Content uploaded by Lars Franzén
Author content
All content in this area was uploaded by Lars Franzén on Feb 01, 2019
Content may be subject to copyright.
http://oncology.thelancet.com Vol 8 July 2007
587
Articles
Bone-targeted radium-223 in symptomatic, hormone-
refractory prostate cancer: a randomised, multicentre,
placebo-controlled phase II study
Sten Nilsson, Lars Franzén, Christopher Parker, Christopher Tyrrell, René Blom, Jan Tennvall, Bo Lennernäs, Ulf Petersson, Dag C Johannessen,
Michael Sokal, Katharine Pigott, Jeff rey Yachnin, Michael Garkavij, Peter Strang, Johan Harmenberg, Bjørg Bolstad, Øyvind S Bruland
Summary
Background The alpha-emitter radium-223 (223Ra) is a bone-seeking radionuclide studied as a new treatment for
patients with bone metastases from hormone-refractory prostate cancer. We aimed to study mature outcomes from a
randomised, multicentre, phase II study of 223Ra.
Methods Patients with hormone-refractory prostate cancer and bone pain needing external-beam radiotherapy were
assigned to four intravenous injections of 223Ra (50 kBq/kg, 33 patients) or placebo (31 patients), given every 4 weeks.
Primary endpoints were change in bone-alkaline phosphatase (ALP) concentration and time to skeletal-related events
(SREs). Secondary endpoints included toxic eff ects, time to prostate-specifi c-antigen (PSA) progression, and overall
survival. All tests were done at a 5% signifi cance level, based on intention to treat.
Findings Median relative change in bone-ALP during treatment was –65⋅6% (95% CI –69⋅5 to –57⋅7) and 9⋅3%
(3⋅8–60⋅9) in the 223Ra group and placebo groups, respectively (p<0⋅0001, Wilcoxon ranked-sums test). Hazard ratio
for time to fi rst SRE, adjusted for baseline covariates, was 1⋅75 (0⋅96–3⋅19, p=0⋅065, Cox regression). Haematological
toxic eff ects did not diff er signifi cantly between two groups. No patient discontinued 223Ra because of treatment
toxicity. Median time to PSA progression was 26 weeks (16–39) versus 8 weeks (4–12; p=0⋅048) for 223Ra versus placebo,
respectively. Median overall survival was 65⋅3 weeks (48⋅7–∞) for 223Ra and 46⋅4 weeks (32⋅1–77⋅4) for placebo
(p=0⋅066, log rank). The hazard ratio for overall survival, adjusted for baseline covariates was 2⋅12 (1⋅13–3⋅98,
p=0⋅020, Cox regression).
Interpretation 223Ra was well tolerated with minimum myelotoxicity, and had a signifi cant eff ect on bone-ALP
concentrations. Larger clinical trials are warranted to study 223Ra on the prevention of SREs and on overall survival in
patients with hormone-refractory prostate cancer. Bone-targeting properties of 223Ra could also potentially be used for
treating skeletal metastasis from other primary cancers.
Introduction
Hormone-refractory prostate cancer has a propensity to
involve the bone marrow at an early stage. Subsequently,
this involvement leads to the development of
symptomatic skeletal metastases with pain, spinal-cord
compression, pathological fracture, and pancytopenia.
Bone-targeted treatments such as bisphosphonates (eg,
zoledronate) and bone-seeking radioisotopes (eg, 89Sr
and ethylenediaminetetramethylene phosphonate
(EDTMP)-153Sm) are commonly used to delay skeletal
disease progression and relieve pain. Zoledronate
reduces the risk of skeletal-related events, but does not
extend survival.1 The beta-emitting radioisotopes 89Sr and
EDTMP-153Sm have been shown to improve pain control
in men with symptomatic, hormone-refractory prostate
cancer, with myelo suppression as their dose-limiting
toxicity.2,3
Beta-emitting radioisotopes produce relatively low-
energy radiation with a track length in tissues of up to
several mm. By contrast, alpha-emitters produce high,
linear energy transfer (LET) radiation with a range of
less than 100 µm. Compared with 89Sr and EDTMP-153Sm,
a bone-seeking alpha-emitter might therefore have an
increased anti-tumour eff ect, by virtue of the densely
ionising abilities of high-LET radiation, but with relative
sparing of the bone marrow because of its short-track
length.
Radium-223 (223Ra) is a bone-seeking alpha-emitter
with a half-life of 11⋅4 days that has been studied
extensively in preclinical animal models. In mice, the
biodistribution of 223Ra has been shown to correspond
with that of 89Sr, targeting of the bony skeleton with
retention of its daughter isotopes in the bone matrix.4
Modelling of the dose deposition in relation to tumour
deposits in the bone marrow suggests a substantial
reduction in dose to the healthy bone marrow with 223Ra
compared with 89Sr.4 In a rat model of breast cancer, 223Ra
showed pronounced anti-tumour eff ects in the absence
of bone marrow toxic eff ects.5 After 67 days’ follow-up
period, two of fi ve animals treated with at least 100 kBq/kg
223Ra survived, whereas none did in the control group. In
this model, treatment with bisphosphonates gave no
survival benefi t.
