- Email: [email protected]
II Study of Direct Intraarterial (Ophthalmic Artery) Chemotherapy with Melphalan for Intraocular Retinoblastoma
A Phase I/II Study of Direct Intraarterial (Ophthalmic Artery) Chemotherapy with Melphalan for Intraocular Retinoblastoma Initial Results David H. Abramson, MD,1,2 Ira J. Dunkel, MD,1 Scott E. Brodie, MD, PhD,1,2,3 Jonathan W. Kim, MD,1 Y. Pierre Gobin, MD2 Objective: To develop a technique that would allow us to cannulate repeatedly the ophthalmic artery of young children with advanced retinoblastoma, to find a dose of melphalan that would be tolerable and tumoricidal for retinoblastoma when given intraarterially, and to study the local ocular and systemic side effects of intraarterial melphalan in these children. Design: Phase I/II clinical trial. Participants: Ten children with advanced retinoblastoma (Reese–Ellsworth V) eyes who were indicated for enucleation were entered into an institutional review board–approved protocol of ophthalmic artery infusion of melphalan to avoid enucleation. Methods: Cannulation of the ophthalmic artery was performed by a femoral artery approach using microcatheters while the children were under anesthesia and anticoagulated. Chemotherapy (melphalan) was infused into the artery over a 30-minute period. Main Outcome Measures: Ophthalmic examinations, retinal photography, and electroretinograms were used to document local toxicity, whereas physical examinations and complete blood counts were used to measure systemic toxicity. Results: The ophthalmic arteries were successfully cannulated in 9 cases (total, 27 times), as many as 6 times in 1 patient. Dramatic regression of tumors, vitreous seeds, and subretinal seeds were seen in each case. No severe systemic side effects (sepsis, anemia, neutropenia, fever, or death) occurred. No transfusions were required (red cells or platelets). Three patients developed conjunctival and lid edema that resolved without treatment. There was no toxicity to the cornea, anterior segment, pupil, or motility. One (previously irradiated) eye developed retinal ischemia; another eye had no toxicity after intraarterial chemotherapy but did develop a radiationlike retinopathy after brachytherapy. Vision stabilized or improved in all but 1 patient after treatment. Electroretinograms were generally poor (advanced eyes were treated), but in 2 cases, the electroretinogram improved after treatment (and resolution of a retinal detachment). Seven eyes avoided enucleation. Two intraarterially treated eyes were enucleated, with no viable tumors identified pathologically. Conclusions: We developed a technique of direct ophthalmic artery infusion of melphalan for children with retinoblastoma. The technique had minimal systemic side effects (one patient had grade 3 neutropenia) and minimal local toxicity. Among the first 9 cases treated with this technique, 7 eyes destined to be enucleated were salvaged. Ophthalmology 2008;115:1398 –1404 © 2008 by the American Academy of Ophthalmology.
This is the first report describing a successful technique of selective ophthalmic artery infusion of chemotherapy for children with retinoblastoma. For many years, there have Originally received: September 27, 2007. Final revision: November 14, 2007. Accepted: December 13, 2007. Available online: March 14, 2008. Manuscript no. 2007-1272. 1 Memorial Sloan–Kettering Cancer Center, New York, New York. 2 New York–Presbyterian Hospital, Weill–Cornell Medical Center, New York, New York.
1398
© 2008 by the American Academy of Ophthalmology Published by Elsevier Inc.
been suggestions that arterial infusion of chemotherapy may be applicable to intraocular retinoblastoma. In 1958, Reese described his experience with direct carotid injection of 3
Mount Sinai School of Medicine, New York, New York. Supported in part by a grant from the Fund for Ophthalmic Knowledge, Inc., New York, New York. This article contains digitally enhanced or processed images. Correspondence to David H. Abramson, MD, Memorial Sloan–Kettering Cancer Center, 1275 York Avenue, Box 343, New York, NY 10021. ISSN 0161-6420/08/$–see front matter doi:10.1016/j.ophtha.2007.12.014
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma chemotherapy (triethylene melamine [TEM]) for retinoblastoma, writing that “it is amazing to see the clinical regression of a lesion following the employment of x-ray together with one injection of TEM by way of the carotid artery.”1 Kiribuchi in 1966 demonstrated that favorable ocular responses could be obtained when 5-fluorouracil was injected into the frontal or supraorbital artery of children with retinoblastoma.2 During the past 15 years, Japanese investigators have published their experience with an interventional radiology technique of infusing melphalan into the carotid artery for retinoblastoma.3– 8 Their technique involved the use of a femoral artery puncture and a (balloon) catheter that was passed into the internal carotid artery and inflated to occlude the internal carotid artery (for as long as 30 minutes) beyond the orifice of the ophthalmic artery, allowing chemotherapy infused into the cervical internal carotid artery to perfuse the eye selectively without perfusing the brain. Utilizing this technique, Suzuki and Kaneko have reported 563 infusions performed in 187 patients, with impressive results.7 Despite the low rate of complications in their series, they pointed out that the infusions were not truly selective, because intracranial vascular territories also received high concentrations of chemotherapy through the cavernous branches of the internal carotid. In addition, all of the eyes received concomitant treatment with hyperthermia, external beam radiation, and/or intraocular (intravitreal) injection of melphalan, making it difficult to ascertain the contribution of the carotid artery infusion in achieving the excellent clinical outcomes in their series. In May 2006, we began an institutional review board– approved clinical protocol for intraocular retinoblastoma that attempted to answer 3 questions: 1. Could we develop a technique that allowed us to cannulate repeatedly the ophthalmic artery of young children with retinoblastoma? 2. Could we selectively infuse a chemotherapeutic agent (e.g., melphalan, carboplatin) through the ophthalmic artery that would be effective against advanced intraocular retinoblastoma? 3. Could we identify a drug and infusion technique with minimal ocular and systemic toxicity? This article is a summary of our experience with our first 10 patients utilizing this novel technique of selective ophthalmic artery infusion of chemotherapy for advanced intraocular retinoblastoma.
