ArticlePDF Available

Dose estimation after a mixed field exposure: Radium-223 and intensity modulated radiotherapy


Abstract and Figures

Introduction Radium-223 dichloride ([²²³Ra]RaCl2), a radiopharmaceutical that delivers α-particles to regions of bone metastatic disease, has been proven to improve overall survival of men with metastatic castration resistant prostate cancer (mCRPC). mCRPC patients enrolled on the ADRRAD clinical trial are treated with a mixed field exposure comprising radium-223 (²²³Ra) and intensity modulated radiotherapy (IMRT). While absorbed dose estimation is an important step in the characterisation of wider systemic radiation risks in nuclear medicine, uncertainties remain for novel radiopharmaceuticals such as ²²³Ra. Methods 24-Colour karyotyping was used to quantify the spectrum of chromosome aberrations in peripheral blood lymphocytes of ADRRAD patients at incremental times during their treatment. Dicentric equivalent frequencies were used in standard models for estimation of absorbed blood dose. To account for the mixed field nature of the treatment, existing models were used to determine the ratio of the component radiation types. Additionally, a new approach (M-FISHLET), based on the ratio of cells containing damage consistent with high-LET exposure (complex chromosomal exchanges) and low-LET exposure (simple exchanges), was used as a pseudo ratio for ²²³Ra:IMRT dose. Results Total IMRT estimated doses delivered to the blood after completion of mixed radiotherapy (after 37 IMRT fractions and two [²²³Ra]RaCl2 injections) were in the range of 1.167 ± 0.092 and 2.148 ± 0.096 Gy (dose range across all models applied). By the last treatment cycle analysed in this study (four [²²³Ra]RaCl2 injections), the total absorbed ²²³Ra dose to the blood was estimated to be between 0.024 ± 0.027 and 0.665 ± 0.080 Gy, depending on the model used. Differences between the models were observed, with the observed dose variance coming from inter-model as opposed to inter-patient differences. The M-FISHLET model potentially overestimates the ²²³Ra absorbed blood dose by accounting for further PBL exposure in the vicinity of metastatic sites. Conclusions The models presented provide initial estimations of cumulative dose received during incremental IMRT fractions and [²²³Ra]RaCl2 injections, which will enable improved understanding of the doses received by individual patients. While the M-FISHLET method builds on a well-established technique for external exposures, further consideration is needed to evaluate this method and its use in assessing non-targeted exposure by ²²³Ra after its localization at bone metastatic sites.
Content may be subject to copyright.
Dose estimation after a mixed eld exposure: Radium-223 and intensity
modulated radiotherapy
Isabella Bastiani
, Stephen J. McMahon
, Philip Turner
, Kelly M. Redmond
, Conor K. McGarry
Aidan Cole
, Joe M. O'Sullivan
, Kevin M. Prise
, Rhona Anderson
Centre forHealth Effects of Radiologicaland Chemical Agents,College of Health,Medicine and Life Sciences, Brunel University London,Kingston Lane, Uxbridge, London UB8 3PH,United Kingdom
of Great Britain and Northern Ireland
Patrick G. Johnston Centre for Cancer Research, Queen's University Belfast, Belfast BT9 7AE,United Kingdom of Great Britain and Northern Ireland
Northern Ireland Cancer Centre, Belfast Health and Social Care Trust, Belfast, United Kingdom of Great Britain and Northern Ireland
Centre for Radiation, Chemical & Environmental Hazards,Public Health England, Didcot OX11 0RQ, United Kingdom of Great Britain and Northern Ireland
abstractarticle info
Article history:
Received 17 February 2021
Received in revised form 4 November 2021
Accepted 9 December 2021
Available online xxxx
Chromosome exchanges
Targeted alpha-particle therapy
Prostate cancer
Introduction: Radium-223 dichloride ([
), a radiopharmaceutical that delivers α-particles to regions of
bone metastatic disease, has been proven to improve overall survival of men with metastatic castration resistant
prostate cancer (mCRPC).mCRPC patients enrolled on the ADRRAD clinical trialare treated with a mixed eld ex-
posure comprising radium-223 (
Ra) and intensity modulated radiotherapy (IMRT). While absorbeddose esti-
mation is an important step in the characterisation of widersystemic radiation risks in nuclear medicine, uncer-
tainties remain for novel radiopharmaceuticals such as
Methods: 24-Colour karyotyping was used to quantify the spectrum of chromosome aberrations in peripheral
blood lymphocytes of ADRRAD patients at incremental times during their treatment. Dicentric equivalent
frequencies were used in standard models for estimation of absorbed blood dose. To account for the mixed
eld nature of the treatment, existing models were used to determine the ratio of the component radiation
types. Additionally, a new approach (M-FISH
), based on the ratio of cells containing damage consistent with
high-LET exposure (complex chromosomal exchanges) and low-LET exposure (simple exchanges), was used as
a pseudo ratio for
Ra:IMRT dose.
Results: Total IMRT estimated doses delivered to the bloodafter completion of mixed radiotherapy (after 37 IMRT
fractions and two [
injections) were in the range of 1.167 ± 0.092 and 2.148 ± 0.096 Gy (dose range
across all models applied). By the last treatment cycle analysed in this study (four [
injections), the
total absorbed
Ra dose to the blood was estimated to be between 0.024 ± 0.027 and 0.665 ± 0.080 Gy,
depending on the model used. Differences between the models were observed, with the observed dose
variance coming from inter-model as opposed to inter-patient differences. The M-FISH
model potentially
overestimates the
Ra absorbed blood dose by accounting for further PBL exposure in the vicinity of metastatic
Conclusions: The models presented provide initial estimations of cumulative dose received during incremental
IMRT fractions and [
injections, which will enable improved understanding of the doses received by
individual patients. While the M-FISH
method builds on a well-established technique for external exposures,
further consideration is needed to evaluate this method and its use in assessing non-targeted exposure by
Ra after its localization at bone metastatic sites.
Crown Copyright © 2021 Published by Elsevier Inc. This is an open access article under the CC BY license
1. Introduction
Metastatic castration resistant prostate cancer (mCRPC) is an incur-
able condition. Prostate cancer has a strong predisposition to forming
bone metastases, with upwards of 90% of patients with advanced dis-
ease being affected, often with bone as the only site of metastasis [1].
As the general standard of care in the UK, patients with mCRPC are of-
fered a number of life-prolonging therapies including chemotherapy,
Nuclear Medicine and Biology 106107 (2022) 1020
Corresponding author.
E-mail addresses: (I. Bastiani), (S.J. McMahon), (P. Turner), (K.M. Redmond),
(C.K. McGarry), (A. Cole), (J.M. O'Sullivan), (K.M. Prise), (L. Ainsbury), (R. Anderson).
0969-8051/Crown Copyright © 2021 Published by Elsevier Inc. This is an open access article under the CC BY license (
Contents lists available at ScienceDirect
Nuclear Medicine and Biology
journal homepage:
novel anti-hormonals and Radium-223 dichloride ([
Recent advances in molecular radiotherapy have generated a great
deal of interest, particularly the use of [
after the landmark
phase III trial ALSYMPCA [24] showed for the rsttimeanoverallsur-
vival advantage associated with [
treatment. Along with
survival prolongation, this trial also demonstrated signicant
improvement in the quality of life by delaying the onset of
symptomatic skeletal related events and alleviating pain. This led to
FDA and EMA approvals and the widespread use of [
symptomatic mCRPC [5,6].
Ra has a half-life of 11.43 days, decaying
through creation of a succession of short-lived nuclides to stable
During its decay chain,
Ra emits four αand two βparticles, with
94% of its decay energy released as high LET α-particles over a short
track length of <100 μm[7,8]. When living tissue is exposed to
this results in the localized induction of clustered DSB lesions which
are difcult to repair, effectively leading to cell death [911].
For symptomatic mCRPC patients,
Ra is administered intrave-
nously as [
under the trade name of Xogo [5].Once
administered, [
immediately solubilises in the blood
resulting in free
Ra, which as a calcium mimetic, is cleared from the
blood within 24 h (h) [12] localising to areas of freshly mineralized
bone. The uptake of
Ra into high-turnover areas of bone is not depen-
dent on a particular malignant signalling process. Thus, [
couldbeofbenet in a range of other malignancies which have a
predisposition to forming bony metastases such as breast, lung, kidney
and myeloma, including those with favourable long-term survival prob-
ability. Due to the short range of α-particles, this effective target cell kill
also minimises direct α-particle exposure to non-target normal cells [7,
1316]. There is also evidence, however, that
Ra can lead to high ab-
sorbed doses in sites adjacent to target bone metastases, including oste-
ogenic cells and the red bone marrow [1720]. It has also been reported
that additional biologicalresponses via bystander effects may also occur
Currently, the number of studies which seek to understand the bio-
logical action of
Ra in vivo in humans is limited [22], indeed, there re-
mains a lack of scientic rationale to underpin current dosing strategies
and there remains a great deal of uncertainty about the heterogeneous
distribution of dose at the cellular and tissue levels and the role of direct
and indirect effects such as bystander responses. Exposure of non-
cancerous cells and tissues surrounding the tumours is of concern in
all radiotherapeutic treatments, due to acute toxicity and the potential
for delayed late effects, sometimes many years after treatment
[2330]. Quantifying the absorbed dose to non-target tissues is an im-
portant step in evaluating the potential short- and longer-term second-
ary radiation risks for patients [31]. This is of particular importance if
is to be used earlier in the treatment schedule. A number
of dose calculation models have been developed, for example, for calcu-
lation of the absorbed dose to blood from radiotherapy [32], however
the actual or estimated consequences of these are still being quantied
Cytogenetic analysis of chromosome aberrations in blood lympho-
cytes is widely used to estimate the dose of ionising radiation received
by an individual following a real or suspected radiation overexposure,
to help inform assessment of future health risks [34,35]. The dicentric
assay is the most common method of biological dosimetry, not least
due to the high radiation specicity and low interindividual variation
between yields of dicentric chromosome aberrations [34]. In so called
criticality situations of mixed eld neutron and gamma exposures
after nuclear reactor incidents, where individuals are irradiated by
both high-LET and low-LET sources, a model for emergency exposure
situations has been designed to estimate external doses for the individ-
ual exposure components [35]. Although originally designed for mixed
neutron:gamma exposures, for medical uses of radiation, the ratios of
delivered doses from internal and externally applied radiations can be
readily derived from patient treatment plans. Meaning that this ap-
proach can be applied to calculate alpha and X-ray doses in cases of
mixed exposures. Complex chromosome aberrations are useful bio-
markers of LET [36], and their identication in exposed individuals
could provide another approach for estimating the doses from mixed
exposures and, thus, to further rene dose estimation methods in
mixed exposure scenarios.
ADDRAD is an approved (NHS REC 15/NI/0074) phase I/II clinical
trial which seeks to address the potential benet of treating metastatic
hormone sensitive prostate cancer patients with androgen deprivation
therapy in addition to 6 cycles of [
and intensity modulated
radiotherapy (IMRT) to the prostate and pelvic lymph nodes which
has recently reported its rst clinical data showing good response and
minimal toxicity [37]. This patient group therefore presents a unique
opportunity to address important radiobiological research questions
pertaining to the doses and cellular and tissue level damage associated
with internal
Ra exposure combined with IMRT. In this study, we
describe the use of 24-colour whole chromosome painting (multiplex
uorescence in situ hybridisation, M-FISH) applied to blood samples
taken from the ADRRAD cohort of patients, to quantify the patterns of
chromosome exchange complexity, with the aimof discriminating dam-
aged cells from each component radiation type and estimating absorbed
radiation dose. In addition, patient specic information is used to calcu-
late absorbed doses to blood on the basis of three different established
or adapted models, and the dose estimates are compared, with the over-
all aim to further understand the uncertainties involved in dose estima-
tion in this mixed radiation exposure scenario.
2. Methods
2.1. Physically derived blood dose estimates
The absorbed blood dose per fraction (D
) was estimated using data
from 13 recruited mCRPC patients. [
was administered over
6 cycles (C1C6), comprised of one intravenous injection every
4 weeks containing 55 kBq kg
Ra [38] (Fig. 1). As part of the
treatment, IMRT is received during the rst 7.5 weeks, coinciding with
two [
cycles. The IMRT treatment plan for ADRRAD targets
the prostate with a dose of 74 Gy, with 60 Gy delivered to lymph
nodes. IMRT was delivered in daily fractions of 2 Gy to the prostate
and 1.6 Gy to the lymph nodes, the latter as a concomitant boost.
