Clinical & Experimental Metastasis 18: 385–390, 2001. © 2001 Kluwer Academic Publishers. Printed in the Netherlands.

385

Dual X-ray absorptiometry detects disease- and treatment-related alterations of bone density in prostate cancer patients Gillian L. Smith1,2 , Alan P. Doherty2,3 , Linda M. Banks4 , Jane Dutton5, Iain W. Hanham6 , Timothy J. Christmas2 & Richard J. Epstein1,7 Departments of 1 Metabolic Medicine, 2 Urology, 3 Medical Oncology, 4 Imaging and 6 Radiotherapy, Imperial College School of Medicine and 5 Department of Radiology, Guy’s Hospital, London, UK; 7 National Cancer Centre, Singapore Received 19 June 2000; accepted in revised form 2 February 2001

Key words: bone density, bone metastasis, dual X-ray absorptiometry, prostate-specific antigen, prostatic neoplasms

Abstract Metastatic bone disease is an important clinical problem which has proven difficult to study because of a lack of noninvasive investigative modalities. Here we show that dual-energy X-ray absorptiometry (DXA) scanning provides clinically useful information about the status of metastatic bone lesions in cancer patients undergoing palliative treatment. In the study group of 21 patients, a significant increase in metastatic bone mineral density (BMD) was confirmed in prostate (n = 14) relative to breast (n = 7) cancer patients. With respect to the prostate cancer cohort, further increases in lesional BMD were evident in all evaluable patients in whom biochemical progression occurred; conversely, lesional BMD declined in patients who had a partial response to therapy. BMD of uninvolved bone decreased with all types of androgen-deprivation therapy regardless of whether patients responded or relapsed. We conclude that BMD changes in both lesional and uninvolved bone are readily detectable in metastatic prostate cancer, and propose that DXA scanning represents a promising new approach to monitoring the natural history and therapeutic course of this disease.

Introduction Patients with metastatic prostate and breast cancer typically develop osteosclerotic and osteolytic lesions respectively [1, 2]. Assessment of therapeutic response in such lesions is difficult for two reasons: (i) response of bony secondaries involves both bone resorption and formation [3], and (ii) bone healing after successful anticancer therapy is slow with radiographic sclerosis of lytic metastases often not evident for up to 6 months after treatment [4]. A further complicating factor is that increased bone resorption may occur in patients with localized tumors [5, 6]. Conversely, the bone microenvironment exerts a critical influence on metastatic tumor behavior [7], as suggested by a recent report of reduced bone metastases in breast cancer patients treated with the antiosteoclastic drug clodronate [8]. Anticancer treatment can also influence bone not directly affected by metastatic lesions. Androgen deprivation therapy for prostate cancer may lead to osteoporosis [9, 10], a problem which may become more relevant with increasing use of early endocrine therapy, while breast cancer treatment may cause either bone loss [11, 12] or gain [13]. Bone mineral density (BMD) reflects the balance between bone formation and resorption. The main function of osteoblasts is to secrete bone matrix proteins, whereas that of osteoclasts is to synthesize proteolytic enzymes [14]; Correspondence to: Dr R.J. Epstein, National Cancer Centre, 11 Hospital Drive, Singapore 169610.

the relative activities of these cell types are influenced by hormones and paracrine factors implicated in tumor growth. In nonmalignant conditions BMD is often assessed using dual-energy x-ray absorptiometry (DXA) [15], a technique based on the principle that two x-ray energy sources are able to separate bone and soft tissue by differential absorption. Precision levels of approximately 1% are routine and the usual DXA radiation dose (<1 µSv) is less than four hours of background radiation [16]. These attributes suggest that DXA may prove ideal for use in noninvasive monitoring of cancer patients. In this study, we have performed serial DXA measurements in a cohort of patients with advanced prostate cancer in order to compare DXA with radionuclide bone scanning and plain radiography, to quantify BMD changes in areas of metastatic disease related to progression or response, and to quantify changes in BMD of uninvolved bone in patients on active treatment. Patients and methods Patient eligibility Fourteen subjects with prostate cancer were recruited to the study. Thirteen patients were enrolled at presentation and had therefore received no prior anticancer treatment. Twelve of these thirteen patients had bony metastases diagnosed on radionuclide bone scan at presentation (stage M1), while one patient had a normal bone scan and no