A phase I trial of one intravenous injection of 223Ra was
done recently in 25 patients with metastatic bone disease
(15 with hormone-refractory prostate cancer, ten with
Lancet Oncol 2007; 8: 587–94
Published Online
June 3, 2007
DOI:10.1016/S1470-
2045(07)70147-X
See Refl ection and Reaction
page 564
Karolinska Hospital,
Stockholm, Sweden
(Prof S Nilsson MD,
J Yachnin MD); Karolinska
Institute, Stockholm, Sweden
(Prof S Nilsson, Prof P Strang MD,
J Harmenberg MD);
Länssjukhuset Sundsvall-
Härnösand, Sundsvall, Sweden
(L Franzén MD); Institute of
Cancer Research and Royal
Marsden Hospital, Sutton, UK
(C Parker FRCR); Plymouth
General Hospital, Plymouth,
UK (C Tyrrell FRCR); University
Hospital Linköping, Linköping,
Sweden (R Blom MD);
University Hospital Lund,
Lund, Sweden
(Prof J Tennvall MD,
M Garkavij MD); Sahlgrenska
University Hospital,
Gothenburg, Sweden
(B Lennernäs MD);
Centrallasarettet i Västerås,
Västerås, Sweden
(U Petersson MD); Haukeland
University Hospital, Bergen,
Norway (D C Johannessen MD);
Nottingham City Hospital,
Nottingham, UK
(M Sokal FRCR); Royal Free
Hospital, London, UK
(K Pigott FRCR); Algeta ASA,
Oslo, Norway (J Harmenberg,
B Bolstad MSc); University of
Oslo, Norwegian Radium
Hospital, Oslo, Norway
(Prof Ø S Bruland MD)
Correspondence to:
Dr Christopher Parker, Institute
of Cancer Research and Royal
Marsden Hospital, Sutton, UK
chris.parker@icr.ac.uk
Articles
588
http://oncology.thelancet.com Vol 8 July 2007
breast cancer).6 Doses of 50–250 kBq/kg were well
tolerated, with grade 3 leucopenia (Common Terminology
Criteria for Adverse Events [CTCAE], version 2.0) in three
of 25 patients and no grade 2+ thrombocytopenia. No
dose-limiting toxic eff ects were recorded. Gastrointestinal
adverse events, most commonly diarrhoea (10 of 25 patients;
highest CTCAE grade 2), were most frequent, especially in
the highest dose groups. Preferential uptake was seen in
skeletal metastases compared with healthy bone, with
excretion mainly through the intestinal tract. Preliminary
evidence of effi cacy was seen with substantial reductions
in serum alkaline phosphatase (ALP) concentrations and
improved pain control across the dose levels.6 Preliminary
data suggested that the eff ect on ALP was similar across
doses, and the lowest tested dose (50 kBq/kg) was selected
for future studies.
In a randomised, double-blind, placebo-controlled,
multicentre phase II study, we aimed to investigate the
eff ect of repeated 223Ra doses in men with symptomatic,
hormone-refractory prostate cancer. Main study objectives
were the effi cacy of 223Ra treatment with respect to the
reduction in bone-specifi c ALP (bone-ALP) concentration,
and time to occurrence of skeletal-related events (SREs).
Methods
Patients
Eligible patients had histologically or cytologically confi rmed
adenocarcinoma of the prostate; multiple bone metastases
or one painful lesion with two consecutive rising amounts
of serum prostate-specifi c antigen (PSA); Eastern
Cooperative Oncology Group performance status 0–2; life
expectancy of longer than 3 months; adequate
haematological (neutrophils ≥1⋅5×109/L; platelets ≥100×109/
L; haemoglobin >100 g/L), renal (creatinine <1⋅5×upper
limit of normal), and hepatic (normal bilirubin [within
institutional limits], aspartate aminotransferase and alanine
aminotransferase <2⋅5×upper limit of normal) function;
and bone pain needing external-beam radiotherapy. All
patients had either bilateral orchidectomy or continued
treatment on a luteinising-hormone-releasing-hormone
agonist throu gh out the study. We excluded patients if they
had another currently active malignant disease; had
received chemo therapy, immunotherapy, or external-beam
radiotherapy within the past 6 weeks; had a change in
hormonal treatment within 6 weeks before study drug use,
bisphosphonates within 3 months; or had any previous
systemic radiotherapy with strontium, samarium, or
rhenium. All patients gave written informed consent.
Procedures
Patients due to receive local-fi eld, external-beam
radiotherapy to relieve pain from bone metastases were
assigned to receive either four repeated monthly
injections of 50 kBq/kg 223Ra (Alpharadin, Algeta ASA,
Oslo, Norway) or repeated injections of saline (fi gure 1).
223Ra was supplied to the hospitals as a ready-to-use
solution for injection, with little shielding needed since
the isotope is an alpha-emitter. Randomisation was
done centrally with a random number generator and
stratifi ed according to study centre. For masked
treatment allocation, an individual from the nuclear
medicine department at every centre was responsible
for study treatment preparation and labelling.
