| Japanese Journal of Clinical Oncology | Pages |
Biochemical Markers of Bone Turnover in Breast Cancer Patients with Bone Metastases: A Preliminary Report
Introduction
Materials And Methods
Patients
Diagnosis of Bone Metastases
Measurement of Bone Metabolic Markers
Statistical Considerations
Results
Discussion
Acknowledgment
References
Biochemical Markers of Bone Turnover in Breast Cancer Patients with Bone Metastases: A Preliminary Report
Methods: A panel of bone turnover markers was assessed in 11 normocalcemic patients with bone metastases from BC and in 19 BC patients without clinical evidence of bone metastases. Bone formation was investigated by measuring serum bone isoenzyme of alkaline phosphatase (BALP), osteocalcin (OC) and carboxy-terminal propeptide of type I procollagen (PICP): Bone resorption was investigated by measuring serum carboxy-terminal telopeptide of type I collagen (ICTP), fasting urinary pyridinoline (Pyr) and deoxypyridinoline (D-Pyr).
Results: PICP was influenced by age and menopausal status. Significant correlations were observed between each of bone turnover markers except between BALP and OC. The mean levels of the six bone turnover markers were higher in patients with bone metastases than in those without them and significance was observed except for OC. The best diagnostic efficiency by receiver-operating characteristic (ROC) analysis was provided by ICTP followed by Pyr or D-Pyr, BALP, PICP and OC and significance was observed between ICTP and OC. Multiple logistic regression analysis adjusted by age revealed that the only significant marker related to bone metastases was ICTP.
Conclusions: Serum ICTP appears to be the leading marker of bone metastases from BC. However, to reveal the clinical usefulness of these markers, further examination will be needed to account for the ease and cost-effectiveness of the measurements.
INTRODUCTION
Bone metastases are frequent in breast cancer (BC) patients in addition to prostate, thyroid and kidney cancer patients (1). However, early and accurate detection is still difficult by clinical symptoms or available imaging techniques such as scintigraphy, roentgenography, computed tomography (CT) or magnetic resonance imaging (MRI). Bone scintigraphy is frequently used during the follow-up of BC and is very sensitive in detecting active benign and malignant bone diseases, but its specificity for metastatic bone tumor is poor (1). Moreover, scintigraphy is costly, time consuming and entails a radiation hazard when performed during follow-up (2,3). As for roentgenography, it is effective only when metastases are more than 1-2 cm in diameter and about 30% of bone mineral has been lost or increased (4). Serum tumor markers such as carcinoembryonic antigen (CEA) and CA 15-3 are sometimes useful for detecting metastatic BC and deciding treatment response in advanced or recurrent BC. However, if patients have other metastatic sites, such tumor markers are not necessarily useful for detecting bone lesions or for deciding treatment response for bone lesions. An appropriate way to determine bone mineral density (BMD) has recently been developed to evaluate age-related bone loss (5). It is also expected to detect metastatic bone tumors and to decide treatment response (6).
Recently, some biochemical markers of bone turnover have been developed and they are expected to reflect the disease activity of metastatic bone tumor. As it is difficult to estimate directly the actual production of paracrine substances involved in bone remodeling, evaluating the consequent changes in rates of bone formation and resorption by means of bone turnover markers might be a suitable approach (7). The following bone turnover markers are now promising: serum bone-specific alkaline phosphatase (BALP), osteocalcin (OC) and carboxy-terminal propeptide of type I procollagen (PICP) as bone formation markers; and serum carboxy-terminal telopeptide of type I collagen (ICTP) and urine pyridinoline (Pyr)/deoxypyridinoline (D-Pyr) as bone resorption markers.
Total serum alkaline phosphatase (ALP) activity does not specifically reflect bone function because its activity in the blood also depends on the liver, gastrointestinal tract, placenta and tumors (1). BALP can now be measured in blood more easily than in the past (8). BALP is present on the surface of osteoblasts, but the mechanism of its release in the circulation still remains unclear (9). OC is a non-collagenous matrix protein of bone that is synthesized by osteoblasts and is secreted directly into the circulation. Serum levels of OC can be viewed as a reliable index of bone matrix synthesis (9). As type I collagen is the most common protein in the skeleton, assays of the turnover of this protein should be good markers of bone turnover. The synthesis of type I collagen can be measured by PICP (10). It is reported that the levels of PICP are higher in patients with prostatic carcinoma than in patients with benign prostatic hyperplasia (11). This result is reasonable since most metastatic bone tumors from prostatic cancer are osteoblastic.
