Eyes are shown before (left) and after treatment (right). All eyes have International Classification of Retinoblastoma group D tumors.
eTable. Intensity of Adjuvant Regimen
Customize your JAMA Network experience by selecting one or more topics from the list below.
Abramson DH, Shields CL, Munier FL, Chantada GL. Treatment of Retinoblastoma in 2015: Agreement and Disagreement. JAMA Ophthalmol. 2015;133(11):1341–1347. doi:10.1001/jamaophthalmol.2015.3108
The management of intraocular retinoblastoma is rapidly changing, and even recent reviews on the subject are behind existing practices. The 4 authors of this report collectively represent their management strategies with an emphasis on areas of agreement and disagreement. Ophthalmic artery chemosurgery and intravitreous chemotherapy have completely replaced external beam radiotherapy, reduced the use of systemic chemotherapy, and diminished enucleations by 90% without evidence of compromising patient survival.
Although the management of intraocular retinoblastoma has evolved steadily during the past 100 years in association with progressive salvage of life, eye, and vision,1 the pace of advancement within the last 10 years has amazed even specialists in the field. The introduction of ophthalmic artery chemosurgery (OAC)2 and intravitreous chemotherapy3 has transformed treatment algorithms completely. Many articles published within the last 5 to 6 years are out of date. For example, in a 2009 literature review and commentary by Lin and O’Brien,4 OAC and intravitreous chemotherapy are not mentioned. Today, most patients in our centers (and in most leading centers worldwide) receive one or both of these modalities. In their review, Lin and O’Brien state that patients with International Classification of Retinoblastoma group D disease “ rarely retain the eye.”4 Today, we generally save most such eyes. Lin and O’Brien further comment that patients with extraocular retinoblastoma have a “poor prognosis for survival,”4 but today, most of these patients survive (with treatment).
In another review of therapy for retinoblastoma from 2011,5 intravitreous melphalan hydrochloride was not listed. In an overview of retinoblastoma published in 2014,6 intra-arterial chemotherapy and intravitreous chemotherapy were not mentioned. In the 2015 National Cancer Institute Physician Data Query recommendations for retinoblastoma treatment,7 intravitreous chemotherapy again was not mentioned.
Articles published in the last 10 years by 3 of us are also outdated. One of us (D.H.A.) coauthored a 2007 review8 that highlighted the use of external beam radiotherapy (which subsequently has been abandoned) and made no mention of OAC or intravitreous chemotherapy. Similarly, one of us (F.L.M.) coauthored a 2006 review9 that made no mention of intra-arterial or intravitreous chemotherapy and highlighted possible future radiotherapy techniques that might be useful. In 2008, one of us (C.L.S.)10 wrote that the treatment of eyes with advanced retinoblastoma was enucleation, but most patients now receive OAC.
Retinoblastoma management strategies evolved without randomized multicenter clinical trials. External beam radiotherapy began within 10 years of the discovery of the x-ray by Roentgen without scientific experiments on retinoblastoma. Stallard11 introduced plaque radiotherapy without human or animal studies. Photocoagulation and cryotherapy were introduced without animal investigation. When systemic chemotherapy replaced radiotherapy for retinoblastoma in the mid-1990s, no comparative trials were performed and, furthermore, the retinoblastoma classification schemes changed, making comparison with historical data nearly impossible. Despite this lack of randomized clinical trials in the past 100 years, patient survival and ocular preservation have improved steadily. Retinoblastoma is now the most curable pediatric cancer and stands alone as the single cancer without previous randomized multicenter clinical treatment trial(s).