Researchers remained masked to treatment allocation.
Treatment lasted for 12 weeks, during which four
injections were given at 4-week intervals, with the fi rst
injection given at the time of external-beam radiotherapy
and no later than 7 days afterwards. External-beam
radiotherapy was given to the most painful site, with
appropriate margins, and to an area not exceeding
400 cm2. Either one fraction of 8 Gy or a fractionated
course of 20 Gy (4 Gy×5) over 1 week, or 30 Gy (3 Gy×10)
over 2 weeks was permitted.
Panel: Defi nitions of SREs
• 25% increase in pain severity index compared with baseline after day 15 (during the
fi rst 16 weeks of the study, the pain increase had to be confi rmed with a second
assessment). Patients indicating pain severity on Brief Pain Inventory forms8
(scale 0–10), with index defi ned as mean score of four questions on pain severity
[worst, least, average, and current])
• Increased analgesic consumption (analgesia classifi ed according to World Health
Organization ladder for cancer pain [L0=no analgesia, L1=non-opioids, L2=weak
opioids, L3=strong opioids]; increase defi ned as change to higher level or ≥50%
increase in strong opioids, with temporary increases during fi rst 2 weeks after fi rst
injection excluded)
• Neurological symptoms secondary to skeletal manifestations of prostate cancer
• New pathological bone fractures (vertebral and non-vertebral)
• Tumour-related orthopaedic surgical intervention
• Subsequent external-beam radiation to relieve skeletal pain
• Use of radioisotopes to relieve new skeletal-related symptoms
• Use of corticosteroids for skeletal pain palliation
• Use of chemotherapy, bisphosphonates, or hormones to treat progression of skeletal
disease
4 injections
q 4 weeks
W12 W16
Bone markers
PSA
N=64
Local external-beam
radiotherapy
Patients with
hormone-refractory
prostate cancer
50 kBq/kg 223Ra every 4 weeks
Saline (placebo) every 4 weeks
M6 M9 M18 M24
M12
Study
unmasked
SREs
Pain
Bone markers
PSA
Safety
Survival
Follow-up
of long-term
toxic effects
and survival
Randomise
Treatment
Figure 1: Overall summary of study treatment
W=week. M=month.
Articles
http://oncology.thelancet.com Vol 8 July 2007
589
Patients were monitored every 2 weeks until 4 weeks
after the last injection, and then at 6, 9, and 12 months.
Every visit included clinical history, concomitant
medication, physical examination, pain and analgesic
assessment, adverse event recording, and blood tests
(including bone-ALP, PSA, full blood count, and markers
of bone turnover). Patients were followed for survival and
long-term toxic eff ects at 18 and 24 months.
Bubley described PSA-confi rmed response and
progression in 1999.7 Confi rmed PSA response was a 50%
reduction from baseline and confi rmed with a second
measurement at least 4 weeks later. PSA progression was
an increase of 25% from nadir in patients without a
confi rmed PSA response and 50% in those with a con-
fi rmed PSA response. SREs were defi ned as one of a
specifi c set of events listed in the panel.
Patients who began non-study treatment (eg, hormonal
therapy, external-beam radiotherapy) for progressive
disease during the study were allowed to continue with
the study drug and be analysed for safety. Effi cacy analyses
were done with and without censoring at the time of
commencement of non-study treatments (corticosteroids,
33 allocated to ²²³Ra treatment
33 received at least one ²²³Ra treatment
5 did not receive all four ²²³Ra treatments
2 with progressive disease or
investigator decision
2 with cardiac disease
1 with confusion
31 allocated to placebo treatment
31 received at least one placebo treatment
10 did not receive all four placebo treatments
4 with progressive disease
4 patient preference
1 with cardiac disease
1 with confusion
64 patients with hormone-refractory
prostrate cancer randomised
29 patients in study at week 16 22 patients in study at week 16
4 patients withdrawn
before week 16
9 patients withdrawn
before week 16
19 patients in study at month 12 12 patients in study at month 12
10 patients withdrawn
between week 16 and
month 12 because of
disease progression or
death
10 patients withdrawn
between week 16 and
month 12 because of
disease progression or
death
14 patients in study at month 18 5 patients in study at month 18
5 patients withdrawn
between month 12 and
month 18 because of
disease progression or
death
7 patients withdrawn
between month 12 and
month 18 because of
disease progression or
death
Figure 2: Study profi le
All patients followed for survival irrespective of study status (ie, in study or withdrawn). Information on the number screened was not obtained.
