CLINICAL STUDY
Quantification of Activity Lost to DeliverySystem Residual and Decay in Yttrium-90
Radioembolization
Nam S. Hoang, Mohamed H. Khalaf, MD, Jarrett K. Rosenberg, PhD,
John Kwofie, MS, Aaron L. Reposar, MD, David S. Wang, MD,
John D. Louie, MD, and Daniel Y. Sze, MD, PhD
ABSTRACT
Purpose: To measure the decay activity loss and delivery system residual activity loss of yttrium-90 (Y90) radioembolization treatments across resin and glass microsphere activities.
Materials and Methods: For Y90 administrations between December 2009 and June 2017 at the study institution, the prescribed
activity, prepared activity, and delivered activity were recorded. Six hundred sixty-two administrations were reviewed—345 glass
(0.21–8.52 GBq) and 317 resin (0.18–3.28 GBq). Twenty-five patients (all resin) were excluded for arterial stasis or catheter
clogging. The percentage and actual losses of activity lost to decay and to delivery system residual were calculated for glass and
resin microspheres.
Results: The median time between activity premeasurement and administration was 2.20 hours, resulting in a median activity lost to
decay of 0.030 GBq or 2.35%, with no significant difference observed between glass and resin despite differences in preparation (P ¼
.0697). Resin showed significantly higher activity lost to delivery system residual than glass (0.039 GBq vs 0.010 GBq, 3.01% vs
0.61%, P < .001). The percent activity lost to residual varied with activity prepared, with a maximum of 20.1% and 16.2% for the
smallest activities of resin and glass, respectively.
Conclusions: Residual activity loss differs between glass and resin microspheres. For resin microspheres in particular, percent residual
activity loss increases with lower prepared activities. Protocols for activity calculation and preparation, patient dosimetry, and regulatory
compliance must take these losses into consideration prospectively.
ABBREVIATION
TARE ¼ transarterial radioembolization, yttrium-90 ¼ Y90
The rate of nonresponse to transarterial radioembolization
(TARE) has been reported to be as high as 80% in some
circumstances, while radiation-induced liver disease, gastrointestinal ulceration, pneumonitis, and other radiation-related
adverse events may further temper clinical success (1–10).
Subtherapeutic absorbed dose may result in suboptimal outcomes, and TARE dosimetry is complicated in part because
of the inexact quantification of the activity of yttrium-90
(Y90) delivered, which is rarely the exact amount of prescribed, prepared, or premeasured activity. There is no current
real-time method for quantifying and mapping the delivered
activity of Y90, and absorbed dose is based on preprescribed
activity and not on angiographic endpoints, as with chemoembolization (1,5,11). There is always loss of activity to residual microspheres retained in the delivery system and to
decay occurring between activity premeasurement or
From the Division of Interventional Radiology (N.S.H., M.H.K., A.L.R., D.S.W.,
J.D.L., D.Y.S.), Health Physics Group (J.K.), and Department of Radiology
(J.K.R.), Stanford University School of Medicine, H3646 Stanford University
Medical Center, 300 Pasteur Drive, Stanford, CA 94305. Received February
22, 2018; final revision received July 11, 2018; accepted July 15,
2018. Address correspondence to D.Y.S.; E-mail: dansze@stanford.edu
Gore (Flagstaff, Arizona), Janssen (Raritan, New Jersey), Koli Medical (Fremont, California), RadiAction Medical (Tel Aviv, Israel), Terumo (Tokyo, Japan),
and Viralytics (Sydney, Australia), and is a shareholder in Confluent Medical
(Fremont, California) and Proteus Digital Health (Redwood City, California).
None of the other authors have identified a conflict of interest.
D.Y.S. is a paid consultant for Amgen (Thousand Oaks, California), AstraZeneca (Cambridge, United Kingdom), Boston Scientific (Marlborough, Massachusetts), Bristol-Myers Squibb (New York City, New York), BTG (London,
United Kingdom), Eisai (Tokyo, Japan), EmbolX (Sunnyvale, California), W. L.
J Vasc Interv Radiol 2018; 29:1672–1677
© SIR, 2018
https://doi.org/10.1016/j.jvir.2018.07.011
Volume 29 ▪ Number 12 ▪ December ▪ 2018
preparation and actual delivery, and inaccurate accounting for
these losses could potentially affect outcomes (12). The aim of
this study was to measure the residual activity loss and decay
activity loss in glass and resin microspheres in patients undergoing TARE to aid in prospective correction of activity
calculations.