Within the last 9 years, changes in the management of retinoblastoma have been rapid, especially with OAC and intravitreous chemotherapy. Given the rapid changes in the management of retinoblastoma, we herein review our current management of retinoblastoma, highlighting areas of agreement and disagreement. Our 4 centers represent the two largest in North America (Ophthalmic Oncology Service, Memorial Sloan Kettering Cancer Center, New York, New York, and Wills Eye Hospital, Philadelphia, Pennsylvania), which collectively treat most patients with retinoblastoma in the United States; the largest clinic in Latin America (Hospital J.P. Garrahan, Buenos Aires, Argentina), which is a pioneer in chemotherapy use for extraocular retinoblastoma; and one of the largest clinics in Europe (Jules-Gonin Eye Hospital, Lausanne, Switzerland), which is the center with the largest experience in intravitreous treatment.
Four international experts in retinoblastoma discuss areas of agreement and disagreement in retinoblastoma management.
Retinoblastoma management has changed dramatically in the past 10 years owing to ophthalmic artery chemosurgery and intravitreous chemotherapy.
Ophthalmic artery chemosurgery has eliminated 90% of enucleations performed just 5 years ago.
Treatment of more advanced disease has not resulted in a compromise of patient survival.
Table 1 summarizes areas of agreement and disagreement among us for the management of unilateral retinoblastoma and high-risk disease. Table 2 summarizes areas of agreement and disagreement among us for the management of bilateral retinoblastoma. Table 3 provides patient survival data for all patients receiving OAC at our centers as of June 1, 2015.
Although all of our centers use the Reese-Ellsworth classification and International Classification of Retinoblastoma, 3 versions of the latter exist, including the Los Angeles version developed by Murphree et al,12 the Philadelphia version by Shields and Shields,13 and the Children’s Oncology Group (COG) version.14 The New York center uses the COG version15 and the Philadelphia center uses the Philadelphia version, whereas the Lausanne and Buenos Aires centers use the Los Angeles version. Although all 3 classifications use the same letters (A through E), with A being the smallest tumor burden and E the largest, important differences remain, especially among groups D and E tumors. In the COG classification, group D tumors include diffuse vitreous and/or subretinal seeding less than 6 mm from tumor and/or subretinal fluid more than 6 mm from tumor margin.15 In the Philadelphia classification, group D tumors are defined as tumor with subretinal and/or vitreous seeds greater than 3 mm from tumor.13 Because the COG classification defines group C tumors as localized vitreous and/or subretinal seeding less than 6 mm from the margin, a tumor with vitreous seeds extending 5 mm from a tumor is classified as group C in the COG version and group D in the Philadelphia version. These and other differences in the definition of group E tumors make comparison of treatment results challenging when different international classifications are used.
Small tumors without seeding or retinal detachment can be managed with laser photocoagulation or cryotherapy (with a single exception).16,17 We use the 810-nm diode laser (with the indirect ophthalmoscope [D.H.A. and C.L.S.] and the microscope [F.L.M.]). The exception to this management consists of tumors impinging on the fovea.18 Although these tumors respond well to laser photocoagulation, the chance of compromising central vision is high, so we prefer to use some form of chemotherapy to shrink the tumor before definitive laser treatment. Systemic chemotherapy with 1 drug (D.H.A.), 2 drugs (F.L.M.), or 3 drugs (C.L.S.) is used, or OAC is used.
In the radiotherapy era (1903 to the mid-1990s), most group D tumors were enucleated. The International Classification of Retinoblastoma did not exist, but the Reese-Ellsworth group V classification is similar to group D, and a report on Reese-Ellsworth group Vb tumors from New York18 emphasized that 60% of such eyes were primarily enucleated and of those undergoing radiotherapy, only 50% were saved. Thus, only 20% of such eyes were salvaged with external beam radiotherapy. With multiple-agent chemotherapy, success rates vary from 0% to 47%.19 Today, we treat most group D tumors with primary OAC, and the globe salvage is 94%1,20,21 (Figure). The choice of drugs is also the same in all 4 centers, with melphalan, carboplatin, and topotecan hydrochloride. Doses are similar (2.5-7.5 mg, 40-70 mg, and 1-2 mg, respectively).22 Doses are based on patient age and dose escalation, and multiple drugs are used for advanced tumors. Two centers (C.L.S. and F.L.M.) reported complications in their first few cases but now experience few (<5%) problems.
