radiological imaging for the diagnosis of bone metastases · radiological imaging for the diagnosis...

Q J NUCL MED 2001;45:53-64

Radiological imaging for the diagnosis of bone metastases

L. D. RYBAK, D. I. ROSENTHAL

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 ofassociated 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 partof treatment planning. The frequency with which metas-tases are detected varies considerably with the type ofprimary tumor and with the methodology utilized fordetection. Four main modalities are utilized clinically: plain filmradiography, CT scan, nuclear imaging and magneticresonance imaging. In this discussion, we will review lit-erature on the radiology of skeletal metastases withrespect to lesion detection, assessment of response totreatment and possible therapeutic implications. Thebulk of the discussion will focus on MRI and nuclearstudies since most of the recent advances have beenmade in these areas. Key words: Neoplasm metastasis - Bone neoplasms secon-dary - Tomography, emission computed - Magnetic resonan-ce imaging - Radiography.

Skeletal metastases are unfortunately common.The frequency with which they are detectedvaries considerably with the type of tumor. It

also varies with the methodology used for detection.For some types of tumor, such as breast cancer, ske-letal metastases are readily detected by imaging stu-

From the Department of Radiology, Harvard Medical SchoolMassachusetts General Hospital, Boston, MA, USA

dies and form an important aspect of the clinical disea-se management because of the symptoms they pro-duce. For other conditions (such as chordoma), disse-minated skeletal metastases are frequently detected atthe time of autopsy, but are less apparent during life.

In general, the prognosis for patients presentingwith bone metastasis is poor. Patients with fewermetastases or solitary lesions appear to have a betteroutlook than those with multiple metastatic deposits.

Mechanisms

Direct invasion of the skeleton may result from anadjacent primary tumor (Fig. 1). Perhaps the mostprevalent example is invasion of the chest wall by alung cancer. Lymphogenous spread to bone is uncom-mon and difficult to document. However, secondaryinvasion of bone from involved lymph nodes is notrare. The spine is the most commonly affected site. Theleft side of the vertebral bodies is more often involvedbecause the left-sided nodes are closer to bone thanthe right.1 Direct skeletal invasion is usually accompa-nied by a detectable soft tissue mass, a feature that isunusual in metastases that arise by hematogenousspread.

Most tumor implants occur through the hematoge-nous route. The venous, rather than arterial routeappears more important, especially Batson’s paraver-tebral plexus. Presumably, the absence of valves in

Address reprint requests to: D. I. Rosenthal, Massachusetts GeneralHospital, Department of Radiology, WAC 515, 15 Parkman Street, Boston,MA 02114 (USA).

Vol. 45 - No. 1 THE QUARTERLY JOURNAL OF NUCLEAR MEDICINE 53

RYBAK RADIOLOGICAL IMAGING FOR THE DIAGNOSIS OF BONE METASTASES

54 THE QUARTERLY JOURNAL OF NUCLEAR MEDICINE March 2001

these veins permits retrograde distribution of tumorcells. Batson’s plexus communicates with the axialskeleton and the proximal long bones resulting in apredilection for these sites. The preference for axialinvolvement is also due to the greater vascularity of thered (hematogenous) marrow found in the axial skele-ton as opposed to the yellow (fatty) marrow found inthe appendicular bones.

Clinical features

Many of the patients complain of bone pain, whichmay be due to various mechanisms including release

of chemical mediators, elevated intra-osseous pressure,and periosteal elevation.2 The probability that a skel-etal metastasis will be painful may partly dependupon the nature of the primary. Metastases from lungand breast primaries are more likely to be symptomat-ic than are those from prostate cancer. Fractures andimpending fractures are also an important source ofpain, particularly in weight-bearing bones. Such frac-tures are more common in areas involved with lyticmetastases rather than blastic ones. They are difficultto manage, and often fail to heal.

Symptoms of arthritis may result from tumor inbone adjacent to a joint, mechanical collapse of anarticular surface due to lytic metastasis, or from syn-ovial implants of tumor. Primary malignancies thathave been reported to present in the latter fashioninclude lung, colon, breast, melanoma, and rhabdom-yosarcoma.3 Synovial implantation is very rarely doc-umented on imaging studies.

Arthritic symptoms may also be due to paraneo-plastic effects such as carcinoma polyarthritis, a con-dition in which there is sudden onset of rheumatoid-like symptoms with an asymmetric distribution in apatient who is rheumatoid-factor negative. Hypertro-phic (pulmonary) osteoarthropathy is another paraneo-plastic syndrome that may result in joint and longbone symptoms without local involvement.

Secondary gout is a known complication in can-cer patients, especially after treatment. Skeletal chang-es due to radiation, such as osteonecrosis and fracturemay be difficult to distinguish from metastatic lesionson imaging studies, but are usually not symptomatic.Finally, symptoms may be due to the onset of carci-noma-related rheumatic conditions such as Sjogrenssyndrome, lupus, and dermatomyositis.4 The mostcommon malignancy associated with arthritic symp-toms is leukemia.5

Detection - radiography

Radiography is commonly used to evaluate symp-tomatic sites and to confirm findings on other imag-ing studies. It is not generally recommended as ascreening method because of poor sensitivity. Theradiographic survey remains valuable in staging ofmultiple myeloma due to the poor sensitivity of theradioisotope scan in this condition.

Fig. 1.—Bone involvement by direct extension. Squamous cell carci-noma of the skin which has invaded the adjacent first metacarpalbone.

