Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization
ORIGINAL RESEARCH
Radiation dose reduction in intra-arterial chemotherapy infusion for intraocular retinoblastoma Daniel L Cooke, Charles E Stout, Warren T Kim, Steven W Hetts, Randall T Higashida, Van V Halbach, Christopher F Dowd, Robert G Gould Department of Radiology and Biomedical Imaging, University California San Francisco, San Francisco, California, USA Correspondence to Dr D L Cooke, Department of Radiology and Biomedical Imaging, University California San Francisco, 505 Parnassus Ave Long 352, Box 0628, San Francisco, CA 941430628, USA; [email protected] Presented electronically at the ASNR 2013 Annual Meeting. Received 29 July 2013 Revised 8 November 2013 Accepted 2 December 2013 Published Online First 2 January 2014
ABSTRACT Background and purpose Retinoblastoma (RB) is a rare malignancy affecting the pediatric population. Intravenous chemotherapy is the longstanding delivery method, although intra-arterial (IA) chemotherapy is gaining popularity given the reduced side effects compared with systemic chemotherapy administration. Given the sensitivity of the target organ, patient age, and secondary tumor susceptibility, a premium has been placed on minimizing procedural related radiation exposure. Materials and methods To reduce patient x-ray dose during the IA infusion procedure, customized surgical methods and fluoroscopic techniques were employed. The routine fluoroscopic settings were changed from the standard 7.5 pulses/s and dose level to the detector of 36 nGy/pulse, to a pulse rate of 4 pulses/s and detector dose to 23 nGy/pulse. The angiographic dose indicators (reference point air kerma (Ka) and fluoroscopy time) for a cohort of 10 consecutive patients (12 eyes, 30 infusions) were analyzed. An additional four cases (five eyes, five infusions) were analyzed using dosimeters placed at anatomic locations to reflect scalp, eye, and thyroid dose. Results The mean Ka per treated eye was 20.1 ±11.9 mGy with a mean fluoroscopic time of 8.5 ±4.6 min. Dosimetric measurements demonstrated minimal dose to the lens (0.18±0.10 mGy). Measured entrance skin doses varied from 0.7 to 7.0 mGy and were 73.4±19.7% less than the indicated Ka value. Conclusions Ophthalmic arterial melphalan infusion is a safe and effective means to treat RB. Modification to contemporary fluoroscopic systems combined with parsimonious fluoroscopy can minimize radiation exposure.
INTRODUCTION
To cite: Cooke DL, Stout CE, Kim WT, et al. J NeuroIntervent Surg 2014;6:785–789.
Intraocular retinoblastoma (RB) is a rare pediatric solid tumor with fewer than 300 cases diagnosed per year in the USA and less than 10 000 worldwide.1 In the developed world, 5 year survival rates exceed 95%, with evidence supporting the use the external beam radiation, systemic chemotherapy, and enucleation.1 2 Despite such therapies, additional strategies have been attempted aimed at not only preserving the affected eye(s), but also reducing the comorbidities related to treatment, such as gastrointestinal toxicity, myelosuppression, and secondary malignancy.
Intra-arterial (IA) infusion for RB has been used for over 20 years, although not until recently had it been performed in a superselective manner.3 Abramson et al described a technique using a flow directed microcatheter to select the ophthalmic artery (OA) for drug delivery.4 Subsequent series have since been published supporting the use of IA therapy documenting the technical reproducibility, minimal procedural or chemotherapeutic related morbidity, and good clinical response.5–8 The success of the technique has expectantly expanded its use and in turn brought needed attention to possible, although unrealized, pit falls. The historical use of external beam radiation for RB treatment provides a cautionary window into the potential complications of high dose ionizing radiation on the affected pediatric eye.9 The radiation exposure related to IA chemotherapy infusion is significantly lower than that of therapeutic external beam radiation, although care still must be taken to minimize dose. X-ray dose, as part of IA treatment, has certainly not been standardized and is an area of controversy.10 11 Given the mandate to reduce imaging related radiation exposure, particularly in the pediatric population,12–16 we present our experience using an interventional protocol customized for IA RB therapy.
METHODS Fluoroscopic protocol All procedures were performed using a biplane angiographic unit (Artis Zee, Siemens Inc, Malvern, Pennsylvania, USA). To minimize the dose, in collaboration with the vendor, the routine fluoroscopic settings were changed from the standard 7.5 pulses/s and a dose level to the detector of 36 nGy/pulse, to a pulse rate of 4 pulses/s and a detector dose to 23 nGy/pulse. The dose rate continued to be controlled by the automatic exposure control. The orientation of the C arm was with the x-ray tube positioned on the side of the eye being treated. If both eyes were treated in a single procedure, the C arm positioning was flipped to be positioned ipsilateral to the eye being treated during each half of the procedure. The detector was placed as close to the patient as possible so that the source to detector distance was minimized. The x-ray beam was collimated tightly to the region of interest, with care being taken to minimize irradiation of the lens of the eye. Fluoroscopy was performed using a single
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
785
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization imaging plane (either anterior-posterior or lateral tubes) in all cases, although in 67% of the treatments only the anteriorposterior tube was used for both anteroposterior and lateral imaging purposes. Laterally, the patient’s head was positioned at the isocenter, while vertically, centering was at the nasion.