The greatest fear of intravitreal injections has always been extraocular spread, which was previously reported by Suzuki et al,23 but which has not occurred in our centers to date. We all use the injection technique promulgated by Munier et al.24 The eye is softened with an anterior chamber tap (using a 34-gauge needle) or digital massage; a fine needle (30, 32, or 33 gauge) is inserted at the pars plana; and 20 to 30 μg of melphalan hydrochloride is injected. In the New York center, we have collected 200 needle washings immediately on removing the needle and found no cancer cells. Cryotherapy is used to protect the site from tumor escape before the needle is removed. Two of us (D.H.A. and G.L.C.) irrigate the eye with distilled water after needle removal because distilled water has been shown to be effective at killing retinoblastoma cells in culture within a few minutes.25 Primary treatment with OAC has not resulted in decreased patient survival (Table 1).
Systemic chemotherapy and radiotherapy are not used for unilateral treatment-naive group D tumors by any of us, with 1 exception. In infants younger than 3 months and weighing less than 7 kg, we choose to defer definitive OAC to allow the femoral artery to grow before attempting catheterization. In that case, 1 or 2 doses of single-agent carboplatin (18.7 mg/kg) are given. When the infant is 3 months of age and weighs at least 7 kg, OAC is used. This system is termed bridge therapy.26Focal consolidation after OAC is commonly performed (although less frequently than after systemic chemotherapy). Occasionally the single-agent chemotherapy causes sufficient shrinkage so that OAC is not needed.27
Eyes with group D tumors that did not resolve with systemic chemotherapy or radiotherapy were always enucleated in the past. In the radiotherapy era, a second course of radiotherapy for recurrent disease salvaged only 2.2% of eyes.28
Today, all 4 of our centers use a combination of OAC29 and intravitreous injections of melphalan. When OAC is used alone, 83% of such eyes are saved (Figure, D) and 90% retain or improve electroretinographic recordings.30 We increasingly use supplemental intravitreous chemotherapy injections to salvage such eyes. Success rates appear greater with the addition of intravitreous chemotherapy, but the electroretinographic recording decreases at a rate of about 5% with each injection (although central vision is rarely affected).31 Most group D tumors are now managed with primary OAC in 3 centers (by D.H.A., C.L.S., and F.L.M.), but the South American center (G.L.C.) still enucleates most eyes with group D tumors. This result has been accomplished without compromising patient survival (Table 3).
Each of the 3 international classifications designated group E tumors as the most advanced intraocular group, and E was thought to be synonymous with enucleation. The reason for this supposition was that eyes with group E tumors could not be salvaged with systemic chemotherapy.32 Numerous reports from the 1990s confirmed that eyes with group E tumors were enucleated.32,33 We still manage virtually all cases of buphthalmos, neovascular or rubeotic glaucoma, aqueous seeding, or pthisis bulbi and all cases of eyes with transcleral extension with primary enucleation (or neoadjuvant or adjuvant systemic chemotherapy). Progressively, however, all of us have been cautiously treating and retaining more eyes with group E tumors than in the systemic chemotherapy era (from the 1990s through 2006). Eyes with group E tumors include those with total retinal detachment and tumor extending up to the lens. Ironically, such eyes are ideal candidates for OAC because eyes with detachment tend to show fairly reliable success.20,34 About 90% of eyes with extensive retinal detachment can be saved, and 20% will have significant improvement in electroretinographic function after therapy. In South America and in other developing countries, enucleation as a primary therapy is performed more commonly.
