A Multi-Institutional Feasibility Study of Intra-Arterial Chemotherapy in Children with Retinoblastoma. A Children’s Oncology Group Study (COG ARET12P1)

Abstract

Background:

Intra-arterial chemotherapy (IA) as a treatment to salvage the eye with advanced retinoblastoma is increasingly utilized based on successes reported by institutions around the world mainly through retrospective studies.

Objective:

To study the feasibility of delivering melphalan directly into the ophthalmic artery in a multi-institutional prospective study in children with newly diagnosed unilateral Group D retinoblastoma

Methods:

The Children’s Oncology Group (COG) initiated study ARET12P1 in 2014 and was open to nine institutions. Eligible patients older than 6 months of age were enrolled. Feasibility of delivering 3 injections of melphalan into the ophthalmic artery every 28 days was assessed.

Results:

Nine institutions participated in this trial. Fourteen patients were enrolled two of whom were unevaluable for feasibility. Four patients experienced a feasibility failure. In two patients the ophthalmic artery could not be accessed for the second IA injection, in one the artery could not be accessed for the first injection, and one patient experienced grade 4 hypotension during the procedure.

Conclusion:

Delivery of prescribed therapy within the context of this study did not meet the feasibility goals of the study with only 67% feasibility success rate. These results should caution centers that plan to initiate this treatment and suggest investment in training to achieve technical expertise or referral to centers with expertise.

Introduction

The management of advanced intraocular unilateral retinoblastoma historically included enucleation as a common intervention. Attempts to salvage the globe with systemic intravenous chemotherapy were successful in approximately 50% of patients with group D disease^1-7. Increasingly, efforts to cure the child and save the globe have rapidly evolved to local delivery techniques like periocular carboplatin in conjunction with systemic chemotherapy as well as IA therapies ^5,8,9. At the time of initiation of this study, retrospective data from single institutions had reported promising results of the intra-arterial technique of delivering chemotherapy to eyes of children with retinoblastoma via the ophthalmic artery^9-12 . This technique is a variation of the original intra-arterial technique used by Japanese investigators^13,14 . However, there were no prospective studies with well-defined study parameters addressing not only the outcome but also the feasibility and toxicities associated with this technique, and the ability to deliver this treatment by the broader pediatric oncology provider community. Therefore, the COG initiated a study to prospectively evaluate the feasibility of IA in a multi-institutional setting. The overall goal of the study was to evaluate how feasible this technique is in a group wide setting within the COG.

Patients and Methods

The primary aim of this multi-institutional study was to evaluate the feasibility of delivering melphalan directly into the ophthalmic artery in children with newly diagnosed unilateral Group D Retinoblastoma¹⁵. Patients older than 6 months with newly diagnosed intraocular disease and no prior therapy were eligible for this study. Diagnostic examination under anesthesia (EUA) and MRI or CT of the brain were performed within 14 days of study entry. Rapid central review confirmation of Group D disease based on RetCam images from diagnostic EUA was obtained before starting treatment. Patients had a performance status corresponding to ECOG scores of 0, 1 or 2 with adequate renal and liver function. They had a life expectancy of ≥ 8 weeks. Patients with bilateral disease and those with unilateral groups A, B, C, and E disease were not included in the study.

All eligible patients were to receive 1 intra-arterial injection of melphalan every 28 days to a maximum of 3 cycles assuming the patients met the criteria to begin the next cycle. Subsequent cycles required an ANC ≥ 1000/μL, platelets ≥ 100,000/μL, Hemoglobin ≥ 8 g/dL, SGPT (ALT) < 2.5 x upper limit of normal (ULN) for age. Central review comments based on the interventional radiology recording from Cycle 1 were required to be obtained for each patient before starting Cycle 2 treatment. Following 3 cycles of intra-arterial injections patients could continue to receive local therapies with cryotherapy and/or laser therapy. Patients who progressed and then received either systemic chemotherapy, additional intra-arterial, radiation therapy or enucleation were considered to have had an event.