RADIOLOGICAL IMAGING FOR THE DIAGNOSIS OF BONE METASTASES RYBAK


Sensitivity depends partly on location. For instance,metastases to dense cortical bone are easier to detectthan those involving trabecular (medullary) bone. Inthe axial skeleton, medullary metastases may not bedetectable until 50% of the trabecular bone has beendestroyed.

Among the advantages of radiography is the fact thatcertain features may help to distinguish metastasesfrom other conditions and aid in identification of theprimary tumor. Radiography may be used to assess therisk of pathological fracture, which is said to be highif 50% of the cortex is destroyed by a lesion. This is acrude measure at best. Recent work using quantitativeanalysis of CT scans for this purpose has shown muchgreater promise.

Detection - CT

CT scanning has had a limited impact upon theclinical detection of skeletal metastases. Althoughmore sensitive than conventional radiography for thedetection of destructive bone lesions, CT is a cumber-some tool for screening the entire skeleton.Interestingly, CT can detect metastases within bonemarrow before bone destruction has occurred (Fig. 2).Tumor within the marrow causes an increase in atten-uation due to fat replacement. An attenuation differ-ence of more than 20 HU between the right and leftextremities is abnormal.6 Such findings are subtle,and easily overlooked by the radiologist. They arefar less apparent than the marrow changes seen onMRI.

Detection - nuclear methods

Since the introduction of technetium-based scanagents, approximately 25 years ago, the radioisotopebone scan has been the standard method for detectionof skeletal metastases. Isotope scanning is more sen-sitive than radiography for detection of most metas-tases. Tracer accumulates in the reactive new bone thatis formed in response to the lesion. Thus, althoughmost metastatic lesions are “hot”, cold lesions due tocomplete absence of reactive bone may be encoun-tered in particularly aggressive metastases (Fig. 3). Inaddition, the amount of accumulation is sensitive tothe level of blood flow. Diffuse accumulation of trac-

er throughout the skeleton due to disseminated skel-etal disease (super scan) may lead to the false impres-sion of a normal scan (Fig. 4).

The bone scan suffers from a lack of specificity.Tracer accumulation may occur in any skeletal loca-tion with an elevated rate of bone turnover and, thus,may accompany trauma, infection or arthropathy.The probability that an abnormal scan representsmetastatic tumor is directly related to the number ofabnormal foci. In a patient with foci of increaseduptake and a known primary tumor, the scan strong-ly suggests metastases. A small number (less than 4)of abnormalities is more likely to represent metastat-ic disease in some locations than others, with riblesions being particularly low-yield.7 Only 50% of sol-itary foci represent metastases, even among patientswith cancer.

This lack of specificity is well known and has leadto recommendations that positive scans be accompa-nied by radiographic correlation. However, given thegreater sensitivity of the bone scan, a positive radio-graph may confirm a finding, but a negative radio-graph does not exclude a metastasis.

Recent advances in isotope scanning methods, par-ticularly single photon emission computer tomography(SPECT), have improved the detection of metastases.SPECT imaging has increased both the sensitivity andspecificity of bone scanning.8 The tomographic pres-entation of SPECT images helps to identify foci ofabnormal uptake especially in the thicker body partssuch as the spine and pelvis. In addition, improvedspatial localization helps to distinguish between met-

Fig. 2.—CT scan of the proximal femora demonstrates increased den-sity of the left-sided marrow. This is due to replacement of the nor-mal marrow fat by tumor, a finding that may precede more obviousbone lysis.

RYBAK RADIOLOGICAL IMAGING FOR THE DIAGNOSIS OF BONE METASTASES


astatic foci and other abnormalities causing increaseduptake, such as spondylosis. Increased uptake thatinvolves the posterior vertebral body is more likely tobe due to metastasis.

A baseline radio-isotope bone scan is no longerrecommended in stage 1 or 2 breast cancer becauseof low yield.9 Several recent studies have suggestedthat the diagnostic yield of bone scan in patients withsmall and well-differentiated prostate carcinomas andPSA values <20 is too low to warrant routine use.10 11

If scanning is withheld from such individuals thenational savings would be $ 38 million/year.12

However, recent data indicates that for patients whohave received androgen depletion therapy, the PSAmay be unreliable in excluding metastases.

At present, the conventional bone scan has no rolein the detection of metastases from renal cell carcino-ma and head and neck cancer.13 14 For non-small celllung cancer, the bone scan is still recommended at thetime of diagnosis to aid in selection of patients for

surgery, although FDG-PET is also promising.8There have been conflicting reports as to the effi-

cacy of PET scanning and, as of yet, a consensushas not been reached. One study of 44 patients withknown metastatic disease from lung, prostate andthyroid primaries showed FDG-PET to be more sen-sitive and specific than conventional bone scan.15

Schirrmeister et al. studied 34 patients with breastcancer and concluded that FDG PET allowed forearlier detection of small marrow metastases thanconventional bone scan and lead to clinically sig-nificant changes in management of four patients 16

(Fig. 5).Other results have not been as encouraging. A

recent publication based on imaging of 98 bonelesions in 24 patients reported better specificity, butlower sensitivity for FDG-PET as compared to conven-tional bone scan.1 Another study similarly concludedthat FDG-PET can detect prostate metastases to bonewith moderate sensitivity (65%), but high specificity

Fig. 3.—A) Technetium-99m MDP scan. There is a focal absence of isotope uptake in the left ninth rib. At the margins of the “cold” area, slightincreased uptake is visible. Very destructive lesions may produce focal cold areas as the result of complete bone destruction in the affectedarea. Increased uptake in the marginal, less destroyed bone is not unusual. B) An oblique radiograph of the chest demonstrates completeabsence of the posterior ninth rib.