386 other manifestations of metastatic spread (stage M0). The remaining patient (stage M1) was recruited after instigation of androgen-deprivation therapy. Seven breast cancer patients were also recruited as controls. These women had all received tamoxifen prior to the study. Four had also received second line hormonal therapy prior to recruitment (one each of goserelin, formestane, limonene and medroxyprogesterone acetate). Three had previously received combination chemotherapy. One patient had received radiotherapy to the thoracic spine and one to the lumbar spine prior to recruitment. Data from patients was compared with age-matched control data provided by the manufacturer of the DXA scanner (Lunar Corporation, Madison, Wisconsin) [17–19]. Informed consent was obtained from all patients prior to enrolment. Clinical monitoring Patients underwent initial evaluation with serum tumor markers, plain radiography of the lumbar spine and 99 Tc MDP bone scanning. DXA scanning of the lumbar spine and total body was carried out at baseline and then at 6 and 12 months using a Lunar DPX-L bone densitometer. DXA scans of the lumbar spine (L1-4) and total body were performed in the standard manner using software version 1.34. Neck of femur and tibial BMDs were measured to represent peripheral uninvolved bone. Tibial BMD was calculated from the total body BMD scan by drawing a region of interest over both tibiae and fibulae on the baseline scan. This region of interest was then transferred to the follow-up scans. Exclusions were made for patients unable to attend because of intercurrent illness, disease progression or death. Patients were followed up at three monthly intervals with clinical assessment. Serum PSA was measured in prostate cancer patients. Biochemical response was defined as a fall in serum PSA of >50% baseline value or two consecutive falls of at least 5% and biochemical progression as a rise in serum PSA on two consecutive occasions. For breast cancer patients, disease status was defined as follows: stable disease – no new lesions and existing lesions unchanged; progressive disease – progression of lesions or appearance of new lesions [20]. Results Results are discussed in relation to the study endpoints: (i) comparison of information obtained from DXA with results of radionuclide bone scanning and plain radiography; (ii) quantification of changes in DXA findings in areas of metastatic disease related to biochemical progression or response; (iii) quantification of changes in BMD of uninvolved bone related to disease status or treatment. Characteristics of subjects are described in Table 1. Correlation of DXA scanning with radionuclide bone scan and plain radiographs Radionuclide bone scanning, plain radiographs and DXA BMD results were compared as illustrated in Figure 1A.

G.L. Smith et al. Table 1. Characteristics of study cohort with respect to treatment, disease progression and monitoring. Prostate cancer (n = 14)

Breast cancer (n = 7)

Mean age (range)

70 (55–84)

52 (28–77)

Racial origin Caucasian Afro-Caribbean Asian

11 3 0

6 0 1

Stage M0 M1

1 13

0 7

Drug treatment during study LHRH analogue Anti-androgen monotherapy Maximal androgen blockade Medroxyprogesterone acetate Chemotherapy Corticosteroids

4 5 4 0 0 0

0 0 0 1 4 1

Skeletal radiotherapy during study Pelvis Lumbosacral spine

1 1

0 0

Disease status at 12 months Biochemical response Biochemical progression Stable disease Progressive disease Deceased

8 6 – – 3

– – 4 3 3

14 8 10

7 5 4

DXA scans completed Baseline 6 month 12 month

DXA confirmed increased BMD compared with age- and sex-matched controls in areas of osteosclerotic metastases in prostate cancer patients that were evident on plain radiographs. This is illustrated by case PC(s), a typical prostate cancer patient presenting with osteosclerotic lumbar vertebral metastases. Lumbar vertebral BMD values, however, were within normal limits in a prostate cancer patient with multiple vertebral metastases diagnosed on radionuclide bone scanning but in whom plain radiographs of the lumbar spine appeared entirely normal (case PC(ns)), indicating that DXA scanning cannot detect all bone metastases. Mean lumbar spine BMD in the patients with prostate cancer was greater than in patients with metastatic breast cancer or in age-matched men (Figure 1B). BMD is of course generally lower in females. Baseline BMD data for lumbar spine, tibia and whole body are shown for prostate cancer patients in Table 2. Six of seven breast cancer patients were diagnosed as having osteolytic metastases at presentation. The other patient presented with the rare osteosclerotic phenotype. By the time of recruitment, however, all had received endocrine

DXA detects disease- and treatment-related alterations of bone density in prostate cancer patients

387

Table 2. Baseline data. Measurement of lumbar vertebrae, tibia, neck of femur and whole body BMD by DXA in prostate cancer patients. M1 = bony metastases diagnosed on radionuclide bone scan at presentation, M0 = normal bone scan and no other manifestations of metastatic spread. DXA BMD = bone mineral density measured by DXA. L = lumbar vertebra. Patient