Materials and Methods Under general anesthesia, the femoral artery (alternatively right or left) was punctured and a 4-French (F) arterial sheath was placed. Anticoagulation was obtained with intravenous heparin (75 IU/kg). The 4-F (1.3-mm diameter) catheter was guided into the ipsilateral internal carotid artery. An arteriogram was performed to visualize the eye and cerebral vasculature and to select the best incidence showing the takeoff of the ophthalmic artery from the internal carotid. Using fluoroscopy and roadmapping, we selectively catheterized the ophthalmic artery using a microcatheter. We used 2 types of microcatheters: flow-directed ones such as the Magic
(Balt Therapeutics, Montgomery, France), whose outer diameter at the distal tip is 1.2 or 1.5 F (respectively, 0.4 mm or 0.5 mm), or guidewire-directed ones such as the Excelsior SL 10 (Boston Scientific, Fremont, CA), whose outer diameter at the distal tip is 1.7 F (0.57 mm). Once the microcatheter was in stable position at the ostium of the ophthalmic artery, a selective angiogram was performed to verify that the artery vascularized the entire vascular crescent. The dose of chemotherapy in successive patients was escalated according to the principles developed by Gobin et al9: (1) the dose was chosen (Table 1 [available at http://aaojournal.org]) according to the territory infused—that is, the size of the eye—and not body surface area or body weight, and (2) the injection was pulsatile to avoid streaming and inhomogeneous drug delivery. The drug is prepared as 5 mg/ml and then dilated in 30 cm3 of saline. In 2 cases, carboplatin was also added (these were advanced, remaining eyes that had failed all prior therapies). After the 30-minute chemotherapy infusion, the catheters were withdrawn, the femoral sheath was pulled, and hemostasis of the femoral artery was obtained by manual compression. The child was awakened from anesthesia, observed for 6 hours, and discharged the same day. In traditional phase I drug trials, doses are escalated to evaluate dose-limiting toxicity. The actual dose chosen here (as pointed out above) depended on the size of the eye, but at a dose of 7.5 mg, we encountered periocular edema and measurable neutropenia. Patients with Reese–Ellsworth V eyes were eligible for the protocol if the only alternative for therapy was enucleation. Exclusion criteria included documented congenital brain anomalies. No patient had both eyes treated. Ophthalmic evaluations were performed every 3 weeks and included external examination, visual acuity testing, pupil and motility evaluation by an independent physician, and a complete fundus examination under anesthesia including RetCam digital photography (Massie Industries, Dublin, CA), B-scan ultrasound with the OTI unit (Ophthalmic Technologies Inc., Toronto, Ontario, Canada) at 12 megahertz and standard electroretinogram testing under both photopic and scotopic conditions. Details of the electroretinogram examinations will be reported separately. Systemic evaluation included an interval history, weight and height measurements, and complete blood counts performed between visits.
Results This report summarizes our experience in the first 10 patients with Reese–Ellsworth V eyes (Vb, 9; Va, 1) who were enrolled on the protocol (Table 1). An attempt at catheterization was unsuccessful in 1 patient because of a vascular anomaly (ophthalmic artery arising from the middle meningeal artery). Every other attempt in every child was successful in cannulating the ophthalmic artery. Overall, 27 separate infusions were performed in 9 patients. The mean follow-up is 8.8 months, and the median, 7.5 months. Representative cases before and after treatment are present in Figures 1 to 6.
Complications Local. There were no complications to the lashes, cornea, anterior chamber, lens, vitreous, orbit, or motility. Three patients had edema to the lids for 3 to 7 days. Three patients had 2⫹ hyperemia of the conjunctiva for 3 to 7 days. One eye developed radiationlike retinopathy after subsequent brachytherapy (performed in the same
1399
Ophthalmology Volume 115, Number 8, August 2008
Figure 1. Two representative Reese–Ellsworth Vb eyes that were entered into the protocol (A, B).
Figure 2. Before (A) and 3 weeks after (B) a single infusion of melphalan (3 mg).
Figure 3. Before (A) and 3 weeks after (B) one dose of melphalan (3 mg).
1400
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma
Figure 4. Before (A) and after (B) two 5-mg doses of melphalan.
year). One patient (previously irradiated) developed retinal ischemia. Case 3 was enucleated after unsuccessful cannulation of the ophthalmic artery because of a vascular anomaly. Cases 6 and 9 were enucleated because of concern regarding persistent retinal detachment (RD), no vision, flat electroretinograms, and suspected continued tumor activity in eyes that had failed all prior therapies. Before intraarterial treatment, each eye had a flat electroretinogram and total RD. In both eyes, pathological analysis demonstrated no viable tumor. Systemic Toxicity. There were no deaths, stroke, fevers, or red blood cell or platelet transfusions. No ports were needed. One patient developed neutropenia (white blood cell count, 2300; absolute neutrophil count, 600).
Discussion This report summarizes our experience with a novel approach for advanced intraocular retinoblastoma utilizing a technique of selective ophthalmic artery infusion with locally high-dose chemotherapy. The standard management
of an eye with advanced intraocular retinoblastoma (Reese– Ellsworth Va and Vb) is enucleation,10 which is effective in preventing progression to clinical metastatic disease in ⬎95% of cases. During the past 30 years, unilateral retinoblastoma has also been managed in selective cases without enucleation, utilizing a combination of radiotherapy, chemotherapy, laser treatment, and cryotherapy.11 There are several valid reasons for pursuing a strategy of globe preservation even in eyes with advanced disease and poor visual potential. Enucleations in young children may compromise orbital bone growth, despite a seemingly adequate replacement of the socket volume with an alloplastic implant and properly fitted prosthesis. Despite the advent of integrated implant systems, ocular motility and overall cosmetic results are typically better with a blind nonphthisical eye when compared with a prosthesis. Psychologically, the loss of an eye can be a lifelong disability for parent and child alike. In some cultures, parents may choose to leave an eye with retinoblastoma in place even if it means the child will develop metastatic disease. In such settings, developing an effective method to save an eye with
Figure 5. Before (A) and after (B) a single 5-mg dose of melphalan.