One hundred years ago, bilateral retinoblastoma was managed with bilateral enucleation (often with exenteration). By the 1940s, bilateral disease was managed with enucleation of the eye with more advanced disease and radiotherapy in the remaining eye. Fifty years ago, physicians began using bilateral radiotherapy successfully for bilateral disease.1 Overall, both eyes were ultimately enucleated in 25% of cases, 1 eye was salvaged in 50% of cases, and 2 eyes were salvaged in 25% of cases. Today, management is very different. Radiotherapy has been abandoned for bilateral cases because of the subsequent risk for second cancers, which is increased with radiotherapy alone,35 radiotherapy in the first year of life,36 radiotherapy combined with systemic chemotherapy,37 and high-dose radiotherapy (a clear dose-response curve exists).38 After the radiotherapy era, multiple-agent systemic chemotherapy was adopted by all of us for bilateral cases. Consolidation with focal treatments (laser and cryotherapy) was important for tumor control. Salvage rates for eyes with group D bilateral tumors were similar to those for eyes with group D unilateral tumors (<50%), which was disappointing. Although radiotherapy and its consequences were avoided, we discovered a host of chemotherapy-related complications, including fever, neutropenia, the need for venous port access, port infections, the need for hematologic transfusions, hearing loss, and rare chemotherapy-induced second cancers.39
The algorithm for the management of bilateral retinoblastoma has changed in the OAC era, and we have areas of agreement and disagreement. We agree that bilateral cases with small, extrafoveal tumors can be managed with focal treatments alone. In addition, 3 of us (D.H.A., F.L.M., and G.L.C.) agree that in asymmetric disease, 1 eye can be managed with focal techniques and the fellow eye can be managed as outlined above for eyes with unilateral advanced disease. One of us (C.L.S.) prefers systemic chemotherapy for asymmetric bilateral disease. We agree that in bilateral cases in which 1 eye is primarily enucleated and the other eye has advanced disease, systemic chemotherapy or primary OAC (sometimes with primary intravitreal chemotherapy) can be used.
We disagree in cases of bilateral advanced disease, in which attempts to salvage both eyes are used. Two of us (D.H.A. and G.L.C.) use bilateral, simultaneous OAC (called tandem therapy).40 Total treatment time for both eyes is 1 hour, and most such eyes are saved without systemic complications.40 Two of us (C.L.S. and F.L.M.) prefer to use systemic chemotherapy first and, if that treatment is unsuccessful, to use OAC.
For bilateral cases that are not advanced but have tumors too large for focal techniques alone (eg, some eyes with group C disease), 3 centers (C.L.S., F.L.M., and G.L.C.) start with systemic chemotherapy, and 1 center (D.H.A.) starts with OAC. In the first few years of OAC use, complications included eyelid edema, cutaneous hyperemia, loss of eyelashes, retinal vascular attenuation with peripheral nonperfusion, optic neuropathy, and choroidal vascular atrophy. More recently, these complications have been reduced to fewer than 5% of eyes based on improved techniques of catheterization and infusion.
The incidence of metastatic disease varies considerably throughout the world. In the United States and Switzerland, fewer than 5% of patients with retinoblastoma develop metastases. In Argentina, the incidence is twice that, but throughout most of the world, the incidence of metastases from retinoblastoma of more than 50% is common. Protocols developed worldwide are now in use by all of us, and survival after metastases (not involving the central nervous system) exceeds 75% in all centers.41 All centers use a combined approach of multiple-agent intensive systemic chemotherapy, autologous hematopoietic stem cell rescue when metastatic disease is present, and radiotherapy to bulky sites, although different centers use the TNM system or the International Retinoblastoma Staging System.42
After enucleation, children with microscopic tumor residua, such as tumor at the cut section of the optic nerve and transscleral invasion into the orbit, have an unequivocal risk for extraocular relapse. We use protocols for adjuvant therapy involving multiple-agent systemic chemotherapy, often combined with focal radiotherapy to the orbit.5,43 About 70% of children worldwide so treated currently survive.44
We disagree about the definition of high-risk pathologic features guiding the use of adjuvant therapy for prevention of extraocular relapse. The following detail these therapies.
