Q JN U CL M ED 2001;45:53-64

Radiological imaging for the diagnosis of bone metastases

L.D .RY BAK , D .I.RO SEN TH AL
Primary neoplasms of the skeleton are rare, but metastat-
ic involvement is, unfortunately, a common occurrence.
This is particularly true for certain primary tumors.
Skeletal metastases are clinically significant because of
associated symptoms, complications such as patholog-
ical fracture and their profound significance for stag-
ing, treatment and prognosis.
Detection of bone metastases is, thus, an important part
of treatment planning. The frequency with which metas-
tases are detected varies considerably with the type of
primary tumor and with the methodology utilized for
detection.
Four main modalities are utilized clinically: plain film
radiography, CT scan, nuclear imaging and magnetic
resonance imaging. In this discussion, we will review lit-
erature on the radiology of skeletal metastases with
respect to lesion detection, assessment of response to
treatment and possible therapeutic implications. The
bulk of the discussion will focus on MRI and nuclear
studies since most of the recent advances have been
made in these areas.
Key words: N eoplasm m etastasis - Bone neoplasm s secon-
dary - Tom ography, em ission com puted - M agnetic resonan-
ce im aging - Radiography.
S
keletal m etastases are unfortunately com m on.
The frequency w ith w hich they are detected
varies considerably w ith the type of tum or. It
also varies w ith the m ethodology used for detection.
For som e types of tum or, such as breast cancer, ske-
letal m etastases are readily detected by im aging stu-
From the Department of Radiology, Harvard Medical School
Massachusetts General Hospital, Boston, MA, USA
dies and form an im portant aspect of the clinical disea-
se m anagem ent because of the sym ptom s they pro-
duce. For other conditions (such as chordom a), disse-
m inated skeletal m etastases are frequently detected at
the tim e of autopsy, but are less apparent during life.
In general, the prognosis for patients presenting
w ith bone m etastasis is poor. Patients w ith few er
m etastases or solitary lesions appear to have a better
outlook than those w ith m ultiple m etastatic deposits.
Mechanisms
D irect invasion of the skeleton m ay result from an
adjacent prim ary tum or (Fig. 1). Perhaps the m ost
prevalent exam ple is invasion of the chest w all by a
lung cancer. Lym phogenous spread to bone is uncom -
m on and difficult to docum ent. H ow ever, secondary
invasion of bone from involved lym ph nodes is not
rare. The spine is the m ost com m only affected site. The
left side of the vertebral bodies is m ore often involved
because the left-sided nodes are closer to bone than
the right.
1
D irect skeletal invasion is usually accom pa-
nied by a detectable soft tissue m ass, a feature that is
unusual in m etastases that arise by hem atogenous
spread.
M ost tum or im plants occur through the hem atoge-
nous route. The venous, rather than arterial route
appears m ore im portant, especially Batsons paraver-
tebral plexus. Presum ably, the absence of valves in
Address reprint requests to: D .I.Rosenthal, M assachusetts G eneral
H ospital, D epartm ent of Radiology, W AC 515, 15 Parkm an Street, Boston,
M A 02114 (U SA).
Vol. 45 - N o. 1 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E 53
RY BAK RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES
54 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E M arch 2001
these veins perm its retrograde distribution of tum or
cells. Batsons plexus com m unicates w ith the axial
skeleton and the proxim al long bones resulting in a
predilection for these sites. The preference for axial
involvem ent is also due to the greater vascularity of the
red (hem atogenous) m arrow found in the axial skele-
ton as opposed to the yellow (fatty) m arrow found in
the appendicular bones.
Clinical features
M any of the patients com plain of bone pain, w hich
m ay be due to various m echanism s including release
of chem ical m ediators, elevated intra-osseous pressure,
and periosteal elevation.
2
The probability that a skel-
etal m etastasis w ill be painful m ay partly depend
upon the nature of the prim ary. M etastases from lung
and breast prim aries are m ore likely to be sym ptom at-
ic than are those from prostate cancer. Fractures and
im pending fractures are also an im portant source of
pain, particularly in w eight-bearing bones. Such frac-
tures are m ore com m on in areas involved w ith lytic
m etastases rather than blastic ones. They are difficult
to m anage, and often fail to heal.
Sym ptom s of arthritis m ay result from tum or in
bone adjacent to a joint, m echanical collapse of an
articular surface due to lytic m etastasis, or from syn-
ovial im plants of tum or. Prim ary m alignancies that
have been reported to present in the latter fashion
include lung, colon, breast, m elanom a, and rhabdom -
yosarcom a.
3
Synovial im plantation is very rarely doc-
um ented on im aging studies.
Arthritic sym ptom s m ay also be due to paraneo-
plastic effects such as carcinom a polyarthritis, a con-
dition in w hich there is sudden onset of rheum atoid-
like sym ptom s w ith an asym m etric distribution in a
patient w ho is rheum atoid-factor negative. H ypertro-
phic (pulm onary) osteoarthropathy is another paraneo-
plastic syndrom e that m ay result in joint and long
bone sym ptom s w ithout local involvem ent.
Secondary gout is a know n com plication in can-
cer patients, especially after treatm ent. Skeletal chang-
es due to radiation, such as osteonecrosis and fracture
m ay be difficult to distinguish from m etastatic lesions
on im aging studies, but are usually not sym ptom atic.
Finally, sym ptom s m ay be due to the onset of carci-
nom a-related rheum atic conditions such as Sjogrens
syndrom e, lupus, and derm atom yositis.
4
The m ost
com m on m alignancy associated w ith arthritic sym p-
tom s is leukem ia.
5
Detection - radiography
Radiography is com m only used to evaluate sym p-
tom atic sites and to confirm findings on other im ag-
ing studies. It is not generally recom m ended as a
screening m ethod because of poor sensitivity. The
radiographic survey rem ains valuable in staging of
m ultiple m yelom a due to the poor sensitivity of the
radioisotope scan in this condition.
Fig. 1.Bone involvem ent by direct extension. Squam ous cell carci-
nom a of the skin w hich has invaded the adjacent first m etacarpal
bone.
RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES RY BAK
Vol. 45 - N o. 1 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E 55
Sensitivity depends partly on location. For instance,
m etastases to dense cortical bone are easier to detect
than those involving trabecular (m edullary) bone. In
the axial skeleton, m edullary m etastases m ay not be
detectable until 50% of the trabecular bone has been
destroyed.
Am ong the advantages of radiography is the fact that
certain features m ay help to distinguish m etastases
from other conditions and aid in identification of the
prim ary tum or. Radiography m ay be used to assess the
risk of pathological fracture, w hich is said to be high
if 50% of the cortex is destroyed by a lesion. This is a
crude m easure at best. Recent w ork using quantitative
analysis of CT scans for this purpose has show n m uch
greater prom ise.
Detection - CT
CT scanning has had a lim ited im pact upon the
clinical detection of skeletal m etastases. A lthough
m ore sensitive than conventional radiography for the
detection of destructive bone lesions, CT is a cum ber-
som e tool for screening the entire skeleton.
Interestingly, CT can detect m etastases w ithin bone
m arrow before bone destruction has occurred (Fig. 2).
Tum or w ithin the m arrow causes an increase in atten-
uation due to fat replacem ent. An attenuation differ-
ence of m ore than 20 H U betw een the right and left
extrem ities is abnorm al.
6
Such findings are subtle,
and easily overlooked by the radiologist. They are
far less apparent than the m arrow changes seen on
M RI.
Detection - nuclear methods
Since the introduction of technetium -based scan
agents, approxim ately 25 years ago, the radioisotope
bone scan has been the standard m ethod for detection
of skeletal m etastases. Isotope scanning is m ore sen-
sitive than radiography for detection of m ost m etas-
tases. Tracer accum ulates in the reactive new bone that
is form ed in response to the lesion. Thus, although
m ost m etastatic lesions are hot, cold lesions due to
com plete absence of reactive bone m ay be encoun-
tered in particularly aggressive m etastases (Fig. 3). In
addition, the am ount of accum ulation is sensitive to
the level of blood flow . D iffuse accum ulation of trac-
er throughout the skeleton due to dissem inated skel-
etal disease (super scan) m ay lead to the false im pres-
sion of a norm al scan (Fig. 4).
The bone scan suffers from a lack of specificity.
Tracer accum ulation m ay occur in any skeletal loca-
tion w ith an elevated rate of bone turnover and, thus,
m ay accom pany traum a, infection or arthropathy.
The probability that an abnorm al scan represents
m etastatic tum or is directly related to the num ber of
abnorm al foci. In a patient w ith foci of increased
uptake and a know n prim ary tum or, the scan strong-
ly suggests m etastases. A sm all num ber (less than 4)
of abnorm alities is m ore likely to represent m etastat-
ic disease in som e locations than others, w ith rib
lesions being particularly low -yield.
7
O nly 50% of sol-
itary foci represent m etastases, even am ong patients
w ith cancer.
This lack of specificity is w ell know n and has lead
to recom m endations that positive scans be accom pa-
nied by radiographic correlation. H ow ever, given the
greater sensitivity of the bone scan, a positive radio-
graph m ay confirm a finding, but a negative radio-
graph does not exclude a m etastasis.
Recent advances in isotope scanning m ethods, par-
ticularly single photon em ission com puter tom ography
(SPECT), have im proved the detection of m etastases.
SPECT im aging has increased both the sensitivity and
specificity of bone scanning.
8
The tom ographic pres-
entation of SPECT im ages helps to identify foci of
abnorm al uptake especially in the thicker body parts
such as the spine and pelvis. In addition, im proved
spatial localization helps to distinguish betw een m et-
Fig. 2.CT scan of the proxim al fem ora dem onstrates increased den-
sity of the left-sided m arrow . This is due to replacem ent of the nor-
m al m arrow fat by tum or, a finding that m ay precede m ore obvious
bone lysis.
RY BAK RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES
56 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E M arch 2001
astatic foci and other abnorm alities causing increased
uptake, such as spondylosis. Increased uptake that
involves the posterior vertebral body is m ore likely to
be due to m etastasis.
A baseline radio-isotope bone scan is no longer
recom m ended in stage 1 or 2 breast cancer because
of low yield.
9
Several recent studies have suggested
that the diagnostic yield of bone scan in patients w ith
sm all and w ell-differentiated prostate carcinom as and
PSA values <20 is too low to w arrant routine use.
10 11
If scanning is w ithheld from such individuals the
national savings w ould be $ 38 m illion/year.
12
H ow ever, recent data indicates that for patients w ho
have received androgen depletion therapy, the PSA
m ay be unreliable in excluding m etastases.
At present, the conventional bone scan has no role
in the detection of m etastases from renal cell carcino-
m a and head and neck cancer.
13 14
For non-sm all cell
lung cancer, the bone scan is still recom m ended at the
tim e of diagnosis to aid in selection of patients for
surgery, although FD G -PET is also prom ising.
8
There have been conflicting reports as to the effi-
cacy of PET scanning and, as of yet, a consensus
has not been reached. O ne study of 44 patients w ith
know n m etastatic disease from lung, prostate and
thyroid prim aries show ed FD G -PET to be m ore sen-
sitive and specific than conventional bone scan.
15
Schirrm eister et al. studied 34 patients w ith breast
cancer and concluded that FD G PET allow ed for
earlier detection of sm all m arrow m etastases than
conventional bone scan and lead to clinically sig-
nificant changes in m anagem ent of four patients
16
(Fig. 5).
O ther results have not been as encouraging. A
recent publication based on im aging of 98 bone
lesions in 24 patients reported better specificity, but
low er sensitivity for FD G -PET as com pared to conven-
tional bone scan.
1
Another study sim ilarly concluded
that FD G -PET can detect prostate m etastases to bone
w ith m oderate sensitivity (65% ), but high specificity
Fig. 3.A) Technetium -99m M D P scan. There is a focal absence of isotope uptake in the left ninth rib. At the m argins of the coldarea, slight
increased uptake is visible. Very destructive lesions m ay produce focal cold areas as the result of com plete bone destruction in the affected
area. Increased uptake in the m arginal, less destroyed bone is not unusual. B) An oblique radiograph of the chest dem onstrates com plete
absence of the posterior ninth rib.
RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES RIBAK
Vol. 45 - N o. 1 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E 57
(98% ), and m ay have som e value in lesion detec-
tion.
17
Yeh et al. reported dism al results in a sm all
series of 13 patients w ith m ultiple bone m etastases
from prostate cancer in w hich FD G -PET only detect-
ed 18% of the lesions apparent by bone scan. The
authors concluded that prostate m etastases m ust have
m eans other than glycolysis for m etabolism and ener-
gy.
18
W ith regards to breast m etastases, one group
reported that PET dem onstrated superiority to bone
scan w ith osteolytic m etastases, but failed to consis-
tently detect osteoblastic m etastases.
19
These observa-
tions suggest a possible role for PET sim ilar to plain
film radiography, as confirm ation of positive results
from conventional technetium scans, rather than a
m eans of initial detection.
FD G appears not to be taken up by Paget disease,
and, therefore, PET im aging m ay be useful to separ-
ate Pagets disease and other benign conditions from
m etastases and/or sarcom atous degeneration.
8 20
H ow ever, despite undeniable utility in certain circum -
stances, presently there is no established role for PET
im aging in the clinical evaluation of bone m etasta-
ses.
M arrow scanning has been reported to be m ore
sensitive than the conventional bone scan in detection
of m etastases from prostate cancer.
21
In one study
com paring m arrow im aging using
99m
Tc anti-N CA-95
m onoclonal antibody w ith conventional M D P, the
m arrow im aging technique detected alm ost double
the num ber of lesions seen on M D P scans. H ow ever,
the num ber of patients identified as having m etasta-
ses w as the sam e: 13 in both instances
22
. In another
study of 23 patients w ith breast m etastases, bone m ar-
row im m unoscintigraphy (anti-N CA 95 M ab 250/183)
dem onstrated better specificity (88 vs75% ) and pos-
itive predictive value (92 vs 85% ) than conventional
bone scan w ith no significant difference in sensitivity.
23
Like PET, m arrow scanning has not yet found a role
in routine clinical practice.
For certain types of prim ary tum ors, (especially
lym phom as and soft tissue sarcom as) G allium scan-
ning m ay be a useful staging tool, detecting m etasta-
ses that are not otherw ise observed.
24
It m ay also be
helpful to follow the effect of treatm ent in these
patients.
25
Detection - MRI
M agnetic resonance im aging (M RI) is highly sensi-
tive to the presence of skeletal m etastases w ithin the
bone m arrow . Since bone m arrow (including hem at-
opoietic or redm arrow ) contains a high percent-
age of fat, T1-w eighted m agnetic resonance im ages
generally reveal m etastases as focal areas of low sig-
nal intensity. Lesions can be often be distinguished
from focal deposits of red m arrow on T1-w eighted
im ages because the latter are m ore focal and m ay
have centrally located fat, giving the appearance of a
bulls eye.
26
O n fat-suppressed T1-w eighted im ages,
m etastases dem onstrate m ixed to high signal inten-
sity.
27
O n T2-w eighted im ages, m etastatic lesions usu-
ally are m uch brighter than norm al m arrow due to
their high w ater-content (Fig. 6). M etastases often,
but not alw ays, have a rim of bright T2 signal around
them (halo sign).
26
A variety of different M RI pulse sequences have been
evaluated. In general, conventional spin-echo pulse
Fig. 4.Super Scan. D iffuse uptake of Technetium -99m M D P in
this patient w ith w idespread breast cancer m etastasis. The use of
inappropriate technical settings m ay result in a relatively norm al
appearance in the absence of focal sites of m ore prom inent uptake.
RIBAK RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES
58 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E M arch 2001
sequences provide the best signal-to-noise ratios and
anatom ical detail. H ow ever, because of the need for
rapid evaluation of large regions, fast spin echo and
inversion recovery im age sequences have been tested
and found to be acceptable.
28
Castillo et al. reported that
diffusion w eighted im aging had no advantage over
non-contrast enhanced T1 w eighted im aging in detec-
tion and characterization of vertebral m etastases, but
w as som ew hat superior to T2 w eighted im aging.
29
U nfortunately, it can be difficult to distinguish
changes due to tum or from the affects of treatm ent,
fracture, and inflam m ation. In one study, M RI scans
w ere com pared w ith histological specim ens at 21
sites. Seven of these contained tum or, 14 did not. All
of the tum or-positive sites show ed abnorm alities on
M RI scans. H ow ever, in the sites show n to be free of
tum or, a significant (m ore than 50% , depending upon
pulse sequence) false-positive rate w as encountered,
presum ably due to the effects of treatm ent.
30
It has becom e clear the m agnetic resonance im ag-
ing can detect m etastases that are not apparent on
radioisotope bone scans.
31
M RI is particularly w ell suit-
ed to detect spinal m etastases, and m ost authorities
agree that it is superior to planar scintigraphy for this
purpose. This m ay be partly due to the difficulty of rec-
ognizing subtle radionuclide abnorm alities, since ret-
rospective review of the isotope scans m ay reveal
m any abnorm alities that w ere initially m issed.
32
O ne
study of breast cancer patients concluded that M RI
w as specifically superior to bone scan in detection of
Fig. 5.The utility of PET in the detection of m etastatic disease w as dem onstrated in this patient
w ith past history of Ew ings sarcom a of the sacrum . A) A bone scan revealed foci of uptake
in the low er right ribs w hich corresponded to a site of prior surgery, but no other suspicious
sites. B) A subsequent PET scan revealed a focus of increased FD G uptake in the proxim al right
hum erus suspicious for m etastasis. C) Plain film s of the area revealed no obvious abnorm al-
ity. D ) A T1 fat saturated coronal im age of the right shoulder dem onstrated an obvious focus
of m etastatic disease dem onstrating diffuse enhancem ent.
RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES RIBAK
Vol. 45 - N o. 1 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E 59
aggressive spinal m etastases that dem onstrated estro-
gen receptor negativity and increased biologic activity.
33
M ost studies that have com pared M RI to bone scanning
have used planar bone scans, not SPECT. Planar scin-
tigraphy detects about 1/3-2/3 of the lesions seen by
M RI. M ultiple investigators have dem onstrated the
superiority of SPECT to planar im aging w ith regards to
detection of vertebral m etastases.
34-36
Som e authors
believe that SPECT m akes nuclear scanning com par-
able to M RI, w ith M RI better for vertebral body lesions
and SPECT better for the posterior elem ents.
37
Advantages of the isotope scan include a large field
of view , inexpensive radiopharm aceutical, low m or-
bidity, and the ability to provide som e functional and
vascular inform ation. M RI m ay be a sim pler and less
expensive m ethod to evaluate the axial skeleton, but
it has been considered less w ell-suited for screening
the long bones. Because of this, som e investigators
have recom m ended that the role of M RI be restricted
to clarifying an equivocal bone scan.
38
H ow ever, one
group took the position that lesions m issed in the
extrem ities w ere not significant. In their analysis of 200
patients w ith breast and prostate carcinom a, 3 of 4
peripherally located skeletal lesions that w ere m issed
by M RI w ere detected by plain film radiography
because they w ere painful.
39
Identification of appendicular lesions by M RI has
been facilitated by developm ent of faster pulse
sequences. Several recent papers have reported that
in com parison to conventional bone scan, w hole body
M RI utilizing w hole body fast short tau inversion
recovery (STIR) sequences has significantly better
sensitivity and specificity.
40-43
O ne of these papers
noted that detectability of rib lesions w as suboptim al
utilizing M RI.
40
W alker et al. further suggest that w hole
body M RI for patients w ith breast cancer m ay prove
to be an effective m eans of detecting skeletal, brain
and liver m etastases w ith one study.
44
There are a num ber of specific situations in w hich iso-
tope scanning m ay be preferred to M RI. They include
patients w ith contraindications to M RI such as claustro-
phobia and pacem akers, and patients w ith thyroid can-
cer in w hich scanning m ay be done w ith iodine. M RI
is preferred w hen the differential diagnosis includes
m arrow diseases such as lym phom a, leukem ia, m yelo-
m a and W aldenstrom m acroglobulinem ia.
45-47
The factors that influence the choice of im aging
m odalities continue to evolve. Progress in M RI has
been particularly rapid, and it appears probable that
its role in screening and staging in reference to skel-
etal m etastases w ill continue to grow and that it m ay
ultim ately replace the isotope scan. U ntil that tim e, iso-
tope scanning, especially w ith SPECT im aging, w ill
continue to have an im portant role.
Identification of the primary tumor
U nfortunately, patients m ay first com e to m edical
attention as the result of skeletal m etastasis from an
Fig. 6.The relative sensitivity of CT and M RI in the detection of skeletal m etastasis. A) A CT scan of the pelvis in this patient w ith know n
colon cancer revealed no obvious area of abnorm ality. B) A T2 w eighted axial im age of the sam e patient dem onstrated innum erous areas
of abnorm al high signal corresponding to m etastatic deposits.
RIBAK RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES
60 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E M arch 2001
unknow n prim ary tum or. For such individuals, im ag-
ing studies m ay be used to help identify the prim ary
lesion. Com m on tum ors w ith a high rate of bone
m etastasis include: breast 72% , prostate 84% , thyroid
50% , lung 31% , kidney 37% , pancreas 33% . Together,
these account for m ore than 80% of prim ary tum ors in
patients presenting w ith m etastases.
48-50
Previous studies have reported lim ited success in
identifying the prim ary tum or w hen a patient presents
w ith skeletal m etastasis of unknow n origin. In gener-
al, the prim ary tum or has been identified in less than
50% , even w hen patients w ere follow ed to autopsy.
In one of the m ost positive reports, the relative val-
ue of the history and physical exam ination, CB C,
erythrocyte sedim entation rate, blood chem istries,
alkaline phosphatase, plain film s of the lesion, radio-
isotope bone scan and com puted tom ography of the
chest, abdom en, and pelvis w ere com pared. The pri-
m ary tum or w as identified in 34/40 (85% ).
Interestingly, the laboratory values w ere unhelpful.
H istory and physical exam ination revealed the pri-
m ary in 4, chest X-ray identified 17, CT of chest add-
ed another 6, and CT of abdom en/pelvis added
another 5. If used alone, CT w ould have diagnosed
75% of all prim aries on the initial evaluation. All oth-
er m odalities, including follow -up CT only added an
additional 10% .
51
CT has an im portant role in providing im aging guid-
ance for tissue sam pling. Biopsy of the skeletal lesion,
follow ed by exam ination of the tissue using sophisti-
cated histologic techniques to lim it the im aging search
m ay be an equally valid approach to this difficult
problem .
48
O ne study, how ever, concluded that in
the m ajority of cases, the com bination of CT scan and
clinical inform ation provided com parable results to CT-
directed biopsy.
52
Radiographic features of the skeletal m etastases
m ay help direct the search for a prim ary lesion. Som e
prim ary tum ors tend to result in m etastases that are
purely lytic in nature, such as lung, renal and thyroid
cancer, others to be associated w ith variable degrees
of sclerosis, especially prostate, breast, carcinoid and
tum ors of endocrine glands (Fig. 7).
Evaluation of treatment
The prognosis for patients presenting w ith bone
m etastases is poor. In one series, only 4/578 patients
w ere free of disease 10 years after diagnosis of bone
disease. M ean survival for patients w ith all prim aries
w as 5 m onths after diagnosis.
48
In general, patients
w ith solitary lesions or a sm all num ber of m etastases
have a better outlook than those w ith m ultiple m eta-
static deposits.
Skeletal m etastases m ay respond to the chem other-
apy or horm one therapy used for the prim ary tum or.
They m ay also respond to radiation, or agents
designed to block bone resorption such as the new
class of bisphosphonate drugs.