Individualised patient treatment schedules are outlined in supplemen-
tary Table 1. The absorbed dose to the blood was estimated for
based on its pharmacokinetic properties, and the IMRT dose by two
C1 C2 C3 C4 C5 C6
Treatment schedule
Blood sampling
Fig. 1. Treatment time line. IMRT daily fractions fo r 7.5 weeks. [
administered every 4 weeks for a total of 6 injections. Blood samples were collected at
each cycle, wit h a control sample collected at w eek 0 prior to treatment start (C 1).
Subsequent blood samples were taken every 4 weeks just prior to the next [
administration, up to C6. Hence bl ood samples from C2 and C3 are representative of
mixed eld exposure while C4C6 are
Ra only. For this study only samples up to C5
were analysed.
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
simple blood dose models originally derived by Moquet et al. [32] with
modications as described.
2.1.1. Pharmacokinetic derived absorbed blood dose of
In the ADDRAD trial, [
was administered at an activity of
55 kBq kg
of total body weight. Following this, the radionuclide will
clear from the blood by 24 h. Previous studies identied the amount of
Ra in the whole blood volume at varying time points within 24 h of
administration [12,3941]. The median percentage of circulating
was averaged at timepoints of 15 min, 4 h and 24 h between all
studies reporting median values with upper and lower ranges [3941].
This gave estimates for the percentage of radium in the blood of 22%
after 15 min (928%), 3% (1.34.95%) after 4 h and 0.8% (0.371%) at
24 h. This information was used to estimate the physical absorbed
dose to the blood by circulating
Firstly, the total activity of
Ra present in the blood was estimated
at each timepoint. This was determined by multiplying the initial
injectedactivity of radium by the estimated percentage of
Ra remain-
ing in the blood and the decay constant of
Ra, 7.02 × 10
The rate of energy deposition was then calculated using the energy re-
lease for each decay of
Ra and its short-lived daughter products
Po and
Bi) [43], with the total α-particle energy release es-
timated to be 4.30 × 10
J. Finally, the dose rate to blood was calcu-
lated by dividing this rate of energy deposition by the mass of blood of
the patient, which for this work was estimated using 75 ml of blood
per kg body weight.
The absorbed dose rates were then plotted as a function of time fol-
lowing injection, and the area under the curve was calculated
(GraphPad Prism 9; GraphPad Software; computed using trapezoid
rule) to estimate the patient specic dose delivered to the blood for
the rst 24 h of each treatment cycle. As less than 1% of the activity re-
mains at 24 h post-irradiation, it was assumed that the additional dose
deposited from 24 h to t = was negligible. The uncertainty on the
Ra absorbed blood dose estimates was dominated by the uncertainty
in the amount of circulating
Ra at each time period, which, as dis-
cussed below, could be up to 50% [3941].
2.1.2. IMRT blood ow model
For this part of the work, data was assessed from 13 ADDRAD pa-
tients, for whom treatment p lan and patient specic data wereavailable.
The blood ow model (BF) enabled estimation of dose within the high
dose organ area. The absorbed blood dose per fraction (D
estimated as follows:
ðÞ ð1Þ
where D
was absorbed dose to blood per fraction ,D
was the prescribed
dose per fraction, V
represents the high dose volume and V
was the
total blood volume.
Two variates of the model were used: BF
which was applied as per
Moquet et al. [32] which estimates V
by assuming75 ml of blood per kg
and, BF
which uses thestatic volume of blood in the prostate and lymph
nodes for V
, estimated by calculating a scaling factor between the
whole-body volume and the area irradiated (prostate and/or lymph
nodes) on the basis of treatment plan information. This scaling factor
was then applied to the whole-body blood volume estimates to achieve
a static blood volume for prostate and lymph nodes. In both cases, D
and V
were taken from the treatment plans. The uncertainty on V
and V
was estimated to be on the order of 10%, and the uncertainty
on D
can be estimated on the basis of Moquet and colleagues to be up
to 20%. Hence a conservative estimate of uncertainty on D
would be
approximately ±25%.
2.1.3. IMRT CT plan model
For this part of the work, data from 13 ADDRAD patients was
assessed, for whom treatment plan and patient specic data were
available. The CT planned volume (CTPV) absorbed blood dose model
enables dose estimation of both high and low dose areas. The CT volume
mapped for each patient was utilized to calculate a scaling factor for
each patient relative to whole-body volume, with whole-body volume
being calculated based on the average patient weight (kg) over the
course of treatment [44]. For the following dose models, the whole CT
volume was considered within the CTPV
along with a high dose
volume only CTPV
. Utilizing the CTPV
model, the whole-body mean
dose was estimated, and it was assumed the blood volume had also re-
ceived this. Firstly, a scalingfactor, S, was estimated as the ratio between
the CT plan volume and the whole-body volume. The following was
then applied to estimate the absorbed blood dose per fraction:
where D
was the absorbed dose to blood per fraction; D
was the
mean dose in Gy to the body volume covered by the CT scan (specic
to each patient, taken from the treatment plans), N
was the number
of fractions of radiotherapy and S was the patient specic scaling factor.
A further estimate, CTPV
, was performed in the same manner but in
this case rstly calculating the average dose per fraction for prostate
and lymph nodes, and then calculating a scaling factor based on their
volume compared to the whole-body volume.
The error associated with the plan volume was considered to be on
the order of 11.5%. The clinical upper acceptable limit in dose delivery
to the treatment plan was within 3%, as anything greater than this
would trigger re-calibration of dose planned area. The scaling factor
was estimated from whole body volume which was estimated by as-
suming that 1.01 g of human body mass ts within 1 cm
and that the
tissue density within the target volume was consistent with this for
the scalingfactor estimation,with an error within 1%.The total error as-
sociated with CTPV estimations was considered to be within ±6%.
2.2. M-FISH dicentric quantication
2.2.1. Sample collection
Whole blood was received from 5 male patients recruited onto the
ADRADD trial (EudraCT 2014-00273-39) with full informed consent
(NHS REC 15/NI/0074) at The Belfast Health and Social Care Trust.
Blood samples were drawn into lithium heparinised anticoagulant
tubes, once every 4 weeks, immediately prior to the rst (C1) and
then within 24 h prior to each subsequent (C2C5) [
administration (Fig. 1). The samples were then shipped at room
temperature for nextday delivery to Brunel University London. Upon ar-
rival, the samples were processed and whole blood stimulated to divide
to enable the collectionof 1st in vitro cell division metaphase cells for cy-
togenetic assessment, as described below.
2.2.2. Cell culture
For each sample, 0.4 ml of whole blood was used to inoculate 2.6 ml
of freshly prepared media (PBMAX Karyotyping Medium)
(ThermoFisher, Cat. Number 12557021) supplemented with 0.5 μg/ml
puried phytohaemagglutinin (PHA) (ThermoFisher, Cat. Number
R30852801), 10 μM 5-bromo-2-deoxyuridine (BrdU, Sigma-Aldrich
Cat. Number 19-160), 10 μl/ml heparin (Sigma-Aldrich Cat. Number
9041-08-1) and cultured in a humidied incubator at 37 °C (95% air/
5% CO
), at a 45° angle, and with the cap left slightly open to allow
gaseous exchange. Cultures were set up to maximise the yield of 1st
cell division of peripheral blood lymphocytes (PBLs) and harvested
using standard cytogenetic techniques after a total of 5060 h. To arrest
cells at the metaphase stage of the cell cycle, 50 μl/ml of Colcemid
KaryoMAX (ThermoFisher, Cat. Number 1521012), a tubulin inhibitor,
was added 4 h prior to harvest. After this time,the cultures were centri-
fuged at 200gfor 10 min and the cell pellet re-suspended before the ad-
dition of 0.075 M KCl hypotonic solution (Fisher ScienticCat
Number10575090) for 8 min at 37 °C. Cells were then centrifuged at
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
200gfor 10 min and xed in 3:1 methanol (Thermo Fisher Catalogue
Number 15654570) acetic acid (Thermo Fisher Catalogue Number
1743468) on ice. The xation process was repeated until the samples
appeared clear (~5 times), and these were then stored in the freezer at
20 °C.
2.2.3. Harlequin stain
Fixed chromosome preparations were dropped onto clean, grease-
free slides and assessed for metaphase quality. Harlequin staining was
used to assess the number of 1st division metaphase cells. For this, slides
were aged on a hot plate for 45 min at 90 °C and immersed in Hoescht
(Thermo Fisher Scientic Cat Number 62249) for 10 min, then trans-
ferred to a at tray and covered in 2× Saline-Sodium Citrate (SSC)
(Thermo Fisher Cat Number 15557036), before being exposed in UV
box to 1.0 J/cm
for 1 h. After exposure slideswere washed with distilled
water twice and air dried. Treated slides were stained in 5% Giemsa
(VWR Cat Number 350864) for 4 min,removed and rinsed with distilled
water. Once dry, the slides were mounted with coverslips with 4 drops
of DPX (Fisher Scientic Cat Number 15538321). Slides were scored
using brighteld microscopy with oil immersion at ×100 magnication.
The fraction of 1st/2nd/3rd division cells was determined based upon
the chromatid staining patterns [45,46]. Samples with 5% 2nd division
cells were assayed by M-FISH.
2.2.4. Multiplex uorescence in situ hybridisation (M-FISH)
M-FISH was carried out utilizing 24XCyte staining probe
(Metasystems Probe Cat NumD-0125-600-DI) as per manufacturer pro-
tocol. Patient slides were selected from 5 patients according to meta-
phase spread quality and whether samples containing 5% 2nd division
cells, a minimum of 3 patient slides were painted per cycle. In brief,
slides were incubated in 2xSSC at 70 °C (±1 °C) for 30 min. After this
time, the cooled slide was transferred into 0.1xSSC at RT for 1 min. Chro-
mosomes were then denatured in 0.07 NaOH at RTfor 1 min followed by
1 min incubation in 0.1xSSC, followed by 2xSSC at 4 °C, and then
dehydrated through immersion in a series of alcohol solutions of as-
cending strength (70%, 95% and 100%). The 24Xcyte probe was dena-
tured by incubating at 75 °C (±1 °C) for 5 min, placed on ice briey
and then incubated at 37 °C for 30 min. The probe was overlaid on to
the slide and left to hybridize in a humidied chamber at 37 °C (±
1 °C) for 23 days. Slides were washed in 0.4xSSC preheated to 72 °C
(±1 °C) for 2 min then incubated in 2xSSCT (containing 0.05%
Tween20) for 30 s. For counterstaining, the slide was rinsed in double
distilled water and left to air dry before application of DAPI/antifade
and sealing.
Slides were visualised utilizing 8-position Zeiss Axioplan 2 micro-
scope containing individual lter sets for 24XCyte probe cocktail plus
DAPI (FITC, Spectrum Orange, Texas red, Cy5, DEAC and DAPI).
Metaphase cells were imaged under x63 oil immersion and captured
by Cool Cube driven by Metafer4 version 3.14.191 software. The image
les were exported and karyotyped in ISIS version 5.8.11.
2.2.5. Chromosome aberration classication
A cell was classied as being apparently normal if all 46 chromo-
somes were present and contained the appropriate uorophore combi-
nation along their entire length. Only metaphase cells with good uoro-
chrome staining were selected for analysis in cells containing 43 chro-
mosomes. Chromosomal aberrations were identied by colour junctions
along the length of each individual chromosome and/or by the presence
of chromosomal fragments (Fig. 2b). A chromosome interchange involv-
ing 2 breaks in 2 chromosomes was categorised as a simple exchange,
and further classied as a reciprocal translocation or dicentric. Ring
chromosomes, which involve 2 breaks in one chromosome were also
classed as simple [47]. Exchanges involving 3 or more breaks in 2 or
more chromosomes were classed as complex and assigned theminimal
number of breaks, arms and breaks involved (CAB) [48].Chromosomes
having breaks only, not involving any additional chromosomes, were
classed as chromosome breaks.When classifying cells with multiple ab-
errations, all aberrations were recorded as independent events and the
chromosomes involved identied. Where homologous chromosomes
were involved, efforts were made to establish whether the homologues
were in the same event or in different independent events, mainly by
consideration of chromosome length [49]. All exchanges were recorded
as either complete (all break-ends re-joined), true incomplete (where
one or more break-ends fail to nd an exchange partner) or one-way
(where one or more elements appear to be missing) [50,51]. The poten-
tial transmissibility of exchanges was also recorded, where a stable
(transmissible) exchange was dened as complete and with no evi-
dence of unstable elements e.g.dicentric or acentric fragments. Each in-
dependent complex event was also determined to be transmissible or
non-transmissible and the presence of insertion-type rearrangements
were noted [49]. Metaphase spreads were categorised as stable only if
all the exchange events detected within that spread were classied as
stable. Unstable complex chromosomal exchanges containing polycen-
tric chromosomes were broken down in to their dicentric equivalents
whereby each additional centromere within a chromosome structure
constituted a dicentric equivalent (dicentric equivalent event = n cen-
tromere 1) (Fig. 2).
