68
The Breast
have been shown in a relatively small study to predict radiological response while more recently preliminary data suggest that the new specific markers of collagen synthesis and breakdown may be useful. 90% of bone matrix is composed of type-l collagen. This complex molecule is produced as procollagen. After cleavage of the C (PlCP) and N (PlNP) terminal ends of procollagen, the collagen chains are bound together in the bone matrix by the crosslinking molecules hydroxypyridinoline (HP) and hydroxylysyl-pyridinoline (LP). The procollagen fragments PlCP and possibly also PlNP are potential measures of collagen synthesis (bone repair) while the pyridinolines and their joining regions with collagen (l-CTP and NTX) reflect bone resorption. The assays for collagen crosslinks and fragments are rapidly becoming more specific and simpler to perform on either serum or urine samples. Critical evaluation of their value in malignancy is now indicated.
13 Osteolysis is mediated by humoral mechanisms in patients with breast cancer and bone metastases N. J. Bundred, J. Walls and A. Howell* Departments of Surgery and *Oncology, University Hospital of South Manchester, Manchester, UK The aim of this prospective study was to establish the mechanisms leading to hypercalaemia and progression of bone metastases in patients with breast cancer, using recently identified humoral and osteolytic markers. 80 women with breast cancer and newly diagnosed bone metastases were studied; (progressive n = 35, stable n = 24, responding n = 21) as determined by UICC criteria. Patients were monitored at monthly intervals, when blood and urine samples were collected. Plasma parathyroid hormone-related protein (PTHrP), the putative cause of hypercalcaemia of malignancy, and parathyroid hormone (PTH), were measured using immunoradiometric assays. In-house radioimmunoassays were used to measure serum 1,25 (OH), vitamin D, and 25(OH)D. Urinary cyclic adenosine monophosphate (CAMP), the second messenger of PTH receptor activation was measured by radioimmunoassay. Three novel urinary markers of bone resorption were measured by high performance liquid chromatography; pyridinoline (Pyd), deoxypridinoline (Depyd) and galactosyl hydroxylysine (Gal Hyl). PTHrP was undetectable in 95% of patients with stable bone metastases, and became detectable when they progressed (p < 0.05). Of 18/80 (23%) who developed hypercalcaemia, median plasma PTHrP was significantly elevated (p < 0.001). Hypercalcaemic patients had significantly increased urine CAMP excretion, compared with stable bone metastases (p < 0.02). PTHrP correlated with urine CAMP (r = 0.41, p cO.O5), and serum calcium (r = 0.45, p < 0.04). PTH became suppressed with the onset of hypercalcaemia (< 15 pg/ml, p c O.OOl), and therefore any CAMP produced was a response to renal PTHrP stimulation, suggesting a humoral mechanism of osteolysis. In animal models of hypercalcaemia, PTHrP can stimulate vitamin D synthesis but its effect in patients with breast cancer is unclear. Women with bone metastases had reduced 1,25 (OH),D levels, which became significantly lower with the onset of hypercalcaemia, (p < 0.001). Women with stable bone metastases, had stable levels of 1,25 (OH),D, however when they progressed a significant decrease in 1,25 (OH),D synthesis was observed (p c 0.04). Serum 1,25 (OH)*D fell when PTI-IrP increased, implying PTI-IrP does not stimulate 1,25 (OH)ID production in patients with breast cancer. Urine Pyd excretion indicated normal bone matrix turnover in women with stable bone metastases, but was increased in hypercalcaemia, (p c 0.01). Depyd excretion was also higher in hypercalcaemia, (p < 0.05). Both Pyd and Depyd excretion increased with progression of bone metastasis (p < 0.05). Gal
Hyl excretion was also increased in hypercalcaemia compared with stable bone metastases, (p < O.Ol), and correlated with Pyd and Depyd (r = 0.56, r = 0.48 respectively). 25% of women with hypercalcaemia, however, had normal indices of bone resorption and therefore the mechanism in these cases was humoral. Finally, in 50% of hypercalcaemic women, renal tubular reabsorption of calcium was increased, indicating humorally mediated calcium reabsorption was present. This study indicates that a substantial proportion of bone resorption in women with breast cancer is humorally mediated, and the mechanisms leading to hypercalcaemia are complex. We have demonstrated four potential novel markers for monitoring bone metastases response.