1673
Table. Interquartile Range for Activity, and Activity Loss for
Glass and Resin Microspheres with Exact Wilcoxon RankSum Test P-Values
Number of Procedures
Glass
Resin
345
318
Median Decay Time (Hours)
MATERIALS AND METHODS
Patient records were retrospectively reviewed for all TARE
procedures performed at the study institution between
December 2009 and June 2017. This study was approved by
the institutional review board, and a waiver for informed
consent was obtained. All data were handled in accordance
with the Health Insurance Portability and Accountability Act.
Six hundred eighty-seven patients received TARE treatment using either resin (SIR-Spheres; Sirtex, Lane Cove,
Australia) or glass (Therasphere; Nordion/BTG, Ottawa,
Canada) microspheres, and cases where arterial stasis or
catheter clogging occurred were censored from analysis.
Each procedure using resin microspheres required withdrawal of the prescribed activity from the unit dose vial by a
radiopharmaceutical technologist on the morning of the
administration. The activity was calculated by body surface
area method and prepared and titrated to within ±10% of the
prescribed activity. Likewise, each procedure using glass
microspheres involved activity calculation by the Medical
Internal Radiation Dose method for a target dose of 80–150
Gy or more than 200 Gy for radiation segmentectomy cases,
and measurement of activity on the morning of administration,
confirmed to be within ±10% of the prescribed activity. For
glass microspheres, per manufacturer recommendations, a 2%
excess was prescribed for each treatment in anticipation of
loss to administration apparatus residual activity. The microspheres with administration apparatuses were brought to the
procedure room at the commencement of the procedure and
administered after digital subtraction angiography and conebeam computed tomography confirmed correct catheter location, complete perfusion of the targeted territory, and lack of
extrahepatic distribution.
Sterile water was used as flush for resin microspheres until
2012, when it was replaced by 5% dextrose solution.
Repeated flushing and agitation of lines and stopcocks were
performed until clearance of all visible microspheres, at least 4
and up to 6 flushes of 20-ml volume. Normal saline was used
as flush for glass microspheres, with at least 3 20-ml flushes,
until portable electronic dosimeters (Rados RAD-60R; Eckert
& Ziegler, Berlin, Germany) mounted on the administration
apparatus read 0.0 mrem/min. For both glass and resin microspheres, 130-cm- or 135-cm-length 0.027” lumen microcatheters were used for administration (Renegade HI-FLO;
Boston Scientific, Marlborough, Massachusetts, or Progreat
Omega; Terumo, Somerset, New Jersey).
Prepared activity was measured by arithmetical mean of 4
different cylinder rotations of the 2-L Nalgene jar containing
the microsphere V-vial at every 90 degrees within a beta shield
on a standard template with source-to-detector distance of
.69
Minimum (0th)
0.22
0.13
10th percentile
0.88
1.16
Median (50th)
2.29
2.20
90th percentile
5.40
4.91
8.50
10.46
Minimum (0th)
0.21
0.18
10th percentile
0.58
0.44
Median (50th)
1.79
1.10
90th percentile
3.88
1.98
Maximum (100th)
8.51
3.28
Minimum (0th)
10th percentile
0.20
0.58
0.22
0.48
Median (50th)
1.79
1.16
90th percentile
3.76
2.10
Maximum (100th)
8.60
3.47
Minimum (0th)
0.20
0.14
10th percentile
0.56
0.43
Median (50th)
90th percentile
1.74
3.68
1.08
1.99
Maximum (100th)
8.41
3.36
Minimum (0th)
0.24
0.14
10th percentile
0.95
1.24
Median (50th)
2.45
2.35
90th percentile
5.66
5.15
8.77
10.68
Minimum (0th)
0.00
0.01
10th percentile
0.10
1.14
Median (50th)
0.61
3.01
90th percentile
2.91
9.91
16.18
20.10
Maximum (100th)
Prescribed Activity (GBq)
<. 001
< .001
Prepared Activity (GBq)
< .001
Administered Activity (GBq)
< .001
Activity Lost to Decay (%)
Maximum (100th)
Activity Lost to Residual (%)
Maximum (100th)
P-value
< .001
30 cm and subtracting the background activity in the absence
of the jar. The activity lost to decay was calculated using the
time elapsed between when the dose activity was measured for
preparation and the time at which administration occurred
(decay time in hours). With a half-life of 64.2 hours for Y90, the
formula for percentage of prepared activity lost to decay was:
Decay Activity Loss ¼
1 e ^ ðð0:6931 ðdecay time in hoursÞ ⁄ 64:2Þ
The decay-corrected prepared activity available at the time
of administration was thus ¼ (prepared activity) – (decay
activity loss).
1674 ▪ Activity Lost in Y90 Radioembolization
Hoang et al ▪ JVIR
Figure 1. The percentage of Y90 activity lost at a given prepared activity due to decay is shown for glass and resin microspheres.
Percent loss is dependent solely on elapsed time between activity premeasurement or preparation and time of administration and is
independent of the microsphere carrier and the prepared activity. The difference in X-axis scaling reflects the differences in the range of
the prescribed activities for glass and resin microspheres.
After administration, the empty V-vial, administration
tubing, microcatheter, and all potentially contaminated
towels, gauze, hemostats, and gloves were inserted into
the Nalgene jar. The average residual activity in the jar and
background activity were measured again to obtain the
post-administration residual activity. The time required to
transport the waste jar to the radioisotope laboratory for
measurement was estimated to be 20 minutes, so the
measured residual activity was corrected for 20 minutes’
worth of decay. The actual administered activity was
calculated by subtracting the decay-corrected measured
residual activity from the decay-corrected prepared
activity.
Administered activity ¼ (decay-corrected prepared
activity – decay-corrected residual activity).
Statistical Analysis
The database was de-identified and analysis was performed
using Stata Release 15 software (StataCorp LP, College
Station, Texas). A nonparametric 2-sample Wilcoxon ranksum test was used to compare activity levels and decay
time; a P-value threshold of .05 was used to indicate a
significant difference.
RESULTS
Six hundred sixty-two TARE administrations were
reviewed—345 glass (prescription range: 0.21–8.52 GBq)
and 317 resin (range: 0.18–3.28 GBq). Twenty-five patients
(all resin, 7.9% of resin patients) were excluded from the
analysis due to stasis (n ¼ 20) or clogging of the delivery
system catheter (n ¼ 5). The median time between prepared
activity measurement and administration was 2.20 hours
(range: 0.13–10.46), resulting in a median activity lost to
decay of 0.030 GBq (range: 0.002–0.489) or 2.35% (range:
0.14%–10.68%), with no significant difference observed
between glass and resin despite differences in preparation
(P ¼ .070). In select resin cases where dose preparation and
administration were expected to be separated by a long interval, extra activity was drawn in anticipation. Likewise, in
select glass cases, dose measurement took place early in the
morning for an afternoon administration, resulting in a long
interval. Table outlines the minimum, median, and maximum
for the prescribed activity, prepared or premeasured activity,
and administered activity, as well as the activity lost to
delivery system residual and to decay.
Percentage loss to decay depended only on elapsed time
between measurement and administration, according to the
physical decay constant of Y90 (Fig 1). For activity lost to
delivery system residual, resin showed higher median activity
loss than glass, both for raw activity and for percentage
(0.039 GBq vs 0.010 GBq, 3.01% vs 0.61%, P < .001). The
percent activity lost to delivery system residual varied with
activity prepared, with a maximum of 20.1% for the smallest
activities of resin and a maximum of 16.2% for the smallest
activities of glass. Figure 2a demonstrates how lower
prepared activities of microspheres had a higher percentage of
residual activity loss. For resin microspheres, the median
Volume 29 ▪ Number 12 ▪ December ▪ 2018
1675
Figure 2. The activity lost to the delivery system residual for a given prepared activity for glass and resin microspheres. (a) Percent loss
is dependent on both the microsphere carrier and the amount of prepared activity in GBq. The percent activity loss is lower for glass
microspheres than for resin microspheres at the smallest activities. (b) The actual activity lost in GBq also varies, despite constant
delivery apparatus, microcatheters, technique, and constant specific activity per microsphere for resin.
activity loss for administrations below 1.08 GBq (which was
the median activity administered for resin) was 5.32%.
Furthermore, the variability in percent activity loss
increased as prepared activities decreased. Despite the
constant of delivery system tubing and microcatheter and
constant specific activity of resin microspheres, the raw
residual activity lost to the delivery system was not constant.