RADIOLOGICAL IMAGING FOR THE DIAGNOSIS OF BONE METASTASES RIBAK


(98%), and may have some value in lesion detec-tion.17 Yeh et al. reported dismal results in a smallseries of 13 patients with multiple bone metastasesfrom prostate cancer in which FDG-PET only detect-ed 18% of the lesions apparent by bone scan. Theauthors concluded that prostate metastases must havemeans other than glycolysis for metabolism and ener-gy.18 With regards to breast metastases, one groupreported that PET demonstrated superiority to bonescan with osteolytic metastases, but failed to consis-tently detect osteoblastic metastases.19 These observa-tions suggest a possible role for PET similar to plainfilm radiography, as confirmation of positive resultsfrom conventional technetium scans, rather than ameans of initial detection.

FDG appears not to be taken up by Paget disease,and, therefore, PET imaging may be useful to separ-ate Paget’s disease and other benign conditions frommetastases and/or sarcomatous degeneration.8 20

However, despite undeniable utility in certain circum-stances, presently there is no established role for PETimaging in the clinical evaluation of bone metasta-ses.

Marrow scanning has been reported to be moresensitive than the conventional bone scan in detectionof metastases from prostate cancer.21 In one studycomparing marrow imaging using 99mTc anti-NCA-95monoclonal antibody with conventional MDP, themarrow imaging technique detected almost doublethe number of lesions seen on MDP scans. However,the number of patients identified as having metasta-ses was the same: 13 in both instances 22. In anotherstudy of 23 patients with breast metastases, bone mar-row immunoscintigraphy (anti-NCA 95 Mab 250/183)demonstrated better specificity (88 vs 75%) and pos-itive predictive value (92 vs 85%) than conventionalbone scan with no significant difference in sensitivity.23Like PET, marrow scanning has not yet found a rolein routine clinical practice.

For certain types of primary tumors, (especiallylymphomas and soft tissue sarcomas) Gallium scan-ning may be a useful staging tool, detecting metasta-ses that are not otherwise observed.24 It may also behelpful to follow the effect of treatment in thesepatients.25

Detection - MRI

Magnetic resonance imaging (MRI) is highly sensi-tive to the presence of skeletal metastases within the

bone marrow. Since bone marrow (including hemat-opoietic or “red” marrow) contains a high percent-age of fat, T1-weighted magnetic resonance imagesgenerally reveal metastases as focal areas of low sig-nal intensity. Lesions can be often be distinguishedfrom focal deposits of red marrow on T1-weightedimages because the latter are more focal and mayhave centrally located fat, giving the appearance of a“bull’s eye”.26 On fat-suppressed T1-weighted images,metastases demonstrate mixed to high signal inten-sity.27 On T2-weighted images, metastatic lesions usu-ally are much brighter than normal marrow due totheir high water-content (Fig. 6). Metastases often,but not always, have a rim of bright T2 signal aroundthem (halo sign).26

A variety of different MRI pulse sequences have beenevaluated. In general, conventional spin-echo pulse

Fig. 4.—“Super Scan”. Diffuse uptake of Technetium-99m MDP inthis patient with widespread breast cancer metastasis. The use ofinappropriate technical settings may result in a relatively “normal”appearance in the absence of focal sites of more prominent uptake.

RIBAK RADIOLOGICAL IMAGING FOR THE DIAGNOSIS OF BONE METASTASES


sequences provide the best signal-to-noise ratios andanatomical detail. However, because of the need forrapid evaluation of large regions, fast spin echo andinversion recovery image sequences have been testedand found to be acceptable.28 Castillo et al. reported thatdiffusion weighted imaging had no advantage overnon-contrast enhanced T1 weighted imaging in detec-tion and characterization of vertebral metastases, butwas somewhat superior to T2 weighted imaging.29

Unfortunately, it can be difficult to distinguishchanges due to tumor from the affects of treatment,fracture, and inflammation. In one study, MRI scanswere compared with histological specimens at 21sites. Seven of these contained tumor, 14 did not. Allof the tumor-positive sites showed abnormalities on

MRI scans. However, in the sites shown to be free oftumor, a significant (more than 50%, depending uponpulse sequence) false-positive rate was encountered,presumably due to the effects of treatment.30

It has become clear the magnetic resonance imag-ing can detect metastases that are not apparent onradioisotope bone scans.31 MRI is particularly well suit-ed to detect spinal metastases, and most authoritiesagree that it is superior to planar scintigraphy for thispurpose. This may be partly due to the difficulty of rec-ognizing subtle radionuclide abnormalities, since ret-rospective review of the isotope scans may revealmany abnormalities that were initially missed.32 Onestudy of breast cancer patients concluded that MRIwas specifically superior to bone scan in detection of

Fig. 5.—The utility of PET in the detection of metastatic disease was demonstrated in this patientwith past history of Ewing’s sarcoma of the sacrum. A) A bone scan revealed foci of uptakein the lower right ribs which corresponded to a site of prior surgery, but no other suspicioussites. B) A subsequent PET scan revealed a focus of increased FDG uptake in the proximal righthumerus suspicious for metastasis. C) Plain films of the area revealed no obvious abnormal-ity. D) A T1 fat saturated coronal image of the right shoulder demonstrated an obvious focusof metastatic disease demonstrating diffuse enhancement.