Bone resorption is currently evaluated by collagen degradation products. Urinary hydroxyproline (Hyp) has long been evaluated. The significance of the diagnostic value of Hyp for solitary bone tumors from breast cancer has been reported (12). Unfortunately, extraskeletal sources, including dietary constituents and serum proteins, can also contribute to the excretion of this substance. Pyr and D-Pyr have also been evaluated as bone resorption markers. As for the difference between Pyr and D-Pyr, D-Pyr is found almost exclusively in bone, while Pyr is found in both cartilage and bone, although both are collagen cross-linkers (13-15). They are different from Hyp in that the urinary concentration of Pyr/D-Pyr is not affected by dieting (16). As for the shortcomings of Pyr and D-Pyr, some researchers have pointed out that the assay, by high-performance liquid chromatography (HPLC), is complicated (13). More recently, an assay of the breakdown of mature type I collagen, ICTP, was developed (17) and has been shown to be a sensitive marker of bone resorption in various kinds of bone disorders (18-20).
Although there are many reports about the diagnostic value of bone turnover markers for bone metastases from BC, the clinical priority of such markers is still controversial. Therefore, in the present study, several biochemical markers were evaluated simultaneously for BC patients with and without bone metastases and appropriate markers were determined for the detection of metastatic bone tumors from BC.
Table 1.
| (a) Patients with bone metastases | Age (years) |
| Mean/range | 57.8/40-46 |
| Menopausal status | |
| Pre/post | 0/11 |
| ECOG performance status | |
| 0/1/2/3 | 0/6/3/2 |
| Concomitant metastatic sites | |
| Lymph nodes | 6 |
| Lung | 6 |
| Pleura | 3 |
| Liver | 3 |
| Bone appearances | |
| Osteoblastic | 0 |
| Mixed | 8 |
| Osteolytic | 3 |
| Treatments | |
| None | 0 |
| Chemotherapy | 0 |
| Methotrexate | 1 |
| Mitomycin C | 1 |
| Endocrine therapy | 2 |
| MPA | 2 |
| Chemoendocrine therapy | 7 |
| Cyclophosphamide + MPA | 3 |
| Methotrexate + MPA | 1 |
| Doxifluridine + MPA | 1 |
| Taxotere + MPA | 1 |
| Carmofur + MPA | 1 |
| (b) Patients without bone metastases | |
| Age (years) | |
| Mean/range | 52.5/37-78 |
| Menopausal status | |
| Pre/post | 7/12 |
| ECOG performance status | |
| 0/1/2/3 | 19/0/0/0 |
| Treatments (adjuvant therapy) | |
| None | 17 |
| Chemotherapy | 0 |
| Endocrine therapy | 2 |
| LH-RH analogue | 1 |
| Tamoxifen | 1 |
| Chemoendocrine therapy | 0 |
MATERIALS AND METHODS
Patients
From June 1995 to April 1996, 11 normocalcemic patients bearing bone metastases from BC and 19 BC patients without apparent skeletal involvement (bone scan negative) were recruited. Characteristics of the subjects are shown in Table 1a and 1b. Patients with bisphosphonate treatment, radiation therapy and with predominantly sclerotic lesions in X-rays were excluded. Serum parathyroid hormone-related protein (PTHrP) was within the normal range in all subjects. None was suffering from liver and/or kidney disease. Patients who had had surgery within 2 months were also excluded. Nine out of 11 patients with bone metastases and two out of 19 patients without bone metastases were treated with hormonal drugs. All subjects were informed of this study's purpose.
