We retrospectively reviewed all breast cancer patients’ data from Aichi Cancer Center Hospital’s database between 2002 and 2012. Patients comprised 144 individuals with bone metastatic breast cancer, who had undergone whole body bone scan examinations in our institution at diagnosis of bone metastasis. The definition of bone metastasis was done by imaging findings such as X-ray examination, CT or MRI scan as well as subjective symptoms and bone scan. Bone scan examinations were performed 3 h after intravenous injection of 740 MBq 99mTc-methyl diphosphonate (FUJIFILM RI Pharma Co., Ltd., Tokyo, Japan), using nuclear medicine imaging procedures to detect radioisotope tracer uptake in the patient’s body (Infinia HE4®, GE healthcare, Tokyo, Japan). A gamma camera equipped with a low-energy high-resolution parallel whole collimator was used at anterior and posterior view scan speeds of 15 cm/min (matrix 256 × 1024). Energy discrimination was provided by a 10 % window centered on the 140 keV of the Tc99m. BSI was calculated first by determining the percentage of each bone that is involved by the tracer in relationship to the total skeletal mass, as determined from reference man, which is using the Bonenavi® automated method (FUJIFILM RI Pharma Co., Ltd., Tokyo, Japan). We retrospectively reviewed imaging data to analyze BSI without having any clinical information of the patient. If patients had taken bone scan examinations more than twice, we analyzed the image at the time of bone metastasis initially. We excluded the scan data if it is before or after three months from the day of diagnosis.

All patients received standard systemic therapy for metastatic breast cancer, including chemotherapy, hormone therapy, and anti-HER2 therapy. For prophylaxis against SRE, bisphosphonates or denosumab were infused initially with analgesics as needed. We defined SRE as either bony, requiring intervention (surgery and/or radiotherapy) for pain or prevention of fracture, or spinal cord compression, and recorded the date on which an SRE was first observed. Patients with an SRE before bone scan examination were excluded. SRE-free survival was recorded from the date of baseline bone scanning to the SRE date. Overall survival (OS) was also recorded from the baseline bone scan to death. Written informed consent was obtained from all patients and this study has been approved in our institution (approved number: 2015-1-058).

The intrinsic subtype of primary tumor was classified using immunohistochemical (IHC) staining of paraffin-embedded thin sections as follows: luminal: ER and/or PgR positive (stained proportion 10 %), HER2 negative; luminal-HER2: ER and/or PgR positive and HER2 positive; HER2: ER negative, PgR negative and HER2 positive; triple negative: ER negative PgR negative and HER2 negative. A diagnosis of HER2 positive cancer was based on the published guidelines (Wolff et al. 2007).

The purpose of this study was to evaluate the rate of SRE and OS according to the baseline BSI, and to identify the cut-off score of BSI for predicting SRE in metastatic breast cancer patients. We performed sensitivity analysis checking P values for SRE at intervals of 0.1 from 0.4 to 2.4 % using a Cox regression model. We used the BSI score with the lowest P values as the cut-off point. To estimate the distribution of SREs and survival data, Kaplan–Meier estimates were used together with the log-rank test. Multivariate analyses were conducted using the Cox proportional hazard model. Factors evaluated in the model included age, metastasis site other than bone, performance status, and intrinsic subtype as well as the cut-off point of BSI. P values less than 0.05 were considered statistically significant. All statistical analyses were conducted using STATA® v.12.1 (StataCorp, College Station, TX, USA).