aggressive spinal metastases that demonstrated estro-gen receptor negativity and increased biologic activity.33Most studies that have compared MRI to bone scanninghave used planar bone scans, not SPECT. Planar scin-tigraphy detects about 1/3-2/3 of the lesions seen byMRI. Multiple investigators have demonstrated thesuperiority of SPECT to planar imaging with regards todetection of vertebral metastases.34-36 Some authorsbelieve that SPECT makes nuclear scanning compar-able to MRI, with MRI better for vertebral body lesionsand SPECT better for the posterior elements.37

Advantages of the isotope scan include a large fieldof view, inexpensive radiopharmaceutical, low mor-bidity, and the ability to provide some functional andvascular information. MRI may be a simpler and lessexpensive method to evaluate the axial skeleton, butit has been considered less well-suited for screeningthe long bones. Because of this, some investigatorshave recommended that the role of MRI be restrictedto clarifying an equivocal bone scan.38 However, onegroup took the position that lesions missed in theextremities were not significant. In their analysis of 200patients with breast and prostate carcinoma, 3 of 4peripherally located skeletal lesions that were missedby MRI were detected by plain film radiographybecause they were painful.39

Identification of appendicular lesions by MRI hasbeen facilitated by development of faster pulsesequences. Several recent papers have reported thatin comparison to conventional bone scan, whole body

MRI utilizing whole body fast short tau inversionrecovery (STIR) sequences has significantly bettersensitivity and specificity.40-43 One of these papersnoted that detectability of rib lesions was suboptimalutilizing MRI.40 Walker et al. further suggest that wholebody MRI for patients with breast cancer may proveto be an effective means of detecting skeletal, brainand liver metastases with one study.44

There are a number of specific situations in which iso-tope scanning may be preferred to MRI. They includepatients with contraindications to MRI such as claustro-phobia and pacemakers, and patients with thyroid can-cer in which scanning may be done with iodine. MRIis preferred when the differential diagnosis includesmarrow diseases such as lymphoma, leukemia, myelo-ma and Waldenstrom macroglobulinemia.45-47

The factors that influence the choice of imagingmodalities continue to evolve. Progress in MRI hasbeen particularly rapid, and it appears probable thatit’s role in screening and staging in reference to skel-etal metastases will continue to grow and that it mayultimately replace the isotope scan. Until that time, iso-tope scanning, especially with SPECT imaging, willcontinue to have an important role.

Identification of the primary tumor

Unfortunately, patients may first come to medicalattention as the result of skeletal metastasis from an

Fig. 6.—The relative sensitivity of CT and MRI in the detection of skeletal metastasis. A) A CT scan of the pelvis in this patient with knowncolon cancer revealed no obvious area of abnormality. B) A T2 weighted axial image of the same patient demonstrated innumerous areasof abnormal high signal corresponding to metastatic deposits.



unknown primary tumor. For such individuals, imag-ing studies may be used to help identify the primarylesion. Common tumors with a high rate of bonemetastasis include: breast 72%, prostate 84%, thyroid50%, lung 31%, kidney 37%, pancreas 33%. Together,these account for more than 80% of primary tumors inpatients presenting with metastases.48-50

Previous studies have reported limited success inidentifying the primary tumor when a patient presentswith skeletal metastasis of unknown origin. In gener-al, the primary tumor has been identified in less than50%, even when patients were followed to autopsy.

In one of the most positive reports, the relative val-ue of the history and physical examination, CBC,erythrocyte sedimentation rate, blood chemistries,alkaline phosphatase, plain films of the lesion, radio-isotope bone scan and computed tomography of thechest, abdomen, and pelvis were compared. The pri-mary tumor was identified in 34/40 (85%).Interestingly, the laboratory values were unhelpful.History and physical examination revealed the pri-mary in 4, chest X-ray identified 17, CT of chest add-ed another 6, and CT of abdomen/pelvis addedanother 5. If used alone, CT would have diagnosed75% of all primaries on the initial evaluation. All oth-er modalities, including follow-up CT only added anadditional 10%.51

CT has an important role in providing imaging guid-ance for tissue sampling. Biopsy of the skeletal lesion,followed by examination of the tissue using sophisti-cated histologic techniques to limit the imaging searchmay be an equally valid approach to this difficultproblem.48 One study, however, concluded that inthe majority of cases, the combination of CT scan andclinical information provided comparable results to CT-directed biopsy.52

Radiographic features of the skeletal metastasesmay help direct the search for a primary lesion. Someprimary tumors tend to result in metastases that arepurely lytic in nature, such as lung, renal and thyroidcancer, others to be associated with variable degreesof sclerosis, especially prostate, breast, carcinoid andtumors of endocrine glands (Fig. 7).

Evaluation of treatment

The prognosis for patients presenting with bonemetastases is poor. In one series, only 4/578 patientswere free of disease 10 years after diagnosis of bone

disease. Mean survival for patients with all primarieswas 5 months after diagnosis.48 In general, patientswith solitary lesions or a small number of metastaseshave a better outlook than those with multiple meta-static deposits.