53
Response of the skeletal lesions m ay result in reac-
tive bone form ation on conventional radiographs
54 55
(Fig. 8).
Sclerosis tends to progress from the periphery of the
lesion tow ard its center. Progressive sclerosis m ay
m ake subtle areas of bone involvem ent m ore visible,
contributing to the false im pression of disease progres-
sion. In one study of the effects of treatm ent, careful
retrospective analysis of bone scans revealed subtle
foci of increased uptake in m any areas w hich show ed
progressive sclerosis, indicating that the lesions had
been present prior to treatm ent.
56
Isotope uptake usually decreases follow ing treat-
m ent of a m etastasis. O ccasionally, increased uptake
is seen, particularly in the early phases of therapy.
This is know n as the flare phenom enon(Fig. 9).
Because of this, it has been suggested that an increas-
ing num ber of lesions is a m ore reliable m arker of
disease progression than increasing intensity of
uptake.
Areas of bone necrosis as a result of chem othera-
py m ay m im ic m etastases, further confusing interpre-
tation of post-treatm ent scans.
57
Bone m arrow scans
using
99m
Tc antigranulocyte m onoclonal antibody have
show n prom ise as an alternative m ethod of assessing
treatm ent response.
50
At least one author has suggested radiopharm aceu-
ticals m ay be used to im prove the tim ing of treatm ent
w ith chem otherapeutic bone seeking agents. In that
study, m any patients w ith prostate cancer receiving
androgen ablation therapy w ere found to dem on-
strate a peak in uptake of
99m
Tc-M D P about 3 w eeks
after institution of horm one treatm ent.
58
Q uantitative m ethods to evaluate the response to
treatm ent have been elusive. In one study, CT density
w as used to evaluate response. Im m ediately after suc-
cessful radiotherapy of vertebral m etastases as judged
by relief of pain, there w as a decrease in density by
25% w ithin the lesions, follow ed by an increase of
61% 3 m onths later. The bone surrounding the lesions
RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES RIBAK
Vol. 45 - N o. 1 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E 61
increased consistently from beginning to end.
59
O ther
studies utilizing dual-energy X -ray absorptiom etry
(D XA) and quantitative bone scintigraphy to m onitor
disease progression and response to treatm ent have
yielded equivocal results and further study in this area
is necessary.
60-62
M RI has show n som e prom ise as a m eans of assess-
ing treatm ent response. O ne group had good results
Fig. 7.Initial clinical presentation of bone
m etastasis w ith subsequent discovery of the
prim ary. A) A plain film of the pelvis in this
patient w ho presented w ith hip pain
revealed an obvious destructive lytic focus
in the right supra-acetabular region. B) This
area w as hot on Technetium -99m M D P bone
scan. C) The cortical destruction and m edi-
al soft tissue extension w as further charac-
terized by CT. D ) CT im ages w ith intrave-
nous contrast m ore proxim ally in the abdo-
m en revealed large bilateral renal cell carci-
nom as. E) The large, necrotic tum ors are
dem onstrated in the coronal plain on a T1
fat saturated postgadolinium im age.
Fig. 8.Progressive sclerosis on plain film s in response to treatm ent. A) Plain film of the pelvis in a m an w ith m etastatic prostate cancer reveals
m ultiple sites of m ixed lytic and sclerotic m etastases. B) O ne year later after radiation treatm ent the lytic areas have been replaced by pro-
gressive sclerosis.
RIBAK RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES
62 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E M arch 2001
using changes in the volum e of the bone and soft tis-
sue com ponents of lesions on T1 w eighted im ages
as criteria in patients w ith breast cancer m etastases. In
four of the patients, M RI revealed a response w hen
blood m arkers w ere equivocal and bone scan results
suggested disease progression.
63
A nother study
revealed sim ilarly encouraging results in utilizing M RI
in prediction of disease progression or stability in
patients w ith breast cancer m etastatic to the spine.
64
Several different M RI m ethods have been devised to
follow the progress of therapy of the prim ary tum or.
O f these, the m ost prom ising is the rate of uptake of
gadolinium (so-called dynam ic scanning). B y this
m ethod, it appears to be possible to distinguish viable
tum or from necrotic tissue and treatm ent effects.
65
Sim ilar m ethods have been applied to the evaluation
of m etastases.
66
These m ethods, w hile show ing prom -
ise in the investigational context, are not yet ready
for routine clinical im plem entation.
Image guided percutaneous biopsy
and other interventions
Progress in im aging technology, and especially the
increased use of CT scanning for guidance, has great-
ly increased the safety and applicability of needle
biopsy for skeletal m etastases. Prior to the era of CT-
guided procedures, needle biopsy w as considered
dangerous for spinal lesions above the lum bar spine.
It is currently rare to encounter a lesion for w hich
needle biopsy is not feasible.
67
Accuracy rates vary depending upon the nature of
the lesions being biopsied. In general, how ever, biop-
sies of m etastatic lesions have higher levels of accu-
racy than infections and prim ary tum ors.
68
A particularly interesting evaluation of im age-guid-
ed biopsy accuracy introduced the concept of effec-
tive accuracy, defined as the ability of the procedure
to replace open biopsy. In this report, the overall
Fig. 9.Flare phenom enon. A) Technetium -99m M D P scan at tim e of presentation dem onstrated m ultiple hotlesions due to m etastasis in
this patient w ith breast cancer. B) Six m onths after the initiation of treatm ent, the patient w as clinically im proved. H ow ever, the bone scan
dem onstrated increased uptake and an increased num ber of lesions.
RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES RIBAK
Vol. 45 - N o. 1 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E 63
accuracy w as approxim ately the sam e for m etasta-
ses, infection and prim ary lesions.Effective accuracy
w as best for m etastases (77% ), since clinicians tend-
ed to disbelieve negative biopsies for infection (72% ),
and pathologists asked for m ore tissue in prim ary
lesions (59% ). The authors noted that if a biopsy (or
other test) has to be repeated, it is w orse than w orth-
less since it adds expense and discom fort w ithout
im proving care.