Dosimeteric measurements In four patients, optically stimulated luminescence dosimeters (OSLD) were placed on the patient’s eyelids, thyroid, and bilateral temporal scalp within the x-ray beam entrance and exit path. Dosimeters were read using the calibration for 80 kVp and have an accuracy of ±5%.
Operative technique All patients were placed under general anesthesia and underwent standard sterile preparation and drape. A quarter inch of nitropaste was placed on the patient’s chest to reduce the risk of catheter induced vasospasm. A baseline activated clotting time was measured and a 70 mg/kg unit IV bolus of heparin administered after ultrasound guided 4 F femoral arterial sheath placement. Via the femoral sheath, a 4 F guide catheter was positioned under x-ray fluoroscopic guidance within the internal carotid artery supplying the affected eye (figure 1A). For those patients undergoing their initial infusion, a 1.5 F Magic flow directed catheter (Balt Extrusion Inc, Montmorency, France) was utilized. The microcatheter was primed with a 0.008 inch Mirage microwire (ev3 Inc, Irvine, California, USA). For those patients undergoing repeat IA treatment, initial microcatheter selection was based on success of the patient’s prior treatment. A single lateral projection guide catheter roadmap (figure 1B) was performed to demonstrate the OA origin. The
microcatheter was then advanced under fluoroscopic visualization into the OA. A single microcatheter angiogram (figure 1C) of the OA was performed in the lateral projection. If the angiogram demonstrated choroidal blush and no significant reflux into the internal carotid artery, melphalan infusion was performed. For those patients with evidence of vasospasm of the OA, an infusion of 1 mg of verapamil (0.5 mg/1.0 mL normal saline) was done prior to melphalan infusion. A roadmap blank image was performed at infusion start to help detect any change in catheter position during melphalan infusion. Melphalan dosing was predetermined by the pediatric oncology service with depots prepared by the pharmacy department and suspended in 30 mL of normal saline. Intermittent (approximately 2 s of fluoroscopy time per each minute of melphalan injection) fluoroscopy was performed to assess any movement of the microcatheter tip. After delivery of approximately half the therapeutic volume, an injection of 1 mL of Omnipaque 300 on a blank roadmap image was performed (figure 1D) to confirm anterograde filling of the OA. If flow was still anterograde in the OA, the remaining 15 mL of melphalan were delivered. For those patients in whom OA access was not successful using a 1.5 F Balt catheter, a Marathon microcatheter (ev3) was used instead, again with an accompanying Mirage microwire. When a patient had variant OA anatomy— for example, arising from the middle meningeal artery—the guide catheter was re-positioned within the ipsilateral external carotid artery. A guide catheter roadmap angiogram was performed. An Excelsior SL10 microcatheter (Boston Scientific Inc., Natick, Massachusetts, USA) was then primed with a Transcend EX Platinum microwire (Boston Scientific Inc). Under fluoroscopic guidance the microcatheter
Figure 1 Composite fluoroscopic images from intra-arterial melphalan infusion. Fluoroscopic screen capture (A), documenting catheterization (large arrowhead) of the right internal carotid artery. Roadmap (B) of the right internal carotid artery demonstrating the ophthalmic artery origin (arrow). Lateral projection DSA image (C) demonstrating anterograde flow in the ophthalmic artery with the catheter tip (small arrowhead) at the ophthalmic artery origin. Fluoroscopic screen capture (D) at the midway point of melphalan infusion re-demonstrating anterograde flow in the ophthalmic artery with the catheter tip (small arrowhead) at the ophthalmic artery origin. 786
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization system was advanced into the targeted external carotid artery branch. Via the microcatheter, an angiogram was performed to confirm OA supply and choroidal blush. The infusion protocol was then performed as described above.
Outcomes assessment All patients underwent routine clinical follow-up, as dictated by the treating ophthalmology and oncology teams. Patients had a complete blood count checked prior to treatment and then again approximately 10 days after each IA treatment to evaluate for myelosuppression. To determine disease response, an ocular oncologist performed an eye examination under anesthesia 3–4 weeks after IA treatment with RETCAM photos. Cases were reviewed by ophthalmologists to classify response to therapy.
Statistical analysis Demographic, dose, anatomic, and technical variables were logged within an Excel spreadsheet (Microsoft Inc, Redmond, Washington, USA). Mean and SD data were calculated within Excel.
