Intra-arterial Chemotherapy for Retinoblastoma: Where Do We Stand?

With more than six years of data, the pros and cons of intra-arterial chemo are clearer than ever before.

Intra-arterial Chemotherapy for Retinoblastoma: Where Do We Stand?

With more than six years of data, the pros and cons of intra-arterial chemo are clearer than ever before.


It is now more than six years since the first publication on superselective ophthalmic artery chemotherapy for intraocular retinoblastoma.1 More than 1,500 treatments have been delivered in 32 countries worldwide, and more than 30 peer-reviewed articles have been published, in addition to stories in the New York Times, JAMA, and non-peer-reviewed magazines.

This article will review that experience, highlight what has been learned (both the good and bad), discuss the published experimental work, and outline the future possibilities and challenges of intra-arterial chemotherapy.


More than 50 years ago Algernon B. Reese, MD, treated intraocular retinoblastoma with arterial injections of chemotherapy by direct puncture of the internal carotid artery, and it was referred to as “intraarterial chemotherapy.”2

Beginning in 1986, Japanese physicians, led by Kaneko and Mori, began catheterizing the internal carotid artery from femoral access and, with a tiny balloon, occluding the internal carotid artery above the exit of the ophthalmic artery, when they then injected chemotherapy into the (temporarily) occluded internal carotid artery. They called this approach “selective ophthalmic artery chemotherapy.”3-8

David H. Abramson, MD, FACS, is Chief of the Ophthalmic Oncology Service at the Memorial Sloan Kettering Cancer Center in New York. He reports no financial interest in any products mentioned here. Dr. Abramson can be reached via e-mail at

Six years ago, we began routinely catheterizing the ophthalmic artery itself and infusing chemotherapy directly into it. We, as well as other researchers, have used different terms to clarify the difference between the Japanese approach and ours. Because the term “selective” was already in use, we called it “super-selective ophthalmic artery infusion of chemotherapy” or “super-selective chemotherapy,” but now we prefer to call it “ophthalmic artery chemosurgery.”

As we have learned, the procedure is more like surgery, in that it requires careful screening of appropriate candidates, assembling a team of anesthesiologists, nurses, technicians, pharmacologists, and pediatric oncologists, acquiring sophisticated equipment, modifying what was planned, and dealing with unexpected consequences. Thus, we now prefer the term “ophthalmic artery chemosurgery” (OAC) (Figure 1).


The basic technique used worldwide is nearly identical for everyone and is based on our original description. Some important details have changed, making the procedure safer and faster. Femoral access is still used in all cases (alternating right and left legs monthly), but we no longer use a guide wire, and we prefer to thread the treating catheter directly from the femoral artery.


Figure 1. Dramatic reduction and complete calcification of advanced retinoblastoma (left, before; right, after) treated with OAC alone.


Figure 2. Treatment of less advanced retinoblastoma with OAC (left, before; right after). Note the normal-appearing retina after treatment.

Successful access into the ophthalmic artery is best approached by allowing the catheter to pass above the exit of the ophthalmic artery and then “pulling back” the catheter under fluoroscopy. Once the tip comes back to the orifice of the ophthalmic artery, it, like normal blood flow, will go into the ophthalmic artery itself.

A common error has been to allow the tip of the catheter to go too far into the ophthalmic artery. This immediately changes blood flow, as the orbital vessels have multiple collaterals with the internal and external carotid artery, and occlusion of the ophthalmic artery by a catheter will result in changes in flow — including even flow reversal in some vessels.

All patients are anticoagulated with heparin, the catheter is kept in a cotton-free environment (as cotton fibers can be embolized into the choroidal vessels, causing loss of vision), and drugs are delivered (after dilution) in a pulsatile fashion. Studies are under way to determine the ideal dilution and duration of infusion.

There have been no deaths reported (from the procedure or from disease) and no strokes worldwide. Even when using the Japanese balloon technique, long-term reports have also seen no strokes or deaths from the procedure. No permanent femoral artery occlusion has been reported either.

The Japanese first noted, and we and others have confirmed, a new, previously undiscovered reflex during the procedure. When the ophthalmic artery is entered, some children developed a sudden decrease in tidal volume that is reversible with intravenous epinephrine. True contrast allergy has developed also.