68
The increased ease and safety of access to lesions
in a variety of different anatom ical locations has also
opened up therapeutic possibilities, including ablative
techniques for palliation, (and potentially cure). At
present, m ost such w ork has been concentrated on
soft tissue m etastases, especially in the liver.
The skeletal im plications of this w ork are just being
realized. In one study, 25 term inally ill patients w ith
painful lesions previously treated unsuccessfully by
radiation and/or chem otherapy, a sm all am ount of
95% alcohol w as injected under CT guidance. Seventy-
four per cent of cases experienced reduction in anal-
gesic needs w ithin 1-2 days after the procedure.
69
Another interesting report used CT guidance to per-
form tum or ablation w ith radiofrequency energy.
70
It
is likely that the near future w ill bring rapid grow th in
the role played by radiologists in the m anagem ent of
patients w ith m etastatic disease.
References
1. Fisher M S. Lum bar spine m etastasis in cervical carcinom a: a char-
acteristic patter. Radiol 1980;134:631-4.
2. Resnick, D onald and N iw ayam a, G en. D iagnosis of bone and joint
disorders. 2
nd
edition. Philadelphia: W B Saunders, 1988:3951.
3. von H ochstetter AR, Cserhati M D . Epitheliod sarcom a presenting
as chronic synovitis and m istaken for osteosarcom a. Skel Radiol
1995;24:636-8.
4. Resnick D . Skeletal m etastatic disease: its articular m anifestations.
O rthop Rev 1985:98-104.
5. Pauzer R, Istom in V, Segal-Lieberm an G , M atetzky S, Farfel Z.
Bilateral patellar m etastases as the clinical presentation of broncho-
genic adenocarcinom a. J Rheum atol 1996;23:939-41.
6. H elm s CA, Cann CE, Brunelle FO , G ilula LA, Chafetz N , G enant H K.
D etection of bone-m arrow m etastases using quantitative com put-
ed tom ography. Radiology 1981;40:745-50.
7. Jacobson AF, Stom per PC, Jochelson M S, Ascoli D M , H enderson
IC, K aplan W D . Association betw een num ber and sites of new
bone scan abnorm alities and presence of skeletal m etastases in
patients w ith breast cancer. J N ucl M ed 1990;31:387-92.
8. Ryan PJ, Fogelm an I. The bone scan: w here are w e now ? Sem
N ucl M ed 1995;25:76-91.
9. Sam ant R, G anguly P. Staging investigations in patients w ith breast
cancer: the role of bone scans and liver im aging. A rch Surg
1999;134(5):551-3.
10. Rydh A, Tom ic R, Tavelin B, H ietala SO , D am ber JE. Predictive val-
ue of prostate-specific antigen, tum our stage and tum our grade for
the outcom e of bone scintigraphy in patients w ith new ly diag-
nosed prostate cancer. Scand J U rol N ephrol 1999;33:89-93.
11. K em p PM , M aguire G A. W hich patients w ith prostatic carcinom a
require a staging bone scan? Br J U rol 1997;79:611-4.
12. Pom eranz SJ, Pretorius H T, Ram singh PS. Bone scintigraphy and
m ultim odality im aging in bone neoplasia: strategies for im aging in
the new health care clim ate. Sem in N ucl M ed 1994;24:188-207.
13. Staudenherz A, Steiner B, Puig S. Is there a diagnostic role for
bone scanning of patients w ith a high pretest probability for m et-
astatic renal cell carcinom a? Cancer 1999;85:153-5.
14. Brow n D H , Leakos M . The value of a routine bone scan in a m et-
astatic survey. J O tolaryngol 1998;27:187-9.
15. Schirrm eister H , G uhlm ann A, Elsner K , K otzerke J, G latting G ,
Rentschler M et al. Sensitivity in detecting osseous lesions depends
on anatom ic localization: planar bone scintigraphy versus 18F
PET. J N ucl M ed 1999;40:1623-9.
16. Schirrm eister H , G uhlm ann A, K otzerke J, Santjohanser C, K uhn
T, K reienberg R et al. Early detection and accurate description of
extent of m etastatic bone disease in breast cancer w ith fluoride ion
and positron em ission tom ography. J Clin O ncol 1999;17:2381-9.
17. Shreve PD , G rossm an H B, G ross M D , W ahl RL. M etastatic prostate
cancer: initial findings of PET w ith 2-deoxy-2-[F18]fluoro-D -glucose.
Radiology 1996;199:751-6.
18. Yeh SD , Im briaco M , Larson SM , G arza D , Zhang JJ, K alaigian H
et al. D etection of bony m etastases of androgen-independent
prostate cancer by PET-FD G . N ucl M ed Biol 1996;23:693-7.
19. Colem an RE. M onitoring of bone m etastases. Eur J Cancer 1998;34:
252-9.
20. D ehdashti F, Siegel BA, G riffeth LK , Fusselm an M J, Trask D D ,
M cG uire AH et al. Benign versusm alignant intraosseous lesions:
discrim ination by m eans of PET w ith 2[F-18]fluoro-2-deoxy-d-glu-
cose. Radiology 1996;200:243-7.
21. Reske SN . Recent advances in bone m arrow scanning skeleton. Eur
J N ucl M ed 1991;3:203-21.
22. Zoeller G , Sandrock D , M unz D L, Ringert RH . Bone m arrow im m u-
noscintigraphy versusconventional bone scintigraphy in the diag-
nosis of skeletal m etastass in urogenital m alignancies. Eur U rol
1994;26:141-4.
23. Lacic M , Bokulic T, Lukac J, K ovacic K , Baum RP, K usic Z. Bone
m arrow im m unoscintigraphy for the detection of skeletal m etas-
tases in patients w ith breast cancer. N ucl M ed Com m un
1999;20:135-43.
24. Finn H A, Sim on M A, M artin W B, D arakjian H . Scintigraphy w ith
gallium -67 citrate in staging of soft-tissue sarcom as of the extrem -
ity. J Bone Joint Surg 1987;69A:886-91.