1401
Ophthalmology Volume 115, Number 8, August 2008
Figure 6. Before (A) and after (B) 3 doses of melphalan (5, 3, and 3 mg).
advanced retinoblastoma (even if blind) may be equivalent to saving a life. Currently, ocular oncologists managing retinoblastoma patients pursue a multimodal treatment approach in an attempt to salvage eyes with advanced disease.12 External beam irradiation has been shown to salvage approximately 50% of eyes with Reese–Ellsworth group Vb disease.13 Well-recognized radiation side effects include cataract, retarded orbital bone growth, and gene-associated second cancers in patients with germinal disease,14 particularly if external beam radiation is given in the first year of life.15 In an effort to minimize the complications of external beam radiation, systemic chemotherapy has emerged as a strategy to treat intraocular retinoblastoma in combination with focal modalities (e.g., laser treatment, cryotherapy). Although success rates with chemoreduction vary between centers, as many as 50% of advanced eyes have been salvaged with 6 to 9 months of treatment with multiagent chemotherapy regimens that may include carboplatin, vincristine, etoposide, and cyclosporine.16 Short-term side effects from the systemic administration of multiple chemotherapeutic agents to young children appear to be manageable, although bone marrow suppression leading to neutropenic fevers, anemia requiring transfusions, severe thrombocytopenia, and sepsis and port infections have all been encountered.17 Recently, the development of secondary acute myelogenous leukemia after the use of systemic chemotherapy for intraocular disease has been reported in 16 children worldwide.18 In addition, there are practical considerations for the families of these children, as systemic chemotherapy is administered for ⱖ9 months in either the outpatient or the inpatient hospital setting. There are few models in clinical oncology for giving aggressive multiagent systemic chemotherapy to a young child without metastatic or presumed micrometastatic disease, particularly when the organ in question (i.e., the globe) represents ⬍1% of the weight and volume of the body. Concern about the short-, intermediate-, and long-term side effects of systemic chemotherapy in young children with retinoblastoma has stimulated the development of
1402
novel approaches for delivering chemotherapy to the globe. Animal experiments have suggested that higher doses of chemotherapy (carboplatin) could be delivered to the eye with negligible systemic absorption and no measurable systemic side effects via periocular delivery.19 Dramatic effects of periocular treatment on vitreous tumor seeds were then demonstrated in humans, and it was discovered that while vitreous levels were 10 times higher than after intravenous administration, systemic blood and tissue levels with this approach were negligible.19,20 Although many eyes responded favorably to this approach, complete regression of vitreous seeding after periocular injection was rarely achieved. Despite the lack of systemic toxicity, local side effects were often induced, including a pseudoorbital cellulitis syndrome in 50% of cases, optic atrophy, and local periorbital scarring.21 In Japan, clinicians face a cultural opposition to the surgical removal of eyes, often dealing with families that prefer to have their children die with an untreated retinoblastoma rather than allow a surgeon to perform a lifesaving enucleation. Motivated by these cultural obstacles, investigators in Japan developed a technique to deliver chemotherapy to the orbit by temporarily occluding the carotid artery distal to the ophthalmic artery and injecting intraarterial melphalan in the cervical internal carotid.22 In their initial reports, they pointed out that other intracranial vessels also received a high concentration of chemotherapy; although anesthesia-related adverse events were common (particularly bradycardia), no deaths or strokes were reported with this technique. Melphalan was chosen as the agent for this technique because it appeared to be the most effective drug when tested in their novel assay system.23 Melphalan is a powerful alkylating agent, but its applications in pediatric oncology have been limited with systemic administration because of its severe bone marrow toxicity. Interestingly, even though Japanese investigators have reported more than 500 infusions performed in 160 children, there are no data published on visual results, local complications, electroretinogram testing, pupil testing, or ocular side effects.
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma This report summarizes our experience with a truly selective ophthalmic artery infusion of chemotherapy. Our group was able to develop a technique that could be reliably, quickly, and efficiently performed in this group of patients. We encountered no adverse anesthesia events, no vascular complications related to the catherization process, no prolonged bleeding from the femoral puncture site, no prolonged episodes of bradycardia, and no stroke events. No child required overnight hospitalization for complications, but 1 child was hospitalized overnight for social reasons. Children were observed for 6 hours after the procedure to ensure that there was no bleeding from the femoral artery puncture site. Only 1 child among our first 10 cases could not be treated due to an anomalous origin of the ophthalmic artery. Anatomic variations in the origin of the ophthalmic artery are thought to occur in approximately 5% of humans, with the most common anomaly being the origin of the ophthalmic artery from the middle meningeal artery. The Japanese experience suggested that 2.5% of eyes could not be treated by arterial infusion due to technical difficulties related to the procedure. In this group of 9 treated patients, we were able to utilize a drug (melphalan or carboplatin) in concentrations that produced a clinical response in all cases. Three weeks after the treatment session, a clinical response of retinal tumors, subretinal tumors, and/or vitreous seeding was documented in all of the treated eyes. In many cases, the treatment response was dramatic; the senior author (DHA) has never observed such an impressive response of vitreous seeding from any other single treatment modality for retinoblastoma. The locally administered dose was approximately one tenth of the usual systemic dose of the chemotherapy agent, although individual dosages were scaled to the eye size and escalated in select cases depending on the clinical situation. It is not known how many separate treatments are necessary to cure an eye with advanced retinoblastoma, but among the 9 treated cases, the number of treatment sessions ranged from 2 to 6 to the same eye. In some of these cases, the patients were also treated with supplementary transpupillary thermotherapy laser, brachytherapy, or external beam irradiation. To date, 2 eyes among the 9 treated cases have been enucleated for suspected tumor recurrence. In both cases, intraarterial chemotherapy was used as salvage therapy for vitreous seeding after multiagent chemotherapy, laser, cryotherapy, and external beam radiation. Because of persistent RD in blind eyes (and a nonrecordable electroretinogram), both eyes were enucleated. On histopathological examination of both enucleated globes, no viable retinoblastoma was found. Despite the high concentrations of drug delivery to the ocular structures, the incidence of local and systemic complications after intraarterial infusion of chemotherapy was low in this group of patients. No child was hospitalized for medical complications during the posttreatment period, and there were no cases of fever, nausea and vomiting, alopecia, sepsis, or bone marrow suppression requiring filgrastim or transfusions with either red cells or platelets. One child (youngest in the group) developed self-limited grade 3 neutropenia during the first week after the initial chemotherapy infusion. Three children developed conjunctival
injection on the treated side for several days after the treatment session. Ocular motility appeared to be unaffected, even in the case that received 6 treatments. In 8 cases, vision was unchanged, whereas in 2 cases an improvement was noted. All of the patients had electroretinograms done during the course of treatment. Almost all of the cases demonstrated a flat electroretinogram before and after treatment; however, 1 case demonstrated definite improvement in electroretinogram potentials, and a second child went from a flat electroretinogram when the retina was completely detached to recordable as the subretinal fluid resolved. In 1 case, the electroretinogram worsened (in an eye previously irradiated). We believe that this report opens the way for an entirely new approach to the treatment of advanced intraocular retinoblastoma. Although we have not yet performed intraarterial chemotherapy in eyes less involved (Reese– Ellsworth groups I–III), this technique has the potential of replacing existing systemic chemotherapy because of the improved ocular effect and markedly diminished systemic toxicity. Similarly, we have not treated two eyes of the same patient, but the technique may allow us to salvage one or both eyes in bilateral advanced Vb eyes in which bilateral enucleations are sometimes necessary. The strategy of direct infusion into the ophthalmic artery allows the delivery of high concentrations of chemotherapy to the eye (and to the cancer), with far lower concentrations to the patient than systemic administration. The promising clinical responses observed in this initial report suggest that intraarterially delivered melphalan may allow many globes that are currently being enucleated to be salvaged with minimal local and systemic toxicity. Only longer follow-up will allow us to know if this technique will replace enucleation and/or standard systemic chemotherapy for intraocular retinoblastoma.