²²³Ra (n=33) Placebo (n=31)
Age, years 73, 73 (57–88) 72, 72 (60–84)
Haemoglobin, g/L 126, 125 (100–153) 129, 126 (99–149)
PSA, ng/mL 167, 511 (10–6000) 233, 480 (1–4002)
Bone-ALP, ng/mL 57, 121 (13–1145) 68, 132 (11–706)
Total ALP, U/L 228, 437 (80–3047) 279, 501 (51–2280)
Albumin, g/L 40, 39 (28–46) 38, 39 (30–47)
Lactate dehydrogenase (U/L) 348, 351 (154–750) 345, 426 (144–1284)
ECOG performance status
09 6
118 20
26 5
Extent of disease
<6 metastases 12 7
6–20 metastases 10 13
>20 metastases 11 11
Pain severity index 3·50, 3·88 (1·00–7·75) 4·00, 3·78 (0·75–7·75)
Data are number of patients (median, mean, range) in intention-to-treat population
Table 1: Baseline patient characteristics
Articles
590
http://oncology.thelancet.com Vol 8 July 2007
external-beam radiotherapy, cytotoxic chemotherapy, add-
itional radionuclides, hormones, and bisphosphonates).
The study design was approved by all the research ethics
committees from the study jurisdictions and centres.
In addition to time to SREs, bone-ALP was chosen as a
primary endpoint because it is used in clinical practice
as a marker of disease extent and prognosis, and because
a bone-targeted therapy would probably not be eff ective
in treating hormone-refractory prostate cancer without
greatly aff ecting bone-ALP. Secondary endpoints
included safety, serum markers of bone turnover (total
ALP, procollagen I N propeptide [PINP], C-terminal
crosslinking telopeptide of type I collagen [S-CTX-I], and
type I collagen crosslinked C-telopeptide [ICTP]), serum
PSA, and overall survival.
Statistical methods
The study was designed to have 80% power to detect an
absolute diff erence at 5% signifi cance level of at least
15% between treatment and control groups with respect
to the mean change in bone ALP from baseline to 4 weeks
after the last injection. Although medians are reported in
the results and are more useful for clinicians, no standard
techniques exist for calculating sample size from
medians, therefore, means were used for this purpose.
Sample size calculations were based on the normal
distribution with the following formula and rounded up:
n=(2*SD2*(Zα/2+Zβ)2/meandiff 2 +Zα/2/4) (Zα/2=value of
normal distribution [1⋅96] for α/2 [α=0⋅05], and Zβ=value
of normal distribution [0⋅84] for 1-power [power=0⋅8]).
We planned a sample size of 60 patients (30 in each
group). Patients who received at least one dose of study
drug were included in the analyses, and all results were
analysed by intention to treat. All tests of signifi cance
were done at the 5% signifi cance level and no adjustment
for multiplicity was done for secondary variables, since
the present study was explorative in nature. We expected
that the relative change of bone ALP and other variables
would not be normally distributed, and would contain
non-positive data, so standard transformation methods
would not be possible. Therefore, we analysed the
diff erence between groups of relative change of bone
ALP and other variables by use of the Wilcoxon ranked-
sums test. For the analysis of time to event, we used a
log-rank test (unadjusted) and a Cox proportional
hazards model (adjusted for covariates). Covariates
were baseline values of albumin, ECOG performance
status, haemoglobin, lactate dehydrogenase, PSA, total
ALP, and age. Proportions were analysed with the
Fisher’s exact test. We used screening values for missing
data at baseline, and a last-value-carried-forward
approach for missing laboratory values. For relative
change and time-to-PSA-progression, we excluded two
patients in the placebo group without any post-baseline
values.
Role of the funding source
This study was sponsored by Algeta ASA (Oslo, Norway)
in collaboration with TFS Trial Form Support (Lund,
Sweden; monitoring and database); Harrison Clinical
Research (Ely, UK; monitoring); and Statisticon AB
(Uppsala, Sweden; statistical advice and analysis). BB
and JH are employees of Algeta. BB was involved in the
study design, and BB and JH contributed to the analysis
²²³Ra (n=33) Placebo (n=30*)
1234 1234
Platelets 6 0 0 0 4 0 1 0
Neutrophils 5 2 1 0 0 0 0 0
White blood cells 9 1 1 0 3 0 0 0
Haemoglobin 26 4 1 0 19 6 0 1
*One patient was not evaluable for toxic eff ects. Numbers in second row are toxicity graded with CTCAE version 3.0.
Data are number of patients.
Table 2: Worst grade for haematological toxic eff ects during treatment
223Ra Placebo
Mild Moderate Severe Mild Moderate Severe
Confusional State 1 0 0 0 0 0
Vomiting 0 1 0 0 0 0
Pseudomonas infection 0 0 1* 0 0 0
Exacerbated chronic
obstructive airways disease
001*000
Pneumonia 0 1 1† 0 0 0
Sepsis 0 0 1† 0 0 1‡
Dehydration 0 0 1† 0 0 0
Spinal cord compression 0 1 0 0 0 1¶
Chest pain 0 1 0 0 0 0
Myocardial infarction 0 0 1§ 0 0 0
Tumour fl are 001§001
Ataxia 0 0 0 0 0 1
Subdural haematoma 0 0 0 0 0 1
Deep-vein thrombosis 0 0 0 0 1 0
Perineal pain 0 0 0 0 0 1
Bone pain 0 0 0 0 1‡ 2¶
Atrial fi brillation 0 0 0 0 0 1‡
Adrenal disorder 0 0 0 0 0 1
Fall 0 0 0 0 0 1**
Cachexia 0 0 0 0 1** 0
Haemoglobin decreased 0 0 0 1 0 0
Pyrexia 0 0 0 0 1†† 0
Malignant neoplasm 0 0 0 0 0 1††
Hypocalcaemia 0 0 0 0 1 0
Intestinal obstruction 0 0 0 0 0 1
Mild=transient and easily tolerated. Moderate=causes discomfort and interrupts usual activities. Severe=considerable
interference with usual activities, and might be incapacitating or life threatening. *One patient had pseudomonas
infection and exacerbated chronic obstructive airways disease. †One patient had pneumonia, sepsis, and dehydration.