25. Southee A E, K aplan W D , Jochelson M S, G onin R, D w yer JA ,
Antm an K H et al. G allium im aging in m etastatic and recurrent
soft tissue sarcom a. J N ucl M ed 1992;33:1594-9.
26. Schw eitzer M E, Levine C, M itchell D , G annon FH , G om ella LG .
Bulls-eyes and halos: useful M R discrim inators of osseous m etas-
tases. Radiology 1993;188:249-52.
27. U chida N , Sugim ura K, Kajitani A, Yoshizako T, Ishida T. M R im ag-
ing of vertebral m etastases: evaluation of fat saturation im aging. Eur
J Radiol 1993;17:91-4.
28. Jones K M , Schw artz RB , M antello M T, A hn SS, K horasani R,
M ukherji S et al. Fast spin-echo M R in the detection of vertebral
m etastases: com parison of three sequences. Am J N euroradiol
1994;15:401-7.
29. Castillo M , Arbelaez A, Sm ith JK , Fisher LL. D iffusion-w eighted
M R im aging offers no advantage over routine noncontrast M R
im aging in the detection of vertebral m etastases. A JN R A m J
N euroradiol 2000;21:948-53.
30. H anna SL, Fletcher BD , Fairclough D L, Jenkins JH 3rd, Le AH .
M agnetic resonance im aging of dissem inated bone m arrow disease
in patients treated for m alignancy. Skeletal Radiol 1991;20:79-84.
31. K attapuram SV, K hurana JS, Scott JA , El-K houry G Y. N egative
scintigraphy w ith positive m agnetic resonance im aging in bone
m etastasis. Skeletal Radiol 1990;19:113-6.
32. Colletti PM , D ang H T, D eseran M W , K err RM , Bosw ell W D , Ralls
RIBAK RAD IO LO G ICAL IM AG IN G FO R TH E D IAG N O SIS O F BO N E M ETASTASES
64 TH E Q U ARTERLY JO U RN AL O F N U CLEAR M ED ICIN E M arch 2001
PW . Spinal M R im aging in suspected m etastases: correlation w ith
skeletal scintigraphy. J M agn Reson Im aging 1991;9:349-55.
33. N ishim ura R, N agao K, M iyayam a H , Yasunaga T, Asao C, M atsuda
M et al. D iagnostic problem s of evaluating vertebral m etastasis
from breast carcinom a w ith a higher degree of m alignancy. Cancer
1999;85:1782-8.
34. Savelli G , Chiti A, G rasselli G , M accauro M , Rodari M , Bom bardieri
E. The role of bone SPET study in diagnosis of single vertebral
m etastases. Anticancer Res 2000;20:1115-20.
35. Sedonja I, Budihna N V. The benefit of SPECT w hen added to pla-
nar scintigraphy in patients w ith bone m etastases in the spine.
Clin N ucl M ed 1999;24:407-13.
36. H an LJ, A u-Yong TK , Tong W C, Chu K S, Szeto LT, W ong CP.
Com parison of bone single-photon em ission tom ography and pla-
nar im aging in the detection of vertebral m etastases in patients w ith
back pain. Eur J N ucl M ed 1998;25:635-8.
37. K osuda Shigeru, K aji Tatsum i, Yokoyam a H isaaki, Yokokaw a
Tokuzo, K atayam a M ichiaki, Iriye Toshiyuke et al. D oes bone
SPECT actually have low er sensitivity for detecting vertebral m etas-
tasis than M RI? J N ucl M ed 1996;37:975-8.
38. Freedm an G M , N egendank W G , H udes G R, Shaer AH , H anks G E.
Prelim inary results of a bone m arrow m agnetic resonance im ag-
ing protocol for patients w ith high-risk prostate cancer. U rology
1999;54:118-23.
39. Traill ZC, Talbot D , G olding S, G leeson FV. M agnetic resonance
im aging versus radionuclide scintigraphy in screening for bone
m etastases. Clin Radiol 1999;54:448-51.
40. Tam ada T, N agai K , Iizuka M , Im ai S, K ajihara Y, Yam am oto S
et al. Com parison of w hole-body M R im aging and bone scintigra-
phy in the detection of bone m etastases from breast cancer
(Japanese). N ippon Acta Radiol 2000;60:249-54.
41. Steinborn M M , H euck AF, Tiling R, Bruegel M , G auger L, Reiser M F.
W hole-body bone m arrow M RI in patients w ith m etastatic dis-
ease to the skeletal system . J Com put A ssist Tom ogr 1999;23:
123-9.
42. Iizuka M , N agai K, Tam ada T, Kajihara Y, Fukunaga M . Fundam en-
tal study of the detection of m etastatic bone tum ors w ith w hole-
body M R im aging (Japanese). N ippon Acta Radiol 1999;59: 470-2.
43. Eustace S, Tello R, D eCarvalho V, Carey J, W roblicka JT, M elhem
ER et al. A com parison of w hole-body turboSTIR M R im aging and
planar
99m
Tc-m ethylene diphosphonate scintigraphy in the exam -
ination of patients w ith suspected skeletal m etastases. A m J
Roentgenol 1997;169:1655-61.
44. W alker R, Kessar P, Blanchard R, D im asi M , H arper K, D eCarvalho
V et al. Turbo STIR m agnetic resonance im aging as a w hole-body
screening tool for m etastases in patients w ith breast carcinom a: pre-
lim inary clinical experience. J M agn Reson Im aging 2000;11:343-50.
45. Stabler A, Baur A, Bartl R, M unker R, Lam erz R, Reiser M . Contrast
enhancem ent and quantitative signal analysis in M R im aging of m ul-
tiple m yelom a: assessm ent of focal and diffuse grow th patterns in
m arrow correlated w ith biopsies and survival rates. A m J
Roentgenol 1996;167:1029-36.
46. K im H J, Ryu K N , Choi W S, Choi BK , Choi JM , Yoon Y. Spinal
involvem ent of hem atopoietic m alignancies and m etastasis: diffe-
rentiation using M R im aging. Clin Im aging 1999;23:125-33.