Melphalan was the single drug used for all 3 cycles in this study. The doses of melphalan in this study were 3 mg for patients between 6 and 12 months of age, 3.5 mg between the ages of 12 months and 24 months and 4 mg for children 24 months and older. This drug schedule was based on growth in volume of the globe with age^16,17.

As in previous COG protocols, a central review mechanism was provided to review eligibility based on initial stage. Three ophthalmologists with one alternate were on the review panel to assess eligibility. Interventional radiology review of radiology recordings of the first IA injection on every patient was required. The purpose of these reviews was to assess the technique of the individual interventional radiologist and the anatomy of the vasculature involving the ophthalmic artery. The expectation was that the reviewer could suggest a dose modification after studying the vasculature of the ophthalmic artery.

Toxicity:

All eligible patients who received protocol therapy were evaluable for toxicity. The toxicities were graded according to the NCI Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. The number and percentage of patients with grade 3 or 4 toxicity during protocol therapy were reported.

Evaluation of Feasibility:

The IA feasibility was assessed over the 3 injections into the ophthalmic artery. A patient was considered to have experienced IA therapy feasibility failure if: a) the interventional radiologist was not able to access the ophthalmic artery for chemotherapy administrations during any of the first three cycles of therapy; b) the patient developed central retinal artery occlusion after the first or second cycle that did not reopen by the time the next injection was due; or c) the patient could not receive all three treatments with IA therapy because of Grade 3 or 4 adverse events that were considered possibly, probably or likely related to treatment. Any patient who received two IA infusions but was removed from protocol therapy prior to the third IA injection was considered unevaluable for the primary outcome measure and replaced for evaluation of the primary goal. The reasons for removal from protocol therapy may have been one of the following: (1) the physician stopped protocol therapy in the best interests of the patient; (2) the parent or guardian refused further therapy; (3) disease was found in the contralateral eye; or (4) the patient did not qualify because repeat eligibility studies were outside enrollment requirements. All other patients were considered evaluable for the feasibility analysis. Any feasibility evaluable patient who did not experience feasibility failure was considered a feasibility success.

Statistics:

Feasibility Analysis:

The study was planned for 18 months of enrollment with a projected enrollment of 31 evaluable patients. The primary goal was to determine the feasibility of delivering melphalan directly into the ophthalmic artery. A two-stage design was used. In the first stage, 12 patients would be accrued. If there were 8 or fewer feasibility successes, the study would be stopped, and treatment would be considered not feasible to deliver. Otherwise, 19 additional patients would be accrued for a total of 31. If 26 or more feasibility successes were observed out of 31 patients, the therapy would be considered feasible to deliver. With this design, if the true feasibility success rate was 72%, the regimen would be identified as not feasible to deliver with a probability 91%. If the true feasibility success rate was 90%, the regimen would be identified as feasible to deliver with a probability 90%.

Ocular Salvage:

Ocular salvage would be assessed over the 2 years following start of protocol therapy. Any patient who was lost to follow-up or for whom follow-up was terminated electively prior to 2 years by patient or parent preference would be considered unevaluable for the ocular salvage endpoint. The number and percentage of ocular salvage success at 2 years would be reported as a secondary objective.

Results:

This study was opened in April 2014 and closed in April of 2017. Fourteen patients, nine of whom were male were enrolled, and their median age was 2 years (0.6 – 3.8). All patients were contributed by large freestanding children’s hospitals in the United States. Of the 14 patients on study three patients did not receive all three injections due to procedure related issues. Of the three patients who failed to receive all three injections, one patient’s ophthalmic artery had an anatomical abnormality and was in spasm and the operator did not proceed with the second injection. Another patient’s ophthalmic artery could not be accessed at the time of the first injection and the procedure was abandoned. The third patient’s ophthalmic artery was in spasm, and did not respond to verapamil at the time of second injection and further infusions were abandoned. These three patients were treated with other interventions. One of these eyes was enucleated, one eye was salvaged with systemic chemotherapy and intravitreal injections and the third one was lost to follow up. A fourth patient did not receive all three injections due to grade 4 toxicity at the time of the first injection with significant hypotension. .