RESULTS Ten consecutive patients (12 affected eyes, 30 infusions) were analyzed (table 1). The majority of patients were of an advanced stage (table 2) with a mean number of treatments of 2.8±0.9 per eye, with the majority (58.3%) demonstrating a successful response. Of the 12 treated eyes, all except one had conventional OA anatomy. ??The mean air kerma (Ka) per treated eye was 20.1 ±11.9 mGy, with a mean fluoroscopic time of 8.5±4.6 min. The majority of cases required a single microcatheter. Device related vasospasm was an uncommon (6.7%) occurrence. There were three cases with elevated x-ray doses relative to the mean dose (figure 2). These outliers represented three different scenarios: (1) unstable catheter position (treatment No 5) requiring usage of multiple microcatheter types; (2) external carotid supply (treatment No 10) to the OA; and (3) mixed use of standard and low dose settings (treatment Nos 17 and 18) to better visualize the arterial anatomy. Subsequent treatments (treatment Nos 15 and 22) in the patient with external carotid origin of the OA regressed toward the mean dose. In the additional four cases (five eyes) treated with the OSLD in place, readings for the lens of the eye (0.18±0.10 mGy) were minimal (table 3), confirming that the lens was seldom in the primary x-ray beam. The entrance skin dose varied from 0.7 to 7.0 mGy. In the patient in whom both eyes were treated, the C arm was rotated to move the x-ray tube to the treatment eye side and thus both
Table 1 Demographics (n=10) Characteristic Weight (kg) (mean±SD) Age (months) (mean±SD) Sex (M/F) Eye(s) OD OS OU OD, oculus dexter; OS, oculus sinister; OU, oculus uterque.
12.9±3.2 24.1±9.6 6/4 5 3 2
Table 2 Tumor staging, arterial anatomy, and number of treatments per eye (n=12) Parameter International classification of retinoblastoma A B C D Ophthalmic artery origin Conventional (ICA) External (ECA) No of treatments 1 2 3 4 5 Response to treatment
No (%)
1 2 1 8
(8.3) (16.7) (8.3) (66.7)
11 (91.7) 1 (8.3) 2 4 2 3 1 7
(16.7) (33.3) (16.7) (25) (8.3) (58.3)
ECA, external carotid artery; ICA, internal carotid artery.
sides of the scalp were x-ray beam entrance surfaces. The x-ray beam exit surface was more than a factor of 10 less than the entrance skin dose. The thyroid dose was negligible. In comparison with the Ka, the entrance skin doses were 73.4±19.7% less, reflecting that the distance of the scalp surface was less than 15 cm from the isocenter, causing Ka to overestimate the skin dose. In all cases, patient dose was dominated by the fluoroscopic imaging, which contributed 62.3±14.0% of the indicated Ka value, reflecting use of only a single angiographic run per treated eye.
DISCUSSION We have described a single institution’s experience with IA melphalan infusion for the treatment of ocular RB. As demonstrated here, and by others, IA chemotherapy has proven to be technically feasible with good response rates as well as relatively benign side effects profiles.1–4 Abramson et al prospectively studied 23 patients (28 eyes) over a 3 year interval that underwent chemotherapeutic infusions.2 The cohort had no deaths, major technical or systemic complications, and only a single enucleation for progressive disease with 89% enucleation free at 2 years. These early efforts are promising although in time we will have a greater understanding of potential IA infusion complications and the durability of clinical success. Dose reduction in imaging has and remains an important part of interventional practice, particularly for pediatric patients. The premium placed on such reduction is even greater in the RB population given the morbidity related to potential injury to the targeted organ, the age of the affected population, and the increased susceptibility to secondary tumor formation incumbent to the disease itself. From the 1950s to the 1980s, RB was often treated using external beam radiation, the delayed clinical outcomes of which revealed the unique susceptibility of these patients to high dose ionizing radiation.17 MacCarthy et al18 quantified delayed deleterious effects of regional radiation therapy, noting an increased (13-fold) rate of secondary tumor formation in RB patients, particularly those with a heritable etiology, relative to the general population, and significantly higher rate (32-fold) in tumors arising within the head and neck.
