Background: Bone Metastases (BM) are responsible of the high morbidity, low quality of life and poor prognosis of patients with BC. The evaluation of bone response to treatment is still controversial. Methods: This is a prospective study on 36 consecutive BC patients with BM for the evaluation of RANK, RANKL and OPG transcripts by Real-Time PCR and NTX levels by ELISA. Patients enrolled had at first diagnosis of bone metastases and had not previously undergone bone targeted treatment. They underwent the standard Zoledronic Acid schedule of a 4-mg infusion every 28 days and standard antitumor therapy. Patients were monitored for about 1 year and blood samples were collected before the first infusion of ZA and every 4 months. The primary aim was to study the predictive role of different circulating markers in the response to ZA respect to objective response (MD Anderson criteria). Results: The NTX levels were not different in patients that underwent to progression respect to those with stable disease or partial response to ZA Treatment. Instead a decrease of 30% and 15% was observed in OPG and RANKL levels in non responders patients. ROC curves evaluation for all markers showed RANKL was the most accurate, with an AUC of 0.70 (95%CI 0.48-0.93). We observed for RANKL/OPG an AUC of 0.64 (95%CI 0.44-0.84) and for NTX an AUC of 0.60 (95%CI 0.38-0.82). Considering as a cut off a variation of markers of at least 25%, the positive predictive value (PPV) for RANK was 26.6%, 9.1 for RANKL, 30 for OPG and 13.3 for the ratio RANKL/OPG. NPV for these markers ranged from 79.2 to 84.6 %. NTX PPV and NPV were respectively 0 and 84.6%. Conclusions: The most accurate marker for the evaluation of bone response was RANKL, the key molecule in the pathogenesis of BM and the target of Denosumab. NTX, the marker most used up to now, failed to show a predictive power of response with aspecific decrease both in responder and non responder patients. Further research should be done to improve our knowledge in this field.