Article Text
PDF
Abstract
Background Intra-arterial chemotherapy has an increasingly prominent role in the management of retinoblastoma. One concern regarding this technique is procedural radiation exposure.
Objectives To examine the effects of our institution’s procedural technique on fluoroscopy parameters for patients undergoing intra-arterial chemotherapy infusions for intraocular retinoblastoma. Secondary goals included describing the effect of anatomical variations of the carotid siphon and ophthalmic artery on radiation dose.
Methods A retrospective review of pediatric patients with retinoblastoma referred to interventional neuroradiology for chemosurgery was performed. Techniques were classified as: A (1.2 Fr or 1.5 Fr microcatheter with continuous verapamil flush, advanced without guide through a 2 Fr sheath) or B (1.5 Fr or 1.7 Fr microcatheter advanced within a 4 Fr base catheter, through a 4 Fr sheath). Statistical analysis was performed to determine if there was a significant difference in fluoroscopy parameters based on technique or due to anatomical variation.
Results 26 patients were treated with 94 intra-arterial chemotherapy infusions. 34 procedures were performed using technique A and 60 using technique B. Mean fluoroscopy time (4.75 min), fluoroscopy dose (23.3 mGy), and dose–area product (DAP; 85.2 μGy.m2) for technique A were significantly lower (p value <0.05) than for technique B, 14.0 min., 191 mGy, and 586 μGy.cm2, respectively.
Conclusions Microcatheter-only technique with continuous verapamil infusion resulted in decreased fluoroscopy times, DAP, and radiation doses at our institution for the treatment of intraocular retinoblastoma. Furthermore, our fluoroscopy times using this technique are the lowest reported in the current literature. Additionally, our anatomical analysis has demonstrated a positive correlation between increasing vessel tortuosity and fluoroscopy times.
- embolic
- malignant
- orbit
- pediatrics
- technique
Statistics from Altmetric.com
Introduction
Intraocular retinoblastoma is a rare childhood malignancy, with approximately 300–350 new cases diagnosed within the USA annually.1 The treatment of retinoblastoma has changed dramatically with the advent of intra-arterial chemosurgery, which has contributed to globe salvage with reduced rates of enucleation.2 3 Furthermore, intra-arterial chemotherapy has been shown to be more efficacious if used as a primary versus secondary or salvage therapy.4 Although the efficacy of intra-arterial chemosurgery has been demonstrated, an optimal technique for delivering treatment while minimizing procedural morbidity has yet to be elucidated.
Patients with retinoblastoma have been shown to be highly sensitive to the ionizing effects of radiation. MacCarthy et al previously documented that patients with heritable forms of retinoblastoma receiving regional radiation treatment have a 32 times greater relative risk of developing secondary head and neck cancers than the general population.5 While the total administered radiation dose associated with intra-arterial chemosurgery is significantly lower than that of whole brain radiation, the deleterious effects of ionizing radiation are still significant.6–8 Thus, research to define the most effective technique to balance effective treatment and risk reduction is of paramount importance to minimize the stochastic risk of ionizing radiation.
This paper aims to describe how a change in our operative technique has resulted in significantly lower fluoroscopy times, total fluoroscopy dose, and dose–area product (DAP). The operative technique at our institution changed in 2017 with the technical changes including a microcatheter-only approach with the use of continuous verapamil infusion instead of a microcatheter and sheath apparatus without the use of verapamil. The concept was initially hypothesized to decrease the rate of instrumentation-associated vasospasm and increase our technical success and chemotherapy delivery. We subsequently observed a correlation between our technique change and fluoroscopy parameters.
Methods and materials
Demographics
Twenty-six patients with retinoblastoma amenable to treatment with intra-arterial chemotherapy were referred to Texas Children’s Hospital multidisciplinary group, led by ophthalmology. Patients with extraocular disease were excluded. Appropriate patients were then subsequently referred to neurointerventional radiology for treatment over the past 5 years. Data collected included age, gender, prior treatment modalities, side of treatment, number of cycles, and retinoblastoma grouping.