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
787
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization Figure 2 Scatterplot of treatment number relative to reference point air kerma (mGy) per treated eye. Treatment Nos 3 and 5 (red), 11, 12, 17, 18, 25, and 26 (green), and 1, 15, and 22 (yellow) represent three respective patients treated at different time points. Treatment Nos 11–12 and 17–18 were performed at the same time given the patient’s bilateral disease. The four highest doses (black box) represented three different scenarios: (1) unstable catheter position (treatment No 5) secondary to an acute ophthalmic artery origin requiring usage of multiple microcatheter types; (2) external carotid supply (treatment No 10) to the ophthalmic artery; and (3) mixed use of the standard and low dose settings (treatment Nos 17 and 18) to better visualize the arterial anatomy. To minimize risk of all types of radiation related injury, essential measures need to be taken to limit the radiation dose to the lowest level consistent with a successful procedural outcome. Trade-offs exist in setting imaging technique factors—for example, using too low a dose can reduce the quality of the image so that catheter guidance is difficult and procedure time can increase, with a subsequent greater radiation level to the patient. Beyond the techniques commonly used in interventional practice—tight and careful collimation, reduced source to image distance, single plane usage, and roadmaps in place of conventional subtraction angiograms—we altered the settings on our x-ray fluoroscopic system to reduce the dose per pulse during fluoroscopy by 35%. These changes in operative and fluoroscopic technique, made in collaborative effort with the equipment vendor, allowed our group to attain dose levels comparable with Gobin et al,10 a group at the forefront of IA RB treatment. To simplify the translation of our results to others, we have reported the reference point air kerma (Ka), as this value is routinely recorded by modern angiographic systems. The recording of fluoroscopic time and Ka values should be common practice.19–22 These values serve as surrogates for dose and predictors of deterministic radiation effects.19 23 In our patients in whom skin dose measurements were made, the dosimeters were placed at the points of greatest risk, specifically to examine the dose to the lens of the eye and the skin dose at the entrance and exit surfaces. The authors recognize the coarseness of this
measure in relation to actual skin dose, mainly because the skin surface of the pediatric patients was 8 cm or less from the isocenter of the unit, not 15 cm as used for the international reference point. The reference point air kerma will then overestimate the skin dose (table 3), although given little need for panning of the C arm, the Ka value does relate to the skin dose and a low value for Ka will reflect lower skin and lens doses to the patient. To put our data in perspective with procedures more commonly encountered in clinical practice, the skin dose for an adult chest radiograph is approximately 0.25 mGy while a non-contrast head CT is approximately 60 mGy.24 Given the levels measured in this study, no deterministic effects attributable to radiation should occur and stochastic effects should be minimal. While it is common with some diagnostic imaging procedures to express the radiation level in terms of an equivalent dose in units of millisieverts, it is a calculated value relating to stochastic risk and requires the absorbed dose to radiosensitive organs to be determined. The absorbed dose is then multiplied by a weighting factor reflecting the relative sensitivity of the organ for carcinogenesis to derive the effective dose. In fluoroscopic procedures, specifically angiography, individual organ doses are difficult to determine even when surface dosimeters are employed. Furthermore, for large organs, including the pediatric brain, the absorbed dose is not constant within the volume being directly irradiated, with a variation of >4 between the x-ray entrance and exit surfaces. In addition, when the beam is collimated tightly, only a fraction of an organ is exposed to
Table 3 Radiation dose measurements in four patients Measured dose (mGy)
Reference point value (mGy)
Patient (treated eye(s))
Lens (OS)
Lens (OD)
Scalp (L)
Scalp (R)
Thyroid
Fluoroscopy (min)
Fluoroscopy air kerma
Total air kerma
1 (OS) 2 (OU)*
0.16 0.29
0.16 0.46
1.68 1.57
0.12 7.02
0.06 0.16
6.2 16.3
5.7 17.9
3 (OD) 4 (OS)
0.11 0.07
0.18 0.02
0.14 0.72
2.35 0.03
0.06 0.02
26.4 4.6
4.2 L=10.6 R=1.7 23.4 4.3
32.5 12.3
*Tube orientation was flipped after treatment of the first eye. The right eye was more difficult to treat, resulting in an unequal dose distribution between the left and right scalp surfaces. Treatment of the left eye produced 78% of the total indicated air kerma dose. OD, oculus dexter; OS, oculus sinister; OU, oculus uterque.
788
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization significant levels of radiation. Another difficulty is that accurate absorbed dose calculations require knowledge of the x-ray beam characteristics, which with automatic exposure control varies with projection angle. Thus calculation of a patient specific effective dose is not possible. However, in interventional angiographic procedures, the potential for a large skin dose and subsequent deterministic effects exists as the x-ray entrance skin surface can receive much higher dose levels than those associated with other non-therapeutic and diagnostic forms of x-ray imaging. The display and recording of the air kerma at the international reference point, now required on all angiographic equipment, is related to the skin dose and reflects the concern for deterministic skin effects. The air kerma value is not usable for effective dose calculations, which accounts for stochastic and not deterministic radiation effects. If the lens of the eye is within or close to the primary x-ray beam, the risk of a future cataract can be elevated. Cataract induction is considered a deterministic effect and, as with other deterministic effects, requires some threshold dose level to elevate cataract risk, although some evidence exists that it should be treated as a stochastic risk, meaning that there is no such threshold.18 The International Commission on Radiological Protection recommends annual equivalent dose limits of 50 mSv (20 mSv averaged over a 5 year interval) and 15 mSv to the lens of the eye for occupational and the general population, respectively. They consider the threshold dose for detrimental tissue reactions (eg, cataract formation) to be 0.5 Gy and note no benefit from protracted exposure. In the pediatric cancer population, cataracts as well as other ocular conditions (dry eye, blindness, diplopia) have been noted to occur with doses