‡One patient had atrial fi brillation, sepsis, and bone pain. ¶One patient had bone pain and spinal cord compression
§One patient had myocardial infarction and tumour fl are.**One patient had a fall and cachexia. ††One patient had
pyrexia and malignant neoplasm. When a serious adverse event was reported more then once by the same patient, the
highest intensity is listed.
Table 3: Serious adverse events
Articles
http://oncology.thelancet.com Vol 8 July 2007
591
and interpretation of the data. Algeta had no other role in
the collection, analysis, or interpretation of the data. The
corresponding author had full access to all the raw data
and the fi nal responsibility to submit for publication. JH,
PS, and SN also had access to the raw data.
Results
64 patients were recruited in 11 centres in Sweden,
Norway, and the UK between Feb 11, 2004, and May 3,
2005 (fi gure 2). 33 patients were assigned external-
beam radiotherapy and 223Ra, and 31 to external-beam
radiotherapy and placebo. Table 1 shows the baseline
values of bone-ALP, haemoglobin, albumin, PSA, age,
extent of disease on bone scan,9 and pain score. We
recorded no signifi cant diff erence in external-beam
radiotherapy dose fractionation at baseline between the
study groups (223Ra, median 8 Gy [range 6–30]; placebo,
8 Gy [8–30]). All patients have been followed for at least
18 months (range 18–24).
All 64 patients received external-beam radiotherapy and
at least one injection of study drug. 28 patients in the
223Ra group and 21 in the placebo group completed all
four injections of study drug. Figure 2 shows the reasons
for not completing all four treatments. During treatment
(from week –1 to week 16), corticosteroids were begun in
24 patients (12 223Ra, 12 placebo), anti-androgens in two
223Ra patients, oestrogen in one placebo patient, and
bisphosphonate in two placebo patients.
The extent of haematological toxic eff ects was at a
minimum (table 2). Thrombocytopenia (CTCAE grade
2+) was not seen in the 223Ra group but recorded in one
placebo patient. We recorded grade 2+ neutropenia in
three patients given 223Ra and none given placebo.
Neutropenia was reversible and most commonly seen
during the fi rst 4 weeks of treatment, with no evidence of
cumulative myelotoxic eff ects recorded. We saw no
substantial diff erences in haematological toxic eff ects
between the two groups, and no patient discontinued
223Ra because of treatment-related toxic eff ects.
12 serious adverse events were reported in eight
patients receiving 223Ra, compared with 19 serious adverse
events in 14 patients receiving placebo (table 3). One
patient’s vomiting needed hospital care 6 h after the fi rst
223Ra injection, which was deemed related to the study
drug. This patient received subsequent injections with
223Ra without further vomiting. The investigators deemed
that vomiting was treatment-related; but they were
uncertain whether sepsis (reported in one patient in each
study group) and tumour fl are (one patient in the 223Ra
group) were treatment-related (one patient in the placebo
group had tumour fl are, but this was deemed not
treatment-related). All other serious adverse events were
deemed unrelated to treatment.
Table 4 lists adverse events that occurred in more than
15% of patients. Apart from constipation, no statistically
signifi cant diff erence existed between treatment groups.
Constipation was mild or moderate in all but one patient.
Median time to onset of constipation was 4 weeks (range
1–12) in the 223Ra group, and actual time of onset from
baseline was 7 and 10 weeks in the two patients in the
placebo group. Unfortunately, no data for duration of
constipation were available.
Median relative change in bone-ALP from baseline to
4 weeks after last study injection was –65⋅6% (95% CI
−200
−100
0
100
200
300
400
²²³Ra
Placebo
Relative change in bone−ALP (%)
Figure 3: Distribution of % change in serum bone-ALP from baseline to 4 weeks after last study drug
Individual results sorted by eff ect size.
223Ra (n=33) Placebo (n=31)
Mild Moderate Severe Total Mild Moderate Severe Total
Diarrhoea 6 3 0 9 5 5 0 10
Constipation 6 5 1 12 0 2 0 2
Vomiting 4 4 0 8 3 3 0 6
Nausea 6 3 0 9 6 3 1 10
Fatigue 3 5 0 8 7 0 0 7
Bone pain 2 7 1 10 7 7 2 16
Myalgia 3 2 0 5 3 1 0 4
Tumour fl are 2 3 1 6 1 5 1 7
Anaemia 0 5 0 5 2 3 2 7
Data are number of patients. Mild=transient and easily tolerated. Moderate=causes discomfort and interrupts usual
activities. Severe=considerable interference with usual activities, and might be incapacitating or life threatening.