Operative technique
Operative techniques used at our institution in the past 5 years were of two types—technique A (used starting in 2017) and technique B (used in 2017 and earlier). These procedures were performed by a total of four interventional radiologists, A, B, C, and D. Technique A was performed solely by operator D, with 10 years of post-fellowship experience, while technique B was performed by operators A, B, and C with a combined post-fellowship experience of 56 years. Parental informed consent was obtained for all patients.
Both techniques are performed with the patient supine and under general endotracheal anesthesia, provided by the anesthesiology team. In technique A, femoral arterial access is obtained using a 21-gauge micropuncture needle, using ultrasound guidance, which is then exchanged for a 4 Fr micropuncture dilator and connected to a tuohy borst, providing a 2.9 Fr sheath equivalent. A 1.2 Fr Balt Magic (Irvine, California, USA) or 1.5 Fr Medtronic Marathon (Irvine) microcatheter is then directly advanced into the internal carotid artery at the level of the ophthalmic artery ostium. A solution of 1 mg/500 cc verapamil in heparinized saline is continuously infused at an estimated rate of 20 cc/h at the level of the ophthalmic artery ostium, throughout the procedure to minimize vasospasm from catheter and wire manipulation. Single agent, melphalan 3.5–5 mg, or dual agent, melphalan 3.5–5 mg and carboplatin 40 mg, infusions are then performed. Treatments initially started with single drug infusions. Dual agent infusions were subsequently administered for future treatments if lack of tumor response was determined via direct ophthalmic examination under anesthesia.
In technique B, femoral arterial access is obtained and a 4 Fr sheath is placed. A 4 Fr Davis or Berenstein catheter is used to select the common carotid artery with a .035" glide wire. After angiography, a 1.5 Fr Medtronic Marathon or 1.7 Fr Medtronic Echelon 10 or 1.7 French Stryker Excelsior Sl-10 (Freemont, California, USA) microcatheter is advanced through the 4 Fr base catheter into the internal carotid artery at the level of the ophthalmic artery ostium. A single agent, melphalan 3.5–5 mg is then infused.
If the ophthalmic artery cannot be selected, either owing to vasospasm or difficulty of access, the middle meningeal artery or other collaterals may be used to attempt adequate delivery. A subsequent angiogram is performed, followed by chemotherapy delivery. Post-therapy ophthalmic arterial angiography is then performed. The catheters and wires are removed, and hemostasis is achieved using manual pressure.
Anatomical variations
We classified the origin of the ophthalmic artery as one of five positions: within the curve of the anterior genu (position 1), past the curve of the genu but before the halfway point to the posterior communicating artery origin (position 2), beyond the halfway point from the genu to the posterior communicating artery origin (position 3), originating from the external carotid artery (position 4), and originating from the cavernous carotid (position 5) (figure 1).
Angiographic examples of ophthalmic artery positions.
The variability of the petrous and cavernous internal carotid artery segments has previously been described and classified.9 There are five categories for the A3 angle, the angle that separates the petrous from the cavernous internal carotid artery. Progressing from type 1 to type 5 indicates increasingly tortuous anatomy, with type 1 including a gentle obtuse curvature, type 3 a right angle, and type 5 a kinked angle (figure 2).
Angiographic examples of carotid siphon types.
Statistical analysis
Fluoroscopy time, fluoroscopy dose, and dose–area product (DAP) were summarized by categorical variables. A univariate mixed model with compound symmetry for the covariance structure was applied for the comparisons. P values <0.05 were considered significant.
The mixed model was chosen for the comparisons because the 94 observations were not independent. Procedures of the same patients had dependent variables and the regular two group comparison tests, such as t test or Wilcoxon rank test, do not consider the dependence structure in the data.