References 1. Reese AB, Hyman G, Tapley N, Forrest AW. The treatment of retinoblastoma by X-ray and triethylene melamine. AMA Arch Ophthalmol 1958;60:897–906. 2. Kiribuchi M. Retrograde infusion of anti-cancer drugs to ophthalmic artery for intraocular malignant tumors [in Japanese]. Nippon Ganka Gakkai Zasshi 1966;70:1829 –33. 3. Kaneko A, Takayama J, Matsuoka H, et al. Chemothermotherapy was successful in two cases of recurrence of intraocular retinoblastoma after irradiation [in Japanese]. Rinsho Ganka 1990;44:289 –92. 4. Mohri M. The technique of selective ophthalmic arterial infusion for conservative treatment of recurrent retinoblastoma [in Japanese]. Keio Igaku (J Keio Med Soc) 1993;70:679 – 87. 5. Ueda M, Tanabe J, Inomata K, et al. Study on conservative treatment of retinoblastoma– effect of intravitreal injection of melphalan on the rabbit retina [in Japanese]. Nippon Ganka Gakkai Zasshi 1995;99:1230 –5. 6. Kaneko A, Suzuki S. Eye-preservation treatment of retinoblastoma with vitreous seeding. Jpn J Clin Oncol 2003;33:601–7. 7. Suzuki S, Kaneko A. Management of intraocular retinoblastoma and ocular prognosis. Int J Clin Oncol 2004;9:1– 6. 8. Kaneko A. Japanese contributions to ocular oncology. Int J Clin Oncol 1999;4:321– 6.
1403
Ophthalmology Volume 115, Number 8, August 2008 9. Gobin YP, Cloughesy TF, Chow KL, et al. Intraarterial chemotherapy for brain tumors by using a spatial dose fractionation algorithm and pulsatile delivery. Radiology 2001;218: 724 –32. 10. Abramson DH, Schefler AC. Update on retinoblastoma. Retina 2004;24:828 – 48. 11. Abramson DH, Marks R, Ellsworth RM, et al. The management of unilateral retinoblastoma without primary enucleation. Arch Ophthalmol 1982;100:1249 –52. 12. Shields CL, Mashayekhi A, Cater J, et al. Chemoreduction for retinoblastoma: analysis of tumor control and risks for recurrence in 457 tumors. Trans Am Ophthalmol Soc 2004; 102:34 – 44. 13. Abramson DH, Beaverson KL, Chang ST, et al. Outcome following initial external beam radiotherapy in patients with Reese-Ellsworth group Vb retinoblastoma. Arch Ophthalmol 2004;122:1316 –23. 14. Abramson DH. Retinoblastoma in the 20th century: past success and future challenges. The Weisenfeld Lecture. Invest Ophthalmol Vis Sci 2005;46:2683–91. 15. Abramson DH, Frank CM. Second nonocular tumors in survivors of bilateral retinoblastoma: a possible age effect on radiation-related risk. Ophthalmology 1998;105:573–9. 16. Abramson DH, Schefler AC, Dunkel IJ. Neoplasms of the eye. In: Kufe DW, Bast RC Jr, Hait WN, et al, eds. Holland-Frei
1404
17.
18.
19.
20.
21.
22.
23.
Cancer Medicine 7. 7th ed. Hamilton, Canada: BC Decker; 2006:1071– 83. Benz MS, Scott IU, Murray TG, et al. Complications of systemic chemotherapy as treatment of retinoblastoma. Arch Ophthalmol 2000;118:577– 8. Gombos DS, Hungerford J, Abramson DH, et al. Secondary acute myelogenous leukemia in patients with retinoblastoma. Is chemotherapy a factor? Ophthalmology 2007;114:1378 – 83. Mendelsohn M, Abramson DH, Madden T, et al. Intraocular concentrations of chemotherapeutic agents after systemic or local administration. Arch Ophthalmol 1998;116:1209 –12. Abramson DH, Frank CM, Dunkel IJ. A phase I/II study of subconjunctival carboplatin for intraocular retinoblastoma. Ophthalmology 1999;106:1947–50. Schmack I, Hubbard GB, Kang SJ, et al. Ischemic necrosis and atrophy of the optic nerve after periocular carboplatin injection for intraocular retinoblastoma. Am J Ophthalmol 2006;142:310 –5. Yamane T, Kaneko A, Mohri M. The technique of ophthalmic artery infusion therapy for patients with intraocular retinoblastoma. Int J Clin Oncol 2004;9:69 –73. Inomata M, Kaneko A. Chemosensitivity profiles of primary and cultured human retinoblastoma cells in a human tumor clonogenic assay. Jpn J Cancer Res 1987;78:858 – 68.