Central review of the first procedure was available for 12 patients. In one of the patients the procedure could not be assessed. In the remaining 11 patients the procedure and the injections of melphalan were deemed successful and did not require dose modification based on anatomy.

Toxicity (Table 1):

TABLE 1:

Summary of Grade 3 or above adverse events for treated patients^*.

Toxicity term in CTCAE 4.0	No. (%)
Anaphylaxis	1 (07.7)
Gastroenteritis	1 (07.7)
Hypotension	1 (07.7)
Hypoxia	2 (15.4)
Neutrophil Count Decreased	3 (23.1)
Platelet Count Decreased	1 (07.7)
Any AE	5 (38.5)

^*This analysis included 13 patients who began receiving the trial treatment.

A total of 43 cycles were successfully administered. One patient developed grade 3 neutropenia after the 1^st and the 3^rd injections, one patient had grade 3 neutropenia and grade 3 platelet count decrease after the first injection. One patient developed grade 4 and grade 3 hypoxia after the 2^nd and 3^rd injections. Grade 3 anaphylaxis occurred in one patient with the 2^nd injection and one patient developed grade 4 hypotension with cycle 1. One patient developed grade 3 gastroenteritis after the third injection.

Feasibility Analysis:

Two patients were excluded from the evaluation of feasibility: The first patient did not receive any protocol therapy. The second patient received one cycle and did not experience a feasibility failure, but the parent or guardian refused further therapy prior to the second injection (Figure 1). Among the 12 patients evaluable for feasibility, 4 experienced feasibility failures including inability to access ophthalmic artery in 3 patients, and grade 4 hypotension in 1 patient. The stopping boundary for stage 1 was crossed, so the study was terminated at the end of stage 1, and we concluded that treatment was considered not feasible to deliver. The feasibility success rate was 67% (two-sided 80% CI: 48%, 81%).

Figure 1:

Consort Diagram for ARET12P1

Ocular Salvage:

Eleven patients out of 14 were evaluable for ocular salvage assessment. Three were excluded from the ocular salvage analysis due to loss to follow-up or whose follow-up was terminated electively prior to 2 years by patient or parent preference following the pre-specified analysis plan. Four out of 11 (36%, 95% CI: 15%-65%) did not have the affected eye enucleated within 2 years of enrollment and were considered a success. Three of these 4 patients in the follow up period received systemic chemotherapy and one of these also received intravitreal therapy along with systemic chemotherapy. Only one of the four patients did not receive any further therapy. Seven patients underwent enucleation. These enucleations were performed at a median of 6 months of follow up (3-9 months). Six patients received all 3 cycles of IA therapy and of these 6 four eventually underwent enucleation.

Discussion:

Retinoblastoma, the most common intraocular tumor is highly curable when confined to the globe. Children with unilateral retinoblastoma often present with advanced disease, groups D or E at the time of initial diagnosis. Vision in the eye that has advanced disease is usually very poor and therefore enucleation which is curative is a widely accepted treatment option. However, this results in loss of a globe which is often difficult to accept for parents of the affected children. Efforts to cure the affected child and preserve the globe with chemotherapy were underway since the 1990s when systemic chemotherapy appeared to prevent enucleation in a modest number of children with advanced disease^1-7,18-21. An improvement in globe salvage occurred when a modified technique of intraarterial chemotherapy injected into the ophthalmic artery was introduced^10;14. This technique involves the use of specialized services of an interventional radiologist and a pediatric anesthesiologist which are not universally available especially in centers with more limited resources. The initial experiences with IA therapy were from individual institutions^9;10;22-24.