Tumor embolization
ORIGINAL RESEARCH
Radiation dose reduction in intra-arterial chemotherapy infusion for intraocular retinoblastoma Daniel L Cooke, Charles E Stout, Warren T Kim, Steven W Hetts, Randall T Higashida, Van V Halbach, Christopher F Dowd, Robert G Gould Department of Radiology and Biomedical Imaging, University California San Francisco, San Francisco, California, USA Correspondence to Dr D L Cooke, Department of Radiology and Biomedical Imaging, University California San Francisco, 505 Parnassus Ave Long 352, Box 0628, San Francisco, CA 941430628, USA; [email protected] Presented electronically at the ASNR 2013 Annual Meeting. Received 29 July 2013 Revised 8 November 2013 Accepted 2 December 2013 Published Online First 2 January 2014
ABSTRACT Background and purpose Retinoblastoma (RB) is a rare malignancy affecting the pediatric population. Intravenous chemotherapy is the longstanding delivery method, although intra-arterial (IA) chemotherapy is gaining popularity given the reduced side effects compared with systemic chemotherapy administration. Given the sensitivity of the target organ, patient age, and secondary tumor susceptibility, a premium has been placed on minimizing procedural related radiation exposure. Materials and methods To reduce patient x-ray dose during the IA infusion procedure, customized surgical methods and fluoroscopic techniques were employed. The routine fluoroscopic settings were changed from the standard 7.5 pulses/s and dose level to the detector of 36 nGy/pulse, to a pulse rate of 4 pulses/s and detector dose to 23 nGy/pulse. The angiographic dose indicators (reference point air kerma (Ka) and fluoroscopy time) for a cohort of 10 consecutive patients (12 eyes, 30 infusions) were analyzed. An additional four cases (five eyes, five infusions) were analyzed using dosimeters placed at anatomic locations to reflect scalp, eye, and thyroid dose. Results The mean Ka per treated eye was 20.1 ±11.9 mGy with a mean fluoroscopic time of 8.5 ±4.6 min. Dosimetric measurements demonstrated minimal dose to the lens (0.18±0.10 mGy). Measured entrance skin doses varied from 0.7 to 7.0 mGy and were 73.4±19.7% less than the indicated Ka value. Conclusions Ophthalmic arterial melphalan infusion is a safe and effective means to treat RB. Modification to contemporary fluoroscopic systems combined with parsimonious fluoroscopy can minimize radiation exposure.
INTRODUCTION
To cite: Cooke DL, Stout CE, Kim WT, et al. J NeuroIntervent Surg 2014;6:785–789.
Intraocular retinoblastoma (RB) is a rare pediatric solid tumor with fewer than 300 cases diagnosed per year in the USA and less than 10 000 worldwide.1 In the developed world, 5 year survival rates exceed 95%, with evidence supporting the use the external beam radiation, systemic chemotherapy, and enucleation.1 2 Despite such therapies, additional strategies have been attempted aimed at not only preserving the affected eye(s), but also reducing the comorbidities related to treatment, such as gastrointestinal toxicity, myelosuppression, and secondary malignancy.
Intra-arterial (IA) infusion for RB has been used for over 20 years, although not until recently had it been performed in a superselective manner.3 Abramson et al described a technique using a flow directed microcatheter to select the ophthalmic artery (OA) for drug delivery.4 Subsequent series have since been published supporting the use of IA therapy documenting the technical reproducibility, minimal procedural or chemotherapeutic related morbidity, and good clinical response.5–8 The success of the technique has expectantly expanded its use and in turn brought needed attention to possible, although unrealized, pit falls. The historical use of external beam radiation for RB treatment provides a cautionary window into the potential complications of high dose ionizing radiation on the affected pediatric eye.9 The radiation exposure related to IA chemotherapy infusion is significantly lower than that of therapeutic external beam radiation, although care still must be taken to minimize dose. X-ray dose, as part of IA treatment, has certainly not been standardized and is an area of controversy.10 11 Given the mandate to reduce imaging related radiation exposure, particularly in the pediatric population,12–16 we present our experience using an interventional protocol customized for IA RB therapy.
METHODS Fluoroscopic protocol All procedures were performed using a biplane angiographic unit (Artis Zee, Siemens Inc, Malvern, Pennsylvania, USA). To minimize the dose, in collaboration with the vendor, the routine fluoroscopic settings were changed from the standard 7.5 pulses/s and a dose level to the detector of 36 nGy/pulse, to a pulse rate of 4 pulses/s and a detector dose to 23 nGy/pulse. The dose rate continued to be controlled by the automatic exposure control. The orientation of the C arm was with the x-ray tube positioned on the side of the eye being treated. If both eyes were treated in a single procedure, the C arm positioning was flipped to be positioned ipsilateral to the eye being treated during each half of the procedure. The detector was placed as close to the patient as possible so that the source to detector distance was minimized. The x-ray beam was collimated tightly to the region of interest, with care being taken to minimize irradiation of the lens of the eye. Fluoroscopy was performed using a single
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
785
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization imaging plane (either anterior-posterior or lateral tubes) in all cases, although in 67% of the treatments only the anteriorposterior tube was used for both anteroposterior and lateral imaging purposes. Laterally, the patient’s head was positioned at the isocenter, while vertically, centering was at the nasion.