Results
From June 2013 to November 2018, a total of 94 intra-arterial chemotherapy infusions were performed on 26 (16 male and 10 female) patients to treat 27 eyes diagnosed with retinoblastoma. The mean age and SD for the patients treated was 2.1±1.1 years. There were 18 dual agent infusions, 74 single agent infusions, and two failed infusions. Of the successful infusions, 82 were performed with catheter tip positioned at the ophthalmic artery ostium and 10 with catheter in the middle meningeal artery.
The most recent 34 cases, beginning from September 2017, were performed using technique A, whereas the preceding 60 cases were performed using technique B. Mean fluoroscopy times, fluoroscopy doses, and DAP are presented in table 1.
Comparison of fluoroscopy parameters for technique A versus technique B
Intraoperative digital subtraction angiograms were reviewed to determine carotid siphon type and ophthalmic artery position. Twenty-one percent of our patients had type 1 carotid siphon, 22% type 2, 12% type 3, 34% type 4, and 11% type 5. Twenty-three percent of our population demonstrated ophthalmic artery position 1, 59% position 2, 11% position 3, and 7.4% position 4. No patients had ophthalmic artery position 5. Fluoroscopy times, doses, and DAP based on ophthalmic artery position are listed in table 2.
Comparison of fluoroscopy parameters based on ophthalmic artery (OA) position
Seventy-four procedures used single agent infusions, 18 used dual agent infusions, and two procedures had failed infusions. Dual agent therapy resulted in significantly lower fluoroscopy times than single agent therapy (3.3 min vs 11.1 min, p=0.0186).
Specifically, within the technique A cohort, dual agent therapy demonstrated a trend for lower fluoroscopy parameters (3.28 min, 21.12 mGy, and 71.02 μGy.m2) in comparison with single agent infusions (6.04 min, 26.23 mGy, and 102.2 μGy.m2) however, these results were not statistically significant.
Discussion
The results of this study demonstrate that our change to technique A significantly reduced fluoroscopy times, fluoroscopy dose, and DAP at our institution. Although the microcatheter-only technique has been previously described and used by Gobin,2 the addition of continuous verapamil infusion has not been described. As a result, we significantly reduced fluoroscopy parameters at our own institution, and also present the lowest reported fluoroscopy times for intra-arterial chemotherapy for intraoculuar retinoblastoma.
Our institution collected data on three separate radiation parameters, fluoroscopy time, fluoroscopy dose, and DAP. The measurement of fluoroscopy time and total fluoroscopy dose are primarily used to predict the deterministic effects of ionizing radiation, such as skin burns, hair loss, and cataract formation.10 Deterministic effects typically include a threshold for the development of the specific disease entity, and higher radiation doses past the threshold also increase disease severity.11 DAP is a very accurate descriptor of effective radiation dose—a measure of the total absorbed dose per unit tissue.8 Thus, DAP is also used to account for potential stochastic effects such as the development of secondary malignancies because it is a measure of the absorbed dose of specific tissue per unit area while also accounting for specific tissue-weighting factors.8
Our technique overall has decreased radiation doses and times, but several other variables have been found to have a significant effect on doses. There is significant variability of the blood supply for retinoblastoma therapy, and Marr et al described the significant differences in radiation dose when comparing ophthalmic artery infusion versus middle meningeal artery infusion.3 Other variables that we have examined that may influence radiation exposure include carotid siphon type and ophthalmic artery origin.
Vijaywargiya et al described significant variability in the carotid siphon tortuosity with the A3 angle ranging from 21 to 134 degrees9. Thus, we hypothesized that carotid siphon tortuosity would have a significant impact on fluoroscopy times and radiation dose as increasingly acute angles can be more difficult to traverse. To objectively categorize carotid siphon anatomy we used the classification system described by Vijaywargiya et al and ophthalmic artery positioning off the internal carotid artery.9
As tortuosity increases, represented by progressing carotid siphon types and ophthalmic artery positions, fluoroscopy times and radiation doses also increase. Carotid siphon types 1, 3, and 5 had average fluoroscopy times of 6.42 min, 9.79 min, and 12.6 min, respectively. Furthermore, ophthalmic artery positions 1, 2, 3, and 4 had average fluoroscopy times of 6.58, 11.5, 12.6, and 13.4 min, respectively. However, given our relatively small sample size of 26 patients, our results are not statistically significant.