Much has been learned about the vascular anatomy of the pediatric orbit, and our experience mirrors the excellent earlier publications from Hayreh on this subject.9-10 No two orbits have the same vascular anatomy, and recognition of this fact is critical, in real time, for the person performing the procedure. Not one of our orbits has had the anatomy depicted in the classic Gray’s Anatomy!11

While the ophthalmic artery usually comes off the internal carotid artery, it may also come off the middle meningeal artery, and in 5% of cases originate from the external carotid artery. In these cases, we have demonstrated that catheterizing the external carotid artery and passing the catheter into the middle meningeal artery from the internal maxillary artery is safe and effective.12

Because of connections through the foramen of Hyrtl, chemotherapy (or dye) injected this manner enters the lacrimal artery, and with injection, it will reverse the flow of the ophthalmic artery, allowing chemotherapy (or dye) to then enter the ophthalmic artery.

While all ophthalmologists recognize that more blood flows into the eye via the choroidal vasculature than the retinal vasculature (and that filling is usually slightly faster in the choroid), most are unaware that the largest flow of blood from the ophthalmic artery is not to the eye itself. The largest flow is usually into the supratrochlear artery, supplying the nasal lid structures and nasal forehead. This is why 15% of treated children develop hyperemia in this distribution.

Most importantly, this vessel branches into the nose so that filling of the nose is often seen simultaneously with choroidal filling, if not just before. Significant but variable blood flow also goes to the lacrimal gland, and many anatomic variants of the middle meningeal artery also steal blood from the eye itself.

The anatomy and blood flow patterns must be studied at the time of treatment, and sometimes treatment must be modified immediately.12 If significant stealing of blood happens because of the middle meningeal artery or lacrimal gland, the dose must be immediately increased. Because of the flow into the nose, we routinely use vasoconstricting nasal drops and, at times, vasoconstrictors on the forehead.


Advanced Eyes Scheduled for Enucleation

Our initial institutional review board allowed us to treat advanced eyes scheduled for enucleation. Enucleation has always been the most common initial management of intraocular retinoblastoma worldwide because the majority of children present with Reese-Ellsworth V (or Vb)/International Classification E disease. These are eyes with extensive disease filling more than half the eye, often with vitreous and subretinal seeding. It is widely felt that they have no potential for vision, and left in, they may cause metastasis.

We quickly learned that such eyes could usually be salvaged. In our first report on nine successful infusions, we saved eight eyes that had been scheduled for enucleation. Longer follow-up and larger numbers of cases confirmed that two-thirds of group V eyes (at two years) could be spared enucleation.13 It is worth emphasizing that this is significantly better than systemic chemotherapy or external beam irradiation for similar disease. OAC has allowed the majority of eyes previously enucleated to be salvaged.

Parameters Predicting Success

It is now clear that the success rate of OAC can be predicted based on just a few features of the tumor. Not surprisingly, success is significantly higher in naïve eyes than those that have already failed previous treatment with multiagent systemic chemotherapy or external beam irradiation, or both.

In our four-year review, using Kaplan-Meier estimates of success, 81% of advanced naïve eyes were salvaged, and 58% of those that had failed treatment were saved.14 Although naïve eyes do better, most (if not all) of the eyes that had failed prior treatment would have been enucleated, and the ability to save more than half of them represents a giant step in ocular oncology.

The second feature that predicts success is the presence or absence of a retinal detachment. The more extensive the retinal detachment, the greater the salvage rate with OAC. Eighty-nine percent of eyes with more than 50% retinal detachment can be salvaged.15

Pharmacologic studies have demonstrated that OAC delivers significantly higher (100-1,000 times) vitreous levels of drug.16 It is thought the drug accumulates in the subretinal space, allowing for higher concentrations for relatively long periods of time (four to six hours).

All studies have demonstrated that despite extensive vitreous seeding, OAC is usually successful in saving the eye. Even more important is the discovery that OAC can cure eyes with subretinal seeding. The Reese-Ellsworth classification scheme did not include subretinal seeds, but the International Classification scheme does. Subretinal seeds are known to predict near complete failure with systemic chemotherapy alone, and that is why they are in the International scheme. Remarkably, 83% of naïve eyes (at two years) with subretinal seeding were salvaged using OAC.13

Bilateral Disease

We introduced the concept of treating each eye in the same session when both eyes needed treatment in bilateral cases, and we called it “tandem therapy.”17 It is performed with the same femoral access by withdrawing the catheter from one internal carotid artery and then going up the fellow carotid artery to treat the second eye.

This approach might require double the dose of melphalan, which approaches hematological toxicity, so we have used different drugs in each eye. Melphalan is used in one eye, and carboplatin in the fellow eye, with topotecan added to the more involved eye. Success rates are similar to those of unilaterally treated cases.

Less Extensive Disease

Because we have had no significant systemic problems and acceptable intraocular complications, we have replaced multiagent systemic chemotherapy by OAC for eyes without extensive disease (but too extensive to be cured with laser, cryotherapy or plaque therapy alone).18 Close to 100% of such eyes can be salvaged with OAC, and an example is given in Figure 2.