1 2 3 4 5 6 7 8 9 10 11 12 13 14

L1

L2

L3

DXA BMD g cm−2 L4 L1-4 Tibia

0.740 1.440 2.455 1.106 0.858 0.905 1.929 0.963 0.773 2.285 2.091 1.150 1.373 1.002

0.730 1.528 1.776 1.153 0.908 1.090 1.317 1.123 1.152 2.403 2.275 1.317 1.442 1.033

0.831 1.352 1.428 1.161 1.107 1.055 1.665 1.348 1.084 2.583 2.342 1.363 1.443 1.191

0.878 1.426 1.725 1.414 1.126 1.047 1.491 1.169 1.089 2.323 2.089 1.368 1.398 1.424

Stage

M1 M1 M1 M1 M1 M1 M1 M1 M1 M1 M1 M1 M1 M0

0.806 1.436 1.833 1.221 1.011 1.028 1.594 1.154 1.036 2.392 2.196 1.307 1.414 1.173

– 1.51 1.138 1.412 1.116 1.278 1.166 1.149 – – 1.169 1.372 1.29 1.274

Neck of femur

Whole body

0.725 0.972 – 1.092 0.788 0.998 0.915 – 0.753 – 1.879 0.947 0.933 –

– 1.188 1.256 1.255 1.082 1.271 1.224 1.092 1.092 1.524 1.361 1.279 1.190 1.144

treatment and/or chemotherapy, and most of the osteolytic lesions therefore had some evidence of treatment-related sclerosis as well as lytic areas. DXA measurements from the lumbar spine in these patients were slightly greater than those from areas of normal bone in an age- and sex-matched population (data not shown). This may reflect the sclerosis that occurs in lytic metastases after treatment. A typical breast cancer patient who presented with lytic metastases which underwent some sclerosis after treatment is illustrated by case BC in Figure 1A. Baseline measurements of lumbar vertebrae, tibia, neck of femur and whole body are shown for all prostate cancer patients in Table 2.

developed biochemical progression over the study period, whereas lumbar BMD fell slightly in three patients with a response to therapy. Six of the breast cancer patients had lumbar vertebral metastases diagnosed on radionuclide bone scanning at baseline. Three of these patients had stable disease over the period of the study and three developed progressive disease. Although all patients with stable disease completed the study, the patients with progressive disease had died by the end of the study. Useful information on changes in lumbar vertebral bone density in patients with progressive breast cancer was thus not obtained in this study.

Changes in BMD with disease progression and response

Changes in non-lesional BMD with treatment

Changes in lesional BMD in relation to disease status were assessed using measurements of lumbar vertebral BMD in prostate cancer patients undergoing treatment for active disease. Some patients did not complete all examinations because of intercurrent illness, disease progression or death (Table 1). Six of the prostate cancer patients who completed the study had metastases in lumbar vertebrae diagnosed on radionuclide bone scan and were therefore included in this analysis. In order to quantify biochemical progression and response, a net change in PSA for the 12 months of the study was calculated as follows:

The effects of anticancer treatment on normal bone were assessed by measuring tibial DXA BMD in thirteen prostate cancer patients who had received no anticancer treatment. All thirteen were started on androgen deprivation therapy at diagnosis (Table 1). Four received luteinising hormonereleasing hormone (LHRH) agonists alone, five antiandrogen monotherapy (four with bicalutamide and one with flutamide) and four combined therapy with both an LHRH agonist and an anti-androgen (maximal androgen blockade). Two patients were excluded from this analysis: one who was unable to lie flat on the scanner table in order for the measurements to be undertaken and one who had only a single baseline scan. As shown in Figure 2A, tibial BMD decreased with all types of androgen-deprivation therapy. Moreover, reduced bone density occurred regardless of whether patients maintained their response to treatment or relapsed during the study period, although the reduction was greater in patients with biochemical progression (Figure 2B). Serial neck of femur BMD measurements were obtained in only five patients. Three of these patients had metasta-

Net change in PSA = PSA at 12 months − PSA nadir (in patients progressing); Net change in PSA = PSA nadir − highest value during study (in responders). Percentage changes in vertebral BMD over 12 months are tabulated with net PSA changes in Table 3. Lumbar vertebral BMD increased over 12 months in three patients who

388

G.L. Smith et al. Table 3. Change in L1-4 vertebral BMD and net PSA change over 12 months in prostate cancer patients with lumbar vertebral metastases.