Table 4: Adverse events in more than 15% of study population during treatment (reported by at least
nine patients)
223Ra Placebo p*
Bone-ALP –65·6% (–92·2 to 124·9) 9·3% (–77·4 to 384·1) <0·0001
Total ALP –46·2% (–89·3 to 102·5) 30·7% (–75·4 to 212·9) <0·0001
PINP –63·2% (–93·7 to 151·0) 38·3% (–72·5 to 602·8) <0·0001
CTX-I –31·4% (–74·3 to 143·3) 31·7% (–57·5 to 395·8) 0·002
ICTP 14·6% (–54·6 to 158·9) 43·2% (–56·3 to 242·1) 0·011
Data are median (range). *Wilcoxon ranked-sums test.
Table 5: Relative change in serum markers of bone metabolism from baseline to 4 weeks after last study
injection
Articles
592
http://oncology.thelancet.com Vol 8 July 2007
–69⋅5 to –57⋅7) in the 223Ra group and 9⋅3% (3⋅8–60⋅9) in
the placebo group (p<0⋅0001, Wilcoxon ranked-sums test).
Figure 3 shows distribution of the change in bone-ALP.
Table 5 shows median relative change in bone-ALP and the
other serum markers of bone metabolism. Compared with
the placebo group, the 223Ra group had a signifi cant
reduction in all fi ve markers (bone-ALP, total-ALP, PINP,
CTX-I, and ICTP).
Median time to fi rst SRE was 14 weeks (95% CI 9–30) in
the 223Ra group and 11 weeks (5–25) in the placebo group
(p=0⋅257, log rank; fi gure 4). Hazard ratio for time to fi rst
SRE adjusted for baseline covariates was 1⋅75 (95% CI
0⋅96–3⋅19, p=0⋅065, Cox regression) where the placebo
group is the reference. By week 16, 17 patients in the 223Ra
group had 34 SREs in total, compared with 18 patients
who had 44 SREs in the placebo group (p=0⋅625, Fisher’s
exact test). If we excluded pain and analgesic consumption
from the SRE defi nitions, six patients in the 223Ra group
had nine SREs in total, compared with 11 patients in the
placebo group who had 21 SREs (p=0⋅159). By week 52,
26 patients in each group had had at least one SRE.
Median relative change in PSA from baseline to 4 weeks
after last study injection was –23⋅8% (range –98⋅6 to
545⋅6) in the 223Ra group and 44⋅9% (–91⋅3 to 563⋅5) in
the placebo group (p=0⋅003, Wilcoxon ranked-sums test;
fi gure 5). A confi rmed PSA response of more than 50%
was seen in 11 of 31 patients assigned 223Ra and fi ve of
28 assigned placebo (p=0⋅153, Fisher’s exact test). Median
time to PSA progression was 26 weeks (95% CI 16–39)
for 223Ra compared with 8 weeks (4–12) for placebo
(p=0⋅048, log rank).
Censoring for concomitant treatment that might aff ect
PSA did not change the overall results, since most
treatments were started after the PSA endpoints were
reached (PSA response or progression). After censoring
of concomitant treatments, nine patients who were
assigned 223Ra showed a confi rmed PSA response of at
least 50% compared with two assigned placebo (p=0⋅045,
Fisher’s exact test). The three censored patients in the
placebo group and two in the experimental group with
confi rmed 50% PSA response had all started cortico-
steroids. Median time to PSA progression remained at
26 weeks for 223Ra-treated patients compared with 8 weeks
for placebo (p=0⋅040, log rank).
Median overall survival was 65⋅3 weeks (95%
CI 48⋅7–∞) for 223Ra and 46⋅4 weeks (32⋅1–77⋅4) weeks
for placebo (p=0⋅066, log rank; fi gure 6). Hazard ratio for
survival adjusted for baseline covariates was 2⋅12 (95%
CI 1⋅13–3⋅98, p=0⋅020, Cox regression), indicating an
increased risk of death in the placebo group. At
18 months’ follow-up, 15 patients assigned 223Ra survived
compared with eight assigned placebo.
Discussion
In our randomised study of patients with symptomatic,
hormone-refractory prostate cancer, 223Ra was well
tolerated with little or no myelotoxic eff ect, and showed
promising evidence of effi cacy. One primary endpoint
was met, with a signifi cant eff ect on bone-ALP 4 weeks
after last treatment. Effi cacy data in this small study
suggested a potential benefi cial eff ect of 223Ra on SRE
risk, time to PSA progression, and overall survival. The
good toxicity profi le seen for 223Ra will allow future
studies to use increased doses and extended treatment
periods.