Centers that have used this technique have reported few and minor systemic complications — no strokes or neurological damage, no deaths, and no second cancers as of August 10, 2012.19-20

As mentioned earlier, hyperemia in the distribution of the supratrochlear artery develops in 16% of cases (with transient loss of nasal lashes).21 Hematomas in the groin, transient decreased pulses in the leg, contrast dye allergy, and decreased tidal volume (discussed earlier) during the procedure have been recorded, but without long-term consequences.

We reported two cases of metastatic disease (both doing well years later), and it is anticipated that deaths from disease will develop. It is the advanced eyes that are associated with metastatic disease, and in those cases, metastatic disease is present, although usually not appreciated, at presentation.

It is also expected that approximately 1% of such cases will develop a second tumor in the first five years after treatment (trilateral disease) because this development occurs in patients who have the genetic form of the disease, even if no radiation is administered.

A notable difference between OAC and systemic chemotherapy is hematologic toxicity. We found that with melphalan doses of less than 0.4 mg/kg, minimal systemic toxicity developed. When the dose exceeded this amount, the chances of grade 4 neutropenia were significantly higher.22 Overall, fewer than 1% of naïve patients have required transfusions or developed fever/neutropenia (vs rates as high as 60% with systemic chemotherapy).23


Systemic chemotherapy has many undesired side effects (including death from treatment and chemotherapy-induced leukemia), but no significant ocular side effects. OAC has a higher potential for ocular damage, and this has been reported. Most of the side effects have been observed in centers early in their experiences, and strikingly, most of the papers reporting side effects have been first publications on experiences from centers worldwide.

Ophthalmic artery chemosurgery requires a coordinated team, and like all surgical procedures, it has the fewest side effects in the hands of those who do it the most. Significant side effects are related to technique. In our review of eyes enucleated after OAC at our center, no significant ocular toxicity was noted pathologically (the eyes were all removed because of progressive disease).

Similarly, no pathologic findings were identified in the Miami24 or California25 studies of eyes that were enucleated (because of progressive disease). It is clear that in the concentrations used, OAC is not toxic to the cornea, lens, or anterior segment.

Among the reported side effects are embolized cotton particles (presumably from contaminants in the field) and particulate material that may be related to drug preparation or solubility.19 Localized vascular occlusion and retinal pigment abnormalities have been identified and are thought to be related to technique (see discussion above).26-27

In a review of the second largest (and longest) experience with this technique28 Italian investigators concluded that “local adverse events are generally reversible.” Though they did have “permanent ptosis, strabismus/exotropia and chorioretinal dystrophy,” they emphasized that alternative management of these eyes is well known to be associated with significant and permanent local toxicity.28


In the six years since we began routinely catheterizing the ophthalmic artery for chemotherapy in patients with retinoblastoma, we have learned many things. For instance, each case must take into account the individual patient’s vascular anatomy because it varies widely. Further, while side effects decrease as chemo teams’ experience increases, the same cautions must be taken as with systemic chemotherapy.

These points notwithstanding, we now have an effective treatment for this devastating and potentially life-threatening disease. More eyes are being saved, and less recurrence is being seen, which provides greater hope for the future of these young patients. RP


1. Abramson DH, Dunkel IJ, Brodie SE, Kim JW, Gobin YP. A phase I/II study of direct intraarterial (ophthalmic artery) chemotherapy with melphalan for intraocular retinoblastoma initial results. Ophthalmology. 2008;115:1398-1404, 1404.e1.

2. Reese AB, Hyman GA, Tapley ND, Forrest AW. The treatment of retinoblastoma by X-ray and triethylene melamine. AMA Arch Ophthalmol. 1958; 60:897-906.

3. Kaneko A, Takayanna J, Matsuoka H, et al. [Chemothermotherapy was successful in two cases of recurrence of intraocular retinoblastoma after irradiation]. Rinsho Ganka. 1990;40:289-292.

4. Mohri M. [The technique of selective ophthalmic arterial infusion for conservative treatment of recurrent retinoblastoma]. Keio Igaku (J Keio Med Soc). 1993;70:679-687.

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]. Nippon Ganka Gakkai Zasshi. 1995;99:1230-1235.

6. Kaneko A, Suzuki S. Eye-preservation treatment of retinoblastoma with vitreous seeding. Jpn J Clin Oncol. 2003;33:601-607.

7. Suzuki S, Kaneko A, Management of intraocular retinoblastoma and ocular prognosis. Int J Clin Oncol. 2004;9:1-6.

8. Suzuki S, Yamane T, Mohri M, Kaneko A. Selective ophthalmic arterial injection therapy for intraocular retinoblastoma: the long-term prognosis. Ophthalmology. 2011;118:2081-2087.