47. M oulopoulos LA, D im opoulos M A, Vourtsi A, G ouliam os A, Vlahos
L. Bone lesions w ith soft-tissue m ass: m agnetic resonance im ag-
ing diagnosis of lym phom atous involvem ent of the bone m arrow
versusm ultiple m yelom a and bone m etastases. Leuk Lym phom a
1999;34:179-84.
48. Alcalay M , Azais I, Brideon B, Babin P, Vanderm arcq P, D ebiais F
et al. Strategy for identifying prim ary m alignancies w ith inaugural
bone m etastases. Rev Rhum 1995;62:632-42.
49. Brage M E, Sim on M A. Evaluation, prognosis, and m edical treat-
m ent considerations of m etastatic bone tum ors. O rthopedics
1992;15:589-96.
50. Berna L, G erm a JR, Estorch M , Torres G , Blanco R, Carrio I. Bone
m arrow regeneration after horm onal therapy in patients w ith bone
m etastases from prostate cancer. J N ucl M ed 1991;32:2295-8.
51. Rougraff BT, Kneisl JS, Sim on M A. Skeletal m etastases of unknow n
origin. J Bone Joint Surg 1993;75A:1276-81.
52. Ciray I, A strom G , Sundstrom C, H agberg H , A hlstrom H .
Assessm ent of suspected bone m etastases. CT w ith and w ithout
clinical inform ation com pared to CT-guided bone biopsy. Acta
Radiol 1997;38:890-5.
53. van H olten-Verzantvoort AT, Bijvoet O LM , H erm ans J, H arinck
H IJ, Elte JW F, Beex LVA et al. Reduced m ortality from skeletal
m etastases in breast cancer patients during long-term bisphos-
phonate (APD ) treatm ent. Lancet 1987;2:983-5.
54. Colem an RE, W oll PJ, M iles M , Rubens RD . Treatm ent of bone
m etastases from breast cancer w ith (3-am ino-1-hydroxypylidene)-
1,1-bisphosphonate (APD ). Br J Cancer 1988;58:621-5.
55. H ortobagyi G N , Theriault RL, Porter L, Blayney D , Lipton A, Sinoff
C et al. Efficacy of pam idronate in reducing skeletal com plica-
tions in patients w ith breast cancer and lytic bone m etastases. N
Engl J M ed 1996;335:1785-91.
56. Ciray I, Astrom G , Andreasson I, Edekling T, H ansen J, Bergh J
et al. Evaluation of new sclerotic bone m etastases in breast can-
cer patients during treatm ent. Acta Radiol 2000;41:178-82.
57. Reuland P, K aiserling E, K lingebiel T, Feine U , N ietham m er D .
D ifferential diagnosis of m etastases in bone scans: chem otherapy
induced bone necrosis. N uklearm edizin 1999;38:38-42.
58. B ushnell D L, M adsen M , K ahn D , N athan M , W illiam s RD .
Enhanced uptake of
99m
Tc-M D P in skeletal m etastases from pros-
tate cancer follow ing initiation of horm one treatm ent: potential for
increasing delivery of therapeutic agents. N ucl M ed Com m un
1999;20:875-81.
59. Reinbold W D , W annenm acher M , H odapp N , Adler CP. O steoden-
sitom etry of vertebral m etastases after radiotherapy using quanti-
tative com puted tom ography. Skeletal Radiol 1989;18:517-21.
60. Shapiro CL, K eating J, Angell JE, Janicek M , G elm an R, H ayes D
et al. M onitoring therapeutic response in skeletal m etastases using
dual-energy X-ray absorptiom etry: a prospective feasibility study
in breast cancer patients. Cancer Invest 1999;17:566-74.
61. Rydh A, Ahlstrom K R, Larsson A, Johansson L, D am ber JE, Tom ic
R et al. Q uantitative bone scintigraphy. A m ethodological evalu-
ation in prostate cancer. Acta Radiol 2000;41:183-8.
62. Im briaco M , Larson SM , Yeung H W , M aw law i O R, Erdi Y,
Venkatram an ES et al. A new param eter for m easuring m etastatic
bone involvem ent by prostate cancer: the Bone Scan Index. Clin
Cancer Res 1998;4:1765-72.
63. Saip P, Tenekeci N , Aydiner A, D incer M , Inanc S, D em ir C et al.
Response evaluation of bone m etastases in breast cancer: value of
m agnetic resonance im aging. Cancer Invest 1999;17:575-80.
64. Brow n AL, M iddleton G , M acVicar AD , H usband JE. T1-w eighted
m agnetic resonance im aging in breast cancer vertebral m etastases:
changes on treatm ent and correlation w ith response to therapy. Clin
Radiol 1998;53:493-501.
65. Erlem ann R, Sciuk J, Bosse A, Ritter J, K usnierz-G laz CR, Peters PE
et al. Response of osteosarcom a and Ew ing sarcom a to preoper-
ative chem otherapy: assessm ent w ith dynam ic and static M R im ag-
ing and skeletal scintigraphy. Radiol 1990;175:791-6.
66. Sugim ura K, Kajitani A, O kizuka H , Sugihara M , M izutani M , Ishida
T. Assessing response to therapy of spinal m etastases w ith gado-
lim ium -enhanced M R im aging. J M agn Reson Im aging 1991;1:
481-4.
67. K attapuram SV, Rosenthal D I. Percutaneous biopsy of the cervical
spine using CT guidance. Am J Roentgenol 1987;149:539-41.
68. Fraser-H ill M A, Renfrew D L, H ilsenrath PE. Percutaneous needle
biopsy of m usculoskeletal lesions. 2 Cost effectiveness. A m J
Roentgenol 1992;158:813-8.
69. G angi A, K astler B, K linkert A, D ietem ann JL. Injection of alcohol
into bone m etastases under CT guidance. J Com put Assist Tom ogr
1994;18:932-5.
70. D upuy D E, H ong R, O liver B, G oldberg S. N ahum . Radiofrequency
ablation of spinal tum ors: tem perature distribution in the spinal
canal. Am J Roentgenol 2000;175(5)1263-6.