Figure 1. Correlation of DXA BMD with radionuclide bone scanning and plain radiography. A. Patient L1-4 BMD
is compared with the value for age- and sex-matched controls  . PC(s) – prostate cancer patient presenting with osteosclerotic metastases; PC(ns) – prostate cancer with non-sclerotic osseous metastases (positive radionuclide bone scan but normal lumbar spine radiographs); BC – patient with metastatic breast cancer previously treated by tamoxifen and chemotherapy. Mixed osteolytic and osteosclerotic lesions were present. B. Mean L1-4 BMD in prostate cancer patients (PC), age-matched men (AMM) and breast cancer patients (BC).

tic disease involving the femur on plain radiographs and/or bone scanning, leaving only two patients in whom the neck of femur represented non-lesional bone. One patient receiving antiandrogen monotherapy underwent a 9% increase in neck of femur BMD over 12 months. The other patient was treated with an LHRH analogue alone and had a 10%reduction in neck of femur BMD over 12 months (data not shown). In view of the small numbers of subjects, the neck of femur measurements were not suitable for the assessment of changes in density of uninvolved bone in this study. The variety of treatments received by the breast cancer patients both prior to and during this study precluded the systematic assessment of the effects of breast cancer treatment on non-lesional BMD.

Discussion The central finding of this study is that changes in lesional BMD parallel the clinical course of metastatic prostate cancer. Serial DXA measurements of lumbar vertebrae revealed that lesional BMD increased in patients with refractory

Patient

Net PSA change in 12 months (ng ml−1 )

Change in L1-4 BMD in 12 months (%)

1 2 3 4 5 6

27 1735 250 −334 −124 −1993

12.5 27.5 22.75 −6.5 −3.25 −1

Figure 2. Changes in uninvolved (tibial) BMD between first and final DXA scans (6 or 12 months) in patients treated for prostate cancer (A) according to treatment and (B) according to disease status. AA = antiandrogen monotherapy, LHRH-A = LHRH agonist, MAB = maximal androgen blockade (LHRH agonist and antiandrogen).

metastatic disease, presumably reflecting progressive osteosclerosis. In contrast, lumbar BMD tended to decrease in patients maintaining a response to androgen deprivation over a 12-month period, suggesting temporary reversal of the osteoblastic response. Although decreases in responders were minimal and could be partly the result of effects on uninvolved bone, the changes observed were in striking contrast with the increases observed in patients with biochemical progression. We conclude that changes in BMD detected by DXA can be used to assess progression and response in metastatic prostate cancer. Our findings are consistent with a recent description of the use of DXA in monitoring skeletal metastases in breast cancer [21]. The osteopenia associated with hormonal treatment is a surprising feature of our results. Osteoporosis is a wellrecognized complication of castration (either surgical or medical) [9, 10]. Although the number of patients involved was small, our results suggest that antiandrogens as well as

DXA detects disease- and treatment-related alterations of bone density in prostate cancer patients LHRH agonists may lead to bone loss. In contrast, other preliminary work comparing the effects of bicalutamide and castration on bone loss suggests that antiandrogens may in fact preserve bone density [22]. Larger controlled studies will therefore be required in order to confirm the effect of these agents on bone mass in prostate cancer patients. Since the early use of hormonal manipulation in patients with locally advanced or metastatic prostate cancer is increasingly favored [23], a growing number of patients may be at risk of clinical manifestations of treatment-related osteoporosis. Routine DXA monitoring of BMD in skeletal sites not directly affected by metastatic disease may therefore become an important aspect of holistic management of patients on long-term hormonal therapy for prostatic cancer. To date, response of osseous metastatic disease to treatment has been assessed by radionuclide bone scanning, plain radiography or surrogate markers such as changes in serum PSA. Certainly, a significant fall in PSA is a good indication of response to treatment [24]. However, patients with certain histological variants of prostate carcinoma have extensive metastases despite a relatively low PSA, and use of PSA is not therefore universally applicable [25]. The extent of osseous disease present can be quantified accurately from a radionuclide bone scan by calculating an index based on the weighted proportion of tumour involvement in individual bones. This index is a good prognostic indicator [26]. However, the use of serial assessments of this index in monitoring disease progress has not been assessed. It is known that the use of standard bone scanning to assess response to treatment may be unreliable. Following successful treatment of metastatic disease, the healing process of new bone formation causes an initial increase in tracer uptake and scans performed during this phase (<6 months) are likely to show increased intensity and number of hot spots despite successful treatment: after 6 months bone scan appearances may improve as increased production of new bone ceases and isotope uptake falls [27–29]. Bone scanning within 6 months of a change in therapy must therefore be interpreted with caution. Novel techniques that allow early prediction of therapeutic response are therefore needed, and DXA may well fulfil this role. The inability of DXA scanning to detect all metastatic lesions in this series indicates that the technique is not sufficiently sensitive for diagnosis of bone metastases. In addition, while DXA can detect abnormal BMD in areas affected by metastatic disease, it cannot reliably distinguish the cause of such abnormalities: for example, osteosclerotic metastases, collapse due to a pathological fracture, healing after chemotherapy and degenerative disease may all produce a similar increase in BMD. DXA scanning therefore cannot identify accurately the pathology responsible for areas of increased uptake on bone scanning when results of the latter investigation are equivocal nor does it contribute to the management of patients with persistent symptoms suggestive of bone secondaries in the absence of diagnostic abnormalities on bone scanning or plain radiography. Magnetic resonance imaging (MRI) or CT scanning therefore remain the investigations of choice for symptomatic patients