2.3. M-FISH dicentric assay dose estimation
2.3.1. Mixed eld absorbed blood dose ratio
For this part of the work, M-FISH analysis was carried out n=5in-
dividuals with a minimum of n= 3 patient samples processed for C2
C5. The
Ra:IMRT ratio was estimated in two ways. Firstly, by physical
Fig. 2. Dicentric equivalent scoring. a) 1st division Giemsa stained metaphase imaged under bright eld microscopy ×100 magnication under oil. Chromosome with two centromeres
highlighted along with acentric fragment b) Pseudo colour processed image painted by M-FISH, captured at ×63 magnication under oil. Complex chromosomal exchange between
three chromosomes, dicentric chromosome highlighted by red arrow along with other components in yellow. c) DAPI channel of same cell (b) highlighting same aberration as a simple
dicentric exchange.
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
dose estimation and secondly, from categorising cells based on the com-
plexity of chromosome exchange observed. For the physical dose ratio,
the absorbed blood dose was calculated independently for [
and IMRT (Section 2.1) for each exposure and was summed over
the 4-week time period prior to each sample being taken. The average
absorbed blood dose per cycle was estimated by averaging theabsorbed
blood dose across all 13 patients for all treatment cycles. For the dose
ratio the overall treatment plan was used (Fig. 1), not the patient spe-
cic plan adaptation. Therefore, C2 relates to the response to one injec-
tion of [
and 20 IMRT fractions, C3 relates to two injections of
and 37 IMRT fractions, with C4 and C5 each including the
additional [
only administrations. Summing these gives the
total absorbed blood dose of [
and IMRT delivered at each
time point from C2 to C5. For the M-FISH derived ratio (M-FISH
cells containing at least one complex chromosome exchange were
classed as being damaged by the traversal of high-LET α-particles
from the
Ra while cells containing only simple chromosome ex-
changes, as from IMRT (Section 2.2.5). The ratio of cells containing at
least one complex chromosome exchange to cells containing simple ex-
changes only was therefore a pseudo ratio for
Ra:IMRT absorbed
blood dose at each cycle point.
2.3.2. Dicentric absorbed blood dose estimation
The dicentric assay traditionally utilizes Giemsa staining where
chromosomes are evenly stained in one colour, (Fig. 2a), thus enabling
the identication of chromosomes containing more than one centro-
mere along with any associated acentric fragment. In this study, dicen-
trics were quantied from M-FISH painted metaphase cells (Fig. 2b
and c).
Dose estimation was rst carried out utilizing the dicentric equiva-
lent frequency as determined by M-FISH. For the IMRT, the
Co calibra-
tion curve of Lloyd and colleagues, 1986, was used [52] where whole
blood was irradiated in vitro utilizing a
Co source with dose range of
05 Gy. The dicentric yield was entered into Dose Estimate V5.2 [53]
with the following coefcients α= 0.0756 ± 0.0031, β=0.0149±0.
0060 and C = 0.0004 ± 0.0009 [52]. This being a well-established cali-
bration curve, it has been utilized in many exposure scenarios for γ-ray
and X-ray dose estimation by Public Heath England and was judged to
be the most comparable curve in terms of type and energy of radiation
exposure. As there is currently no
Ra calibration data, a calibration
curve based on a
Pu, which emits α-particles of a similar energy to
Ra (5.16 MeV per α-particle), was selected. The
Pu calibration
curve of Purrott et al. [54] was used, the curve coefcients were: β=
0.3696 ± 0.0322, C = 0.0019 ± 0.0126 [54].Thedecaychainof
Pu does differ, however, due to the perceived lack of risk of
a radiation accident involving alpha exposures and the complexity
of the experiments, alpha curves for biodosimetry are rare. This
curve was established by irradiating whole blood with
Pu in the
range of 01.6 Gy. It is important to note, however, that in this
study, the likely non-homogeneous, partial body, nature of both the
Ra and IMRT exposures beyond the treatment plan details, was
not further considered, neither was the microdosimetric heterogene-
ity of radium in areas of high bone turnover, and thus this represents
a key source of unquantied uncertainty. This aspect will need to be
incorporated into further development of absorbed blood dose
models. To calculate the absorbed blood dose in this mixed exposure
scenario, the criticalitymodel was used [35]. In brief, all aberra-
tions were rstly assumed to be attributed to
Ra and from the di-
centric equivalent yield the dose was estimated. The absorbed blood
dose ratio (
Ra:IMRT) calculated according to Section 2.3.1.was
then used to estimate the IMRT dose and then the gamma calibration
curve used to estimate the dicentric equivalent yield. This IMRT yield
was then subtracted from the total yield to give a new
Ra dicen-
tric equivalent yield. This iterative process was repeated until self-
consistent estimates were obtained.
2.4. Other statistical analysis
Statistical analysis was performed using GraphPad Prism 9
(GraphPad Software). Descriptive statistics are presented as mean ±
SE for pooled data. Standard propagation of errors was applied to esti-
mate the uncertainty in the derived dose estimates. Normality testing
indicated ANOVA was appropriate in order to test for differences in ab-
erration frequencies between treatment rounds and for differences in
estimated doses.
3. Results
3.1. Estimating the blood dose from the treatment schedule
Patient data related to the planned treatment for 13 patients was
available for physical dose estimation. Table 1 gives the absorbed
blood dose per fraction predicted by the BF model, estimated for each
patient for the prostate, lymph nodes, and total high dose region com-
bined. The majority of the absorbed blood dose was estimated as
being from lymph node exposure, due to the larger volume of irradia-
tion for this tissue (BF
). The total prostate-only absorbed blood dose
was found to be in the range of 0.8801.962 Gy by the end of IMRT treat-
ment (3037 fractions; 37 fractions assumed for dose ratio calculations
in Section 3.3), comparable to the 0.381.92 Gy reported by Moquet
et al. To consider the differences in lymphatic uid shift, the static vol-
ume was calculated (BF
) and, by combining the BF models (BF
), the cumulative absorbed blood dose was estimated to be between 1.
131 and 2.717 Gy.
The absorbed blood dose was estimated for the whole CT plan area,
including high dose and low dose regions (CTPV
) and for high dose
regions only (CTPV
) to enable comparison with the BF model.
PreviousCTPV approaches haveused a xed scaling factorof 2.5 with
a reported uncertainty of 20%. In this work we instead estimated a pa-
tient specic scaling factor between 2.08 and 2.64, with the largest devi-
ation from the scaling factor for Patient 10 being 16% smaller than the
published 2.5 scaling factor.The increased absorbed blooddose CTPV es-
timates compared to the BF estimates observed here are likely to be re-
ective of the large lymph node irradiation volume [32].Tocomparethe
BF and CTPV models, CTPV was also calculated using the high dose vol-
umes of prostate and lymph nodes, termed as CTPV
. To do so, a scaling
factor for each component as detailed for BF
, was used. The resulting
absorbed blood doses for CTPV
were found to be comparable to that
of BF
within the uncertainty estimates (Table 2). The average
Table 1
BF model. V
blood volume, D
blood dose per fr action. BF
from BF
prostate dose
combined with BF
LN dose (BF
). Patients for whom treatment plan and patien t
specic data were available n= 13. Reported estimated unce rtainty of 25% for al l BF
model variants.
Total dose D
1 9030 12.9 54.5 0.038 0.132 0.170 0.007 0.045
2 7615 11.7 56.8 0.041 0.164 0.205 0.010 0.050
3 6288 9.5 49.2 0.040 0.171 0.212 0.009 0.049
4 6560 12.2 57.2 0.060 0.231 0.291 0.014 0.073
5 6461 13.8 63.9 0.056 0.209 0.265 0.015 0.071
6 6985 10.8 58.9 0.041 0.184 0.225 0.011 0.052
7 7650 9.4 52.0 0.033 0.149 0.181 0.008 0.040
8 5700 7.0 32.4 0.032 0.124 0.157 0.004 0.037
9 6290 11.5 49.1 0.050 0.154 0.204 0.007 0.058
10 6629 14.0 52.3 0.056 0.171 0.227 0.009 0.065
11 5559 9.1 48.4 0.043 0.188 0.231 0.009 0.052
12 7883 7.1 49.2 0.024 0.135 0.159 0.007 0.031
13 6486 10.2 54.8 0.042 0.161 0.203 0.009 0.051
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
absorbed blood dose per fraction of IMRT wasestimated as 0.101 Gy for
and 0.016 Gy for CTPV
(average n= 13 patients), each with an
estimated uncertainty on the order of 6%.
and CTPV
cannot be directly compared as they use different
irradiation volumes. The BF
model assumes the blood ow within
prostate and lymph node regions was identical and in doing so,
it appears to signicantly overestimate the lymph node blood volume.
On the basis of this and the simple nature of the other model
assumptions, this method likely has the largest uncertainty.
Accordingly, only CTPV
and CTPV
are considered further as these
models represent an estimate of combined absorbed blood dose for
high dose regions and low dose regions (CTPV
) and high dose regions
only (CTPV
). Supplementary Table 1 details the patient specic
injected activities and CT planning volumes. Table 2 shows the individ-
ual absorbed blood dose estimates for each
Ra administration, and
the IMRT absorbed blood dose estimates for each fraction. The average
Ra absorbed blood dose per treatment cycle was 0.012 ± 0.002 Gy.
The estimated
Ra absorbed blood dose per fraction was not found
to be statistically different between patients (P= 0.097). To estimate
Ra absorbed blood dose per treatment cycle, the median values
of the activity kinetics reported in the literature were utilized. Based
on the highest and lowest limits reported for each timepoint the
blood dose estimates per treatment cycle could vary by up to 50%. This
value being reported as a conservative estimate of uncertainty for dose
estimates, with more work needed to quantify the variations between
patient clearance.
3.2. Frequency and type of chromosome aberration in ADRADD patients
M-FISH analysis was carried out on blood samples received from 5
patients for C1C5. The control samples of all 5 patients were analysed
and following this a minimum of 3 patients were analysed per treat-
ment cycle. The background frequency of chromosomal aberrations
was found to be within the expected normal range for individuals in
the 50+ age bracket [55,56]. Specically, frequencies of 0.0, 0.024 ± 0.
011 and 0.020 ± 0.009 were observed for simple dicentrics, reciprocal
translocations and break-only aberrations, respectively (sample C1;
pooled for 5 patients). One cell containing an unstable complex rear-
rangement was found in Patient 2 (0.004 ± 0.004), for which the origin
was not clear.
Simple chromosomal exchanges were observed to signicantly in-
crease in frequency between C1 and C2 (0.024 ± 0.011 to 0.319 ± 0.
042 (P< 0.001)) and further at C3 to 0.484 ± 0.042 (P<0.031)(Supple-
mentary Table 2). No other statistical differences were noted between
treatment cycles (C34, P= 0.966 or C4C5, P= 0.996), suggesting
the majority of simple aberrations formed early in the treatment regime
persist over the period sampled, and/or the induction of new aberra-
tions was balanced by cell death of unstable (e.g. dicentric) types. The
frequency of complex exchanges increased from 0.058 ± 0.016 at C2
to 0.174 ± 0.022 at C3 (P= 0.007), rising further to 0.265 ± 0.037 at
C4 and 0.210 ± 0.030 at C5 (Supplementary Table 2).
In terms of classication of complexity, the frequency of damaged
PBLs assigned to each exposure type (
Ra or IMRT) was reported in
Fig. 3 and Table 3. This categorisation was based upon the presence or
absence of a complex chromosome exchange and shows that the major-
ity of damaged PBLs sampled, with combined IMRT and
Ra contained
mostly simple exchanges only at C2 (cells containing simple exchanges
0.204 ± 0.025 and containing at least one complex exchange 0.050 ± 0.
014, P< 0.001) and C3 (0.243 ± 0.021 and 0.150 ± 0.018, P=0.006).
After this time, when patients continue to receive
Ra only, the frac-
tion of damaged PBLs with at least one complex exchange increased.
In the
Ra only cycles, no signicant difference between cells contain-
ing a simple aberration and those containing at least one complex ex-
change was observed (C4 P=0.992,C5P=0.558).