Higher activities involving higher numbers of microspheres
resulted in higher residual loss than lower activities involving
microspheres of the same specific activity.
1676 ▪ Activity Lost in Y90 Radioembolization
Of note, select resin administrations of prescribed activities more than the standard calibrated unit dose vial activity
of 3.0 GBq were made possible by shipment and usage of
the microspheres a day before official calibration (“daybefore dose”).
DISCUSSION
Dosimetry calculations for TARE have evolved in the past
few decades, from crude empiric methods to emerging
methods based on voxel-based segmentation and radiobiological concepts (13,14). However, activity prescription is
complicated by the inherent variability in activity loss to
periprocedural decay and to residual microspheres within
the delivery system and microcatheters. Tolerance of normal
liver parenchyma to radiation exposure varies by patient and
is impossible to measure prospectively. Prior data from
external-beam radiotherapy suggest a tolerance of about 30
Gy to the noncirrhotic whole liver, but tolerances to internal
point sources as with TARE appear to be much higher due to
the heterogeneity of dose distribution. The absorbed dose
necessary to eradicate solid tumor is also highly variable and
impossible to measure prospectively, but it is estimated to be
at least 60–80 Gy for conventionally fractionated externalbeam radiotherapy and in the 100–200 Gy range for resin
and glass microsphere TARE for most tumor cell types
(5,15–17). Increased accuracy of each dosimetric step of the
TARE treatment process, including prediction of activity
lost to decay and to delivery system residual, should allow
improved analysis and prediction of efficacy and adverse
events, including radioembolization-induced liver disease,
which may occur in up to 4% of treated patients (1,3,4).
Because of the different specific activities and unit doses
of each of the microspheres and the different dose preparation techniques, recommendations also differ for correction for anticipated loss to residual microspheres. In our
cohort, the number of microspheres delivered varied from
about 1 to 16 million for glass and from 3 to 80 million for
resin, and the differences in specific density and surface
chemistry between glass and resin may also affect the proportion adhered or trapped in the delivery system even after
thorough flushing. Although the glass microspheres package
insert does not include recommendations, the “interactive
Dose Ordering Calculator” (BTG International, London,
United Kingdom) stipulates correction of an anticipated
percentage, up to 5%. Prescription of activity and timing of
treatment are then planned reflecting this anticipated percentage. No official correction recommendation is included
in the resin microsphere package insert or in the online
“SIR-Spheres Microspheres Activity Calculator.”
Meticulous flushing and agitation resulted in delivery
efficiencies averaging 97% for resin and 99% for glass.
Only 25 patients (7.9% of resin patients) were censored for
catheter clogging or stasis, which has been previously reported in as many as 28% of patients receiving resin microspheres (18). In current practice, many centers
preemptively estimate empirically a 2% loss to residual
Hoang et al ▪ JVIR
microspheres for both glass and resin, which was shown in
this study to be adequately accurate for high activities such
as lobar administrations. However, for smallest activities,
such as for increasingly popular segmental treatments, 2% is
usually an underestimate and can lead to underdosage,
particularly with resin microspheres, and the percentage lost
to residual becomes increasingly variable for smaller activities. In addition, despite the constant surface chemistry,
constant specific density, constant specific activity, and
relatively constant delivery technique of resin microspheres,
a constant absolute amount of residual microspheres was not
found. In contrast, loss to decay was verified to be independent of microsphere type and can be calculated solely
from the elapsed time between initial measurement and
administration. By pinpointing the actual delivered activity,
physicians can decrease the margin of error for the prescribed tumor and liver doses and thus work toward
improvement of outcomes of TARE.
This study was limited by small variabilities in the times
and durations of administration and elapsed times between
conclusion of administration and measurement of delivery
system residual, which was estimated to be 20 minutes.
Measurement of beta radiation is also subject to errors, for
instance from different attenuation and scatter from heterogeneous shielding from the walls of the V-vials, microcatheters, other waste material, and Nalgene jar walls.
Operator dependence may be a potential confounder, since 3
different operators and multiple trainees were responsible
for the administrations.
When using glass microspheres, anticipation of a 1%–2%
loss of activity due to residual microspheres in the delivery
system allows preemptive accommodation. For resin microspheres, however, there is greater percentage and variability in delivery system residual activity loss, especially
for lower prepared activities, such that additional prepared
activity should be drawn based on the magnitude of activity
prescribed, with up to an additional 5% for prepared activities of less than 1 GBq to accommodate anticipated loss.