This group includes combined prelaminar and focal choroidal invasion, with which patients have a 2% risk for extraocular relapse if not given adjuvant chemotherapy,45 and massive choroidal invasion,46 with which patients have a 6% risk for extraocular relapse if not associated with postlaminar invasion.47 The literature suggests that invasion into the anterior chamber alone does not increase the chance of metastasis,48 and that it does.49 Two centers (D.H.A. and G.L.C.) do not treat these cases with adjuvant chemotherapy because overall survival is not affected (because treatment when relapse occurs is usually curative). Two centers (C.L.S. and F.L.M.) use moderate-intensity systemic chemotherapy as adjuvant therapy to prevent metastases.
Because adjuvant therapy is used in most cases of tumor extension beyond the lamina cribrosa and those with intrascleral invasion, the risk for extraocular relapse in these children is unknown. Although we agree that the features discussed below increase the risk for metastases, we do not agree completely about management, because the treatments have toxic effects and most of these patients do well without additional therapy. One center (D.H.A.) does not treat these cases with adjuvant chemotherapy and only treats if or when metastatic disease is evident. One center (G.L.C.) treats only those patients with deeper invasion (>1 mm) and/or those with combined choroidal and/or scleral invasion. The remaining centers (C.L.S. and F.L.M.) treat all patients with postlaminar invasion, massive choroidal (>3-mm) invasion, or any combination of optic nerve and choroidal invasion. The intensity of the adjuvant regimen is also different (eTable in the Supplement). Disease-free survival ranges from greater than 95% to 100% in all these subgroups using all these strategies.
Periocular delivery of chemotherapy increases the amount of drug in the eye by approximately 10-fold and can definitely affect intraocular retinoblastoma.50 Because of the toxic effects and inability to cure intraocular retinoblastoma, all of us rarely use it.51
Although intravitreous chemotherapy has been tried for more than 50 years (with several different agents), concern about iatrogenic spread has limited its use.52 Observations in Japan in the early 2000s suggested that this technique was safe and effective.23 One of us (F.L.M.) introduced the modern day use of melphalan for vitreous seeds.3,53 The use of small-volume doses of 20 to 30 μg, lowering of intraocular pressure before injection, sealing the site before removal of the needle with cryotherapy, and irrigating the surface of the eye has led to an explosion of its use, there now being more than 2000 injections worldwide. To date, no case of extraocular extension has been found by us. This record has enabled the success rate for eyes with vitreous seeds to approach 100%. All 4 of us use this technique in eyes with vitreous seeds, including unilateral, bilateral, treatment-naive, and recurrent disease (Figure, D).
In the last 10 years, standard treatments for intraocular retinoblastoma have undergone major changes. Present-day treatment of most unilateral cases and some bilateral cases involves combinations of OAC and intravitreous chemotherapy that have resulted in higher success rates and lower complication rates, eliminated the use of external beam radiotherapy, and reduced the use of systemic chemotherapy. Most metastatic disease is now curable. All 4 of us manage retinoblastoma similarly, with minor differences. The progress and remarkable achievement in retinoblastoma care has led to retinoblastoma ranking first as the most curable cancer in children in developed nations.
Submitted for Publication: May 24, 2015; final revision received July 16, 2015; accepted July 22, 2015.
Corresponding Author: David H. Abramson, MD, Ophthalmic Oncology Service, Memorial Sloan Kettering Cancer Center, 1275 York Ave, Room A330, New York, NY 10065 (firstname.lastname@example.org).
Published Online: September 17, 2015. doi:10.1001/jamaophthalmol.2015.3108.
Author Contributions: Dr Abramson had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Abramson, Shields.
Acquisition, analysis, or interpretation of data: Shields, Munier, Chantada.
Drafting of the manuscript: Abramson, Shields, Chantada.
Critical revision of the manuscript for important intellectual content: Shields, Munier.
Statistical analysis: Shields.
Obtained funding: Abramson, Shields.
Administrative, technical, or material support: Abramson, Shields.
Study supervision: Abramson, Shields.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Additional Contributions: Armida Fabius, PhD, Ophthalmic Oncology Service, Memorial Sloan Kettering Cancer Center, provided editing assistance. No compensation was received by any person involved in this study.
Create a personal account or sign in to: