Intravitreal Lower-Dose (20 µg) Melphalan for Persistent or Recurrent Retinoblastoma Vitreous Seeds

Introduction

Intravenous chemotherapy (IVC) and intra-arterial chemotherapy (IAC) are the two most commonly used globe-conserving therapies in the management of retinoblastoma and have played a pivotal role in the management of retinoblastoma worldwide.^1–5 Based on the International Classification of Retinoblastoma (ICRB), IVC is successful for 100% of group A, 93% of group B, 90% of group C, 48% of group D, and 25% of group E eyes.⁶ Studies on IAC for retinoblastoma have demonstrated remarkable efficacy;^7,8 for example, Gobin et al.⁹ reviewed a 4-year experience with IAC, demonstrating ocular survival rates at 2 years in 82% for eyes with primary treatment and 58% for secondary treatment. More recently, Shields et al.¹⁰ found globe salvage in 72% of eyes treated with primary IAC and 62% of eyes with secondary IAC. In primary-treated eyes, globe salvage was 100% for group B, 100% for group C, 94% for group D, and 36% for group E eyes.¹⁰

The major cause of IVC and IAC failure is the persistence or recurrence of vitreous seeds (VS).^5,10–13 In recent years, intravitreal administration of chemotherapy has allowed for remarkable control of vitreous seeding.^11–15 Munier et al.¹¹ demonstrated 87% VS control at 22 months and Shields et al.¹² documented 100% control at 9 months of follow-up. In both reports, a 20 µg to 30 µg intermediate dose of melphalan (Alkeran; GlaxoSmithKline, Germany) was employed. Herein, we have studied the use of intravitreous chemotherapy using single, standard lower-dose melphalan (20 µg/0.1 mL) for VS.

Patients and Methods

A retrospective review was performed on all patients treated with low-dose intravitreal chemotherapy at the Ocular Oncology Service of Istanbul University, Istanbul Faculty of Medicine, Department of Ophthalmology, between August 2012 and April 2014. Institutional review board approval from Istanbul University, Istanbul Medical Faculty (approval # 14/1871) and written informed consent from parents of all children were obtained for this retrospective study. The trial conformed to the tenets of the Declaration of Helsinki. Inclusion criteria were eyes with viable VS (persistent or recurrent) following standard intravenous or intra-arterial chemotherapy, in which the remainder of the eye was stable with completely regressed solid intraretinal tumor and subretinal seeds. All eyes were followed for a minimum of 12 months after the last injection. Three eyes with less than 12 months of follow-up were excluded from the study. Exclusion criteria were eyes that showed additional viable solid tumor or subretinal seeds and those deemed at risk for metastatic disease with uveal or optic nerve invasion. A written informed consent of the parents was taken before every injection and potential risks of the procedure, including vitreous hemorrhage, retinal detachment, intraocular infection, and extraocular extension were explained in detail. All patients received outpatient intravitreal melphalan injection as an alternative to enucleation or external beam radiotherapy.

Each patient was examined under general anesthesia (ST), with detailed fundus drawings and Retcam digital photography (Massie Industries, Dublin, CA) to document tumors and related features in each eye. A comprehensive ocular examination under anesthesia included anterior segment evaluation, intraocular pressure measurement with Schiotz tonometry, and indirect ophthalmoscopy.

For the purpose of the study, data were collected regarding demographic patient features, including age at initial presentation (months), age at intravitreal melphalan injection (months), and gender. Data regarding tumor features were collected at the time of initial presentation, including laterality, retinoblastoma stage (ICRB group in both non-injected and injected eyes), largest tumor size in basal dimension (millimeter [mm]), thickness (mm), and initial treatment methods. Data regarding tumor features at the time of intravitreal melphalan injection included the status of the main (solid) retinoblastoma, the status of subretinal seeds and of VS (regressed, persistent, or recurrent), and number of quadrants involved with viable VS.

Results

The study included seven eyes of seven consecutive patients with viable (persistent or recurrent) VS. The demographic features are shown in the Table. At the time of initial presentation, based on the ICRB, the injected eyes were classified as Group D (n = 4; 57%) and Group E (n = 3; 43%). The mean largest tumor basal diameter was 17 mm (median: 17 mm; range: 12 mm-24 mm) and tumor thickness was 9.1 mm (median: 9.5; range: 5.5 mm-12.8 mm). Primary treatment included intra-arterial chemotherapy in four eyes (57%) or intravenous chemotherapy with vincristine, etoposide, and carboplatin for six cycles in three eyes (43%; Table). These therapies led to complete solid tumor and subretinal seed control in all eyes, but viable VS showed persistence (n = 5; 71%) or recurrence (n = 2; 29%), necessitating intravitreal chemotherapy.

Table: Summary of Seven Patients With Retinoblastoma and Persistent or Recurrent Vitreous Tumor Seeds Managed With Intravitreal Melphalan Injection

The mean patient age at the time of intravitreal injection was 37 months (median: 37 months; range: 18–67 months). The VS involved one (n = 3; 43%), two (n = 3; 29%), or four (n = 2; 29%) quadrant(s) of the vitreous cavity. Each patient received injection of a standard lower-dose (20 µg/0.1 mL) melphalan until complete VS control was achieved. A total of 14 injections were administered in seven eyes, with a mean of two injections (median: 2; range: 1–4; Figures 1 and 2).

Figure 1. Case 1: A 55-month-old child with group D retinoblastoma (A–F). (A) A large, inferonasal endophytic retinoblastoma with vitreous seeds (VS) is seen inferotemporally. (B) After four cycles of intra-arterial chemotherapy, tumor regression was achieved but VS persisted. (C) Following intravitreal melphalan (IVM) No. 1. (D) Following IVM No. 2. (E,F) Following IVM No. 3 showing complete VS regression and diffuse retinal pigment epithelial (RPE) mottling. Case 2: A 12-month-old child with group D retinoblastoma (G–I). (G) The mass was superotemporal, endophytic, and with diffuse VS. (H) After two cycles of intra-arterial chemotherapy, the mass responded but VS persisted. (I) Following four IVM injections, the VS completely regressed. Moderate RPE mottling was noted. Case 3: An 8-month-old child with recurrent VS (J–L). (J) After six cycles of intravenous chemotherapy, marginal tumor recurrence (white arrow) with two viable VS (black arrow) were detected. This area was managed with cryotherapy. (K) Following cryotherapy, there were 10 new viable VS (white inbox). (L) One month following one IVM injection, the VS showed complete regression.

Figure 2. Case 4: A 19-month-old child with group D retinoblastoma in the left eye. (A–B) After six cycles of intravenous chemotherapy, multiple recurrent vitreous seeds (VS) were detected (white arrows,white box, and white outline). (C–D) Following two sessions of intravitreal melphalan, the VS were completely regressed. Radiation maculopathy from previous plaque radiotherapy was also noted.

After a median follow-up period of 20 months (range: 12–32 months), complete regression of VS was achieved in seven cases (100%) and globe salvage in six cases (86%). One eye required enucleation for solid tumor recurrence despite regression of VS. In three cases (43%), VS control was achieved with a single injection, and follow-up of more than a year showed no recurrence. In four cases (58%), two to four injections were necessary.

The complications included local retinal pigment epithelial (RPE) mottling in two cases (29%). In one patient (Table, case no. 2; Figures 1G-1I), two (superotemporal and inferotemporal) quadrants were involved with the VS. RPE changes appeared after the second injection (ie, 2 months after the first injection) and involved the superotemporal quadrant. These changes showed progression with subsequent two injections. Fortunately, VS regressed completely after the fourth injection, and fovea was spared despite extension of these RPE changes into the macular region. In the second patient (Table, case no. 6), one (superotemporal) quadrant was involved with the VS. RPE changes appeared 1 month after the first injection in the superonasal quadrant. There was no case of vitreous hemorrhage, retinal detachment, endophthalmitis, cataract, hypotony, phthisis bulbi, extraocular tumor invasion, metastasis, or death.

Discussion

The most challenging part of the management of retinoblastoma is the control of VS. Due to their distant location from the blood supply, it is obvious that intravenous chemotherapy does not sufficiently eradicate these VS in most of the cases. In 2002, an analysis of 158 eyes with retinoblastoma found 54 eyes with active VS at presentation.¹⁶ Following standard IVC, VS recurrence was found in 50% by 5-year follow-up, generally leading to plaque radiotherapy, external beam radiotherapy, or enucleation for tumor control.¹⁶ Pharmacokinetic studies have shown that periocular and intra-arterial chemotherapy have greatly improved the ocular penetration of drugs compared to IVC. However, the achieved vitreous concentration is barely tumoricidal and does not last long enough for tumor control.¹⁷ In 2011, an analysis of 17 eyes with retinoblastoma treated with IAC showed that nine eyes had vitreous seeding at presentation and VS control was achieved in six eyes (67%), but persistence/recurrence occurred in three eyes (33%).⁷ These reports indicated that control of vitreous seeding can be challenging with both IVC and IAC. To overcome this problem, intravitreal chemotherapy has emerged as an effective drug for seed control, but the ideal concentration is not yet clear.

In the 1990s, the first study on intravitreal melphalan, by Kaneko and Suzuki, employed 8 µg melphalan combined with ocular hyperthermia for vitreous seeding in 41 eyes, and the globe salvage rate was 51% at 50 months’ follow-up.¹⁸ This early very-low–dose melphalan was found to be ineffective for complete control of VS. Thus, an intermediate dose of 20 µg to 40 µg was explored in later studies.^11,12 Munier et al.¹¹ reported 23 cases with viable VS treated with up to eight weekly intravitreal melphalan injections at an intermediate dose of 20 µg/0.1 mL to 30 µg/0.1 mL and found 87% control at 22 months. Ghassemi and Shields¹³ subsequently evaluated 12 eyes treated with a variable dose (8 µg/0.05 mL-50 µg/0.05 mL) of intravitreal melphalan for recurrent VS following previous therapies of IVC and IAC. Supporting the aforementioned findings in those two studies,^11,18 they demonstrated that eyes treated with very-low–dose melphalan (8 µg /0.1 mL-10 µg /0.1 mL) showed less control (43%) and minimal side effects, whereas those treated with higher doses (30 µg/0.1 mL-50 µg/0.1 mL) demonstrated excellent control (100%); however, the 50-µg dose was toxic with severe complications, including hypotony and phthisis bulbi.¹³ Later, Shields et al.¹² studied 11 eyes with viable vitreous seeding managed with a similar intermediate dose (20 µg/0.1 mL-30 µg/0.1 mL) melphalan with complete VS control and ocular survival in all eyes (100%), but their protocol of six consecutive injection lead to a total of 55 monthly injections (mean: 5; range: 2–6).

Complications of intravitreal melphalan need to be considered. Ueda et al.¹⁹ studied the effects of intravitreal melphalan injections in albino rabbits using electroretinography (ERG) and found that the retinal structure remained unchanged after a 10-µg injection, moderate change was observed after a 20-µg injection, and deterioration was observed after 90-µg injection. In a recent study by Francis et al.,²⁰ retinal and systemic toxicity of 30 µg intravitreal melphalan were evaluated in a rabbit model, demonstrating that weekly injections resulted in a decreased ERG response, confirming retinal toxicity. Histopathology revealed severely atrophic retina. Systemic toxicity in humans or rabbits was not detected.²⁰

From a clinical standpoint, Munier et al.¹⁵ evaluated 30 eyes of 30 children with retinoblastoma using 135 intravitreal injections and found complications of transient localized vitreous hemorrhage (n = 3/30 eyes) and localized peripheral salt-and-pepper retinopathy (n = 10/23 eyes) at the site of injection. Shields et al.¹² studied 11 consecutive eyes managed with intravitreal melphalan at a dose of 20 µg to 30 µg and noted minor complications of focal retinal pigment epithelial mottling near the site of injection in two eyes and nonaxial posterior lens opacity in two eyes. In our study, there was no case of extraocular extension, vitreous hemorrhage, retinal detachment, endophthalmitis, cataract, hypotony, or phthisis bulbi. Similar to previous studies,^20,21 we noted localized peripheral salt-and-pepper retinopathy in two eyes (29%) at the site of injection.

In the current literature, there is no clear consensus in the use of post-injection antibiotic prophylaxis. Munier et al.¹⁵ reported the use of postoperative topical antibiotics (chloramphenicol) and steroids (dexamethasone phosphate); on the other hand, some authors^12,13 did not prescribe any further medication postoperatively. In this study, we used postoperative antibiotic drops in our patients. However, the current data suggest such treatment is likely unnecessary.

In this study, we selectively explored the efficacy of lower-dose melphalan (20 µg/0.1 mL). We achieved 100% VS control in all 14 eyes with minimal toxic effects. Previous protocols^11–13 have called for six to eight injections of melphalan. However, we achieved VS control with a median of two injections. Therefore, perhaps tumor control can be reached with fewer injections than previously thought.

There are limitations to our study, including the small cohort size, but it should be considered that this is a rare cancer, and our seven patients demonstrated 100% response. Additionally, this cohort has a relatively short follow-up of a minimum 12 months (median: 20 months), and longer follow-up can address long-term stability of our observations. However, most active VS demonstrate recurrence within 12 months; therefore, we believe our follow-up to be sufficient.

In conclusion, intravitreal chemotherapy is currently reserved for eyes with active VS and we have observed that a standard lower-dose dose (20 µg/0.1 mL) intravitreal melphalan is a safe and remarkably effective therapy, requiring only two injections in most cases to achieve control.