3.3. M-FISH dicentric assay absorbed blood dose estimates
The absorbed blood dose ratios were estimated for C2C5, this was
equivalent to 4 intravenous injections of
Ra and completed IMRT
schedule of 37 fractions by C5. For the physical dose ratio, this was
based on the average absorbed blood dose per fraction across all 13 pa-
tients for both the IMRT and
Ra dose. The M-FISH
ratios described
in Section 2.3.1, was obtained from blood samples received from 5
patients with a minimum of 3 patient samples analysed per cycle. The
absorbed blood dose ratios for all models can be seen in Table 3.
Applying these blood dose ratios to the dicentric assay, the absorbed
blood dose was estimated in three ways (Table 4). The CTPV methods
were derived from the physical absorbed blood dose estimates as per
method Section 2.1. with CTPV
estimating the absorbed blood dose
across the planned volume including both high and low dose volumes
while CTPV
including the high dose regions only. The M-FISH
absorbed blood dose was instead estimated from the ratio of cells
consistent with high LET exposure (
Ra) and low LET exposure
The absorbed blood dose estimates from the CTPV models suggest
the dose after 20 IMRT fractions, measured at C2 (Table 4), was between
1.327 ± 0.115 Gy and 1.395 ± 0.115 Gy. The M-FISH
method was
found to be in similar range at 0.956 ± 0.114 Gy at C2. The IMRT ab-
sorbed blood dose after the end of fractionation (C3 assuming 37 com-
pleted fractions) was estimated for CTPV
as 2.148 ± 0.096 Gy and 2.
073 ± 0.096 Gy for CTPV
, with M-FISH
blood dose estimates of 1.
167 ± 0.092 Gy.
Table 2
Physically derived absorbed blooddose, in Gy.
Ra absorbedblood dose was representa-
tive of an average dose to the blood during localization period. The absorbed blood dose
after all treatment cycles was calculatedfor n= 6 injections with theexception of Patient
5 where n= 5. IMRT absorbed blood dose estimated for whole CT plan area CTPV
high dose regions only CTPV
. Scaling factor est imated from patient body volume
derived from ave rage weight during treatment. Patient numbers n=13patientsfor
whom treatment plan and patient specic data were available. Uncertainties of up to 6%
for CTPV models and up to 50% considered for the
Ra estimates.
Patient ID
dose (Gy)
1 0.0116 0.105 0.013
2 0.0114 0.104 0.015
3 0.0116 0.101 0.016
4 0.0118 0.113 0.022
5 0.0116 0.100 0.020
6 0.0116 0.112 0.017
7 0.0117 0.092 0.014
8 0.0118 0.095 0.012
9 0.0117 0.089 0.015
10 0.0115 0.113 0.017
11 0.0114 0.108 0.018
12 0.0116 0.094 0.012
13 0.0114 0.088 0.015
Fig. 3. Frequency of cells containing either si mple () or at least one complex ()
chromosome exchange. Data pooled from 5 patients total, n= 3 pat ients/cycle (C2, 3
and 5) and n= 4 for C3. Error bars represent standard error of the mean for pooleddata.
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
For the
Ra absorbed blood dose at C2, the CTPV
and CTPV
found to be 0.008 ± 0.025 and 0.048 ± 0.032, both lower than the M-
estimate of 0.150 ± 0.046 Gy. For C3 the
Ra absorbed blood
dose estimated from CTPV methods was between 0.013 ± 0.025 and
0.082 ± 0.033 Gy (CTPV
and CTPV
respectively) with the M-FISH
absorbed blood dose estimated at 0.719 ± 0.073 Gy. By the nal
sample point studied here (C5), the absorbed blood dose from
was estimated as 0.024 ± 0.027 Gy by CTPV
and 0.141 ± 0.042 Gy
for CTPV
with the M-FISH
absorbed blood dose estimate at 0.
665 ± 0.080 Gy (see Fig. 4).
4. Discussion
In the future, α-particle emitters, such as
Ra, alone or in combina-
tion, are likely to be adopted in the treatment of a wide range of cancers,
contributing to more targeted, personalized medicine with the potential
to impact a large number of patients [17,57,58]. To date, however, there
are only limited studies using biological endpoints which seek to under-
stand the biological action of
Ra in vivo in humans, to improve theop-
timal delivery of these and, to minimise the risk of adverse effects
through radiation exposure of normal tissues [5963]. Quantifying the
absorbed blood dose delivered from the treatment is an important
step in maximising clinical efcacy and evaluating radiation risks, esti-
mates of which are currently based on population studies [31]. In this
study, whole blood was sampled from patients recruited onto the
ADRRAD trial [37]. The treatment includes a planned mixed eld expo-
sure from daily IMRT fractions, 37 × 2 Gy to prostate with concomitant
boost (37 × 1.6 Gy) to lymph nodes, over a period of 7.5 weeks, together
with 6 intravenous injections of [
over a period of 20 weeks.
Dose estimation for non-targeted tissues in mixed exposure scenarios is
challenging. To account for the mixed eld nature of the treatment,
existing models were used to determine the ratio of the component ra-
diation types. Additionally, a new approach (M-FISH
), based on the
ratio of cells containing damage consistent with high-LET exposure
(complex chromosomal exchanges) and low-LET exposure (simple ex-
changes), was used as a pseudo ratio for
Ra:IMRT absorbed blood
Ra exposure was dictated by its unique pharmacokinetic proper-
ties and its ability to target calcium-dependent bone turnover. Once in-
travenously administered,
Ra is rapidly cleared through the gastroin-
testinal tract and the remainder through the kidneys [39,41].Previous
studies on rodents found minimal uptake of
Ra in non-targeted
areas such as kidneys and the spleen [13,64] with the highest absorbed
doses being observed in humans in neighbouring sites of target bone
metastases, including osteogenic cells and the red bone marrow [17,
19,20]. Here, the
Ra absorbed blood dose was estimated from the
injected activity using an existing clearance model, based on the quan-
tity of
Ra in circulation at three time points (15 min, 4 h and 24 h
post intravenous administration). The uncertainty for the
Ra physical
absorbed blood dose estimations per cycle was based on the errors asso-
ciated with the injected activity, patient weight and clearance models
used. The injected activity per patient was estimated from the known
activity in the syringe prior to administration and the remaining activity
after injection, the error associated with this was considered negligible.
Patient individual weights were found to uctuate during treatment by
8% (up to 6.2 kg lost by C6). Therefore, the
Ra absorbed blood dose
was estimated for each chosen timepoint from the data available at
the specic treatment cycle rather than an average across treatment
Table 3
Absorbed blood dose ratios. M-FISH derived ratio based on the complexity of cellular damage with
Ra assigned cells to include those with at least 1 complex exchange and, IMRT
assigned cells dened as containing only simple exchanges. Frequency per total cells scored pooled per cycle, with calculated SE. D
per cycle of patient average (n= 13), complexity
of cellular damage from pooled patient data (n=5).
Cycle Cellular classication D
per cycle
Ra cells (f) IMRT cells (f)
C2 0.050 ± 0.014 0.204 ± 0.025 0.012 ± 0.002 2.021 ± 0.050 0.318 ± 0.017 1:174 1:27 1:4
C3 0.150 ± 0.018 0.243 ± 0.021 0.023 ± 0.003 3.739 ± 0.092 0.587 ± 0.031 1:161 1:25 2:3
C4 0.198 ± 0.025 0.229 ± 0.026 0.035 ± 0.003 3.739 ± 0.092 0.587 ± 0.031 1:108 1:17 1:1
C5 0.162 ± 0.021 0.220 ± 0.024 0.046 ± 0.004 3.739 ± 0.092 0.587 ± 0.031 1:81 1:13 3:4
Table 4
Absorbed blood doses during treatment (Gy) calculated on the basis of CTPV
and M-FISH
models. Patients sampled duringtreatment cycles were coded and identier number
(ID) assigned, n= 5 patients were assayedby M-FISH to determine the dicentric equivalent yield.The absorbed blooddose was calculatedwith use of dicentricassay for each patientwith
the dose ratios estimated from CT plan and M-FISH
derived models. The SE was reported for the absorbed blood dose estimation, this being the largest uncertainty for M-FISH
estimation. CTPV
and CTPV
reported SE of absorbed blood dose estimation with a propagated conservative uncertainty estimated up to 50% not included.
Cycle Patient ID Cells scored Dicentric equivalents
(yield ± SE)
2 1 55 8 (0.145 ± 0.066) 0.007 ± 0.033 1.268 ± 0.246 0.044 ± 0.054 1.201 ± 0.246 0.207 ± 0.105 0.845 ± 0.245
3 101 16 (0.158 ± 0.067) 0.008 ± 0.029 1.329 ± 0.182 0.046 ± 0.043 1.261 ± 0.182 0.221 ± 0.081 0.899 ± 0.181
4 104 21 (0.202 ± 0.048) 0.009 ± 0.029 1.516 ± 0.180 0.053 ± 0.045 1.448 ± 0.180 0.263 ± 0.086 1.070 ± 0.179
Total 260 45 (0.173 ± 0.035) 0.008 ± 0.025 1.395 ± 0.115 0.048 ± 0.032 1.327 ± 0.115 0.235 ± 0.055 0.959 ± 0.114
3 1 100 30 (0.30 ± 0.0820) 0.012 ± 0.031 1.874 ± 0.184 0.071 ± 0.051 1.799 ± 0.184 0.584 ± 0.128 0.947 ± 0.182
2 101 38 (0.376 ± 0.076) 0.013 ± 0.031 2.113 ± 0.184 0.080 ± 0.053 2.038 ± 0.184 0.701 ± 0.139 1.138 ± 0.182
4 101 45 (0.446 ± 0.111) 0.014 ± 0.032 2.310 ± 0.185 0.088 ± 0.055 2.234 ± 0.184 0.801 ± 0.149 1.300 ± 0.182
5 105 45 (0.429 ± 0.065) 0.014 ± 0.031 2.264 ± 0.181 0.086 ± 0.053 2.188 ± 0.181 0.777 ± 0.144 1.261 ± 0.179
Total 407 158 (0.388 ± 0.042) 0.013 ± 0.025 2.148 ± 0.096 0.082 ± 0.033 2.073 ± 0.096 0.719 ± 0.073 1.167 ± 0.092
4 1 48 15 (0.313 ± 0.099) 0.018 ± 0.042 1.908 ± 0.264 0.106 ± 0.082 1.797 ± 0.264 0.680 ± 0.198 0.789 ± 0.262
2 102 53 (0.520 ± 0.062) 0.023 ± 0.035 2.497 ± 0.184 0.141 ± 0.066 2.384 ± 0.184 1.048 ± 0.169 1.216 ± 0.181
3 103 32 (0.311 ± 0.116) 0.018 ± 0.033 1.902 ± 0.182 0.106 ± 0.058 1.791 ± 0.182 0.677 ± 0.136 0.785 ± 0.179
Total 253 100 (0.395 ± 0.052) 0.020 ± 0.028 2.162 ± 0.119 0.121 ± 0.043 2.050 ± 0.119 0.833 ± 0.097 0.966 ± 0.116
5 1 106 47 (0.443 ± 0.083) 0.028 ± 0.037 2.290 ± 0.180 0.169 ± 0.070 2.142 ± 0.180 0.864 ± 0.151 1.175 ± 0.178
4 101 35 (0.347 ± 0.084) 0.025 ± 0.036 2.009 ± 0.184 0.147 ± 0.068 1.862 ± 0.183 0.705 ± 0.140 0.959 ± 0.182
5 102 18 (0.176 ± 0.043) 0.017 ± 0.033 1.394 ± 0.182 0.099 ± 0.057 1.254 ± 0.181 0.392 ± 0.105 0.533 ± 0.180
Total 309 100 (0.324 ± 0.042) 0.024 ± 0.027 1.937 ± 0.108 0.141 ± 0.042 1.791 ± 0.108 0.665 ± 0.080 0.905 ± 0.105
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
along with the median values of
Ra in the blood at each timepoint
(averaged between reported studies). The average absorbed blood
dose per cycle was estimated as 0.012 ± 0.002 Gy (average of Table 2
as per methods Section 2.3.1). In order to estimate the uncertainty for
the clearance model, the reported ranges in terms of upper and lower
range limits were taken from the literature [3941] with a variation
from the median of up to 50%. The average uncertainty on the delivered
Ra activity was thus estimated to be as high as50%. The additional ab-
sorbed blood dose delivered after 24 h period was not considered in this
model. Although the additional absorbed blood dose is likely to be
within the 50% uncertainty described, it may not be implicitly negligible
and does form an additional contribution to the uncertainty not quanti-
ed here.