Skeletal metastases may respond to the chemother-apy or hormone therapy used for the primary tumor.They may also respond to radiation, or agentsdesigned to block bone resorption such as the newclass of bisphosphonate drugs.53

Response of the skeletal lesions may result in reac-tive bone formation on conventional radiographs 54 55

(Fig. 8).Sclerosis tends to progress from the periphery of the

lesion toward its center. Progressive sclerosis maymake subtle areas of bone involvement more visible,contributing to the false impression of disease progres-sion. In one study of the effects of treatment, carefulretrospective analysis of bone scans revealed subtlefoci of increased uptake in many areas which showedprogressive sclerosis, indicating that the lesions hadbeen present prior to treatment.56

Isotope uptake usually decreases following treat-ment of a metastasis. Occasionally, increased uptakeis seen, particularly in the early phases of therapy.This is known as the “flare phenomenon” (Fig. 9).Because of this, it has been suggested that an increas-ing number of lesions is a more reliable marker ofdisease progression than increasing intensity ofuptake.

Areas of bone necrosis as a result of chemothera-py may mimic metastases, further confusing interpre-tation of post-treatment scans.57 Bone marrow scansusing 99mTc antigranulocyte monoclonal antibody haveshown promise as an alternative method of assessingtreatment response.50

At least one author has suggested radiopharmaceu-ticals may be used to improve the timing of treatmentwith chemotherapeutic bone seeking agents. In thatstudy, many patients with prostate cancer receivingandrogen ablation therapy were found to demon-strate a peak in uptake of 99mTc-MDP about 3 weeksafter institution of hormone treatment.58

Quantitative methods to evaluate the response totreatment have been elusive. In one study, CT densitywas used to evaluate response. Immediately after suc-cessful radiotherapy of vertebral metastases as judgedby relief of pain, there was a decrease in density by25% within the lesions, followed by an increase of61% 3 months later. The bone surrounding the lesions



increased consistently from beginning to end.59 Otherstudies utilizing dual-energy X-ray absorptiometry(DXA) and quantitative bone scintigraphy to monitordisease progression and response to treatment have

yielded equivocal results and further study in this areais necessary.60-62

MRI has shown some promise as a means of assess-ing treatment response. One group had good results

Fig. 7.—Initial clinical presentation of bonemetastasis with subsequent discovery of theprimary. A) A plain film of the pelvis in thispatient who presented with hip painrevealed an obvious destructive lytic focusin the right supra-acetabular region. B) Thisarea was hot on Technetium-99m MDP bonescan. C) The cortical destruction and medi-al soft tissue extension was further charac-terized by CT. D) CT images with intrave-nous contrast more proximally in the abdo-men revealed large bilateral renal cell carci-nomas. E) The large, necrotic tumors aredemonstrated in the coronal plain on a T1fat saturated postgadolinium image.

Fig. 8.—Progressive sclerosis on plain films in response to treatment. A) Plain film of the pelvis in a man with metastatic prostate cancer revealsmultiple sites of mixed lytic and sclerotic metastases. B) One year later after radiation treatment the lytic areas have been replaced by pro-gressive sclerosis.



using changes in the volume of the bone and soft tis-sue components of lesions on T1 weighted imagesas criteria in patients with breast cancer metastases. Infour of the patients, MRI revealed a response whenblood markers were equivocal and bone scan resultssuggested disease progression.63 Another studyrevealed similarly encouraging results in utilizing MRIin prediction of disease progression or stability inpatients with breast cancer metastatic to the spine.64

Several different MRI methods have been devised tofollow the progress of therapy of the primary tumor.Of these, the most promising is the rate of uptake ofgadolinium (so-called dynamic scanning). By thismethod, it appears to be possible to distinguish viabletumor from necrotic tissue and treatment effects.65

Similar methods have been applied to the evaluationof metastases.66 These methods, while showing prom-ise in the investigational context, are not yet readyfor routine clinical implementation.

Image guided percutaneous biopsyand other interventions

Progress in imaging technology, and especially theincreased use of CT scanning for guidance, has great-ly increased the safety and applicability of needlebiopsy for skeletal metastases. Prior to the era of CT-guided procedures, needle biopsy was considereddangerous for spinal lesions above the lumbar spine.It is currently rare to encounter a lesion for whichneedle biopsy is not feasible.67

Accuracy rates vary depending upon the nature ofthe lesions being biopsied. In general, however, biop-sies of metastatic lesions have higher levels of accu-racy than infections and primary tumors.68

A particularly interesting evaluation of image-guid-ed biopsy accuracy introduced the concept of effec-tive accuracy, defined as the ability of the procedureto replace open biopsy. In this report, the overall

Fig. 9.—Flare phenomenon. A) Technetium-99m MDP scan at time of presentation demonstrated multiple “hot” lesions due to metastasis inthis patient with breast cancer. B) Six months after the initiation of treatment, the patient was clinically improved. However, the bone scandemonstrated increased uptake and an increased number of lesions.



accuracy was approximately the same for metasta-ses, infection and primary lesions. Effective accuracywas best for metastases (77%), since clinicians tend-ed to disbelieve negative biopsies for infection (72%),and pathologists asked for more tissue in primarylesions (59%). The authors noted that if a biopsy (orother test) has to be repeated, it is worse than worth-less since it adds expense and discomfort withoutimproving care.68

The increased ease and safety of access to lesionsin a variety of different anatomical locations has alsoopened up therapeutic possibilities, including ablativetechniques for palliation, (and potentially cure). Atpresent, most such work has been concentrated onsoft tissue metastases, especially in the liver.

The skeletal implications of this work are just beingrealized. In one study, 25 terminally ill patients withpainful lesions previously treated unsuccessfully byradiation and/or chemotherapy, a small amount of95% alcohol was injected under CT guidance. Seventy-four per cent of cases experienced reduction in anal-gesic needs within 1-2 days after the procedure.69

Another interesting report used CT guidance to per-form tumor ablation with radiofrequency energy.70 Itis likely that the near future will bring rapid growth inthe role played by radiologists in the management ofpatients with metastatic disease.