389

[30–32]. Since intervals between DXA measurements were relatively long (6 months) in this study, it is not yet possible to draw firm conclusions as to whether DXA scanning is superior to serial radionuclide bone scans in the assessment of response of bone metastases to therapeutic intervention. Control data used in comparison of baseline BMD with age- and sex-matched subjects were obtained from the manufacturer of the DXA scanner. It is of course possible that this data does not accurately reflect the racial and socioeconomic background of the small number of patients in this study. Apparent differences between BMD of prostate cancer patients at presentation and age-matched men must therefore be interpreted with caution. There are also limitations of the breast cancer control data. As females, these individuals would be expected to have a different baseline bone density. Moreover, in this study, all of the breast cancer patients had previously been treated and consequently they constitute a rather heterogeneous group. This was a pilot study of a promising new technology and it will of course be necessary to confirm the findings in a larger cohort of patients. We also recognize that newer scanners such as the Expert may be more sensitive than the model used in this study, and such technological advances may enhance the value of DXA scanning in the evaluation of metastatic bone disease. In conclusion, our results suggest that measurement of lesional BMD using DXA methodology may play a valuable role in monitoring the biological behavior of metastatic prostate cancer. DXA may also prove valuable in screening patients receiving anticancer treatment over long periods for bone loss in an attempt to prevent osteoporotic complications. Further studies are now underway to confirm and extend these preliminary findings in larger numbers of patients.

Acknowledgements We are grateful to Steve McNealy and Robert Higgins for performing the DXA scans.

References 1.

2.

3. 4. 5.

6. 7.

Belliveau RE, Spencer RP. Incidence and sites of bone lesions detected by 99mTc-polyphosphate scans in patients with tumors. Cancer 1975; 36: 359–63. Cook G, Watson F. Events in the natural history of prostate cancer: using salvage curves, mean age distributions and contingency coefficients. J Urol 1968; 96: 87–96. Hayward JL, Carbone PP, Heuson JC et al. Assessment of response to therapy in advanced breast cancer. Cancer 1977; 3: 1389–94. Coleman RE. Evaluation of bone disease in breast cancer. Breast 1994; 3: 73–8. Nguyen-Pamart M, Bonneterre J, Hecquet B. Urinary excretion of deoxypyridinoline in patients with breast cancer. Anticancer Res 1995; 15(4): 1601–3. Delmas PD, Fontana A. Bone loss induced by cancer treatment and its management. Eur J Cancer 1998; 34: 260–2. Hauschka PV, Manrakos AE, Iafrati MD et al. Growth factors in bone matrix. Isolation of multiple types of affinity chromatography on heparin sepharose. J Biol Chem 1986; 261: 12665–74.