The physical absorbed blood doses reported here are higher than
those reported by Stephan et al. using
Ra and in a similar range to
those reported by Schumann et al. for
Ra [65,66]. The decay kinetics
and energies of
Ra are broadly comparable to those of
Ra, how-
ever, as Stephan and colleagues report, there are a number of limitations
associated with the application of the ICRP 67 model [67], including the
lack of information regarding the local distribution of the activity. The
aim of the
Ra clearance calculations in this work was not to test the
ICRP model, rather, to provide a very simple kinetic method for valida-
tion of the newly proposed biological absorbed blood dose estimation
methods presented. Nevertheless, further work is required to assess
the most appropriate means of estimating dose, and this will include re-
assessment of more detailed kinetic models including those from ICRP.
IMRT utilizes conformal beams to accurately target a designated area
minimising exposure to non-target organs/tissues. Unlike
Ra, this
treatment is tailored to the disease burden of each patient with the
number of fractions, dose and area covered dependent upon patient
specic information (Supplementary Table 1). Using this, estimates
were carried out employing two models previously described by
Moquet and colleagues, which were designed to provide a relatively
simple assessment of RT doses [32].Therst model, the blood ow
model (BF), was based on the time taken for blood to ow through
the high dose planned area. Moquet et al., assumed that 6 L of whole
blood ows through the high dose region within a 1-min IMRT exposure
time [32]. For this study, individual patient weights were available,
therefore blood volume was more accurately estimated by assuming
each kg of body weight contains 75 ml of blood [68],(BF
). The
absorbed blood dose per fraction was then estimated to range
between 0.159 and 0.291 Gy and this was equivalent to a cumulative
dose of 5.8728.715 Gy at IMRT treatment completion (see supplemen-
tary Table1). Although lymph nodes are in proximity ofvascularized tis-
sue, they do not circulate blood, instead they circulate lymph uid. BF
was therefore likely to overestimate the lymph node absorbed dose.
The uid shift within lymph nodes is approximated to ~4 L per day
[69] which may be negligible during treatment. To model the lymph
node exposure, the static absorbed blood dose was calculated (BF
using similar principles, subsequently scaled down to lymph node
planned volume. Lymph node absorbed blood dose per fraction
represents 0.0040.015 Gy which was equivalent to 0.1520.508 Gy
by IMRT completion (patient individualised schedule). This BF
was likely to be a more accurate representation of static PBL irradiation
in this tissue that may then be ltered to the circulatory system. By com-
bining thetwo models, the cumulative absorbed blood dose to high dose
volume likely was between 1.131 and 2.470 Gy by completion of
individualised IMRT schedule. Neither of these models consider the
low dose regions that are also exposed during treatment; therefore, so
both will underestimate the total blood dose. The original work of
Moquet and colleagues [32] was relatively simple and requires further
validation to ascertain exposure circumstances in which such models
can be applied. In this study, patient specic data was utilized to rene
these models, with limited success. The uncertainties in CTPV absorbed
blood dose were derived by propagation of the uncertainties associated
with factors in Eqs. (1) and (2). For CTPV physical absorbed blood dose
these were the planned volume (considered minimal at 11.5% varia-
tion), the dose to planned volume and the estimated scaling factors.
The dose to the plan region was expected to have an uncertainty of
less than 3%, based on published estimates of the upper bound of
inter-treatment dosimetric uncertainty [70] and greater errors would
be detected by treatment QA and trigger re-planning and re-
validation. To estimate the whole-body volume, it was assumed that 1.
01 g of human body mass ts within 1 cm
and that the tissue density
within the target volume was consistent with this for the scaling factor
estimation, with an error within 1%. The total error associated with CTPV
estimations was considered to be 6%. It is important to note, however,
that the estimates of uncertainty are themselves uncertain and further
work is needed here to better understand how patient specicdata
can contribute to such estimates. In addition, more detail with respect
to beam-on times are required to improve theBF dose model, in partic-
ular, for larger dose volumes, therefore, despite its use previously, this
model was not considered further.
To enable the absorbed blood dose estimation by dicentric assay in
this mixed (
Ra and IMRT) exposure scenario, the absorbed blood
dose ratio between each source was estimated per treatment cycle.
This then facilitated use of the criticalitydose estimation technique,
originally designed to separate and quantify neutron and gamma expo-
sures following a nuclear accident or incident [35], but here used to as-
sess the IMRT and α-particle absorbed doses, on the basis of either the
Fig. 4. Comparison of
Ra dose estimates. M-FISH
ratio was implemented in the dicentric assay estimates of blood dose for n= 5 patients (patient ID 15), the estimates for these
patients were compared across all other models. CTPV
and CTPV
calculated fromthe dicentric assay utilizing the physical blood dose estimates by
Ra and respective IMRT models.
Simple linear regression plotted for CTPV
Y = 0.019 X + 0.004, R
= 0.559 and for CTPV
Y = 0.115 X + 0.025, R
= 0.600. Physical
Ra dose estimated through clearance of
Ra from circulation 24 h post administration. Simple linear regression plotted for Y = 0.028 X + 0.012, R
= 0.301. The error associated with the physical
Ra was conservatively
estimated to 50% and propagated to CTPV
. The error for M-FISH
dose was propagated as 30% for C2 and 14% for C3C5.
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
treatment planning information (CTPV ratios), or the categorisation of
complexity of aberrations observed using M-FISH. The dicentric assay
was applied to two separate models, the rst utilizing the ratio derived
from physical models,
Ra clearance and IMRT CTPV models, and the
second utilizing the M-FISH
ratio based on the categorisation of
high LET to low LET exposed cells. The absorbed blood dose estimations
were carried out for 5 patients for which blood samples were drawn and
the dicentric equivalent frequency estimated. The CTPV derived ratios
assume that IMRT induced aberrations accumulated in the circulatory
blood pool by C3 are not cleared from the peripheral pool in the follow-
ing C4 and C5 treatment cycles and similarly for the
Ra, it was as-
sumed the aberrations accumulate through treatment with no clear-
ance. This results in a plateau of IMRT dose for CTPV
as the dose ratio
was based on a large IMRT component with a small
Ra dose. For
, as the model was based on high dose regions areas, the IMRT es-
timated dose was lower than CTPV
, therefore when expressed as a ratio
Ra, this assumes a higher proportion of absorbed blood dose to
be attributed to
Ra, which decreases the estimated dose for C4 and
C5. The dicentric assay was estimated from the ratio of physical
and CTPV dose estimates with a conservative error of 50% being attrib-
uted to the
Ra dose, this was considered to be the largest uncertainty.
Future studies will aim to reduce this uncertainty by increasing the
number of patients in the study and by including later timepoints.
model was based on the chromosomal aberration
spectrum in PBL sampled 4 weeks after each [
administration. Following IMRT completion, the dose was estimated at
1.167 ± 0.092 Gy. This was based on the assumption that all IMRT in-
duced aberrations would be of simple type while
Ra aberrations
were of complex type. As a larger proportion of cells containing complex
aberrations than simple chromosomal aberrations was observed, this
was reected in the dose ratio. Accordingly, the dose attributed to
Ra was proportionally larger than that attributed to IMRT. The result-
ing IMRT absorbed blood dose estimated by M-FISH
was therefore
lower than both CTPV dose estimates. The physically derived
Ra esti-
mates were representative of the period taken for
Ra to clear from the
blood. Due to the sampling schedule being every 4 weeks, it cannot be
excluded that absorbed dose from
Ra was also received by circulating
PBLs in the vicinity of metastatic sites, especially as metastatic sites tend
to be highly vascularized [7173]. The M-FISH
absorbed blood dose
estimates may better account for this, with the
Ra dose by C5
estimated to be 0.665 ± 0.080 Gy. This estimate was signicantly
larger than CTPV
0.024 ± 0.027 Gy and CTPV
0.141 ± 0.042 Gy. The
largest uncertainty in the M-FISH
absorbed blood dose estimation
was found to be also in C2 whereby the variation in the frequency of
cells consistent with IMRT exposure(cells containing simple aberrations
only) was up to 13% and for those consistent with
Ra exposure of up
to 28%. The error from the calibration curve used was estimated to be12
and 23% for IMRTand
Ra absorbed blood doseestimates, respectively.
The total propagated error on absorbed dose was within 30% for the M-
derived estimates.
In this study, the complexity of chromosome exchange observed in
PBLs was used as a biomarker of radiation quality [36] from which to
make estimates of absorbed blood dose ratio, termed the M-FISH
Based upon principles of radiation track structure and PBL cell geometry
[36,74], all cells which contained at least one complex chromosome ex-
change [75,76] were categorised as having been traversed by high LET
α-particles emitted from the
Ra, while damaged cells containing
only simple chromosome exchanges (reciprocal translocation, dicen-
trics and rings) were categorised as being exposed to low LET radiation
from IMRT. Although in vitro studies do show the majority of high-LET
induced damage to result in complex chromosome aberrations largely
independent of dose [7781] it is also the case that the simple ex-
changes can be directly induced after α-particles of lower incident LET
[74] and, exposure to low-LET radiation will result in the formation of
complex exchanges, in a manner strongly dependent upon dose [50,
82,83]. For instance, an increasing fraction of complex exchanges of up
to 2040% have previously been attributed to exposure of a large target
eld in IMRT treated prostate cancer patients [84,85]. Therefore, it is
likely that IMRT absorbed blood dose may be underestimated using
the M-FISH
reported here. Given the potential usefulness of this
ratio in cases of unknown exposures where physical information is not
available, further work to determine frequencies of complex exchange
occurrence in IMRT only and
Ra only treated patients is required.
As the M-FISH
absorbed dose ratio and the dicentric
quantication was based on cytogenetic observations of sampled PBL,
the resulting doses estimated will be directly affected by haematopoietic
cell death and repopulation dynamics. The IMRT dose estimates were
found to plateau between C3C5 suggesting cells containing unstable
aberrations remain over the time course studied. However, IMRT has
been shown to signicantly decrease the number of PBLs in circulation,
and therefore the clearance and repopulation dynamics should also be
taken into consideration for the CTPV models [25,86,87]. White blood
cell counts have been found to increase within 68 weeks of therapy
completion with a signicant increase in lymphocyte population after
3 months [26,88]. The 4-week period between C3C4 (after end of
IMRT) could provide a sufcient break for haematopoietic cells to
boost PBL repopulation, which if the case, would have the effect of dilut-
ing the frequency of persisting unstable chromosomal events. Newly in-
duced aberrations by subsequent
Ra treatment cycles would then
add to this aberrant cell pool. There is also the potential of bystander re-
sponses playing a role both in cell turnover and aberration formation
[89]. Further work on blood samples representative of later [
administrations (C6) and follow up samples (up to 1 year post
start of treatment), together with patient hematological counts, will
help elucidate the cellular dynamics of damaged PBL. An assessment of
the occurrence and type of stable chromosome exchange from these
samples will also offer the potential to make estimates of absorbed
dose delivered to the bone marrow.
In conclusion, in this study we have evaluated a number of absorbed
blood dose methods for mixed eld exposure, in a unique population of
patients receiving external beam photons and a systemically delivered
Ra. We highlight key observations and limitations to estab-
lish an approach from which we can make dose assessments to better
understand mixed eld exposures. The models presented provide an
initial estimation of cumulative absorbed dose received to the blood
during incremental IMRT fractions and [
injections, all of
which move towards assessing patient specic dose information for
mixed eld treatment to help optimise treatment outcomes and
minimise patient risk in the future.
This project received funding from Public Health England, as part of
the PHE PhD studentship scheme. This work was supported by the
Movember Prostate Cancer UK Centre of Excellence (CEO13_2-004)
and the Research and Development Division of the Public Health Agency
of NI (COM/4965/14).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
[1] Halabi S, Small EJ, Kantoff PW, Kattan MW, Kaplan EB, Dawson NA, et al. Prognostic
model for predicting survival in men with hormone-refractory metastatic prostate
cancer. J Clin Oncol. 2003;21(7):12327.
[2] Parker C, Finkelstein SE, Michalski JM, O'Sullivan JM, Bruland Ø, Vogelzang NJ, et al.
Efcacy and safety of Radium-223 dichloride in symptomatic castration-resistant
prostate cancer patients with or without baseline op ioid use from the phase 3
ALSYMPCA trial. Eur Urol. 2016;70(5):87583.