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma Table 1. Data on First 10 Retinoblastomas Treated with Chemosurgery Age at First No. of IA Follow-up IA Prior Patient Laterality (mos) (mos) Sessions Treatment 1 2
B U
22 25
14 11
3 6
Yes No
3
U
27
11
0
No
4 5 6 7 8 9
U U B U B B
14 6 22 51 34 30
10 9 6 6 4 4
2 2 3 2 3 3
No No Yes Yes Yes No
10
U
10
3
3
No
Dose at Each Session (mg) (Melphalan, Carboplatin) 3, 5, 5 5, 5, 5 15, 30, 50 Failed catheterization 5, 5 3, 3 7.5, 7.5, 7.5 7.5, 7.5 7.5, 7.5, 7.5 7.5, 7.5, 7.5 30, 30, 30 5, 3, 3
Vision Before Vision After APD APD (Fix, Follow) (Fix, Follow) Before After ERG Enucleation No F, F No F, F
Same Same
F, no F
Yes No
Same Same
Yes
NA 1
No No
↔
Yes
No F, F ? F, F No F, F No F, F No F, F No F, F
Same F, F Same Same No F, F No F, F
No No No No No Yes
Same Same Same Same Same Same
↔ 1 ↔ ↔ 2 ↔
No No Yes No No Yes
No F, F
F, F
No
Same
↔
No
APD ⫽ afferent pupillary defect; B ⫽ bilateral retinoblastoma; ERG ⫽ electroretinogram (1, improved;↔, same;2, worse); IA ⫽ intraarterial
treatment; U ⫽ unilateral retinoblastoma.
1404.e1
This is the first report describing a successful technique of selective ophthalmic artery infusion of chemotherapy for children with retinoblastoma. For many years, there have Originally received: September 27, 2007. Final revision: November 14, 2007. Accepted: December 13, 2007. Available online: March 14, 2008. Manuscript no. 2007-1272. 1 Memorial Sloan–Kettering Cancer Center, New York, New York. 2 New York–Presbyterian Hospital, Weill–Cornell Medical Center, New York, New York.
1398
© 2008 by the American Academy of Ophthalmology Published by Elsevier Inc.
been suggestions that arterial infusion of chemotherapy may be applicable to intraocular retinoblastoma. In 1958, Reese described his experience with direct carotid injection of 3
Mount Sinai School of Medicine, New York, New York. Supported in part by a grant from the Fund for Ophthalmic Knowledge, Inc., New York, New York. This article contains digitally enhanced or processed images. Correspondence to David H. Abramson, MD, Memorial Sloan–Kettering Cancer Center, 1275 York Avenue, Box 343, New York, NY 10021. ISSN 0161-6420/08/$–see front matter doi:10.1016/j.ophtha.2007.12.014
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma chemotherapy (triethylene melamine [TEM]) for retinoblastoma, writing that “it is amazing to see the clinical regression of a lesion following the employment of x-ray together with one injection of TEM by way of the carotid artery.”1 Kiribuchi in 1966 demonstrated that favorable ocular responses could be obtained when 5-fluorouracil was injected into the frontal or supraorbital artery of children with retinoblastoma.2 During the past 15 years, Japanese investigators have published their experience with an interventional radiology technique of infusing melphalan into the carotid artery for retinoblastoma.3– 8 Their technique involved the use of a femoral artery puncture and a (balloon) catheter that was passed into the internal carotid artery and inflated to occlude the internal carotid artery (for as long as 30 minutes) beyond the orifice of the ophthalmic artery, allowing chemotherapy infused into the cervical internal carotid artery to perfuse the eye selectively without perfusing the brain. Utilizing this technique, Suzuki and Kaneko have reported 563 infusions performed in 187 patients, with impressive results.7 Despite the low rate of complications in their series, they pointed out that the infusions were not truly selective, because intracranial vascular territories also received high concentrations of chemotherapy through the cavernous branches of the internal carotid. In addition, all of the eyes received concomitant treatment with hyperthermia, external beam radiation, and/or intraocular (intravitreal) injection of melphalan, making it difficult to ascertain the contribution of the carotid artery infusion in achieving the excellent clinical outcomes in their series. In May 2006, we began an institutional review board– approved clinical protocol for intraocular retinoblastoma that attempted to answer 3 questions: 1. Could we develop a technique that allowed us to cannulate repeatedly the ophthalmic artery of young children with retinoblastoma? 2. Could we selectively infuse a chemotherapeutic agent (e.g., melphalan, carboplatin) through the ophthalmic artery that would be effective against advanced intraocular retinoblastoma? 3. Could we identify a drug and infusion technique with minimal ocular and systemic toxicity? This article is a summary of our experience with our first 10 patients utilizing this novel technique of selective ophthalmic artery infusion of chemotherapy for advanced intraocular retinoblastoma.
Materials and Methods Under general anesthesia, the femoral artery (alternatively right or left) was punctured and a 4-French (F) arterial sheath was placed. Anticoagulation was obtained with intravenous heparin (75 IU/kg). The 4-F (1.3-mm diameter) catheter was guided into the ipsilateral internal carotid artery. An arteriogram was performed to visualize the eye and cerebral vasculature and to select the best incidence showing the takeoff of the ophthalmic artery from the internal carotid. Using fluoroscopy and roadmapping, we selectively catheterized the ophthalmic artery using a microcatheter. We used 2 types of microcatheters: flow-directed ones such as the Magic
(Balt Therapeutics, Montgomery, France), whose outer diameter at the distal tip is 1.2 or 1.5 F (respectively, 0.4 mm or 0.5 mm), or guidewire-directed ones such as the Excelsior SL 10 (Boston Scientific, Fremont, CA), whose outer diameter at the distal tip is 1.7 F (0.57 mm). Once the microcatheter was in stable position at the ostium of the ophthalmic artery, a selective angiogram was performed to verify that the artery vascularized the entire vascular crescent. The dose of chemotherapy in successive patients was escalated according to the principles developed by Gobin et al9: (1) the dose was chosen (Table 1 [available at http://aaojournal.org]) according to the territory infused—that is, the size of the eye—and not body surface area or body weight, and (2) the injection was pulsatile to avoid streaming and inhomogeneous drug delivery. The drug is prepared as 5 mg/ml and then dilated in 30 cm3 of saline. In 2 cases, carboplatin was also added (these were advanced, remaining eyes that had failed all prior therapies). After the 30-minute chemotherapy infusion, the catheters were withdrawn, the femoral sheath was pulled, and hemostasis of the femoral artery was obtained by manual compression. The child was awakened from anesthesia, observed for 6 hours, and discharged the same day. In traditional phase I drug trials, doses are escalated to evaluate dose-limiting toxicity. The actual dose chosen here (as pointed out above) depended on the size of the eye, but at a dose of 7.5 mg, we encountered periocular edema and measurable neutropenia. Patients with Reese–Ellsworth V eyes were eligible for the protocol if the only alternative for therapy was enucleation. Exclusion criteria included documented congenital brain anomalies. No patient had both eyes treated. Ophthalmic evaluations were performed every 3 weeks and included external examination, visual acuity testing, pupil and motility evaluation by an independent physician, and a complete fundus examination under anesthesia including RetCam digital photography (Massie Industries, Dublin, CA), B-scan ultrasound with the OTI unit (Ophthalmic Technologies Inc., Toronto, Ontario, Canada) at 12 megahertz and standard electroretinogram testing under both photopic and scotopic conditions. Details of the electroretinogram examinations will be reported separately. Systemic evaluation included an interval history, weight and height measurements, and complete blood counts performed between visits.