14 Gross cystic disease fluid protein (CP-15) identifies breast cancer bone metastases response to treatment as accurately as the tumour marker CAE-3 N. J. Bundred, J. Walls and A. Howell* Departments of Surgery and *Oncology, University Hospital of South Manchester, Manchester, UK Gross cystic disease fluid protein- 15 (CP- 15) has been claimed to be a useful marker of disease response in breast cancer. The aim of this study was to compare CP-15 with other standard markers of tumour response in patients with breast cancer and bone metastases. Plasma CP- 15 was measured by a 2-stage solid phase enzyme immunoassay, in women with bone metastases (responding n = 30, progressing n = 22), early breast cancer (EBC, n = 72), and locally advanced breast cancer (n = 8). Carcinoembryonic antigen (CEA) was measured by fluoroimmunoassay, and CA15-3 by an immunoradiometric assay. The plasma levels of CP-15 did not differ between patients with EBC, locally advanced breast cancer, or patients with bone metastases. CP- 15 levels were significantly higher in progressing bone metastases, compared to those with responding bone metastases, (p c 0.001). CP-15 levels sequentially fell when bone metastases responded to treatment, and sequentially rose when bone metastasesprogressed(p < 0.04, ANOVA). In patients with progressing bone metastases, 59% of CEA, 68% of CA15-3 and 77% of CP-15 levels were elevated at presentation. At monthly intervals, however, median CEA levels did not change, CA-15 levels sequentially increased by a median of
32% and CP-15 by 50%. Plasma CP-15 levels appear more sensitive than either CA15-3 or CEA in predicting response of patients with bone metastases to changes in treatment. For individual patients, changes in CP-15 levels were easier to monitor, and were greater in magnitude. Plasma CP-15 levels identify response to therapy in bone metastases patients with greater frequency and equal accuracy to the tumour marker CA15-3. Statistics by Kruskal-Wallis analysis of variance (ANOVA) Mann-Whitney, and Chi-squared test.
15 C-erbB2 oncoprotein with breast cancer
in the serum of patients
P. Willsher and J. F. R. Robertson Department of Surgery, City Hospital, Nottingham, UK C-erbB2 DNA amplification has been shown to correlate with decreased survival in breast cancer patients. Tissue detection of the c-erbB2 protein has been assessed in a number of studies. However, investigation of the significance of serum c-erbB2 protein is limited. We have studied sera from normal individuals (24), patients
Abstracts 69 with benign breast disease (47) and breast cancer patients (Stage I/II n = 46, Stage III n = 34, Stage IV n = 37) using a quantitative ELISA kit. There were no significant differences between these groups in terms of either the proportion of each group with serum levels > 20 units/ml (norm 12.5%, Benign 14.9%, Stage I and II 15.2%, Stage III 26.5%, Stage IV 21.6%; p = ns) or between the absolute levels of c-erbB2 protein detected (Median Values: Norm 3.8 U/ml (< 3-61), Benign 78 U/
ml (< 3-155), Stage I/II 51 U/ml (< 3-75), Stage III 55 U/ml (< 3-104) Stage IV 11 U/ml (< 3 - > 300) p = ns) (Stage I and II, Stage III, Stage IV). Nor did the presence of c-erbB2 protein in sera predict response to treatment (p = 0.74). This study suggests that c-erbB2 serum protein is not helpful for diagnosis or in predicting patient outcome. The potential for monitoring disease behaviour is yet to be assessed.