390 8. Diel IJ, Solomayer E-F, Costa SD et al. Reduction in new metastases in breast cancer with adjuvant clodronate treatment. N Engl J Med 1998; 339(6): 357–63. 9. McGrath SA, Diamond T. Osteoporosis as a complication of orchiectomy in 2 elderly men with prostatic cancer. J Urol 1995; 154: 535–6. 10. Townsend MF, Sanders WH, Northway RO et al. Bone fractures associated with luteinizing hormone-releasing hormone agonists used in the treatment of prostate carcinoma. Cancer 1997; 79: 545–50. 11. Bruning PF, Pit MJ, de Jong-Bakker M et al. Bone mineral density after adjuvant chemotherapy for premenopausal breast cancer. Br J Cancer 1990; 61: 308–310. 12. Powles TJ, Hickish T, Kanis JA et al. Effect of tamoxifen on bone mineral density measured by dual energy X-ray absorptiometry in healthy premenopausal and postmenopausal women. J Clin Oncol 1996; 14: 78–84. 13. Love RR, Mazess RB, Barden HS, al e. Effects of tamoxifen on bone mineral density in postmenopausal women with breast cancer. N Engl J Med 1992; 326: 852–6. 14. Rodan GA. Introduction to bone biology. Bone 1992; 13 (Suppl 1): S3–S6. 15. Wahner HW, Fogelman I. The Evaluation of Osteoporosis. Dual XRay Absorptiometry in Clinical Practice. London: Martin Dunitz 1994. 16. Banks LM. Dual energy X-ray absorptiometry (DXA). In Grainger RG, Allison DJ (eds) Diagnostic Radiology – a Textbook for Radiologists, 3rd ed. London: Churchill Livingstone 1997; 125–36. 17. Kroger H, Laituieu K. Bone mineral density measured by dual energy X-ray absorptiometry in normal men. Eur J Clin Nutr 1992; 22: 454– 60. 18. Krall EA, Dawson-Hughes B, Hirst K et al. Bone mineral density and biochemical markers of bone turnover in healthy elderly men and women. J Gerontol 1997; 52A: M61–M67. 19. Burgher H, van Daele PLA, Algra D et al. The association between age and bone mineral density in men and women aged 55 years and over. The Rotterdam study. J Bone Miner Res 1994; 25: 1–13. 20. Hayward JL, Rubens RD, Carbone PP. Assessment of response to therapy in advanced breast cancer. Br J Cancer 1977; 35: 292–8.

G.L. Smith et al. 21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Shapiro CL, Keating J, Angell JE et al. Monitoring therapeutic response in skeletal metastases using dual-energy X-ray absorptiometry: A prospective feasability study in breast cancer patients. Cancer Invest 1999; 17: 566–74. Tyrell CJ, Blake GM, Iversen P et al. ‘Casodex’ may preserve bone mineral density of patients with advanced prostate cancer. BJU Int BJU Intl 2000; 86 (Suppl 3): 220–1. Group MRCPCWPI. Immediate versus deferred treatment for advanced prostate cancer: Initial results of the Medical Research Council trial. Br J Urol 1998; 79(2): 235–46. Scher HI, Kelly WMK, Zhang Z-F et al. Post-therapy serum prostate-specific antigen level and survival in patients with androgenindependent prostate cancer. J Natl Cancer Inst 1999; 91: 244–51. Schwartz LH, LaTrenta LR, Bonaccio E, Kelly WK, Scher HI, Panicek DM. Small cell and anaplastic prostate cancer: Correlation between CT findings and prostate-specific antigen level. Radiology 1998; 208: 735–8. Sabbatini P, Larson SM, Kremer A, Zhang ZF, Sun m, Yeung H, et al. Prognostic significance of extent of disease in bone in patients with androgen-independent prostate cancer. J Clin Oncol 1999; 17: 948– 57. Pollen JJ, Witzkun KF, Ashburn WL. The flare phenomenon on radionuclide bone scan in metastatic prostate cancer. Am J Radiother 1984; 142: 773–6. Rossleigh MA, Lovegrove FTA, Reynolds PM, Byrne MJ, Whitney BP. The assessment of response to therapy of bone metastases in breast cancer. Aust NZ J Med 1984; 14: 19–22. Coleman RE, Rubens RD, Fogelman I. The bone scan flare following systemic treatment for bone metastases. J Nucl Med 1988; 29: 1354– 9. Yuh WT, Zacher CK, Barloon TJ, Sato Y, Sickels WJ, Hawes DR. Vertebral compression fractures: A distinction between benign and malignant causes with MR imaging. Radiology 1989; 172: 215–8. Jones AL, Williams MP, Powles TJ, et al. Magnetic resonance imaging in the detection of skeletal metastases in patients with breast cancer. Br J Cancer 1990; 62: 296–8. Coleman RE. Monitoring of bone metastases. Eur J Cancer 1998; 34(2): 252–9.