Bone-specific alkaline phosphatase (BSAP) and serum N-telopeptide (sNTX) as predictors of bone loss in postmenopausal women with early breast cancer receiving letrozole as adjuvant therapy: a 5-year study (Z-FAST).

Abstract

Abstract Abstract #2067 Background: Changes in biochemical markers of bone turnover are expected to be predictors of long-term changes in bone mineral density (BMD). We studied the potential of bone-specific alkaline phosphatase (BSAP) and serum N-telopeptide (sNTX) at 6 and 12 months (mo) in predicting BMD of the lumber spine (LS) and total hip (TH) at 48 mo in Z-FAST study.
 Methods: Postmenopausal women with ER+ and/or PR+ early stage breast cancer were randomized 1:1 into Up-front (U) or Delayed (D) arms (N=301 in each). Patients (pts) in U received zoledronic acid (ZOL) 4 mg IV q6 mo starting at baseline; pts in D started ZOL q6 mo when their LS or TH T-score fell below -2. All pts started Letrozole 2.5 mg/day for 5 years. Levels of BSAP and sNTX were assessed at screening and every 6 mo on a subset of pts. LS and TH BMD were assessed by dual-energy x-ray absorptiometry (DEXA) at screening, 6 mo and every 12 mo thereafter for all available pts. This is an exploratory analysis to investigate the correlation between short-term % changes from baseline in BSAP and sNTX and long-term % changes from baseline in BMD. Pearson correlations and Receiver Operating Characteristics (ROC) analysis were performed. We report the results based on 48 mo data.
 Results: The Pearson correlations between % change in BSAP at 6 mo and LS BMD at 48 mo were -0.13 [(N=38, not significant (NS)] and -0.17 (N=39, NS) in arms U and D respectively. The correlations between % change in BSAP at 12 mo and LS BMD at 48 mo were -0.15 (N=39, NS) in U and -0.43 (N=38, p=0.0077) in D. Very similar correlations were obtained for % change in sNTX at 6, 12 mo and LS BMD at 48 mo in D, but in U the correlations were smaller and positive. The correlations between % change of both markers at 6, 12 mo and TH BMD at 48 mo were all NS at 0.05. ROC analysis showed that at best performance (maximum of sensitivity plus specificity), for any increase in sNTX at 12 mo, the sensitivity (SE), specificity (SP), positive predictive value (PPV) and negative predictive value (NPV) to predict LS BMD loss at 48 mo was 74%, 77%, 85% and 63% in D. For BSAP change of 3%, the four ROC parameters were 71%, 71%, 81% and 59% respectively. For prediction of TH BMD loss, the SE, SP and NPV were lower for both markers with higher PPV. In U, the SE, SP and PPV were all much lower with very high NPV for both markers and BMD loss.
 Conclusion: Interim results of Z-FAST data show that short-term changes in BSAP and sNTX have the potential for predicting long-term bone loss. The weaker correlations in U also suggest that zoledronic acid may have a direct preventive effect on bone loss independent of changes in the markers. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 2067.