Results This report summarizes our experience in the first 10 patients with Reese–Ellsworth V eyes (Vb, 9; Va, 1) who were enrolled on the protocol (Table 1). An attempt at catheterization was unsuccessful in 1 patient because of a vascular anomaly (ophthalmic artery arising from the middle meningeal artery). Every other attempt in every child was successful in cannulating the ophthalmic artery. Overall, 27 separate infusions were performed in 9 patients. The mean follow-up is 8.8 months, and the median, 7.5 months. Representative cases before and after treatment are present in Figures 1 to 6.
Complications Local. There were no complications to the lashes, cornea, anterior chamber, lens, vitreous, orbit, or motility. Three patients had edema to the lids for 3 to 7 days. Three patients had 2⫹ hyperemia of the conjunctiva for 3 to 7 days. One eye developed radiationlike retinopathy after subsequent brachytherapy (performed in the same
1399
Ophthalmology Volume 115, Number 8, August 2008
Figure 1. Two representative Reese–Ellsworth Vb eyes that were entered into the protocol (A, B).
Figure 2. Before (A) and 3 weeks after (B) a single infusion of melphalan (3 mg).
Figure 3. Before (A) and 3 weeks after (B) one dose of melphalan (3 mg).
1400
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma
Figure 4. Before (A) and after (B) two 5-mg doses of melphalan.
year). One patient (previously irradiated) developed retinal ischemia. Case 3 was enucleated after unsuccessful cannulation of the ophthalmic artery because of a vascular anomaly. Cases 6 and 9 were enucleated because of concern regarding persistent retinal detachment (RD), no vision, flat electroretinograms, and suspected continued tumor activity in eyes that had failed all prior therapies. Before intraarterial treatment, each eye had a flat electroretinogram and total RD. In both eyes, pathological analysis demonstrated no viable tumor. Systemic Toxicity. There were no deaths, stroke, fevers, or red blood cell or platelet transfusions. No ports were needed. One patient developed neutropenia (white blood cell count, 2300; absolute neutrophil count, 600).
Discussion This report summarizes our experience with a novel approach for advanced intraocular retinoblastoma utilizing a technique of selective ophthalmic artery infusion with locally high-dose chemotherapy. The standard management
of an eye with advanced intraocular retinoblastoma (Reese– Ellsworth Va and Vb) is enucleation,10 which is effective in preventing progression to clinical metastatic disease in ⬎95% of cases. During the past 30 years, unilateral retinoblastoma has also been managed in selective cases without enucleation, utilizing a combination of radiotherapy, chemotherapy, laser treatment, and cryotherapy.11 There are several valid reasons for pursuing a strategy of globe preservation even in eyes with advanced disease and poor visual potential. Enucleations in young children may compromise orbital bone growth, despite a seemingly adequate replacement of the socket volume with an alloplastic implant and properly fitted prosthesis. Despite the advent of integrated implant systems, ocular motility and overall cosmetic results are typically better with a blind nonphthisical eye when compared with a prosthesis. Psychologically, the loss of an eye can be a lifelong disability for parent and child alike. In some cultures, parents may choose to leave an eye with retinoblastoma in place even if it means the child will develop metastatic disease. In such settings, developing an effective method to save an eye with
Figure 5. Before (A) and after (B) a single 5-mg dose of melphalan.
1401
Ophthalmology Volume 115, Number 8, August 2008
Figure 6. Before (A) and after (B) 3 doses of melphalan (5, 3, and 3 mg).