References

1. Fisher MS. Lumbar spine metastasis in cervical carcinoma: a char-acteristic patter. Radiol 1980;134:631-4.

2. Resnick, Donald and Niwayama, Gen. Diagnosis of bone and jointdisorders. 2nd edition. Philadelphia: WB Saunders, 1988:3951.

3. von Hochstetter AR, Cserhati MD. Epitheliod sarcoma presentingas chronic synovitis and mistaken for osteosarcoma. Skel Radiol1995;24:636-8.

4. Resnick D. Skeletal metastatic disease: its articular manifestations.Orthop Rev 1985:98-104.

5. Pauzer R, Istomin V, Segal-Lieberman G, Matetzky S, Farfel Z.Bilateral patellar metastases as the clinical presentation of broncho-genic adenocarcinoma. J Rheumatol 1996;23:939-41.

6. Helms CA, Cann CE, Brunelle FO, Gilula LA, Chafetz N, Genant HK.Detection of bone-marrow metastases using quantitative comput-ed tomography. Radiology 1981;40:745-50.

7. Jacobson AF, Stomper PC, Jochelson MS, Ascoli DM, HendersonIC, Kaplan WD. Association between number and sites of newbone scan abnormalities and presence of skeletal metastases inpatients with breast cancer. J Nucl Med 1990;31:387-92.

8. Ryan PJ, Fogelman I. The bone scan: where are we now? SemNucl Med 1995;25:76-91.

9. Samant R, Ganguly P. Staging investigations in patients with breastcancer: the role of bone scans and liver imaging. Arch Surg1999;134(5):551-3.

10. Rydh A, Tomic R, Tavelin B, Hietala SO, Damber JE. Predictive val-ue of prostate-specific antigen, tumour stage and tumour grade for

the outcome of bone scintigraphy in patients with newly diag-nosed prostate cancer. Scand J Urol Nephrol 1999;33:89-93.

11. Kemp PM, Maguire GA. Which patients with prostatic carcinomarequire a staging bone scan? Br J Urol 1997;79:611-4.

12. Pomeranz SJ, Pretorius HT, Ramsingh PS. Bone scintigraphy andmultimodality imaging in bone neoplasia: strategies for imaging inthe new health care climate. Semin Nucl Med 1994;24:188-207.

13. Staudenherz A, Steiner B, Puig S. Is there a diagnostic role forbone scanning of patients with a high pretest probability for met-astatic renal cell carcinoma? Cancer 1999;85:153-5.

14. Brown DH, Leakos M. The value of a routine bone scan in a met-astatic survey. J Otolaryngol 1998;27:187-9.

15. Schirrmeister H, Guhlmann A, Elsner K, Kotzerke J, Glatting G,Rentschler M et al. Sensitivity in detecting osseous lesions dependson anatomic localization: planar bone scintigraphy versus 18FPET. J Nucl Med 1999;40:1623-9.

16. Schirrmeister H, Guhlmann A, Kotzerke J, Santjohanser C, KuhnT, Kreienberg R et al. Early detection and accurate description ofextent of metastatic bone disease in breast cancer with fluoride ionand positron emission tomography. J Clin Oncol 1999;17:2381-9.

17. Shreve PD, Grossman HB, Gross MD, Wahl RL. Metastatic prostatecancer: initial findings of PET with 2-deoxy-2-[F18]fluoro-D-glucose.Radiology 1996;199:751-6.

18. Yeh SD, Imbriaco M, Larson SM, Garza D, Zhang JJ, Kalaigian Het al. Detection of bony metastases of androgen-independentprostate cancer by PET-FDG. Nucl Med Biol 1996;23:693-7.

19. Coleman RE. Monitoring of bone metastases. Eur J Cancer 1998;34:252-9.

20. Dehdashti F, Siegel BA, Griffeth LK, Fusselman MJ, Trask DD,McGuire AH et al. Benign versus malignant intraosseous lesions:discrimination by means of PET with 2[F-18]fluoro-2-deoxy-d-glu-cose. Radiology 1996;200:243-7.

21. Reske SN. Recent advances in bone marrow scanning skeleton. EurJ Nucl Med 1991;3:203-21.

22. Zoeller G, Sandrock D, Munz DL, Ringert RH. Bone marrow immu-noscintigraphy versus conventional bone scintigraphy in the diag-nosis of skeletal metastass in urogenital malignancies. Eur Urol1994;26:141-4.

23. Lacic M, Bokulic T, Lukac J, Kovacic K, Baum RP, Kusic Z. Bonemarrow immunoscintigraphy for the detection of skeletal metas-tases in patients with breast cancer. Nucl Med Commun1999;20:135-43.

24. Finn HA, Simon MA, Martin WB, Darakjian H. Scintigraphy withgallium-67 citrate in staging of soft-tissue sarcomas of the extrem-ity. J Bone Joint Surg 1987;69A:886-91.

25. Southee AE, Kaplan WD, Jochelson MS, Gonin R, Dwyer JA,Antman KH et al. Gallium imaging in metastatic and recurrentsoft tissue sarcoma. J Nucl Med 1992;33:1594-9.

26. Schweitzer ME, Levine C, Mitchell D, Gannon FH, Gomella LG.Bull’s-eyes and halos: useful MR discriminators of osseous metas-tases. Radiology 1993;188:249-52.