9. Hayreh SS, Dass R. The ophthalmic artery: II. Intra-orbital course. Br J Ophthalmol. 1962;46:165-185.

10. Hayreh SS, Dass R. The ophthalmic artery: I. Origin and intra-cranial and intra-canalicular course. Br J Ophthalmol. 1962;46:65-98.

11. Gray’s Anatomy (30th ed). Clemente CD, ed. Philadelphia, PA; Lea and Febiger; 1985

12. Klufas MA, Gobin YP, Marr B, et al. Intra-arterial chemotherapy as a treatment for intraocular retinoblastoma: alternatives to direct ophthalmic artery catheterization. ANJR Am J Neuroradiol. 2012;33:1608-1614.

13. Abramson DH, Marr BP, Dunkel IJ, et al. Intra-arterial chemotherapy for retinoblastoma in eyes with vitreous and/or subretinal seeding: 2-year results. Br J Ophthalmol. 2012;96:499-502.

14. Gobin YP, Dunkel IJ, Marr BP, Brodie SE, Abramson DH. Intra-arterial chemotherapy for the management of retinoblastoma four year experience. Arch Ophthalmol. 2011;129:732-737.

15. Palioura S, Gobin P, Brodie S et al. Ophthalmic artery chemosurgery for the management of retinoblastoma in eyes with extensive (>50%) retinal detachment. Ped Blood Cancer. 2012 Apr 10. [Epub ahead of print]

16. Schaiquevich P, Buitrago E, Taich P, et al. Pharmacokinetic analysis of melphalan after superselective ophthalmic artery infusion in preclinical models and retinoblastoma patients. Invest Ophthalmol Vis Sci. 2012;53:4205-4212.

17. Abramson DH, Dunkel IJ, Brodie SE, Marr B, Gobin YP. Bilateral superselective ophthalmic artery chemotherapy for bilateral retinoblastoma tandem therapy. Arch Ophthalmol. 2010;128:370-372.

18. Abramson DH, Marr BP, Brodie SE, Dunkel I, Palioura S, Gobin YP. Ophthalmic artery chemosurgery for less advanced intraocular retinoblastoma: five year review. PLoS One. 2012;7:e34120.

19. Shields CL, Bianciotto CG, Jabbour P, et al. Intra-arterial chemotherapy for retinoblastoma report no. 2, treatment complications. Arch Ophthalmol. 2011;129:1407-1415.

20. Abramson DH, Dunkel IJ, Brodie SE, Marr B, Gobin YP. Superselective ophthalmic artery chemotherapy as primary treatment for retinoblastoma (chemosurgery). Ophthalmology. 2010;117:1623-1629.

21. Marr B, Gobin PY, Dunkel IJ, Brodie SE, Abramson DH. Spontaneous resolving periocular erythema and ciliary madarosis following intra-arterial chemotherapy for retinoblastoma. Middle East Afr J Ophthalmol. 2010;17:207-209.

22. Abramson DH. Chemosurgery for retinoblastoma: what we know after 5 years. Arch Ophthalmol. 2011;129:1492-1494.

23. Rizzuti AE, Dunkel IJ, Abramson DH. The adverse events of chemotherapy for retinoblastoma: what are they? Do we know? Arch Ophthalmol. 2008; 126;862-865.

24. Vajzovic LM, Murray TG, Aziz-Sultan MA, et al. Clinicopathologic review of enucleated eyes after intra-arterial chemotherapy with melphalan for advanced retinoblastoma. Arch Ophthalmol. 2010;128:1619-1623.

25. Kim J, Do H, Egbert P. Enucleated eyes after failed intra-arterial infusion of chemotherapy for unilateral retinoblastoma: histopathologic evaluation of vitreous seeding. Clin Ophthalmol. 2011;5:1655-1658.

26. Eagle RC, Shields CL, Bianciotto C, et al. Histopathologic observations after intra-arterial chemotherapy for retinoblastoma. Arch Ophthalmol.2011; 129:1416-1421.

27. Munier FL, Beck-Popovic M, Balmer A, et al. Occurrence of sectoral choroidal occlusive vasculopathy and retinal arteriolar embolization after superselective ophthalmic artery chemotherapy for advanced intra-ocular retinoblastoma. Retina. 2011;31:566-573.

28. Venturi C, Bracco S, Cerase A, et al. Superselective ophthalmic artery infusion of melphalan for intraocular retinoblastoma: preliminary results from 140 treatments. Acta Ophthalmol. 2012 Jan 23. [Epub ahead of print]