16 TPS in breast cancer J. F. R. Robertson, J. Beaver, D. Pearson, A. Gilmore, P. Willsher, K. O’Neill and P. G. McKenna* Department of Surgery, City Hospital, Nottingham & *University of Ulster, Northern Ireland, UK TPS is a new assay for the quantitative measurement of a specific epitope, M3, of tissue polypeptic antigen, TPA, which was first reported over 30 years ago. TPA was identified as a protein molecule with 35 different epitopes. The M3 epitope (TPS) is reported to be related to tumour cell proliferation. We have examined serum TPS in the following patient groups - normal controls (n = 47) benign breast disease (n =
84), Stage I/II (n = 79) Stage III (n = 57) and systemic metastases (n = 66). TPS was elevated in the group of patients with systemic metastases compared to all other groups (p < 0.001, ANOVA). There was no elevation of TPS in patients with primary breast cancer compared to normal controls. 28 patients receiving endocrine therapy for metastatic breast cancer were sequentially monitored for TPS in addition to carcinoembryonic antigen (CEA), CAL-3 and erythrocyte sedimentation rate (ESR). TPS was the least accurate of the 4 markers in predicting clinical outcome. TPS accurately predicted response but in one-third of patients with progressive disease TPS rose after the diagnosis of progressive disease by UICC criteria. The authors are unable to offer an explanation for this delay in rise in TPS of these patients. The most common combination of markers currently used is CEA and CA15-3. We have previously reported that the addition of ESR to this combination increases the sensitivity to tumour markers. We compared the combination CEA, CA15-3 & ESR with the following two combinations - CEA, CA15-3, ESR, & TPS and then CEA, CA15-3 & TPS. CEA, CA15-3 and ESR measured therapeutic response better than either combination which included TPS. The reason for this difference was again that in a third of patients with progressing disease the TPS values remained low even when progressive disease had become apparent by UICC criteria. Although TPS is an accurate predictor of response to systemic therapy, in one-third of patients developing progressive disease TPS does not appear to give an early indication of disease progression. Thymidine kinase (TK) is a pyrimidine nucleotide salvage pathway enzyme. We previously reported that serum TK levels were elevated both in patients with early (Stage I/II) and advanced breast cancer. TK rises in the cell during the late GI and S-phase of the cell cycle. We have compared TPS, a putative marker of proliferation, with TK in the same serum samples (n = 63) using the Pearson correlation test. The r value was 0.4037 (p = 0.001). It would appear from this result that only 16% of the variation in TPS is due to variation in TK. If one very high value of TPS was excluded the Pearson correlation for the remaining for the remaining samples (n = 62) was r = 0.5083
(p < 0.001). At this value 25% of the variation in TPS is due to variation in TK. It would appear therefore that in both analyses the variation between TPS and TK must be explained by other factors.
17 Assessment of therapy by the serum markers TPS and CA15-3 - related to the UICC criteria in breast cancer Stage IV P-E. Jonsson, M. MaImberg and G. Nordin Departments of Surgery and Clinical Chemistry, Helsingborg Hospital, S-251 87 Helsingborg, Sweden TPS is a tumour marker that gives a quantification of tumour cell proliferation activity while CA15-3 reflects the tumour mass. In an earlier report the preliminary result of S-TPS showed that it was a suitable marker for response evaluation in advanced breast cancer. In our ongoing prospective registration of treatment effects in breast cancer 46 patients have been included for marker analysis (median age 56, range 4&84 years). Both markers have been analysed by an immunoradioassay technique (STPS/Beki Diagnostics, Sweden, and S-CAlS-3/CIS, France). The cut-off level for S-TPS was 80 U/l and for CA15-3 30 U/l. An increase of the markers by more than 25% indicated progressive disease and a decrease of S-TPS by 50% and CA 15-3 by 25% indicated response. Evaluation of patients were performed at start or change of therapy and then at 3 and 6 months together with UICC criteria. Patients were evaluated if both markers were increased at start of therapy and therefore the further follow-up included 39 patients. 32 patients were given chemotherapy and 7 hormonal therapy. Metastatic sites were bone alone in 18, bone and liver in 9, bone, liver and lung in 6, and lung and soft tissue in 6 patients. Of 46 patients the S-TPS was increased in 80% and CA15-3 in 85% at commencement of therapy. The assessment at 3 months showed figures of partial response (PR) according to UICC in 20%, to TPS in 50% and to CA15-3 in 29% of the patients. At six months the figures were 34%, 52% and 48% respectively. At this time progressive disease (PD) was found in 20% according to UICC criteria, to S-TPS in 24% and to CA15-3 in 28% of the patients. When evaluated by both markers at the same time the figures for PR at 3 months were for UICC 18%, TPS 59%, CA15-3 29% and at 6 months 24%, 43% and 44% respectively. We try to answer the question which marker shows response or progression before the others. S-TPS showed response before CA15-3 in 48% of the patients compared to 14% for S-CA15-3 before STPS. S-CA15-3 showed progressive disease before S-TPS in 44% of the patients while S-TPS showed PD first in 21%. Increased values of S-TPS and CA15-3 correlated well with both values positive in 29139 patients and both negative in four patients. Four patients with negative S-TPS had positive values of S-CA15-3 and two S-TPS were positive while S-CA15-3 was negative. To conclude S-TPS has been shown to be a good indicator of responsive disease, however S-CA15-3 seems to be better for progressive disease. The correlation at 6 months between STPS and S-CA15-3 and tumour response was good. S-TPS and S-CA15-3 are both markers of great value for monitoring therapy in breast cancer stage IV.
18 Tumour markers vs UICC - the health economists view D. K. Whynes Reader in Health Economics, University of Nottingham, UK