Diagnosis of Bone Metastases
Diagnosis of bone involvement was performed with a bone scintigraphy followed by confirmation with a plain X-ray. MRI was performed to discriminate lesions that appeared positive at scintigraphy and negative at X-ray. Out of 11 patients with bone metastases, three were diagnosed as having lytic lesions and eight had mixed lesions.
Measurement of Bone Metabolic Markers
In the early morning after an overnight fast, blood specimens for the assessment of BALP, OC, PICP and ICTP and spot urine specimens for determination of Pyr and D-Pyr were collected. Urine samples were collected as follows: after several hours' overnight fast, patients were instructed to discard the first urine, then a second urine sample was collected for the analyses. Serum and urine samples were stored at -20°C until analysis. BALP was specifically measured by immunoselective enzyme assay (ISEA) using a monoclonal antibody (Mitsubishi Kagaku BCL, Tokyo, Japan) (21). The cross-reaction to liver isozyme of ALP is <5% (21). Intact OC was measured using commercially available immunoradiometric assay (IRMA) kits (Mitsubishi Kagaku BCL). PICP was assayed by radioimmunoassay (RIA) (Orion Diagnostica, Oulunsalo, Finland) (reference value: 37-177 ng/ml) (10). ICTP was assayed with an RIA (Orion Diagnostica) (reference value: 0.76-5.24 ng/ml) (17). Urinary Pyr and D-Pyr were measured by HPLC (Mitsubishi Kagaku BCL) (13). The Pyr and D-Pyr data obtained in this study were adjusted by urinary creatinine.
Statistical Considerations
The correlation was tested using Spearman's rank-order correlation coefficient test. Differences between groups were tested by the non-parametric Mann-Whitney test. A value of P < 0.05 was regarded as statistically significant. The statistical significance of the differences in receiver-operating characteristic (ROC) analysis was determined using the method developed by Hanley and McNeil (22). The statistical analyses including multiple logistic regression analysis were carried out by SPSS general-purpose statistical software.
RESULTS
Figure 1. Relationship between serum carboxy-terminal propeptide of type I procollagen (PICP) and menopausal status in patients without bone metastases. PICP in postmenopausal patients was significantly higher than that in premenopausal patients (P < 0.05).
First, we investigated the influence of age or menopausal status on bone turnover markers for 19 patients without bone metastases. Only PICP was significantly related to both age (Table 2) and menopausal status (Fig. 1). That is, PICP in elderly or postmenopausal patients was significantly higher than that in younger or premenopausal patients. Second, the correlation between each of the two bone turnover markers was examined for the 30 patients. Significant correlations were observed between any of the two markers except between BALP and OC for all subjects (Table 3). The correlations were also examined in patients with bone metastases separately in patients without bone metastases. The differences of the results in the two groups were (i) whereas BALP significantly correlated with four markers except OC in the metastasis group, it significantly correlated only with PICP in the control group; (ii) whereas OC did not significantly correlate with any markers in the metastasis group, it significantly correlated with three markers (PICP, Pyr and ICTP) in the control group; (iii) whereas PICP significantly correlated with three markers except OC in the metastasis group, it significantly correlated only with BALP and OC in the control group (Table 3a and 3b). Third, the six bone turnover markers were compared between patients with bone metastases and those without them. The mean levels of bone metastatic patients were higher than those without them in all six bone turnover markers. Statistical significance was observed except for OC (Fig. 2). Fourth, by plotting the sensitivity versus (1 - specificity), ROC curves were constructed to compare the discriminative power of the tested markers. The results are shown separately for markers of bone formation and bone resorption (Fig. 3). The best diagnostic efficiency by comparing the area under the curves was provided by ICTP followed by Pyr or D-Pyr, BALP, PICP and OC. Statistical significance was observed between ICTP and OC (P = 0.014). Finally, to reveal the relative significance of the markers of bone metastases, multiple logistic regression analysis was performed by setting the existence of bone metastasis as the dependent variable and the five bone turnover markers (BALP, OC, PICP, D-Pyr and ICTP) and age as the independent variables. Pyr was excluded from the independent variables since the correlation between Pyr and D-Pyr was very high (rs = 0.957). On the other hand, age was always included as the independent variable to adjust for its possible confounding effect. The only significant factor related to bone metastasis was ICTP (P = 0.019) (Table 4). Figure 2. The mean levels of bone metastatic patients were higher than those without them in all six bone turnover markers. Statistical significance was observed except for OC. References values, which were calculated from mean ± 2SD in healthy control studies indicated from each assay kit, are shown by oblique lines. The vertical bar represents the standard deviation of the data. The mid-point of the vertical line indicates the mean of the data. Figure 3. Receiver-operating characteristic (ROC) curves comparing the discriminative power of the tested markers. The results are shown separately for markers of bone formation and bone resorption. The best diagnostic efficiency using this analysis was provided by ICTP followed by Pyr or D-Pyr, BALP, PICP and OC. Statistical significance was observed between ICTP and OC (P < 0.014). Table 2. . Table 3. Table 4. 
