unilateral or bilateral), total number of retinal tumors per eye, and Reese-Ellsworth classification.1,2 Each eye was assessed for the development of a new retinal tumor and its location and treatment. The total number of new retinal tumors per eye was recorded. The interval from initiation of chemoreduction to the new tumor development was calculated. The follow-up was continued until the date the patient was last examined or until the date of enucleation of the eye.
The clinical data were analyzed with regard to the outcome of first new tumor development. The effect of each individual clinical variable recorded at the time the patient was first examined on the Ocular Oncology Service on the development of this outcome was analyzed by a series of univariate Cox proportional hazards regressions. The correlation among the variables was determined by using Pearson correlations. All variables were analyzed as discrete variables (continuous variables were analyzed by grouping them into discrete categories). Variables that were significant on a univariate level (P≤.05) were entered first into the multivariate Cox regression analysis. For variables that showed a high degree of correlation, only one variable from the set of associated variables was entered at a time into subsequent multivariate models. A final multivariate model tested variables that were identified as significant predictors (P≤.05, Wald statistic or 95% confidence interval of the relative risk) from the initial model as well as variables deemed clinically important for the outcomes of development of new tumor. In the final model, a predictor was considered as a significant risk factor if the 95% confidence interval of its relative risk did not contain a risk value of 1.
One hundred sixty-two eyes of 106 patients were treated with 6 cycles of this chemoreduction protocol between January 1, 1995, and May 31, 2002. The demographic features are listed in Table 1. The hereditary pattern of the retinoblastoma was sporadic in 75% and familial in 25%. The disease was unilateral in 26% and bilateral in 74%. The affected eye was the right in 83 cases (51%) and the left in 79 cases (48%). The Reese-Ellsworth classification of each eye is listed in Table 2. The mean follow-up was 26 months(median, 20 months; range, 1-83 months). During or after chemoreduction, new retinal tumors were discovered in 25 (24%) of the 106 patients at a mean of 5 months (median, 4 months; range, 1 to 23 months) after initiation of chemoreduction. According to Kaplan-Meier analysis, new tumors were discovered in 23% of patients at 1-year follow-up and 24% at 5 years (Figure 1 and Table 3). More specifically, new retinal tumors were found in 2 (9%) of the 23 patients with unilateral sporadic retinoblastoma, 4 (80%) of the 5 patients with unilateral familial disease, 11 (19%) of the 57 patients with bilateral sporadic disease, and 8 (38%) of the 21 patients with bilateral familial disease. The new tumors were 2 mm or less in basal dimension and thickness in all cases. Treatment of new tumors was conservative in all cases, with cryotherapy, thermotherapy, or both. The clinical features of the new tumors for children with unilateral and bilateral disease are listed in Table 4. Clinical features of eyes with new tumors vs eyes without new tumors are compared in Table 5.Univariate and multivariate risk factors for new tumor development are listed in Table 6.
Chemoreduction has assumed an important role in the management of retinoblastoma. Our study represents the first large group evaluation, to our knowledge, of new tumors after chemoreduction for retinoblastoma. From a historical perspective, however, the development of new tumors has been recognized after various therapies for retinoblastoma. In 1971, Bedford and associates13 reported that 5 (8%) of 58 eyes treated with 4000 rad (40 Gy) of external beam radiotherapy to the whole eye developed new primary tumors after treatment. On the other hand, they noted that 12 (20%) of eyes treated with focal measures such as cryotherapy, light coagulation, or cobalt plaque radiotherapy developed new primary tumors. Salmonsen and coworkers14 later evaluated new tumors after external beam radiotherapy, but they excluded eyes with Reese-Ellsworth group Vb to avoid confusion with implantation growth of seeds. Their data included information on patients (not eyes) with unilateral and bilateral retinoblastoma. They noted that 38 (11%) of 360 patients with retinoblastoma treated with external beam radiotherapy showed the development of new tumors. They found that most new tumors were small, at 2.2 mm, and the geographic distribution of new tumors was mostly in the retinal periphery(90%) rather than the posterior pole (10%). Despite radiotherapy to the whole eye, they speculated that the new tumors could have been previously overlooked by the observer, missed by the radiation technique, or related to slower maturation of the retinal periphery, allowing new tumors to develop at a later time. Their latter speculation was supported by information from Hollyfield's work with the tadpole eye in which retinal maturation occurred from posterior to anterior.15 Similar results were later published by Messmer and coworkers16 in their analysis of 127 patients with hereditary retinoblastoma treated with external beam radiotherapy. Using life-table analysis, they found that 27% of eyes developed new tumors, usually anterior to the equator of the eye.
In 1994, Abramson and associates17 reported new tumor development in the same or opposite eye in 25 (6%) of 427 eyes with unilateral retinoblastoma treated with various modalities including radiotherapy, chemotherapy, cryotherapy, photocoagulation, enucleation, and other combination methods. Children with unilateral retinoblastoma in the first 6 months of life and with family history of the disease were at greatest risk for development of new tumors. Abramson and coworkers18 reported on 355 eyes with bilateral retinoblastoma, treated with various nonenucleation measures, and found new tumor development in 88 eyes (25%). New tumor development was greater in children younger than 6 months (45%), whereas those older than 6 months showed new tumors in 14%. Merrill and coworkers19 found new tumor development in germinal mutation retinoblastoma to be similar when external beam radiotherapy (7%) and focal therapies (11%) were compared, whereas Hadjistilianou and coworkers20 found the rates quite different at 19% and 85%, respectively.