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
[3] Parker C, Nilsson S, Heinrich D, Helle SI, O'Sullivan JM, Fosså SD, et al. Alpha emitter
radium-223 and survival in metastatic prostate cancer. N Engl J Med. 2013;369(3):
[4] Parker C, Heinrich D, O'Sullivan JM, Fossa SD, Chodacki A, Demkow T, et al. Overall
survival benet and safety prole of radium-223 chloride, a rst-in-class alpha-
pharmaceutical: results from a phase III randomized trial (ALSYMPCA) in patients
with castration-resistant prostate cancer (CRPC) with bone metastase s. J Clin
Oncol. 2012;30(5_suppl):8.
[5] Xogo.; 2018.
[Accessed 17 February 2021]..
[6] Postmarket Drug and Biologic Safety Evaluations. Completed from January 2017
March 2017. rveillance/p ostmarket-drug-and-
biologic-safety-evaluations-completed-january-2017-march-2017; 2017. [Accessed
17 February 2021]..
[7] Nilsson S, ProfFranzén L, MD Parker C, Dr Tyrrell C, FRCRBlom R, MD Tennvall J, Prof.
Bone-targeted radium-223 in symptomatic, hormone-refractory prostate cancer: a
randomised, multicentre, placebo-controlled phase II study. Lancet Oncol. 2007;8
[8] Nilsson S, Strang P, Aksnes AK, Franzèn L, Olivier P, Pecking A, et al. A randomized,
doseresponse, multicenter phase II study of radium-223 chloride for the palliation
of painful bonemetastases in patientswith castration-resistant prostate cancer. EurJ
Cancer. 2012;48(5):67886.
[9] Nickoloff JA, Sharma N, Taylor L. Clustered DNA double-strand breaks: biological ef-
fects and releva nce to cancer radi otherapy. Genes. 2020;11(1):99 10.3390%
[10] Bannik K, Madas B, Jarzombek M, Sutter A, Siemeister G, Mumberg D, et al. Radiobi-
ological effects of the alpha emitter ra-223 on tumor cells. Sci Rep. 2019;9(1):111.
[11] Ritter MA, Tobias CA, Cleaver JE. High-LET radiations induce a large proportion of
non-rejoining DNA breaks. Nature. 1977;266(5603):6535. htt ps://
[12] Nilsson S, Larsen RH, Fosså SD, Balteskard L, Borch KW, Westlin J, et al. First clinical
experience with α-emitting Radium-223 in the treatment of skeletal metastases.
Clin Cancer Res. 2005;11(12):44519.
[13] Henriksen G, Breistøl K, Bruland ØS, Fodstad Ø, Larsen RH. Signicant antitumor ef-
fect from bone-seeking, alpha-particle-emitting (223)Ra demonstrated in an exper-
imental skeletal metastases model. Cancer Res. 2002;62(11):31205.
[14] Pandit-Taskar N, Larson SM, Carrasquillo JA. Bone-seeking radiopharmaceuticals for
treatment of osseous metastases, part 1: therapy with 223Ra-dichloride. J NuclMed.
[15] Heinrich D, Bektic F, Bergman AM, Caffo O, Cathomas R, Chi KN, et al. The contempo-
rary use of Radium-223 in metasta tic castratio n-resistant pr ostate cancer. Clin
Genitourin Cancer. 2017;16(1):22331.
[16] Kerr C. (223)Ra targets skeletal metastases and spares normal tissue. Lancet Oncol.
2002.;453. Sect. 3.
[17] Flux GD. Imaging and dosimetry for radium-223: the potential for personalize d
treatment. Br J Ra diol. 2017;90(1077):20160748.
[18] Pacilio M, Ventroni G, De Vincentis G, Cassano B, Pellegrini R, Di Castro E, et al. Do-
simetry of bone metastases in targeted radionuclide therapy with alpha-emitting
(223)Ra-dichloride. Eur J Nucl Med Mol Imaging. 2016;43(1):2133. https://doi.
[19] Lassmann M, Nosske D. Dosimetry of 223Ra-chloride: dose to normal organsand tis-
sues. Eur J Nucl Med Mol Imaging. 2013;40(2):20712.
[20] Hobbs RF, Song H, Watchman CJ, Bolch WE, Aksnes A,Ramdahl T, et al. A bone mar-
row toxicity model for 223Ra alpha-emitter radiopharmaceutical therapy. Phys Med
Biol. 2012;57(10):320722.
[21] Leung CN, Canter BS, Rajon D, Bäck TA, Fritton JC, Azzam EI, et al. Dose-dependent
growth delay of breast cancer xenografts in the bone marrow of mice treated with
223Ra: the role of bystander effects and their potential for therapy. J Nucl Med.
[22] FDC Guerra-Liberal, O'Sullivan JM, SJ McMahon, Prise KM. Targeted alpha therapy:
current clinical applications. 2020;35(6):40417.
3576 June 16.
[23] Moghaddam FF, Bakhshandeh M, Ghorbani M, Mod B. Assessing the out-of-eld
dose calculation accuracy by eclipse treatment planning system in sliding window
IMRT of prostate cancer patients. Comput Biol Med. 2020.;127:104052. https://doi.
[24] Doležel M, Odrážka K, Vaňásek J, Štuk J, HlávkaA, Vítková M, et al. Long-term clinical
results of IGRT in prostate cancer treatment. Klin Onkol. 2020;33(1):4954https://
[25] Miszczyk M, Majewski W. Hematologic toxicity of conformal radiotherapy and in-
tensity modulated radiotherapy in prostate and bladder cancer patients. Asian Pac
J Cancer Prev. 2018;19(10):28036
[26] Sini C, Fiorino C, PernaL, Noris Chiorda B, Deantoni CL, Bianchi M, et al. Dose-volume
effects for pelvic bone marrow in predicting hematological toxicity in prostate can-
cer radiotherapy with pelvic node irradiation. R adiother Oncol . 2016;118(1) :
[27] Quan EM, Li X, Li Y, Wang X, Kudchadker RJ, Johnson JL, et al. A comprehensive com-
parison of IMRT and VMAT plan quality for prostate cancer treatment. Int J Radiat
Oncol Biol Phys. 2012;83(4):116978.
[28] Teoh M, Clark CH, Wood K, Whitaker S, Nisbet A. Volumetric modulated arctherapy:
a review of current literature and clinical use in practice. Br J Radiol.2011;84(1007):
[29] Gershkevitsh E, Clark C, Staffurth J, Dearnaley D, Trott K. Dose to bone marrow using
IMRT techniques in prostate cancer patients . Strahlenther Onkol. 2005;1 81(3):
[30] Vlachaki MT, Teslow TN, Amosson C, Uy NW, Ahmad S. IMRT versus conventional
3DCRT on prostate and normal tissue dosimetry using an endorectal balloon for
prostate immobilization. Med Dosim. 2005;30(2):6975.
[31] International Commission on Radiological Protection. Annals of the ICRP radiological
protection in medicine. Elsevier; 2007..
[32] MoquetJ, Higueras M, Donovan E, BoyleS, Barnard S, Bricknell C, et al.Dicentric dose
estimates for patients undergoing radiotherapy in the RTGene study to assess blood
dosimetric models and the new bayesian method for gradient exposure. Radiat Res.
2018;190(6):596 (10.1667%2FRR15116.1).
[33] Taprogge J, Murray I, Gear J, Chittenden SJ, Parker CC, Flux GD. Compartm ental
model for 223Ra-dichloride in patients with metastatic bone disease from
castration-resistant prost ate cancer. Int J Radiat Oncol Biol Ph ys. 2019;105(4):
[34] Kulka U, Wojcik A. Special issue: networking in biological and EPR/OSL dosimetry:
the europeanRENEB platform for emergency preparedness and research. Int J Radiat
Biol. 2017;93(1):1.
[35] International Atomic Energy Agency. Cytogenetic dosimetry: applications in pre-
paredness for and response to radiation emergencies. Vienna: IAEA; 2011..
[36] Anderson RM. Cytogenetic biomarkers of radiation exposure. Clin Oncol. 2019;5:
[37] Turner PG, Jain S, Cole A, Grey A, Mitchell D, Prise KM, et al. Toxicity and efcacy of
concurrent androgen deprivation therapy, pelvic radiotherapy, and radium-223 in
patients with de novo metastatic hormone-sensitive prostate cancer. Clin Cancer
Res. 2021.. Epub ahead of print.
[38] Zimmerman BE, Bergeron DE, Cessna JT, Fitzgerald R, Pibida L. Revision of the NIST
standard for (223)Ra: new measurements and review of 2008 data. J Res Natl Inst
Stand Technol. 2015;120:3757 (10.6028%2Fjres.120.004).
[39] Yoshida K,Kaneta T, Takano S, Sugiura M, Kawano T, Hino A, et al. Pharmacokinetics
of single dose radium-223 dichloride (BAY 888223) in japanese patients with
castration-resistant prostate cancer and bone metastases. Ann Nucl Med. 2016;7:
45360 (10.6028%2Fjres.120.004).
[40] Chittenden SJ, Hindorf C, ParkerCC, Lewington VJ, Pratt BE, Johnson B, et al. A phase
1, open-label s tudy of the biodistribution, pharmacokinetics, and dosim etry of
223Ra-dichloride in patients with hormone-refractory prostate cancer and skeletal
metastases. J Nucl Med. 2015;56(9):13049.
[41] Carrasquillo J, ODonoghue J, Pandit-Taskar N, Humm J, Rathkopf D, Slovin S, et al.
Phase I pharmacokinetic and biodistribution study with escalating doses of 223Ra-
dichloride in men with castration-resistant metastatic prostate cancer. Eur J Nucl
Med Mol Imaging. 20 13;40(9):138493.
[42] Attix FH. Introduction to radiological physics and radiation dosimetry. 1. Au. ed.
Wiley-VCH; 2008..
[43] Schumann S, Eberlein U, Muhtadi R, Lassmann M, Scherthan H. DNA damage in leu-
kocytes after internal ex-vivo irradiationof blood with the α-emitter ra-223. SciRep.
[44] Human body weight to volume conversion. Internet.
calculate/weight-to-volume/substance/human-blank-body; 2020. 67. Ann. ICRP.
[45] Sumption ND, Goodhead DT, Anderson RM. No increase in radiation-induced chro-
mosome aberration complexity detected by m-FISH after culture in the presence
of 5'-bromodeoxyuridine. Mutat Res. 2006;594(12):308.
[46] Perry P, Wolff S. New giemsa method for the differential staining of sister chroma-
tids. Nature. 1974;251(5471):1568.
[47] Themis M, Garimberti E, Hill MA, Anderson RM. Reduced chromosome aberration
complexity in normal human bronchial epithelial cells exposed to low-LET γ-rays
and high-LET α-particles. Int J Radiat Biol. 20 13;89(11):93443 (10.3109 %
[48] Savage JR, Simpson PJ. FISH "painting" patterns resulting from complex exchanges.
Mutat Res. 1994;312(1):5160.
[49] Anderson RM, Marsden SJ, Paice SJ, Bristow AE, Kadhim MA, Grifn CS, et al. Trans-
missible and nontransmissible complex chromosome aberrations characterized by
three-color and mFISH dene a biom arker of exposure to high-LET αparticles.
Radiat Res. 2003;159(1):408.[0040:
TANCCA]2 .0.CO;2.
[50] Cornforth MN. Analyzing radiation-induced complex chromosome rearrangements
by combinatorial painting. Rad iat Res. 2001;1 55(5):64359.
[51] Boei JJ, Natarajan AT. Combined use of chromosome painting and telomere de-
tection to analyse radiation-induced chromosomal aberrations in mouse
splenocytes. Int J Radiat Biol. 1998;73(2):12533.
[52] Lloyd DC, Edwards AA, Prosser JS. Chromosome aberrations induced in human lym-
phocytes by in vitro acute X and gamma radiation. Radiat Prot Dosimetry. 1986;15
[53] Ainsbu ry EA, Lloyd D. Dose esti mati on sof twar e for ra diat ion bi odos imet ry.
Health Phys. 2010;98(2):2905.
[54] Purrott RJ, Edwards AA, Lloyd DC, Stather JW. The induction of chromosome
aberrations in human lymphocytes by in vitro irradiation with α-particles
from Plutonium-239. Int J Radiat Biol. 1980;38(3):27784.
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
[55] Sigurdson AJ, Ha M, Hauptmann M, Bhatti P, Sram RJ, Beskid O, et al. International
study of factors affecting human chromosome translocations, 6 52. ; 2008. p.
11221. 2.