Dosimeteric measurements In four patients, optically stimulated luminescence dosimeters (OSLD) were placed on the patient’s eyelids, thyroid, and bilateral temporal scalp within the x-ray beam entrance and exit path. Dosimeters were read using the calibration for 80 kVp and have an accuracy of ±5%.
Operative technique All patients were placed under general anesthesia and underwent standard sterile preparation and drape. A quarter inch of nitropaste was placed on the patient’s chest to reduce the risk of catheter induced vasospasm. A baseline activated clotting time was measured and a 70 mg/kg unit IV bolus of heparin administered after ultrasound guided 4 F femoral arterial sheath placement. Via the femoral sheath, a 4 F guide catheter was positioned under x-ray fluoroscopic guidance within the internal carotid artery supplying the affected eye (figure 1A). For those patients undergoing their initial infusion, a 1.5 F Magic flow directed catheter (Balt Extrusion Inc, Montmorency, France) was utilized. The microcatheter was primed with a 0.008 inch Mirage microwire (ev3 Inc, Irvine, California, USA). For those patients undergoing repeat IA treatment, initial microcatheter selection was based on success of the patient’s prior treatment. A single lateral projection guide catheter roadmap (figure 1B) was performed to demonstrate the OA origin. The
microcatheter was then advanced under fluoroscopic visualization into the OA. A single microcatheter angiogram (figure 1C) of the OA was performed in the lateral projection. If the angiogram demonstrated choroidal blush and no significant reflux into the internal carotid artery, melphalan infusion was performed. For those patients with evidence of vasospasm of the OA, an infusion of 1 mg of verapamil (0.5 mg/1.0 mL normal saline) was done prior to melphalan infusion. A roadmap blank image was performed at infusion start to help detect any change in catheter position during melphalan infusion. Melphalan dosing was predetermined by the pediatric oncology service with depots prepared by the pharmacy department and suspended in 30 mL of normal saline. Intermittent (approximately 2 s of fluoroscopy time per each minute of melphalan injection) fluoroscopy was performed to assess any movement of the microcatheter tip. After delivery of approximately half the therapeutic volume, an injection of 1 mL of Omnipaque 300 on a blank roadmap image was performed (figure 1D) to confirm anterograde filling of the OA. If flow was still anterograde in the OA, the remaining 15 mL of melphalan were delivered. For those patients in whom OA access was not successful using a 1.5 F Balt catheter, a Marathon microcatheter (ev3) was used instead, again with an accompanying Mirage microwire. When a patient had variant OA anatomy— for example, arising from the middle meningeal artery—the guide catheter was re-positioned within the ipsilateral external carotid artery. A guide catheter roadmap angiogram was performed. An Excelsior SL10 microcatheter (Boston Scientific Inc., Natick, Massachusetts, USA) was then primed with a Transcend EX Platinum microwire (Boston Scientific Inc). Under fluoroscopic guidance the microcatheter
Figure 1 Composite fluoroscopic images from intra-arterial melphalan infusion. Fluoroscopic screen capture (A), documenting catheterization (large arrowhead) of the right internal carotid artery. Roadmap (B) of the right internal carotid artery demonstrating the ophthalmic artery origin (arrow). Lateral projection DSA image (C) demonstrating anterograde flow in the ophthalmic artery with the catheter tip (small arrowhead) at the ophthalmic artery origin. Fluoroscopic screen capture (D) at the midway point of melphalan infusion re-demonstrating anterograde flow in the ophthalmic artery with the catheter tip (small arrowhead) at the ophthalmic artery origin. 786
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization system was advanced into the targeted external carotid artery branch. Via the microcatheter, an angiogram was performed to confirm OA supply and choroidal blush. The infusion protocol was then performed as described above.
Outcomes assessment All patients underwent routine clinical follow-up, as dictated by the treating ophthalmology and oncology teams. Patients had a complete blood count checked prior to treatment and then again approximately 10 days after each IA treatment to evaluate for myelosuppression. To determine disease response, an ocular oncologist performed an eye examination under anesthesia 3–4 weeks after IA treatment with RETCAM photos. Cases were reviewed by ophthalmologists to classify response to therapy.
Statistical analysis Demographic, dose, anatomic, and technical variables were logged within an Excel spreadsheet (Microsoft Inc, Redmond, Washington, USA). Mean and SD data were calculated within Excel.