This study shows promising evidence for the safety
and tolerability of 223Ra 50 kBq/kg at 4 weekly intervals.
0·0
0·2
0·4
0·6
0·8
1·0
Proportion without SRE
Numbers at risk
²²³Ra
Placebo
05 504540353025201510
Time (weeks)
33 23 4567811131519
31 23 11245781315
²²³Ra
Placebo
Figure 4: Survival function of time to fi rst SRE
Special instructions (PLEASE MARK WITH RED SPOT IF URGENT)
€$£¥∆Ωµ∏π∑Ωαβχδεγηκλμτ†‡§¶√+−±×÷≈<>≤≥
Ref number
Editor
Author
Created by Helen
²²³Ra
Placebo
−200
0
200
400
600
Relative change in PSA (%)
Figure 5: Distribution of % change in serum PSA from baseline to 4 weeks after last study drug
Individual results sorted by eff ect size.
Articles
http://oncology.thelancet.com Vol 8 July 2007
593
The lack of haematological toxic eff ects contrasts with
those in previous studies of beta-emitting radioisotopes
in the treatment of hormone-refractory prostate cancer.
For example, 89Sr was studied in a trial of similar design
(n=126),10 in which grade 2+ thrombocytopenia was
seen in 41 (61%) of 67 patients assigned 89Sr versus six
(10%) of 58 assigned placebo. Notably, grade 4
thrombocytopenia was seen in seven (10%) patients in
the 89Sr group versus one (2%) in the placebo group.
89Sr was shown to signifi cantly reduce alkaline
phosphatase and PSA concentrations, but no eff ect was
seen on overall survival. On the basis of the present
study, 223Ra at 50 kBq/kg given every 4 weeks seems to
be at least as eff ective as conventional radioisotopes,
but with fewer haematological toxic eff ects. This result
supports the hypothesis that, compared with beta-
emitters such as 89Sr and HEDP-153Sm, the short-range,
high LET alpha radiation from 223Ra has an increased
anti-tumour eff ect, with relative sparing of the bone
marrow.11 Since myelosuppressive chemotherapy is
now a standard treatment for hormone-refractory
prostate cancer and pancytopenia is a feature of
progressive disease, this result represents a major
advantage for the future development of 223Ra. Docetaxel
is the standard fi rst-line chemotherapy for hormone-
refractory prostate cancer.12 In view of the favourable
toxicity profi le of 223Ra in the present trial, combination
therapy with docetaxel could be considered, either
sequentially or concurrently. Furthermore, whereas
89Sr and HEDP-153Sm are used for pain palliation, and
repeat treatment is not possible until bone-marrow
recovery, the toxicity profi le of 223Ra will allow repeat
treatment of asymptomatic patients and could have a
disease-modifying eff ect.
For non-haematological toxic eff ects, constipation was
the only adverse event seen substantially more often in
the 223Ra group than in the placebo group, for which the
explanation is unclear. Constipation was mild or
moderate in all but one patient, and no clear temporal
relation was recorded with study drug treatment.
Notably, the phase I study6 showed that 223Ra was
associated with diarrhoea and not constipation.
A signifi cant benefi t for 223Ra was recorded with
respect to all serum bone markers. These endpoints are
not in themselves clinically meaningful, but suggest
that 223Ra has a real biological eff ect on bone metabolism.
Favourable trends were seen in all PSA endpoints,
including a signifi cant eff ect on time to PSA progression,
which is consistent with the hypothesis that eff ective
treatment of bone metastases can substantially delay
overall disease progression in hormone-refractory
prostate cancer.
Interpretation of the clinically meaningful endpoints
is restricted by the sample size, and by the fact that all
patients received external-beam radiotherapy at
baseline. Too few SREs were recorded for fi rm
conclusions to be drawn regarding the clinical
eff ectiveness of 223Ra. However, in the pivotal trial of
zoledronate,13 the proportion of patients having an SRE
fell from 44% to 33%, a relative reduction of 25%.
The overall survival data in this study of only
64 patients should be interpreted with caution. The
eff ects of known baseline prognostic factors on time to
SRE, time to PSA progression, and overall survival time
have been tested using a Cox proportional hazards
model, which, if anything, increased rather than
decreased the treatment eff ect of 223Ra. Furthermore,
our small randomised trial is the third to show a
signifi cant advantage in overall survival for radioisotope
treatment in hormone-refractory prostate cancer.14,15
Thus, the survival benefi t of 223Ra in this trial could be a
genuine treatment eff ect. However, the benefi ts of 223Ra
could be due to an imbalance in important but unknown
baseline characteristics (eg, doubling time of
pretreatment PSA). Effi cacy data might also be
confounded by concomitant use of other anticancer
treatments. However, the main results remain unaltered
or strengthened after patients were censored at the time
of receiving concomitant treatments. Therefore the
benefi ts of 223Ra seen in this study are unlikely to be
explained by concomitant treatments. Further studies
of 223Ra should explore the potential for escalation of
dose and for increased duration of treatment by more
than four injections. The bone-targeting properties of
223Ra could also be applicable to the treatment of skeletal
metastasis from other primary cancers.