[56] Whitehouse C, Edwards A, Tawn E, Stephan G, Oestreicher U, Moquet J, et al. Trans-
location yields in peripheral blood lymphocytes from control pop ulations. Int J
Radiat Biol. 2005;81(2):13945.
[57] Coleman R, Brown J, Rathbone E, Flanagan L, Reid A, Kendall J, et al. CApecitabine
plus Radium-223 (Xogo) in breast cancer patients with BONe metastases (CAR-
BON): study protocol for a phase IB/IIA randomised controlled trial. Trials. 2020;21
[58] Geva R, Lopez J, Danson S, Joensuu H, Peer A, Harris SJ, et al. Radium-223 in combi-
nation with paclitaxel in cancer patients with bone metastases: safety results from
an open-label, multicenter phase ib study. Eur J Nucl Med Mol Imaging. 2018;46
[59] Huynh-Le M, Shults RC, Connor MJ, Hattangadi-Gluth JA. Adverse events associated
with Radium-223 in metastatic prostate cancer: disproportionality analysis of FDA
data reecting worldwide utilization. Clin Genitourin Cancer. 2020;18(3):192200. e2.
[60] Soldatos TG, Iakovou I, Sachpekidis C. Retrospective toxicolog ical proling of
Radium-223 dichloride for the treatment of bone metastases in prostate cancer
using adverse event data. Medicina (Kaunas). 2019.;55(5).
[61] Jacene H, Gomella L, Yu EY,Rohren EM. Hematologic toxicity from Radium-223 ther-
apy for bonemetastases in castration-resistant prostatecancer: risk factors and prac-
tical considerations. Clin Genitourin Cancer. 2018;16(4):e91926.
[62] Vark aris A, Gunturu K, Kewalramani T, Tre tter C. Acute myeloid leukemia afte r
Radium-223 therapy: case report. Clin Genit ourin Cancer. 2017;15( 4):e7236.
[63] Odo U, Vasudevamurthy AK, SartorO. Acute promyelocytic leukemia after treatment
of metastatic castration-resistant prostate cancer with Radium-223. Clin Genitourin
Cancer. 2017;15(3):e5012.
[64] Abou DS, Ulmert D, Doucet M, Hobbs RF, Riddle RC, Thorek DLJ. Whole-body and mi-
croenvironmental localization of Radium-223 in Naïve and mouse models of pros-
tate cancer meta stasis. J Natl Cancer Inst. 2016;108(5):1 (10. 1093%2Fjnci%
[65] Stephan G, Kampen WU, Noßke D, Roos H. Chromosomal aberrations in peripheral
lymphocytes of patients treated with radium-224 for ankylosing spondylitis. Radiat
Environ Biophys. 2005;44:238.
[66] Schumann S, Eberlein U, LapaC, Müller J, Sering S, Lassmann M, et al. α-particle-in-
duced DNA damage tracks in peripheral blood mononuclear cells of [223Ra]RaCl2-
treated prostate cancer patients. Eur J Nucl Med Mol Imaging. 2021;48(9):
[67] ICRP. Age-dependent doses to members of the public from intake of radionuclides -
part 2 ingestion dose coefcients. ICRP Publication; 1993. p. 34 67. Ann. ICRP.
[68] Hilberath JN, ThomasME, Smith T, Jara C, Fitzgerald DJ, Wilusz K, et al. Blood volume
measurement by hemodilution: association with valve disease and re-evaluation of
the Allen formula. Perfusion. 201 5;30(4):305 11. https://doi.o rg/10.1177/
[69] Moore JE, Bertram CD. Lymphatic system ows. Annu Re v Fluid Mech. 2018;50:
45982. 16-045259.
[70] Thwaites D. Accuracy required and achievable in radiotherapy dosimetry: ha ve
modern technology and techniques changed o ur views? J Phys Conf Ser. 2013.;
[71] Bielenberg DR, Zetter BR. The contribution of angiogenesis to the process of
metastasis. Cancer J. 2015;21(4):26773.
[72] Chavez-MacGregor M, Aviles-Salas A, Green D, Fuentes-Alburo A, Gómez-Ruiz C,
Aguayo A. Angiogenesis in the bone marrow of patients with breast cancer. ClinCan-
cer Res. 2005;11(15):5396400.
[73] Mundy GR. Metastasis to bone: causes, consequences and therapeutic opportunities.
Nat Rev Cancer. 2002;2(8):58493.
[74] Anderson RM, StevensDL, Sumption ND, Townsend KMS, Goodhead DT, Hill MA. Ef-
fect of linear energy transfer(LET) on the complexity of α-particle-induced chromo-
some aberrations in human CD34+ cells. Radiat Res. 2007;167(5):54150. https://
[75] Savage JRK, Simpson PJ. Fish paintingpatterns resulting from complex exchanges.
Mutat Res. 1994;312(1):5160.
[76] Savag e JRK, Simpson P. On th e scoring of FISH-paintedchromosome-type ex-
change aberrations. Mutat Res Fundam Mol Mech Mutagen. 1994;307(1):34553.
[77] Barquinero JF, Stephan G, Schmid E. Effect of americium-241 α-particles on the
doseresponse of chromosome aberrations in human lymphocytes analysed by uo-
rescence in situ hybridization. Int J Radiat Biol. 2004;80(2):15564.
[78] Moquet JE, Fernández JL, Edwards AA, Lloyd DC. Lymphocyte chromosomal aberra-
tions and their complexity induced in vitro by plutonium-239 alpha-particles and
detected by FISH. Cell Mol Biol (Noisy-le-Grand). 2001;47(3):54956.
[79] Anderson RM, Marsden SJ, Wright EG, Kadhim MA, Goodhead DT, Grifn CS. Com-
plex chromosome aberrations in peripheral blood lymphocytes as a potential bio-
marker of exposure to high-LET alpha-particle s. Int J Radiat Biol. 2000;76(1):
[80] Testard I, Dutrillaux B, Sabatier L. Chromosomal aberrations induced in human lym-
phocytes by high-LET irradiation. Int J Radiat Biol. 1997;72(4):42333. https://doi.
[81] Grifn CS, Marsden SJ, Stevens DL, Simpson P, Savage JRK. Frequencies of complex
chromosome exchange aberrations induced by 238Pu α-particles and detected by
uorescence in situ hybridization using single chromosome-specic probes. Int J
Radiat Biol. 1995;67(4):4319.
[82] Loucas BD, Durante M, Bailey SM, Cornforth MN. Chromosome damage in human
cells by γrays, αparticles and heavy ions: track interactions in basic dose-
response relationships. Radiat Res. 2013;179(1):920 (10.1667%2FRR3089.1).
[83] Loucas BD, Cornforth MN. Complex chromosome exchanges induced by gamma rays
in human lymphocytes: an mFISH study. Radiat Res. 2001;155(5):66071. https://[0660:cceibg];2.
[84] Pignalosa D, Lee R, Hartel C, Sommer S, Nikoghosyan A, Debus J, et al. Chromosome
inversions in lymphocytes of prostate cancer patients treated with X-rays and car-
bon ions. Radiother Oncol. 2013;109(2):25661.
[85] Hartel C, Nikoghosyan A, Durante M, Sommer S, Nasonova E, Fournier C, et al. Chro-
mosomal aberrations in peripheral blood lymphocytes of prostate cancer patients
treated with IMRT and carbon ions. Radiother Oncol. 2009;95(1):738. https://doi.
[86] Sanguineti G, Giannarelli D, Petrongari MG, Arcangeli S, Sangiovanni A, Saracino B,
et al. Leukotoxicity after moderately hypofractionated radiotherapy versus conven-
tionally fractionateddose escalated radiotherapy for localized prostatecancer: a sec-
ondary analysis from a randomized study. Radiat Oncol. 2019.;14.
[87] Landoni V, Fiorino C, Cozzarini C, Sanguineti G, Valdagni R,Rancati T. Predictingtox-
icity in radiotherapy for prostate cancer. Phys Med. 2016;32(3):52132. https://doi.
[88] Pinkawa M, Djukic V, Klotz J, Petz D, Piroth MD, Holy R, et al. Hematologic changes
during prostate cancer radiation therapy are dependent on the treatment volume.
Future Oncol. 2014;10(5):83543.
[89] Canter BS, Leung CN, Fritton JC, BäckT, Rajon D, Azzam EI, et al.Radium-223-induced
bystander effects cause DNA damage and apoptosis in disseminated tumor cells in
bone marrow. Mol Cancer Res. 2021..
0005 Epub ahead of print.
I. Bastiani, S.J. McMahon, P. Turner etal. Nuclear Medicine and Biology 106107 (2022) 1020
... The research papers cover studies in cytogenetics after radium-223 treatment [9], the role of antioxidants in the radiosensitivity to [ 177 Lu]Lu-DOTATATE [10], relating the biological effects of radionuclides such as Auger electron-emitters gallium-67 and indium-111 to that of EBRT [11] and determining the DNA damaging capacity of Auger electron-emitter thallium-201 [12]. The future of Auger electron-emitting radiopharmaceuticals was also discussed in a commentaries [13]. ...
Full-text available
Purpose One therapy option for prostate cancer patients with bone metastases is the use of [ ²²³ Ra]RaCl 2 . The α-emitter ²²³ Ra creates DNA damage tracks along α-particle trajectories (α-tracks) in exposed cells that can be revealed by immunofluorescent staining of γ-H2AX+53BP1 DNA double-strand break markers. We investigated the time- and absorbed dose-dependency of the number of α-tracks in peripheral blood mononuclear cells (PBMCs) of patients undergoing their first therapy with [ ²²³ Ra]RaCl 2 . Methods Multiple blood samples from nine prostate cancer patients were collected before and after administration of [ ²²³ Ra]RaCl 2 , up to 4 weeks after treatment. γ-H2AX- and 53BP1-positive α-tracks were microscopically quantified in isolated and immuno-stained PBMCs. Results The absorbed doses to the blood were less than 6 mGy up to 4 h after administration and maximally 16 mGy in total. Up to 4 h after administration, the α-track frequency was significantly increased relative to baseline and correlated with the absorbed dose to the blood in the dose range < 3 mGy. In most of the late samples (24 h – 4 weeks after administration), the α-track frequency remained elevated. Conclusion The γ-H2AX+53BP1 assay is a potent method for detection of α-particle-induced DNA damages during treatment with or after accidental incorporation of radionuclides even at low absorbed doses. It may serve as a biomarker discriminating α- from β-emitters based on damage geometry.
Full-text available
Cells manage to survive, thrive, and divide with high accuracy despite the constant threat of DNA damage. Cells have evolved with several systems that efficiently repair spontaneous, isolated DNA lesions with a high degree of accuracy. Ionizing radiation and a few radiomimetic chemicals can produce clustered DNA damage comprising complex arrangements of single-strand damage and DNA double-strand breaks (DSBs). There is substantial evidence that clustered DNA damage is more mutagenic and cytotoxic than isolated damage. Radiation-induced clustered DNA damage has proven difficult to study because the spectrum of induced lesions is very complex, and lesions are randomly distributed throughout the genome. Nonetheless, it is fairly well-established that radiation-induced clustered DNA damage, including non-DSB and DSB clustered lesions, are poorly repaired or fail to repair, accounting for the greater mutagenic and cytotoxic effects of clustered lesions compared to isolated lesions. High linear energy transfer (LET) charged particle radiation is more cytotoxic per unit dose than low LET radiation because high LET radiation produces more clustered DNA damage. Studies with I-SceI nuclease demonstrate that nuclease-induced DSB clusters are also cytotoxic, indicating that this cytotoxicity is independent of radiogenic lesions, including single-strand lesions and chemically “dirty” DSB ends. The poor repair of clustered DSBs at least in part reflects inhibition of canonical NHEJ by short DNA fragments. This shifts repair toward HR and perhaps alternative NHEJ, and can result in chromothripsis-mediated genome instability or cell death. These principals are important for cancer treatment by low and high LET radiation.
Full-text available
Background: A substantial proportion of breast cancer patients develop metastatic disease, with over 450,000 deaths globally per year. Bone is the most common first site of metastatic disease accounting for 40% of all first recurrence and 70% of patients with advanced disease develop skeletal involvement. Treatment of bone metastases currently focusses on symptom relief and prevention and treatment of skeletal complications. However, there remains a need for further treatment options for patients with bone metastases. Combining systemic therapy with a bone-targeted agent, such as radium-223, may provide an effective treatment with minimal additional side effects. Methods/design: CARBON is a UK-based, open-label, multi-centre study which comprises an initial safety phase to establish the feasibility and safety of combining radium-223 given on a 6-weekly schedule in combination with orally administered capecitabine followed by a randomised extension phase to further characterise the safety profile and provide preliminary estimation of efficacy. Discussion: The CARBON study is important as the results will be the first to assess radium-223 with chemotherapy in advanced breast cancer. If the results find acceptable rates of toxicity with a decrease in bone turnover markers, further work will be necessary in a phase II/III setting to assess the efficacy and clinical benefit. Trial registration: ISRCTN, ISRCTN92755158, Registered on 17 February 2016.