RESULTS Ten consecutive patients (12 affected eyes, 30 infusions) were analyzed (table 1). The majority of patients were of an advanced stage (table 2) with a mean number of treatments of 2.8±0.9 per eye, with the majority (58.3%) demonstrating a successful response. Of the 12 treated eyes, all except one had conventional OA anatomy. ??The mean air kerma (Ka) per treated eye was 20.1 ±11.9 mGy, with a mean fluoroscopic time of 8.5±4.6 min. The majority of cases required a single microcatheter. Device related vasospasm was an uncommon (6.7%) occurrence. There were three cases with elevated x-ray doses relative to the mean dose (figure 2). These outliers represented three different scenarios: (1) unstable catheter position (treatment No 5) requiring usage of multiple microcatheter types; (2) external carotid supply (treatment No 10) to the OA; and (3) mixed use of standard and low dose settings (treatment Nos 17 and 18) to better visualize the arterial anatomy. Subsequent treatments (treatment Nos 15 and 22) in the patient with external carotid origin of the OA regressed toward the mean dose. In the additional four cases (five eyes) treated with the OSLD in place, readings for the lens of the eye (0.18±0.10 mGy) were minimal (table 3), confirming that the lens was seldom in the primary x-ray beam. The entrance skin dose varied from 0.7 to 7.0 mGy. In the patient in whom both eyes were treated, the C arm was rotated to move the x-ray tube to the treatment eye side and thus both
Table 1 Demographics (n=10) Characteristic Weight (kg) (mean±SD) Age (months) (mean±SD) Sex (M/F) Eye(s) OD OS OU OD, oculus dexter; OS, oculus sinister; OU, oculus uterque.
12.9±3.2 24.1±9.6 6/4 5 3 2
Table 2 Tumor staging, arterial anatomy, and number of treatments per eye (n=12) Parameter International classification of retinoblastoma A B C D Ophthalmic artery origin Conventional (ICA) External (ECA) No of treatments 1 2 3 4 5 Response to treatment
No (%)
1 2 1 8
(8.3) (16.7) (8.3) (66.7)
11 (91.7) 1 (8.3) 2 4 2 3 1 7
(16.7) (33.3) (16.7) (25) (8.3) (58.3)
ECA, external carotid artery; ICA, internal carotid artery.
sides of the scalp were x-ray beam entrance surfaces. The x-ray beam exit surface was more than a factor of 10 less than the entrance skin dose. The thyroid dose was negligible. In comparison with the Ka, the entrance skin doses were 73.4±19.7% less, reflecting that the distance of the scalp surface was less than 15 cm from the isocenter, causing Ka to overestimate the skin dose. In all cases, patient dose was dominated by the fluoroscopic imaging, which contributed 62.3±14.0% of the indicated Ka value, reflecting use of only a single angiographic run per treated eye.
DISCUSSION We have described a single institution’s experience with IA melphalan infusion for the treatment of ocular RB. As demonstrated here, and by others, IA chemotherapy has proven to be technically feasible with good response rates as well as relatively benign side effects profiles.1–4 Abramson et al prospectively studied 23 patients (28 eyes) over a 3 year interval that underwent chemotherapeutic infusions.2 The cohort had no deaths, major technical or systemic complications, and only a single enucleation for progressive disease with 89% enucleation free at 2 years. These early efforts are promising although in time we will have a greater understanding of potential IA infusion complications and the durability of clinical success. Dose reduction in imaging has and remains an important part of interventional practice, particularly for pediatric patients. The premium placed on such reduction is even greater in the RB population given the morbidity related to potential injury to the targeted organ, the age of the affected population, and the increased susceptibility to secondary tumor formation incumbent to the disease itself. From the 1950s to the 1980s, RB was often treated using external beam radiation, the delayed clinical outcomes of which revealed the unique susceptibility of these patients to high dose ionizing radiation.17 MacCarthy et al18 quantified delayed deleterious effects of regional radiation therapy, noting an increased (13-fold) rate of secondary tumor formation in RB patients, particularly those with a heritable etiology, relative to the general population, and significantly higher rate (32-fold) in tumors arising within the head and neck.