27. Uchida N, Sugimura K, Kajitani A, Yoshizako T, Ishida T. MR imag-ing of vertebral metastases: evaluation of fat saturation imaging. EurJ Radiol 1993;17:91-4.

28. Jones KM, Schwartz RB, Mantello MT, Ahn SS, Khorasani R,Mukherji S et al. Fast spin-echo MR in the detection of vertebralmetastases: comparison of three sequences. Am J Neuroradiol1994;15:401-7.

29. Castillo M, Arbelaez A, Smith JK, Fisher LL. Diffusion-weightedMR imaging offers no advantage over routine noncontrast MRimaging in the detection of vertebral metastases. AJNR Am JNeuroradiol 2000;21:948-53.

30. Hanna SL, Fletcher BD, Fairclough DL, Jenkins JH 3rd, Le AH.Magnetic resonance imaging of disseminated bone marrow diseasein patients treated for malignancy. Skeletal Radiol 1991;20:79-84.

31. Kattapuram SV, Khurana JS, Scott JA, El-Khoury GY. Negativescintigraphy with positive magnetic resonance imaging in bonemetastasis. Skeletal Radiol 1990;19:113-6.

32. Colletti PM, Dang HT, Deseran MW, Kerr RM, Boswell WD, Ralls



PW. Spinal MR imaging in suspected metastases: correlation withskeletal scintigraphy. J Magn Reson Imaging 1991;9:349-55.

33. Nishimura R, Nagao K, Miyayama H, Yasunaga T, Asao C, MatsudaM et al. Diagnostic problems of evaluating vertebral metastasisfrom breast carcinoma with a higher degree of malignancy. Cancer1999;85:1782-8.

34. Savelli G, Chiti A, Grasselli G, Maccauro M, Rodari M, BombardieriE. The role of bone SPET study in diagnosis of single vertebralmetastases. Anticancer Res 2000;20:1115-20.

35. Sedonja I, Budihna NV. The benefit of SPECT when added to pla-nar scintigraphy in patients with bone metastases in the spine.Clin Nucl Med 1999;24:407-13.

36. Han LJ, Au-Yong TK, Tong WC, Chu KS, Szeto LT, Wong CP.Comparison of bone single-photon emission tomography and pla-nar imaging in the detection of vertebral metastases in patients withback pain. Eur J Nucl Med 1998;25:635-8.

37. Kosuda Shigeru, Kaji Tatsumi, Yokoyama Hisaaki, YokokawaTokuzo, Katayama Michiaki, Iriye Toshiyuke et al. Does boneSPECT actually have lower sensitivity for detecting vertebral metas-tasis than MRI? J Nucl Med 1996;37:975-8.

38. Freedman GM, Negendank WG, Hudes GR, Shaer AH, Hanks GE.Preliminary results of a bone marrow magnetic resonance imag-ing protocol for patients with high-risk prostate cancer. Urology1999;54:118-23.

39. Traill ZC, Talbot D, Golding S, Gleeson FV. Magnetic resonanceimaging versus radionuclide scintigraphy in screening for bonemetastases. Clin Radiol 1999;54:448-51.

40. Tamada T, Nagai K, Iizuka M, Imai S, Kajihara Y, Yamamoto Set al. Comparison of whole-body MR imaging and bone scintigra-phy in the detection of bone metastases from breast cancer(Japanese). Nippon Acta Radiol 2000;60:249-54.

41. Steinborn MM, Heuck AF, Tiling R, Bruegel M, Gauger L, Reiser MF.Whole-body bone marrow MRI in patients with metastatic dis-ease to the skeletal system. J Comput Assist Tomogr 1999;23:123-9.

42. Iizuka M, Nagai K, Tamada T, Kajihara Y, Fukunaga M. Fundamen-tal study of the detection of metastatic bone tumors with whole-body MR imaging (Japanese). Nippon Acta Radiol 1999;59: 470-2.

43. Eustace S, Tello R, DeCarvalho V, Carey J, Wroblicka JT, MelhemER et al. A comparison of whole-body turboSTIR MR imaging andplanar 99mTc-methylene diphosphonate scintigraphy in the exam-ination of patients with suspected skeletal metastases. Am JRoentgenol 1997;169:1655-61.

44. Walker R, Kessar P, Blanchard R, Dimasi M, Harper K, DeCarvalhoV et al. Turbo STIR magnetic resonance imaging as a whole-bodyscreening tool for metastases in patients with breast carcinoma: pre-liminary clinical experience. J Magn Reson Imaging 2000;11:343-50.

45. Stabler A, Baur A, Bartl R, Munker R, Lamerz R, Reiser M. Contrastenhancement and quantitative signal analysis in MR imaging of mul-tiple myeloma: assessment of focal and diffuse growth patterns inmarrow correlated with biopsies and survival rates. Am JRoentgenol 1996;167:1029-36.

46. Kim HJ, Ryu KN, Choi WS, Choi BK, Choi JM, Yoon Y. Spinalinvolvement of hematopoietic malignancies and metastasis: diffe-rentiation using MR imaging. Clin Imaging 1999;23:125-33.

47. Moulopoulos LA, Dimopoulos MA, Vourtsi A, Gouliamos A, VlahosL. Bone lesions with soft-tissue mass: magnetic resonance imag-ing diagnosis of lymphomatous involvement of the bone marrowversus multiple myeloma and bone metastases. Leuk Lymphoma1999;34:179-84.

