Background: When monitoring patients with metastatic breast cancer (mBC), the optimal strategies for imaging and utilization of tumor markers (TM) are uncertain. Patients and Methods: We used a retrospective cohort of 302 patients with de novo mBC treated from 2000 to 2012 at Dana-Farber Cancer Institute to describe the type and timing of imaging and TM testing during the first line of treatment (baseline, first, and subsequent testing). Results: At baseline, all patients had staging scans, with increasing use of PET/PET-CT (17.5% from 2000–2002; 40.3% from 2009–2012). PET/PET-CT was used by itself in only 12.5% of cases. Overall, 30.1% of patients, of whom 80.2% had no neurologic symptoms, underwent central nervous system (CNS) screening; 78.2% of patients had baseline TM testing. Over the course of treatment, 23.5% of patients had TM retested once a month or more. Time-to-first reimaging varied by disease site (hazard ratios for shorter time-to-first reimaging [95% CI] vs bone: brain, 4.27 [1.46–12.50]; liver, 2.19 [1.39–3.46]; lung, 2.75 [1.66–4.57]), but was not associated with tumor subtype or baseline TM testing, regardless of test results. First reimaging was prompted by an elevation in TM in only 1.4% of cases. There was weak correlation between frequency of imaging and TM tests (r=0.33; R2 =0.11; P<.001). Discussion: Over time, we found an increased utilization of more sophisticated imaging staging techniques, such as PET/PET-CT scan, which was mostly requested in addition to other radiographic studies. CNS evaluations were frequently performed to screen asymptomatic patients. TM testing was often ordered, both at baseline and after treatment initiation. However, patterns of imaging utilization, although appropriately influenced by clinicopathologic factors such as disease site, did not appear to be impacted by TM testing. Conclusions: Studies focused on optimizing disease monitoring, including better integration of TM testing with imaging, are encouraged.