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
787
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization Figure 2 Scatterplot of treatment number relative to reference point air kerma (mGy) per treated eye. Treatment Nos 3 and 5 (red), 11, 12, 17, 18, 25, and 26 (green), and 1, 15, and 22 (yellow) represent three respective patients treated at different time points. Treatment Nos 11–12 and 17–18 were performed at the same time given the patient’s bilateral disease. The four highest doses (black box) represented three different scenarios: (1) unstable catheter position (treatment No 5) secondary to an acute ophthalmic artery origin requiring usage of multiple microcatheter types; (2) external carotid supply (treatment No 10) to the ophthalmic artery; and (3) mixed use of the standard and low dose settings (treatment Nos 17 and 18) to better visualize the arterial anatomy. To minimize risk of all types of radiation related injury, essential measures need to be taken to limit the radiation dose to the lowest level consistent with a successful procedural outcome. Trade-offs exist in setting imaging technique factors—for example, using too low a dose can reduce the quality of the image so that catheter guidance is difficult and procedure time can increase, with a subsequent greater radiation level to the patient. Beyond the techniques commonly used in interventional practice—tight and careful collimation, reduced source to image distance, single plane usage, and roadmaps in place of conventional subtraction angiograms—we altered the settings on our x-ray fluoroscopic system to reduce the dose per pulse during fluoroscopy by 35%. These changes in operative and fluoroscopic technique, made in collaborative effort with the equipment vendor, allowed our group to attain dose levels comparable with Gobin et al,10 a group at the forefront of IA RB treatment. To simplify the translation of our results to others, we have reported the reference point air kerma (Ka), as this value is routinely recorded by modern angiographic systems. The recording of fluoroscopic time and Ka values should be common practice.19–22 These values serve as surrogates for dose and predictors of deterministic radiation effects.19 23 In our patients in whom skin dose measurements were made, the dosimeters were placed at the points of greatest risk, specifically to examine the dose to the lens of the eye and the skin dose at the entrance and exit surfaces. The authors recognize the coarseness of this
measure in relation to actual skin dose, mainly because the skin surface of the pediatric patients was 8 cm or less from the isocenter of the unit, not 15 cm as used for the international reference point. The reference point air kerma will then overestimate the skin dose (table 3), although given little need for panning of the C arm, the Ka value does relate to the skin dose and a low value for Ka will reflect lower skin and lens doses to the patient. To put our data in perspective with procedures more commonly encountered in clinical practice, the skin dose for an adult chest radiograph is approximately 0.25 mGy while a non-contrast head CT is approximately 60 mGy.24 Given the levels measured in this study, no deterministic effects attributable to radiation should occur and stochastic effects should be minimal. While it is common with some diagnostic imaging procedures to express the radiation level in terms of an equivalent dose in units of millisieverts, it is a calculated value relating to stochastic risk and requires the absorbed dose to radiosensitive organs to be determined. The absorbed dose is then multiplied by a weighting factor reflecting the relative sensitivity of the organ for carcinogenesis to derive the effective dose. In fluoroscopic procedures, specifically angiography, individual organ doses are difficult to determine even when surface dosimeters are employed. Furthermore, for large organs, including the pediatric brain, the absorbed dose is not constant within the volume being directly irradiated, with a variation of >4 between the x-ray entrance and exit surfaces. In addition, when the beam is collimated tightly, only a fraction of an organ is exposed to
Table 3 Radiation dose measurements in four patients Measured dose (mGy)
Reference point value (mGy)
Patient (treated eye(s))
Lens (OS)
Lens (OD)
Scalp (L)
Scalp (R)
Thyroid
Fluoroscopy (min)
Fluoroscopy air kerma
Total air kerma
1 (OS) 2 (OU)*
0.16 0.29
0.16 0.46
1.68 1.57
0.12 7.02
0.06 0.16
6.2 16.3
5.7 17.9
3 (OD) 4 (OS)
0.11 0.07
0.18 0.02
0.14 0.72
2.35 0.03
0.06 0.02
26.4 4.6
4.2 L=10.6 R=1.7 23.4 4.3
32.5 12.3
*Tube orientation was flipped after treatment of the first eye. The right eye was more difficult to treat, resulting in an unequal dose distribution between the left and right scalp surfaces. Treatment of the left eye produced 78% of the total indicated air kerma dose. OD, oculus dexter; OS, oculus sinister; OU, oculus uterque.
788
Cooke DL, et al. J NeuroIntervent Surg 2014;6:785–789. doi:10.1136/neurintsurg-2013-010905
Downloaded from http://jnis.bmj.com/ on May 24, 2015 - Published by group.bmj.com
Tumor embolization significant levels of radiation. Another difficulty is that accurate absorbed dose calculations require knowledge of the x-ray beam characteristics, which with automatic exposure control varies with projection angle. Thus calculation of a patient specific effective dose is not possible. However, in interventional angiographic procedures, the potential for a large skin dose and subsequent deterministic effects exists as the x-ray entrance skin surface can receive much higher dose levels than those associated with other non-therapeutic and diagnostic forms of x-ray imaging. The display and recording of the air kerma at the international reference point, now required on all angiographic equipment, is related to the skin dose and reflects the concern for deterministic skin effects. The air kerma value is not usable for effective dose calculations, which accounts for stochastic and not deterministic radiation effects. If the lens of the eye is within or close to the primary x-ray beam, the risk of a future cataract can be elevated. Cataract induction is considered a deterministic effect and, as with other deterministic effects, requires some threshold dose level to elevate cataract risk, although some evidence exists that it should be treated as a stochastic risk, meaning that there is no such threshold.18 The International Commission on Radiological Protection recommends annual equivalent dose limits of 50 mSv (20 mSv averaged over a 5 year interval) and 15 mSv to the lens of the eye for occupational and the general population, respectively. They consider the threshold dose for detrimental tissue reactions (eg, cataract formation) to be 0.5 Gy and note no benefit from protracted exposure. In the pediatric cancer population, cataracts as well as other ocular conditions (dry eye, blindness, diplopia) have been noted to occur with doses