advanced retinoblastoma (even if blind) may be equivalent to saving a life. Currently, ocular oncologists managing retinoblastoma patients pursue a multimodal treatment approach in an attempt to salvage eyes with advanced disease.12 External beam irradiation has been shown to salvage approximately 50% of eyes with Reese–Ellsworth group Vb disease.13 Well-recognized radiation side effects include cataract, retarded orbital bone growth, and gene-associated second cancers in patients with germinal disease,14 particularly if external beam radiation is given in the first year of life.15 In an effort to minimize the complications of external beam radiation, systemic chemotherapy has emerged as a strategy to treat intraocular retinoblastoma in combination with focal modalities (e.g., laser treatment, cryotherapy). Although success rates with chemoreduction vary between centers, as many as 50% of advanced eyes have been salvaged with 6 to 9 months of treatment with multiagent chemotherapy regimens that may include carboplatin, vincristine, etoposide, and cyclosporine.16 Short-term side effects from the systemic administration of multiple chemotherapeutic agents to young children appear to be manageable, although bone marrow suppression leading to neutropenic fevers, anemia requiring transfusions, severe thrombocytopenia, and sepsis and port infections have all been encountered.17 Recently, the development of secondary acute myelogenous leukemia after the use of systemic chemotherapy for intraocular disease has been reported in 16 children worldwide.18 In addition, there are practical considerations for the families of these children, as systemic chemotherapy is administered for ⱖ9 months in either the outpatient or the inpatient hospital setting. There are few models in clinical oncology for giving aggressive multiagent systemic chemotherapy to a young child without metastatic or presumed micrometastatic disease, particularly when the organ in question (i.e., the globe) represents ⬍1% of the weight and volume of the body. Concern about the short-, intermediate-, and long-term side effects of systemic chemotherapy in young children with retinoblastoma has stimulated the development of
1402
novel approaches for delivering chemotherapy to the globe. Animal experiments have suggested that higher doses of chemotherapy (carboplatin) could be delivered to the eye with negligible systemic absorption and no measurable systemic side effects via periocular delivery.19 Dramatic effects of periocular treatment on vitreous tumor seeds were then demonstrated in humans, and it was discovered that while vitreous levels were 10 times higher than after intravenous administration, systemic blood and tissue levels with this approach were negligible.19,20 Although many eyes responded favorably to this approach, complete regression of vitreous seeding after periocular injection was rarely achieved. Despite the lack of systemic toxicity, local side effects were often induced, including a pseudoorbital cellulitis syndrome in 50% of cases, optic atrophy, and local periorbital scarring.21 In Japan, clinicians face a cultural opposition to the surgical removal of eyes, often dealing with families that prefer to have their children die with an untreated retinoblastoma rather than allow a surgeon to perform a lifesaving enucleation. Motivated by these cultural obstacles, investigators in Japan developed a technique to deliver chemotherapy to the orbit by temporarily occluding the carotid artery distal to the ophthalmic artery and injecting intraarterial melphalan in the cervical internal carotid.22 In their initial reports, they pointed out that other intracranial vessels also received a high concentration of chemotherapy; although anesthesia-related adverse events were common (particularly bradycardia), no deaths or strokes were reported with this technique. Melphalan was chosen as the agent for this technique because it appeared to be the most effective drug when tested in their novel assay system.23 Melphalan is a powerful alkylating agent, but its applications in pediatric oncology have been limited with systemic administration because of its severe bone marrow toxicity. Interestingly, even though Japanese investigators have reported more than 500 infusions performed in 160 children, there are no data published on visual results, local complications, electroretinogram testing, pupil testing, or ocular side effects.
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma This report summarizes our experience with a truly selective ophthalmic artery infusion of chemotherapy. Our group was able to develop a technique that could be reliably, quickly, and efficiently performed in this group of patients. We encountered no adverse anesthesia events, no vascular complications related to the catherization process, no prolonged bleeding from the femoral puncture site, no prolonged episodes of bradycardia, and no stroke events. No child required overnight hospitalization for complications, but 1 child was hospitalized overnight for social reasons. Children were observed for 6 hours after the procedure to ensure that there was no bleeding from the femoral artery puncture site. Only 1 child among our first 10 cases could not be treated due to an anomalous origin of the ophthalmic artery. Anatomic variations in the origin of the ophthalmic artery are thought to occur in approximately 5% of humans, with the most common anomaly being the origin of the ophthalmic artery from the middle meningeal artery. The Japanese experience suggested that 2.5% of eyes could not be treated by arterial infusion due to technical difficulties related to the procedure. In this group of 9 treated patients, we were able to utilize a drug (melphalan or carboplatin) in concentrations that produced a clinical response in all cases. Three weeks after the treatment session, a clinical response of retinal tumors, subretinal tumors, and/or vitreous seeding was documented in all of the treated eyes. In many cases, the treatment response was dramatic; the senior author (DHA) has never observed such an impressive response of vitreous seeding from any other single treatment modality for retinoblastoma. The locally administered dose was approximately one tenth of the usual systemic dose of the chemotherapy agent, although individual dosages were scaled to the eye size and escalated in select cases depending on the clinical situation. It is not known how many separate treatments are necessary to cure an eye with advanced retinoblastoma, but among the 9 treated cases, the number of treatment sessions ranged from 2 to 6 to the same eye. In some of these cases, the patients were also treated with supplementary transpupillary thermotherapy laser, brachytherapy, or external beam irradiation. To date, 2 eyes among the 9 treated cases have been enucleated for suspected tumor recurrence. In both cases, intraarterial chemotherapy was used as salvage therapy for vitreous seeding after multiagent chemotherapy, laser, cryotherapy, and external beam radiation. Because of persistent RD in blind eyes (and a nonrecordable electroretinogram), both eyes were enucleated. On histopathological examination of both enucleated globes, no viable retinoblastoma was found. Despite the high concentrations of drug delivery to the ocular structures, the incidence of local and systemic complications after intraarterial infusion of chemotherapy was low in this group of patients. No child was hospitalized for medical complications during the posttreatment period, and there were no cases of fever, nausea and vomiting, alopecia, sepsis, or bone marrow suppression requiring filgrastim or transfusions with either red cells or platelets. One child (youngest in the group) developed self-limited grade 3 neutropenia during the first week after the initial chemotherapy infusion. Three children developed conjunctival
injection on the treated side for several days after the treatment session. Ocular motility appeared to be unaffected, even in the case that received 6 treatments. In 8 cases, vision was unchanged, whereas in 2 cases an improvement was noted. All of the patients had electroretinograms done during the course of treatment. Almost all of the cases demonstrated a flat electroretinogram before and after treatment; however, 1 case demonstrated definite improvement in electroretinogram potentials, and a second child went from a flat electroretinogram when the retina was completely detached to recordable as the subretinal fluid resolved. In 1 case, the electroretinogram worsened (in an eye previously irradiated). We believe that this report opens the way for an entirely new approach to the treatment of advanced intraocular retinoblastoma. Although we have not yet performed intraarterial chemotherapy in eyes less involved (Reese– Ellsworth groups I–III), this technique has the potential of replacing existing systemic chemotherapy because of the improved ocular effect and markedly diminished systemic toxicity. Similarly, we have not treated two eyes of the same patient, but the technique may allow us to salvage one or both eyes in bilateral advanced Vb eyes in which bilateral enucleations are sometimes necessary. The strategy of direct infusion into the ophthalmic artery allows the delivery of high concentrations of chemotherapy to the eye (and to the cancer), with far lower concentrations to the patient than systemic administration. The promising clinical responses observed in this initial report suggest that intraarterially delivered melphalan may allow many globes that are currently being enucleated to be salvaged with minimal local and systemic toxicity. Only longer follow-up will allow us to know if this technique will replace enucleation and/or standard systemic chemotherapy for intraocular retinoblastoma.