48. Alcalay M, Azais I, Brideon B, Babin P, Vandermarcq P, Debiais Fet al. Strategy for identifying primary malignancies with inauguralbone metastases. Rev Rhum 1995;62:632-42.

49. Brage ME, Simon MA. Evaluation, prognosis, and medical treat-ment considerations of metastatic bone tumors. Orthopedics1992;15:589-96.

50. Berna L, Germa JR, Estorch M, Torres G, Blanco R, Carrio I. Bonemarrow regeneration after hormonal therapy in patients with bonemetastases from prostate cancer. J Nucl Med 1991;32:2295-8.

51. Rougraff BT, Kneisl JS, Simon MA. Skeletal metastases of unknownorigin. J Bone Joint Surg 1993;75A:1276-81.

52. Ciray I, Astrom G, Sundstrom C, Hagberg H, Ahlstrom H.Assessment of suspected bone metastases. CT with and withoutclinical information compared to CT-guided bone biopsy. ActaRadiol 1997;38:890-5.

53. van Holten-Verzantvoort AT, Bijvoet OLM, Hermans J, HarinckHIJ, Elte JWF, Beex LVA et al. Reduced mortality from skeletalmetastases in breast cancer patients during long-term bisphos-phonate (APD) treatment. Lancet 1987;2:983-5.

54. Coleman RE, Woll PJ, Miles M, Rubens RD. Treatment of bonemetastases from breast cancer with (3-amino-1-hydroxypylidene)-1,1-bisphosphonate (APD). Br J Cancer 1988;58:621-5.

55. Hortobagyi GN, Theriault RL, Porter L, Blayney D, Lipton A, SinoffC et al. Efficacy of pamidronate in reducing skeletal complica-tions in patients with breast cancer and lytic bone metastases. NEngl J Med 1996;335:1785-91.

56. Ciray I, Astrom G, Andreasson I, Edekling T, Hansen J, Bergh Jet al. Evaluation of new sclerotic bone metastases in breast can-cer patients during treatment. Acta Radiol 2000;41:178-82.

57. Reuland P, Kaiserling E, Klingebiel T, Feine U, Niethammer D.Differential diagnosis of metastases in bone scans: chemotherapyinduced bone necrosis. Nuklearmedizin 1999;38:38-42.

58. Bushnell DL, Madsen M, Kahn D, Nathan M, Williams RD.Enhanced uptake of 99mTc-MDP in skeletal metastases from pros-tate cancer following initiation of hormone treatment: potential forincreasing delivery of therapeutic agents. Nucl Med Commun1999;20:875-81.

59. Reinbold WD, Wannenmacher M, Hodapp N, Adler CP. Osteoden-sitometry of vertebral metastases after radiotherapy using quanti-tative computed tomography. Skeletal Radiol 1989;18:517-21.

60. Shapiro CL, Keating J, Angell JE, Janicek M, Gelman R, Hayes Det al. Monitoring therapeutic response in skeletal metastases usingdual-energy X-ray absorptiometry: a prospective feasibility studyin breast cancer patients. Cancer Invest 1999;17:566-74.

61. Rydh A, Ahlstrom KR, Larsson A, Johansson L, Damber JE, TomicR et al. Quantitative bone scintigraphy. A methodological evalu-ation in prostate cancer. Acta Radiol 2000;41:183-8.

62. Imbriaco M, Larson SM, Yeung HW, Mawlawi OR, Erdi Y,Venkatraman ES et al. A new parameter for measuring metastaticbone involvement by prostate cancer: the Bone Scan Index. ClinCancer Res 1998;4:1765-72.

63. Saip P, Tenekeci N, Aydiner A, Dincer M, Inanc S, Demir C et al.Response evaluation of bone metastases in breast cancer: value ofmagnetic resonance imaging. Cancer Invest 1999;17:575-80.

64. Brown AL, Middleton G, MacVicar AD, Husband JE. T1-weightedmagnetic resonance imaging in breast cancer vertebral metastases:changes on treatment and correlation with response to therapy. ClinRadiol 1998;53:493-501.

65. Erlemann R, Sciuk J, Bosse A, Ritter J, Kusnierz-Glaz CR, Peters PEet al. Response of osteosarcoma and Ewing sarcoma to preoper-ative chemotherapy: assessment with dynamic and static MR imag-ing and skeletal scintigraphy. Radiol 1990;175:791-6.

66. Sugimura K, Kajitani A, Okizuka H, Sugihara M, Mizutani M, IshidaT. Assessing response to therapy of spinal metastases with gado-limium-enhanced MR imaging. J Magn Reson Imaging 1991;1:481-4.

67. Kattapuram SV, Rosenthal DI. Percutaneous biopsy of the cervicalspine using CT guidance. Am J Roentgenol 1987;149:539-41.

68. Fraser-Hill MA, Renfrew DL, Hilsenrath PE. Percutaneous needlebiopsy of musculoskeletal lesions. 2 Cost effectiveness. Am JRoentgenol 1992;158:813-8.

69. Gangi A, Kastler B, Klinkert A, Dietemann JL. Injection of alcoholinto bone metastases under CT guidance. J Comput Assist Tomogr1994;18:932-5.

70. Dupuy DE, Hong R, Oliver B, Goldberg S. Nahum. Radiofrequencyablation of spinal tumors: temperature distribution in the spinalcanal. Am J Roentgenol 2000;175(5)1263-6.

radiological imaging for the diagnosis of bone metastases · radiological imaging for the diagnosis...

Documents