0·0
0·2
0·4
0·6
0·8
1·0
Proportion alive
Numbers at risk
0 10 1009080706050403020
Time (weeks)
33 32 779161820252931
31 30 335111314192227
²²³Ra
Placebo
²²³Ra
Placebo
Figure 6: Overall survival
Articles
594
http://oncology.thelancet.com Vol 8 July 2007
Contributors
Ø
B, PS, BB, JY, and SN contributed to study design. SN, LF, CP, CT, RB,
JT, BL, UP, DJ, MS, KP, JY, and MG contributed to data collection.
Ø
B,
CP, BB, JH, PS, and SN had access to the raw data, and contributed to
analysis and interpretation of the data. CP wrote the fi rst draft of the
manuscript and had fi nal editorial control. All authors contributed to
revision of the manuscript.
Confl icts of interest
CP,
Ø
B, and PS are external consultants of Algeta ASA, Norway
(sponsors of the study and manufacturers of 223Ra). BB and JH are
employees of Algeta ASA.
Ø
B, LF, BB, and JH own stock or options in
Algeta ASA. CP received payment from Algeta ASA for writing the
manuscript. No other confl ict of interest has been declared.
Acknowledgments
We thank our Nuclear Medicine colleagues and the Central Laboratory in
Synarc (Lyon, France) for analysing serum markers of bone metabolism
and PSA, and Marcus Thuresson for statistical support.
References
1 Saad F, Gleason DM, Murray R, et al. Long-term effi cacy of
zoledronic acid for the prevention of skeletal complications in
patients with metastatic hormone-refractory prostate cancer.
J Natl Cancer Inst 2004; 96: 879–82.
2 Finlay IG, Mason MD, Shelley M. Radioisotopes for the palliation of
metastatic bone cancer: a systematic review. Lancet Oncol 2005;
6: 392–400.
3 Lewington V. Bone-seeking radionuclides for therapy. J Nucl Med
2005; 46 (suppl 1): 38–47S.
4 Henriksen G, Fisher DR, Roeske JC, et al. Targeting of osseous sites
with alpha-emitting 223Ra: comparison with the beta-emitter 89Sr in
mice. J Nucl Med 2003; 44: 252–59.
5 Henriksen G, Breistol K, Bruland OS, et al. Signifi cant antitumor
eff ect from bone-seeking, alpha-particle-emitting (223)Ra
demonstrated in an experimental skeletal metastases model.
Cancer Res 2002; 62: 3120–25.
6 Nilsson S, Larsen RH, Fossa SD, et al. First clinical experience with
alpha-emitting radium-223 in the treatment of skeletal metastases.
Clin Cancer Res 2005; 11: 4451–59.
7 Bubley GJ, Carducci M, Dahut W, et al. Eligibility and response
guidelines for phase II clinical trials in androgen-independent
prostate cancer: recommendations from the Prostate-Specifi c
Antigen Working Group. J Clin Oncol 1999; 17: 3461–67.
8 Klepstad P, Loge JH, Borchgrevink PC, et al. The Norwegian brief
pain inventory questionnaire: translation and validation in cancer
pain patients. J Pain Symptom Manage 2002; 24: 517–25.
9 Soloway MS, Hardeman SW, Hickey D, et al. Stratifi cation of
patients with metastatic prostate cancer based on extent of disease
on initial bone scan. Cancer 1988; 61: 195–202.
10 Porter AT, McEwan AJB, Powe JE, et al. Results of a randomized
phase-III trial to evaluate the effi cacy of strontium-89 adjuvant to
local fi eld external beam irradiation in the management of
endocrine resistant metastatic prostate cancer. Int J Radiat Oncol
Biol Phys 1993; 25: 805–13.
11 Bruland OS, Nilsson S, Fisher DR, Larsen RH. High-linear energy
transfer irradiation targeted to skeletal metastases by the
alpha-emitter 223Ra: adjuvant or alternative to conventional
modalities? Clin Cancer Res 2006; 12: 6250–57s.
12 Tannock IF, de Wit R, Berry WR, et al. Docetaxel plus prednisone or
mitoxantrone plus prednisone for advanced prostate cancer.
N Engl J Med 2004; 351: 1502–12.
13 Saad F, Gleason DM, Murray R, et al. A randomized,
placebo-controlled trial of zoledronic acid in patients with hormone-
refractory metastatic prostate carcinoma. J Natl Cancer Inst 2002;
94: 1458–68.
14 Tu SM, Millikan RE, Mengistu B, et al. Bone-targeted therapy for
advanced androgen-independent carcinoma of the prostate: a
randomised phase II trial. Lancet 2001; 357: 336–41.
15 Palmedo H, Manka-Waluch A, Albers P, et al. Repeated
bone-targeted therapy for hormone-refractory prostate carcinoma:
randomized phase II trial with the new, high-energy radio-
pharmaceutical rhenium-188 hydroxyethylidene diphos phonate.
J Clin Oncol 2003; 21: 2869–75.