References 1. Reese AB, Hyman G, Tapley N, Forrest AW. The treatment of retinoblastoma by X-ray and triethylene melamine. AMA Arch Ophthalmol 1958;60:897–906. 2. Kiribuchi M. Retrograde infusion of anti-cancer drugs to ophthalmic artery for intraocular malignant tumors [in Japanese]. Nippon Ganka Gakkai Zasshi 1966;70:1829 –33. 3. Kaneko A, Takayama J, Matsuoka H, et al. Chemothermotherapy was successful in two cases of recurrence of intraocular retinoblastoma after irradiation [in Japanese]. Rinsho Ganka 1990;44:289 –92. 4. Mohri M. The technique of selective ophthalmic arterial infusion for conservative treatment of recurrent retinoblastoma [in Japanese]. Keio Igaku (J Keio Med Soc) 1993;70:679 – 87. 5. Ueda M, Tanabe J, Inomata K, et al. Study on conservative treatment of retinoblastoma– effect of intravitreal injection of melphalan on the rabbit retina [in Japanese]. Nippon Ganka Gakkai Zasshi 1995;99:1230 –5. 6. Kaneko A, Suzuki S. Eye-preservation treatment of retinoblastoma with vitreous seeding. Jpn J Clin Oncol 2003;33:601–7. 7. Suzuki S, Kaneko A. Management of intraocular retinoblastoma and ocular prognosis. Int J Clin Oncol 2004;9:1– 6. 8. Kaneko A. Japanese contributions to ocular oncology. Int J Clin Oncol 1999;4:321– 6.
1403
Ophthalmology Volume 115, Number 8, August 2008 9. Gobin YP, Cloughesy TF, Chow KL, et al. Intraarterial chemotherapy for brain tumors by using a spatial dose fractionation algorithm and pulsatile delivery. Radiology 2001;218: 724 –32. 10. Abramson DH, Schefler AC. Update on retinoblastoma. Retina 2004;24:828 – 48. 11. Abramson DH, Marks R, Ellsworth RM, et al. The management of unilateral retinoblastoma without primary enucleation. Arch Ophthalmol 1982;100:1249 –52. 12. Shields CL, Mashayekhi A, Cater J, et al. Chemoreduction for retinoblastoma: analysis of tumor control and risks for recurrence in 457 tumors. Trans Am Ophthalmol Soc 2004; 102:34 – 44. 13. Abramson DH, Beaverson KL, Chang ST, et al. Outcome following initial external beam radiotherapy in patients with Reese-Ellsworth group Vb retinoblastoma. Arch Ophthalmol 2004;122:1316 –23. 14. Abramson DH. Retinoblastoma in the 20th century: past success and future challenges. The Weisenfeld Lecture. Invest Ophthalmol Vis Sci 2005;46:2683–91. 15. Abramson DH, Frank CM. Second nonocular tumors in survivors of bilateral retinoblastoma: a possible age effect on radiation-related risk. Ophthalmology 1998;105:573–9. 16. Abramson DH, Schefler AC, Dunkel IJ. Neoplasms of the eye. In: Kufe DW, Bast RC Jr, Hait WN, et al, eds. Holland-Frei
1404
17.
18.
19.
20.
21.
22.
23.
Cancer Medicine 7. 7th ed. Hamilton, Canada: BC Decker; 2006:1071– 83. Benz MS, Scott IU, Murray TG, et al. Complications of systemic chemotherapy as treatment of retinoblastoma. Arch Ophthalmol 2000;118:577– 8. Gombos DS, Hungerford J, Abramson DH, et al. Secondary acute myelogenous leukemia in patients with retinoblastoma. Is chemotherapy a factor? Ophthalmology 2007;114:1378 – 83. Mendelsohn M, Abramson DH, Madden T, et al. Intraocular concentrations of chemotherapeutic agents after systemic or local administration. Arch Ophthalmol 1998;116:1209 –12. Abramson DH, Frank CM, Dunkel IJ. A phase I/II study of subconjunctival carboplatin for intraocular retinoblastoma. Ophthalmology 1999;106:1947–50. Schmack I, Hubbard GB, Kang SJ, et al. Ischemic necrosis and atrophy of the optic nerve after periocular carboplatin injection for intraocular retinoblastoma. Am J Ophthalmol 2006;142:310 –5. Yamane T, Kaneko A, Mohri M. The technique of ophthalmic artery infusion therapy for patients with intraocular retinoblastoma. Int J Clin Oncol 2004;9:69 –73. Inomata M, Kaneko A. Chemosensitivity profiles of primary and cultured human retinoblastoma cells in a human tumor clonogenic assay. Jpn J Cancer Res 1987;78:858 – 68.
Abramson et al 䡠 Intraocular Chemotherapy for Retinoblastoma Table 1. Data on First 10 Retinoblastomas Treated with Chemosurgery Age at First No. of IA Follow-up IA Prior Patient Laterality (mos) (mos) Sessions Treatment 1 2
B U
22 25
14 11
3 6
Yes No
3
U
27
11
0
No
4 5 6 7 8 9
U U B U B B
14 6 22 51 34 30
10 9 6 6 4 4
2 2 3 2 3 3
No No Yes Yes Yes No
10
U
10
3
3
No
Dose at Each Session (mg) (Melphalan, Carboplatin) 3, 5, 5 5, 5, 5 15, 30, 50 Failed catheterization 5, 5 3, 3 7.5, 7.5, 7.5 7.5, 7.5 7.5, 7.5, 7.5 7.5, 7.5, 7.5 30, 30, 30 5, 3, 3
Vision Before Vision After APD APD (Fix, Follow) (Fix, Follow) Before After ERG Enucleation No F, F No F, F
Same Same
F, no F
Yes No
Same Same
Yes
NA 1
No No
↔
Yes
No F, F ? F, F No F, F No F, F No F, F No F, F
Same F, F Same Same No F, F No F, F
No No No No No Yes
Same Same Same Same Same Same
↔ 1 ↔ ↔ 2 ↔
No No Yes No No Yes
No F, F
F, F
No
Same
↔
No
APD ⫽ afferent pupillary defect; B ⫽ bilateral retinoblastoma; ERG ⫽ electroretinogram (1, improved;↔, same;2, worse); IA ⫽ intraarterial
treatment; U ⫽ unilateral retinoblastoma.
1404.e1