In comparison with Cooke et al, we report lower average fluoroscopy times of 4.75 to 8.5 min. However, we demonstrate higher fluoroscopy doses of 23.3 to 20.1 mGy. Although intuitively, lower fluoroscopy times should result in lower fluoroscopy doses, we postulate that differences in the fluoroscopy settings of our respective institutions account for this disparity. For example, Cooke reports specific vendor collaboration to minimize radiation dosing and as a result, changing fluoroscopy settings from standard 7.5 pulses/s and 36 nGy/pulse to 4 pulses/s and 23 nGy/pulse.10 Furthermore, similar to our practice before late 2017, the Cooke study used only a single agent, melphalan.10
In comparison with Gobin et al, we demonstrate lower fluoroscopy times, fluoroscopy dose, and DAP. Although our techniques are similar, the routine use of continuous verapamil infusion is novel to our institution. However, our results include a relatively small sample size of middle meningeal artery infusion (11%) and no procedures using balloon occlusion of the internal carotid artery distal to the ophthalmic artery, which can be associated with longer fluoroscopy times and larger radiation doses. Furthermore, our percentage of dual agent infusions, 19%, is lower than that reported by Gobin who used multiagent infusions in 89% of procedures.12
Although theoretically, the use of dual agent therapy could increase fluoroscopy if the catheter position needed to be checked between switching agents, our results demonstrated a decreased in fluoroscopy time in dual agent infusions in comparison with single agent infusions. We believe that this trend is probably secondary to our method of dose escalation. If a patient was found to have no evidence of tumor regression following single agent treatment, subsequent treatments either increased the dose or used dual agent therapy. Since patient anatomy had been characterized during the initial procedure, subsequent procedures using dual agent therapy were performed with increased anatomic familiarity, reducing fluoroscopy utilization.
Limitations of our study include a single primary operator for technique A and a relatively small sample size of 26 patients and 94 treatments. Although having a single primary operator for technique A does potentially bias our results, the 56 years of experience of the three neurointerventionalists who performed technique B infusions should ensure the best possible results were obtained for that technique.
The number of independent neurointerventionalists performing intra-arterial chemotherapy at our institution for retinoblastoma over the past 5 years provided us with a unique opportunity to study changes in technique. This study showed that the microcatheter-only approach, when coupled with verapamil, can decrease the fluoroscopy time and dose associated with ophthalmic artery chemotherapy infusion. We also demonstrate that anatomical variations may alter the technical complexity and resultant fluoroscopy times and radiation dose. Further analysis will be performed to determine the effect of this technique and other variables on clinical outcomes, including preservation of retinal function, tumor regression, and time to recurrence.
References
Footnotes
Contributors Each of the listed authors directly contributed to the study design, acquisition or interpretation of data, manuscript revision or final approval of the work.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests CH reports personal fees from Merck, outside the submitted work. DG reports non-financial support and other from Children’s Oncology Group, during the conduct of the study; grants from Houseman/Wilkins Ophthalmological Foundation, personal fees, and non-financial support from Abbvie, other from 3T Ophthalmics, other from Aura Biosciences, outside the submitted work.
Ethics approval Institutional review board for Baylor College of Medicine and Affiliated Hospitals H-43720.
Provenance and peer review Not commissioned; externally peer reviewed.
Correction notice Since this paper was first published online, the corresponding author has been updated to Stephen Chen. The author Dan Gambos has been updated to Dan Gombos.
Patient consent for publication Not required.