The current study was an attempt to evaluate the feasibility of administering IA therapy in a multi-institutional setting to further assess and understand the ability to generalize this treatment outside centers of excellence. At the time of the initiation of the study, the drug melphalan was considered the single most effective drug for this treatment modality and was therefore selected. In addition to this, melphalan does not require passage through the liver for activation and has an immediate peak followed by decease in concentration and elimination. Based on the growth characteristics of the globe, melphalan doses were chosen according to age^16,17. Initial reports suggested that there was evidence of systemic toxicity although the therapy was local. Our study also showed that systemic toxicity can occur in terms of grade 3 hematologic toxicity. In addition, hypoxia, “anaphylaxis” and hypotension were also seen in this study. At the initiation of the study, there was limited information about the potential risks and incidence of cardiovascular decompensation. Currently, data has accumulated to show that there is an approximately 30% incidence of hypoxia, hypotension and bradycardia associated with manipulation of the catheter within the carotid artery^25-28. Today, these events are anticipated and addressed by the anesthesiologists when they occur and the procedures are completed without prolonged interruptions.

Grade 4 hypotension occurred in the last patient on the study associated with first injection and given that this was an unexpected grade 4 toxicity and related to IA therapy this was considered as a feasibility failure. Our study did not meet the feasibility goal with the majority of feasibility failures due to technical failures to access the ophthalmic arteries. This highlights the importance of limiting the delivery of this complex treatment in high volume centers with technical expertise. Retinoblastoma is a rare disease and one individual institution will not see a substantial number of patients to acquire the required experience with the delivery technique in a short period of time; extensive training and programmatic focus are needed for success.²⁹ This study was initiated in 2014 and therefore the results reported here need to be interpreted in its due historical context as one can expect greater experience with IA therapy at the current time. Commensurate with feasibility failure, it is not surprising that the globe salvage rate with IA therapy alone at 2 years in this study was only 36%. However, this salvage rate should be considered in the context of the prescribed three IA injections only with a single drug melphalan and without intravitreal therapy.

The management of advanced retinoblastoma has evolved from using one drug to up to three drugs for IA therapy with topotecan and carboplatin in addition to melphalan^24,30-33. In addition, intravitreal injections of melphalan, cryotherapy, diode laser thermotherapy, and brachytherapy along with IA therapy are increasingly part of the initial therapy in patients with advanced retinoblastoma. However, there are no established criteria for the use one vs two vs three drugs for IA therapy and the need and timing for concomitant intravitreal injections. Therefore, it is increasingly difficult to evaluate the contribution of IA therapy alone to ocular salvage. Likewise, there are very few prospective randomized studies to establish efficacy, documentation of toxicities, and evaluation of responses to IA therapy (clinicaltrials.gov-NCT04681417). The use of IA therapy is accepted in many parts of the world, even in places where there are no interventional radiologists and anesthesiologists specifically trained to treat pediatric patients, and the outcomes and toxicities are not always published. The feasibility limitations encountered in this trial may be a good reminder of the reality of the generalized use of this therapy beyond centers of excellence. A few centers in developed and middle income countries have reported excellent outcomes in globe salvage in eyes with advanced disease, but this has not been validated convincingly in multi-institutional studies with careful monitoring of efficacy and toxicity. The impediments to initiating a multidisciplinary multi-institutional study are several:1) There are not enough patients for a randomized trial in an individual country or a consortium. 2) given the variations in approach to management, gathering a consensus for a protocol is difficult and 3) it is difficult to restrict members of the multidisciplinary team to the confines of a protocol when the end point of globe salvage justifies the means (with protocol violations) to achieve it.

Although the current study was a feasibility failure, considerable experience has been achieved in the years since the initiation of this study. The time is now ripe to join forces to design an international study of IA therapy with intravitreal injections to establish evidence-based guidelines which can lead to more precise understanding of the contributions of the various treatments to globe salvage with expected but minimal toxicities. Such a study should involve large Retinoblastoma Centers to which patients should be referred.

Abbreviation

COG

Children’s Oncology Group

IA

Intra-arterial chemotherapy

EUA

Examination under anesthesia