Bone turnover markers
Spearman's correlation coefficient
BALP
0.222
OC
0.420
PICP
0.507*
Pyr
0.089
D-Pyr
-0.060
ICTP
0.429
OC
PICP
Pyr
D-Pyr
ICTP
(a) All patients (n = 30)
BALP
0.246
0.522**
0.482**
0.483**
0.513**
OC
0.441*
0.430*
0.424*
0.413*
PICP
0.552**
0.540**
0.648**
Pyr
0.957**
0.802**
D-Pyr
0.761**
(b) Patients without bone metastases (n = 19)
BALP
0.264
0.491*
0.071
0.125
0.425
OC
0.534*
0.499*
0.408
0.484*
PICP
0.310
0.285
0.418
Pyr
0.910**
0.568*
D-Pyr
0.464*
(c) Patients with bone metastases (n = 11)
BALP
-0.141
0.664*
0.656*
0.627*
0.647*
OC
0.196
-0.009
0.064
0.132
PICP
0.755**
0.755**
0.752**
Pyr
0.655*
0.916**
D-Pyr
0.943**
Regression coefficient
Standard error
z value
P
ICTP
1.448
0.616
2.35
0.019*
BALP
0.132
0.098
1.35
0.176
Age
0.025
0.062
0.400
0.687
DISCUSSION
The results of this study suggest that serum ICTP appears to be the leading marker of bone metastases from BC among six bone turnover markers.
First, when evaluating and discussing data of bone turnover markers from women, generally the effect of age or menopausal status on such markers is very important. PICP in elderly or postmenopausal patients was significantly higher than that in younger or premenopausal patients in this study. Similar results for PICP in healthy control studies have been reported by other researchers (24,25). Fukunaga et al. (26) in a study of healthy controls found that PICP was lowest in patients in their 30s and 40s and increased after 50. From these reports, we should consider carefully age or menopausal status when interpreting the PICP data. As for other markers, urine Pyr/D-Pyr (27) and BALP (8) were reported to be significantly higher in postmenopausal women than in premenopausal women in healthy control studies. On the other hand, there was a report of no significant change in ICTP in patients from their 40s to 60s (28).
Second, significant correlations were observed between each of the markers except between BALP and OC for all subjects in this study and exceptionally high correlations were seen both between Pyr and D-Pyr and between Pyr and ICTP (rs = 0.957 and 0.802, respectively). However, there were some discrepancies of the results between patients with bone metastasis and without bone metastasis. Blomqvist et al. (29) in a study on BC reported that there was a positive correlation between the concentration of ICTP and PICP. Berruti et al. (30) reported that significant correlations were found between BALP and PICP, but not OC, in patients with lytic appearances and between BALP and OC, but not PICP, in patients with mixed/sclerotic lesions in a study on patients with heterogeneous cancer. In addition, they reported that the reasons for these discrepancies were unclear but that the control of osteoblast activity was multifactorial even in the absence of bone involvement from neoplastic cells. Unfortunately, as our subjects included only three patients with lytic appearances out of 11 patients with bone metastasis, we could not elucidate the difference in the results between lytic lesions and mixed/sclerotic lesions. Nevertheless, the results of these studies might suggest that to detect bone metastases from BC we need not assess so many markers simultaneously.