Full-text available
Targeted alpha therapy is an emerging innovative approach for the treatment of advanced cancers, in which targeting agents deliver radionuclides directly to tumors and metastases. The biological effects of α-radiation are still not fully understood - partly due to the lack of sufficiently accurate research methods. The range of α-particles is <100 μm, and therefore, standard in vitro assays may underestimate α-radiation-specific radiation effects. In this report we focus on α-radiation-induced DNA lesions, DNA repair as well as cellular responses to DNA damage. Herein, we used Ra-223 to deliver α-particles to various tumor cells in a Transwell system. We evaluated the time and dose-dependent biological effects of α-radiation on several tumor cell lines by biological endpoints such as clonogenic survival, cell cycle distribution, comet assay, foci analysis for DNA damage, and calculated the absorbed dose by Monte-Carlo simulations. The radiobiological effects of Ra-223 in various tumor cell lines were evaluated using a novel in vitro assay designed to assess α-radiation-mediated effects. The α-radiation induced increasing levels of DNA double-strand breaks (DSBs) as detected by the formation of 53BP1 foci in a time- and dose-dependent manner in tumor cells. Short-term exposure (1–8 h) of different tumor cells to α-radiation was sufficient to double the number of cells in G2/M phase, reduced cell survival to 11–20% and also increased DNA fragmentation measured by tail intensity (from 1.4 to 3.9) dose-dependently. The α-particle component of Ra-223 radiation caused most of the Ra-223 radiation-induced biological effects such as DNA DSBs, cell cycle arrest and micronuclei formation, leading ultimately to cell death. The variable effects of α-radiation onto the different tumor cells demonstrated that tumor cells show diverse sensitivity towards damage caused by α-radiation. If these differences are caused by genetic alterations and if the sensitivity could be modulated by the use of DNA damage repair inhibitors remains a wide field for further investigations.
Purpose: Radium-223 is an alpha-emitting radionuclide associated with overall survival (OS) improvement in metastatic castration resistant prostate cancer (mCRPC). External beam radiotherapy (EBRT) to prostate extends OS in men with metastatic hormone sensitive prostate cancer (mHSPC) limited to <4 metastases. We hypothesised that combination radium-223 + pelvic EBRT could safely deliver maximal radiotherapy doses to primary and metastatic prostate cancer and may improve disease control. Experimental design: Thirty patients with de-novo bone metastatic mHSPC who had commenced androgen deprivation therapy (ADT) and docetaxel were recruited to this single-arm, open-label, prospective clinical trial: ADRRAD (Neo-adjuvant Androgen Deprivation Therapy, Pelvic Radiotherapy and RADium-223 for new presentation T1-4 N0-1 M1B adenocarcinoma of prostate). Study treatments were: ADT, 6 cycles of radium-223 q28 days, conventionally fractionated prostate radiotherapy (74 Gy) and simultaneous integrated boost to pelvic lymph nodes (60 Gy). Results: No grade 4/5 toxicity was observed. Three patients experienced grade 3 leucopenia and 1 each experienced grade 3 neutropenia and thrombocytopenia, all were asymptomatic. One patient each experienced grade 3 dysuria and grade 3 urinary infection. No grade 3 gastrointestinal toxicity was observed. On treatment completion, there was a signal of efficacy; 24 (80%) patients had whole-body MRI evidence of tumour response or stability. Twenty-seven (90%) patients showed a reduction in ALP compared to pre-treatment levels. Median progression free survival was 20.5 months. Conclusions: This is the first trial of combination ADT, radium-223 and EBRT to pelvis, post docetaxel. The combination was safe, with an efficacy signal. Multi-centre RCTs are warranted.
Radiation-induced bystander effects have been implicated in contributing to the growth delay of disseminated tumor cells (DTC) caused by 223RaCl2, an alpha particle–emitting radiopharmaceutical. To understand how 223RaCl2 affects the growth, we have quantified biological changes caused by direct effects of radiation and bystander effects caused by the emitted radiations on DTC and osteocytes. Characterizing these effects contribute to understanding the efficacy of alpha particle–emitting radiopharmaceuticals and guide expansion of their use clinically. MDA-MB-231 or MCF-7 human breast cancer cells were inoculated intratibially into nude mice that were previously injected intravenously with 50 or 600 kBq/kg 223RaCl2. At 1-day and 3-days postinoculation, tibiae were harvested and examined for DNA damage (γ-H2AX foci) and apoptosis in osteocytes and cancer cells located within and beyond the range (70 μm) of alpha particles emitted from the bone surface. Irradiated and bystander MDA-MB-231 and MCF-7 cells harbored DNA damage. Bystander MDA-MB-231 cells expressed DNA damage at both treatment levels while bystander MCF-7 cells required the higher administered activity. Osteocytes also had DNA damage regardless of inoculated cancer cell line. The extent of DNA damage was quantified by increases in low (1–2 foci), medium (3–5 foci), and high (5+ foci) damage. MDA-MB-231 but not MCF-7 bystander cells showed increases in apoptosis in 223RaCl2-treated animals, as did irradiated osteocytes. In summary, radiation-induced bystander effects contribute to DTC cytotoxicity caused by 223RaCl2. Implications This observation supports clinical investigation of the efficacy of 223RaCl2 to prevent breast cancer DTC from progressing to oligometastases.
Abstract Aim The objective of this study was to evaluate out-of-field dose distribution calculation accuracy by the Anisotropic Analytical Algorithm (AAA), version 13.0.26, in Eclipse TPS, (Varian Medical Systems, Palo Alto, Ca, USA) for sliding window IMRT delivery technique in prostate cancer patients. Materials and methods Prostate IMRT plans with nine coplanar were calculated with the AAA Eclipse treatment planning system. To assess the accuracy of dose calculation predicted by the Eclipse in normal tissue and OARs located out of radiation field areas, including the rectum, bladder, right and left head of the femur, absolute organ dose value, and dose distribution were measured using the Delta4+ IMRT phantom. Results In the out-of-field areas, underestimation of −0.66% in organs near the field edge to −39.63% in organs far from the field edge (2.5 and 7.3 cm respectively) occurred in the TPS calculations. The percentage of dose deviation for the femoral heads was 95.7 on average while for the organ closer to the target (rectum) it was 79.81. Conclusions AAA dosimetry algorithm (used in Eclipse TPS) showed poor dose calculation in areas beyond the treatment fields border where underestimation varies with the distance from the field edges. A significant underestimation was found for the AAA algorithm in the sliding window IMRT technique (P-value > 0.05).
α-Emitting radionuclides have been approved for cancer treatment since 2013, with increasing degrees of success. Despite this clinical utility, little is known regarding the mechanisms of action of α particles in this setting, and accurate assessments of the dosimetry underpinning their effectiveness are lacking. However, targeted alpha therapy (TAT) is gaining more attention as new targets, synthetic chemistry approaches, and α particle emitters are identified, constructed, developed, and realized. From a radiobiological perspective, α particles are more effective at killing cells compared to low linear energy transfer radiation. Also, from these direct effects, it is now evident from preclinical and clinical data that α emitters are capable of both producing effects in nonirradiated bystander cells and stimulating the immune system, extending the biological effects of TAT beyond the range of α particles. The short range of α particles makes them a potent tool to irradiate single-cell lesions or treat solid tumors by minimizing unwanted irradiation of normal tissue surrounding the cancer cells, assuming a high specificity of the radiopharmaceutical and good stability of its chemical bonds. Clinical approval of 223RaCl2 in 2013 was a major milestone in the widespread application of TAT as a safe and effective strategy for cancer treatment. In addition, 225Ac-prostate specific membrane antigen treatment benefit in metastatic castrate-resistant prostate cancer patients, refractory to standard therapies, is another game-changing piece in the short history of TAT clinical application. Clinical applications of TAT are growing with different radionuclides and combination therapies, and in different clinical settings. Despite the remarkable advances in TAT dosimetry and imaging, it has not yet been used to its full potential. Labeled 227Th and 225Ac appear to be promising candidates and could represent the next generation of agents able to extend patient survival in several clinical scenarios.
Background: The combination of intensity modulated radiation therapy (IMRT) and image guided radiotherapy (IGRT) plays a significant role in sparing normal tissue during prostate cancer treatment. We report the clinical outcomes of 260 patients treated with high-dose IGRT as well as the toxicity of high-dose IGRT in these patients. Materials and methods: From September 2008 to February 2012, 260 men with clinically localized prostate cancer underwent radical radiotherapy. Two hundred patients were treated with IMRT (78 Gy in 39 fractions) to the prostate and base of seminal vesicles using an adaptive protocol combining cone-beam computed tomography (CBCT) and kilovoltage image matching with individualized safety margin calculation. Sixty patients underwent treatment with the same prescribed dose using RapidArc with a reduced safety margin of 6 mm and daily online matching using CBCT. Late toxicity was scored prospectively according to the RTOG/FC-LENT scale. Results: Eighteen patients (6.9%) experienced acute grade 2 gastrointestinal toxicity. There was no acute grade 3 or 4 gastrointestinal toxicity. Thirty-nine patients (15%) experienced acute grade 2 genitourinary toxicity and 6 patients (2.3%) had grade 3 gerourinary toxicity. Genitourinary toxicity grade 4 was observed in 5 (1.9%) patients, due to installation of a urinary catheter. At a median follow up of 84.2 months, the estimated 7-year cumulative incidences of grade 2 gastrointestinal and genitourinary toxicity were 4.4 and 7.1% respectively. The estimated 7-year prostate specific antigen relapse free survival was 97.1% for low-risk disease, 83.6% for intermediate-risk disease and 75% for high-risk patients. Conclusion: The use of IMRT in combination with IGRT results in a low rate of late toxicity. The authors declare they have no potential conflicts of interest concerning drugs, products, or services used in the study. The Editorial Board declares that the manuscript met the ICMJE recommendation for biomedical papers. Submitted: 8. 9. 2019 Accepted: 25. 10. 2019.
Background: Radionuclide radium-223 improves survival in men with metastatic castrate-resistant prostate cancer. The United States (US) Food and Drug Administration Adverse Events Reporting System (FAERS) is a post-market pharmacovigilance database valuable for adverse event (AE) assessments. We analyzed FAERS to identify disproportionate AE signals related to radium-223, and to explore radium-223's international utilization. Materials and methods: We identified 2182 radium-223 cases associated with AE(s) from 2013 to 2018. The duration of radium-223 therapy was calculated. Reporting odds ratio (ROR) and proportional reporting ratio (PRR), with 95% confidence intervals (CIs), were calculated for AEs of interest. ROR shows disproportionate signals if the lower limit of the 95% CI > 1. PRR shows disproportionate signals if PRR ≥ 2, χ2 statistic ≥ 4, and ≥ 3 AEs were reported. We identified any US Food and Drug Administration enforcement actions for radium-223. Results: A majority (60.8%) of events occurred outside the US. Among patients with radium-223-associated AEs, the median therapy duration was only 56 days (corresponding to 2-3 treatment cycles). Disproportionate signals were detected for general health deterioration (ROR, 5.03; 95% CI, 4.23-5.98 and PRR, 4.94; 95% CI, 4.16-5.87), bone pain (ROR, 4.53; 95% CI, 3.67-5.59 and PRR, 4.48; 95% CI, 3.63-5.53), and hematologic AEs including anemia (ROR, 2.89; 95% CI, 2.55-3.27 and PRR, 2.80; 95% CI, 2.48-3.17), thrombocytopenia (ROR, 3.22; 95% CI, 2.77-3.74 and PRR, 3.16; 95% CI, 2.72-3.67), and pancytopenia/bone marrow failure (ROR, 4.83; 95% CI, 4.11-5.67 and PRR, 4.73; 95% CI, 4.03-5.55). There were no enforcement actions for radium-223. Conclusions: Patients with metastatic castrate-resistant prostate cancer experiencing AEs are only receiving one-half the prescription dose of radium-223 required for survival benefit. Radium-223 is associated with health deterioration, bone pain, and hematologic AEs. Real-world analyses are important for ongoing radium-223 risk-benefit assessments.