92 Data from the NCI SEER Registries Since 1973, the SEER registries of the NCI have been collecting population-based information on cancer cases and the initial course of treatment. These registries include 9 states, 5 metropolitan areas, and the Alaskan Native Tumor Registry. Together they represent about 28% of the entire US population, broadened in the past 20 years to offer a truly representative cross-section of the country with regard to our ethnic, immigrant, racial, educational, and socioeconomic diversity. Analyzing SEER data enables researchers and policy makers to monitor cancer trends and gather data on incidence, the extent of disease at diagnosis, initial therapy, mortality, and survival. Unfortunately, because only incidence, initial treatment, and mortality are captured in the SEER registries, and recurrent cancer is not tracked, the data on all metastatic cancers, including MBC are limited. Incidence The actual number of new cases of MBC diagnosed each year is unknown. This is because SEER only records the 5% of newly diagnosed breast cancer patients who have de novo MBC. However, most patients with MBC were first diagnosed at earlier stages of breast cancer that then recurs, months to years later[30] . An estimated 20% to 30% of early stage breast cancer patients will develop MBC sooner or later. The SEER registries do not capture this much larger percentage. As a result, the actual annual incidence of MBC remains unknown. Prevalence The prevalence of breast cancer is increasing. Today, an estimated 3.1 million women living in the US already have a history of invasive breast cancer, and in 2014, an estimated 232,670 women will be newly diagnosed[99] . However, we have no way of knowing how many of these people are actually living with MBC as a chronic, progressive, and ultimately fatal disease or how many are “cured” of the disease, meaning they will go on to die of other causes. After early stage breast cancer is treated, it can lie dormant for as many as 20 or more years, with no way of determining whether it is actually cured or in a temporary state where there is “no evidence of disease.” This complicates the already challenging assessment of MBC prevalence. Neither the total number of people living with MBC nor its burden in society can currently be determined. Creative methods have therefore been used to estimate the prevalence of MBC. The duration of survival of patients with MBC (itself an estimate based on data from clinical trials involving highly selected patients), multiplied by the annual number of breast cancer deaths, has been used to approximate MBC prevalence. Estimating survival duration is complicated by significant variability related to the type of MBC and the treatment received. With good access to care and favorable tumor biology, some MBC patients can live for a decade or more. Using more sophisticated techniques, Australian biostatisticians have modeled the prevalence of MBC using the New South Wales cancer registry, estimating the prevalence as 3 to 4 times the number of annual deaths from breast cancer[100] . This approach is based on the fact that at least 90% of breast cancer deaths occur as a result of complications related to MBC. An estimated 3.1 million women in the US have a history of invasive breast cancer. We have no way of knowing how many of these people are actually living with MBC.