Third, between patients with and without bone metastases, we found no significant difference in OC. Plebani et al. (1) reported that the sensitivity of OC to detect bone metastases was low (0.40), while the specificity was 0.88 and the efficiency 0.64. Some other problems which have been reported for the assessment of OC are as follows: small fragments are also secreted into the circulation during osteolysis, especially when bone turnover is high (26), its stability on keeping the blood in a freezer is poor (9) and it is strongly affected by kidney function of the patients (9) and chemotherapy itself (31).
Fourth, ROC curves constructed in this study indicated that the best diagnostic efficiency was provided by ICTP followed by Pyr or D-Pyr, BALP, PICP and OC and statistical significance was observed only between ICTP and OC. OC did not appear to be an appropriate marker of bone metastases from BC.
Fifth, multiple logistic regression analysis adjusted by age revealed that serum ICTP was the only significant marker related to bone metastases, although an experimental analysis in which ICTP was excluded from the independent variables revealed that D-Pyr was the only significant factor (data not shown). Some researchers have also reported a significant relationship between the number of bone metastases and the concentration of serum ICTP or urinary Pyr/D-Pyr (25,32,33). Koizumi et al. (33) reported that ICTP was good in detecting bone metastases from BC and was more sensitive than PICP. On the other hand, some shortcomings of ICTP in clinical application have been indicated as follows: it is affected by kidney function (17) and hypercalcemia, which are sometimes found in advanced cancer patients (16), and it is elevated in other diseases such as osteoporosis (34), rheumatoid arthritis (19) and Graves' disease (35).
Recently, some researchers reported that serum ICTP significantly related to treatment response. Blomqvist et al. (29) in a study of BC reported that reduction in ICTP correlated significantly with the treatment response at 3 months, whereas PICP showed a borderline negative correlation with it. They also pointed out that of all the biochemical parameters studied, the changes in ICTP showed the best correlation with the treatment response and the diagnostic accuracy of both ICTP and PICP in assessing the treatment response was at the same level as published results on the classical tumor markers CA 15-3 and CEA (29).
Finally, some limitations of this study were that the differences in some backgrounds in the two groups (patients with and without bone metastases) might affect the results. For example, there was a large difference in the distribution of performance status (PS) in the two groups, the PS of the patients without metastases being much better. Another limitation was the difference in the methods of treatment. In the control group, only two of 19 patients received adjuvant therapy and these two patients received hormonal therapy. On the other hand, in the bone metastasis group, all the patients received hormonal therapy and/or chemotherapy.
Hormonal therapy such as tamoxifen might affect the bone turnover of patients. Although tamoxifen is considered to be an antiestrogen compound, it also possesses a partial estrogenic effect in some tissues (36). Therefore, the influence of tamoxifen on bone turnover of BC patients is controversial. Kalef-Ezra et al. (37) reported that tamoxifen has neither an adverse nor a protective effect on the amount and composition of primarily cortical bones in the appendicular skeleton. Moreover, we should consider that chemotherapy and radiation therapy might have adverse effects on the skeleton. Cytotoxic chemotherapy with drugs such as methotrexate and cyclophosphamide have been reported to associate with alterations in mineralized tissue metabolism (38).
In conclusion, this study revealed that serum ICTP appears to be the leading marker of bone metastases from BC among six bone turnover markers. However, to reveal the clinical usefulness of these markers, further examination will be needed to account for the ease and cost-effectiveness of measurements.
Acknowledgment
This work was supported in part by a Research Project Grant (VI, 8-604) from Kawasaki Medical School.
References
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 22 Feb 1999
Copyright© 1999 Foundation for Promotion of Cancer Research.
CiteULike 
Connotea 
Del.icio.us What's this?
This Article 
Abstract
Alert me when this article is cited
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in ISI Web of Science
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Search for citing articles in:
ISI Web of Science (20)
Request Permissions
Google Scholar
Articles by Shimozuma, K
Articles by Tanaka, K
Search for Related Content
PubMed
PubMed Citation
Articles by Shimozuma, K
Articles by Tanaka, K
Social Bookmarking
What's this?