Currently, chemoreduction is the most important conservative therapy in the management of retinoblastoma, particularly bilateral disease.21,22 External beam radiotherapy is generally reserved for patients in whom chemoreduction fails. Scott and associates23 noted new tumor development in 4 children treated with chemoreduction, 3 of whom were 7 months or younger at diagnosis. They found new tumors after 6, 7, 5, and 2 cycles of intensive chemotherapy. In our study, we evaluated new tumors after chemoreduction for retinoblastoma in a large cohort of patients.
In our analysis, we found new tumor development during or after 6 planned cycles of chemoreduction in 24% of 106 patients, similar to the incidence of new tumors after external beam radiotherapy and other methods.13,14,16- 20 The mean time interval for new tumor development was 5 months (median, 4 months; range, 1-23 months) after initiation of chemoreduction. By Kaplan-Meier analysis, new tumors were discovered in 23% of patients by 1 year and 24% by 5 years of follow-up. More specifically, new retinal tumors were found in 9% of the 28 patients with unilateral sporadic retinoblastoma, 80% of the 5 patients with unilateral familial disease, 19% of the 57 patients with bilateral sporadic disease, and 38% of the 21 patients with bilateral familial disease. Similar to the observations of Salmonsen and associates, 14 new tumor development was found rarely in the macula (3%) and more commonly between the macula and equator (54%) or equator to ora serrata (43%). The disparity in location of new tumors could represent the relative area of retinal surface occupied by these regions, or it could reflect the slightly delayed maturation of the retinal periphery as compared with the posteriorly located macula.
Patients with new tumor development were more likely to be younger at diagnosis (median age, 2 months) and have a family history of retinoblastoma(Table 3 and Table 6). On the other hand, those without new tumors after chemoreduction were more likely to be older (median age, 9 months) and to have sporadic disease. Bilateral retinoblastoma was found equally in both groups, representing 76% of those with new tumors and 72% of those without new tumors. Similar findings were noted for unilateral retinoblastoma, as it represented 24% of those with new tumors and 27% of those without new tumors. Of the 6 children with unilateral retinoblastoma who developed new tumors, 4 had familial cases and 2 had sporadic cases. Thus, children with both unilateral and bilateral retinoblastoma should be followed up for the development of new primary tumors within the retina. Children with bilateral disease are presumed to have the retinoblastoma mutation in every retinal cell, and any one of those cells might undergo the second event before differentiation and could develop into a new tumor. Approximately 10% of patients with unilateral tumors, regardless of family history, are also presumed to have the gene mutation and would, therefore, be at risk for new tumors.
Fortunately, all of the new tumors were small at detection and all were treatable with focal measures such as cryotherapy and/or thermotherapy. We did not administer additional chemotherapy or provide radiotherapy. Cautious follow-up, especially of the equatorial and peripheral retina, by means of indirect ophthalmoscopy and scleral depression technique, was vital for such early detection. Salmonsen and coworkers14 also noted small size of new tumors at detection after radiotherapy, but in their series a wider range of size was found, as the largest new primary tumor was 15 mm.
Screening for new tumors and tumor recurrence should be intensive during the first 3 years after chemoreduction. It is speculated that differentiation of retinoblasts into mature cells incapable of tumor formation is nearly complete after this period. In our series, new tumors occurred at a mean of 5 months from initiation of chemoreduction, at a time while the child was still receiving the planned 6 monthly cycles. We speculate that these new tumors develop despite chemotherapy because they receive an inadequate dose of chemotherapy as a result of their minute blood supply, or they may, in fact, be resistant to chemotherapy. Tumors that develop after the discontinuation of chemotherapy may not have been adequately exposed to chemotherapy. The latest new tumor occurred 23 months after initiation of chemoreduction. Salmonsen and associates14 found new tumor development after external beam radiotherapy at a mean of 3.5 months from the previous examination, and the latest new tumor occurred at an interval of 17 months from the previous examination. However, in that report, there was no information on the interval from radiotherapy to development of new tumor. This information on new tumor development within 2 years from chemoreduction, coupled with the previously published information on tumor and seed recurrence found to be greatest in the first 3 years after chemoreduction, 10 underscores the importance of cautious follow-up of such children and include examination with the patient under anesthesia every few months during the first 3 to 4 years after therapy.
There are limitations to our analysis that should be realized. It can be difficult to differentiate new tumors from recurrent seeds, as mentioned previously. Some of the tumors that we classified as "new" could have been a recurrent seed or a tumor that was overlooked on previous examination. In addition, as with any study on a new therapy, longer follow-up will be important in delineating the long-term findings.
In summary, children with retinoblastoma treated with chemoreduction should be followed up both systemically and ophthalmologically.24 From an ocular standpoint, they are at risk for recurrence of tumor and recurrence of subretinal and/or vitreous seeds, and they should also be carefully watched for the development of new primary intraretinal tumors. Such new tumors are generally found in patients who are very young at diagnosis (median age, 2 months) and those with the genetic form of the disease. With early detection of these tumors, focal treatment with cryotherapy or thermotherapy can be successful for tumor control.
Corresponding author and reprints: Carol L. Shields, MD, Ocular Oncology Service, Wills Eye Hospital, 840 Walnut St, Philadelphia, PA 19107 (e-mail: email@example.com).
Submitted for publication February 12, 2003; final revision received June 16, 2003; accepted June 20, 2003.
This study was supported by the Eye Tumor Research Foundation, Philadelphia(Dr C. L. Shields); the Macula Foundation, New York, NY (Dr C. L. Shields); the Rosenthal Award of the Macula Society, Barcelona, Spain (Dr C. L. Shields); and the Paul Kayser International Award of Merit in Retina Research, Houston, Tex (Dr J. A. Shields).