REFERENCES
1. Kennedy A, Nag S, Salem R, et al. Recommendations for radioembolization of hepatic malignancies using yttrium-90 microsphere
brachytherapy: a consensus panel report from the radioembolization
brachytherapy oncology consortium. Int J Radiat Oncol Biol Phys 2007;
68:13–23.
2. Kallini JR, Gabr A, Salem R, Lewandowski RJ. Transarterial radioembolization with Yttrium-90 for the treatment of hepatocellular carcinoma. Adv Ther 2016; 33:699–714.
3. Padia SA, Lewandowski RJ, Johnson GE, et al. Radioembolization of
hepatic malignancies: background, quality improvement guidelines, and
future directions. J Vasc Interv Radiol 2017; 28:1–15.
4. Habib A, Desai K, Hickey R, Thornburg B, Lewandowski R, Salem R.
Transarterial approaches to primary and secondary hepatic malignancies.
Nat Rev Clin Oncol 2015; 12:481–489.
5. Tong AKT, Kao YH, Too CW, Chin KFW, Ng DCE, Chow PKH. Yttrium-90
hepatic radioembolization: clinical review and current techniques in
interventional radiology and personalized dosimetry. Br J Radiol 2016; 89:
20150943.
6. Kouri BE, Abrams RA, Al-Refaie WB, et al. ACR appropriateness criteria
radiologic management of hepatic malignancy. J Am Coll Radiol 2016; 13:
265–273.
Volume 29 ▪ Number 12 ▪ December ▪ 2018
7. Braat AJ, Smits ML, Braat MN, et al. 90Y hepatic radioembolization: an
update on current practice and recent developments. J Nucl Med 2015;
56:1079–1087.
8. Rognoni C, Ciani O, Sommariva S, et al. Trans-arterial radioembolization in
intermediate-advanced hepatocellular carcinoma: systematic review and
meta-analyses. Oncotarget 2016; 7:72343–72355.
9. Riaz A, Lewandowski RJ, Kulik LM, et al. Complications following radioembolization with yttrium-90 microspheres: a comprehensive literature
review. J Vasc Interv Radiol 2009; 20:1121–1130; quiz 1131.
10. Sangro B, Martínez-Urbistondo D, Bester L, et al. Prevention and treatment of complications of selective internal radiation therapy: expert
guidance and systematic review. Hepatology 2017; 66:969–982.
11. Kao YH, Steinberg JD, Tay YS, et al. Post-radioembolization yttrium-90
PET/CT - part 1: diagnostic reporting. EJNMMI Res 2013; 3:56.
12. Kallini JR, Gabr A, Thorlund K, et al. Comparison of the adverse event
profile of TheraSphere(®) with SIR-Spheres(®) for the treatment of
unresectable hepatocellular carcinoma: a systematic review. Cardiovasc
Intervent Radiol 2017; 40:1033–1043.
1677
13. Smits ML, Elschot M, Sze DY, et al. Radioembolization dosimetry: the
road ahead. Cardiovasc Intervent Radiol 2015; 38:261–269.
14. Mahnken AH, Spreafico C, Maleux G, Helmberger T, Jakobs TF. Standards of practice in transarterial radioembolization. Cardiovasc Intervent
Radiol 2013; 36:613–622.
15. Chiesa C, Mira M, Maccauro M, et al. Radioembolization of hepatocarcinoma with (90)Y glass microspheres: development of an individualized treatment planning strategy based on dosimetry and radiobiology.
Eur J Nucl Med Mol Imaging 2015; 42:1718–1738.
16. Garin E. Radioembolization with (90)Y-loaded microspheres: high clinical
impact of treatment simulation with MAA-based dosimetry. Eur J Nucl
Med Mol Imaging 2015; 42:1189–1191.
17. Strigari L, Sciuto R, Rea S, et al. Efficacy and toxicity related to treatment
of hepatocellular carcinoma with 90Y-SIR spheres: radiobiologic considerations. J Nucl Med 2010; 51:1377–1385.
18. Ahmadzadehfar H, Meyer C, Pieper CC, et al. Evaluation of the delivered
activity of yttrium-90 resin microspheres using sterile water and 5%
glucose during administration. EJNMMI Res 2015; 5:54.