Mammographic screening is under debate because it is associated with so-called overdiagnosis. Overdiagnosis relates to the fact that cancers may take a “benign” course, i.e. would not, even if left un-diagnosed and thus untreated, progress to a lethal disease. Although there is little disagreement about the fact that overdiagnosis does exist, there is substantial debate about the actual fraction of breast cancer that would behave this way. Recent analyses suggest a rate between 1% and 10% of cancers diagnosed by mammographic screening are prognostically unimportant. Specifically the increasing number of low-grade-DCIS that is picked up by screen-detected calcifications will contribute to overdiagnosis. For another, mammographic screening is associated with significant “underdiagnosis” of prognostically important breast cancer. The simple fact that proves this statement to be true is the fact that, in spite of the well-established correlation between early diagnosis and prognosis, and in spite of decades of mammographic screening, breast cancer continues to represent the second most important cause of cancer death in women. On pathophysiological grounds, overdiagnosis, but also underdiagnosis of breast cancer due to mammographic screening is plausible. Radiographic breast imaging (digital mammography, but also digital breast tomosynthesis) is based mainly on the depiction of regressive changes. Mammographic hallmarks of breast cancer are architectural distortions, spiculated masses, and calcifications. This reflects pathophysiological changes such as fibrosis and necrosis, i.e. effects that are caused by cancer hypoxia, and that lead to slowed growth, and cell death. Even before the discussion around overdiagnosis, it was well established that mammography preferably detects slowly-growing cancers. Cancers detected through mammographic screening are known to enjoy a better prognosis than cancers of the same size and stage that were not diagnosable through mammography, an effect known as “length time bias”. Overdiagnosis is a length time bias put to extreme. On the other hand, if a cancer is successful in maintaining its need for perfusion, it will not develop necrosis or calcifications, and will not cause architectural distortions. Biologically important breast cancers are therefore frequently occult on mammography, and if they are detectable on mammography or ultrasound, may mimic fibroadenomas or even cysts. Accordingly, over- but also underdiagnosis is an unavoidable and logical consequence of the way we diagnose breast cancer with mammography. Modern approaches to breast cancer screening should strive to account for both issues. Radiologists must learn that the aim of breast cancer screening is not to detect all breast cancers and their precursors by all means. Rather, the goal must be to develop imaging methods that combine a maximum sensitivity for prognostically relevant disease with a desirable lack of sensitivity for disease that is prognostically unimportant.

The major difference between mammography or ultrasound, and MRI, is the fact that in breast MRI, cancer is detected due to local contrast enhancement. MRI is not only a diagnostic tool – but is indeed an effective in-vivo biomarker for disease activity or tumor biology. Enhancement of a DCIS or of an invasive cancer depends on a locally increased vessel density, an increased vessel permeability and – in the case of DCIS – an increased permeability of the ductal basal membrane. Accordingly, breast cancer detection in MRI is based on pathophysiological changes that are indicative of cancer proliferation, infiltrative growth and metastasis. In fact, the more angiogenesis or protease activity a cancer or DCIS exhibit, the higher the likelihood that it will be detected by MRI. In line with this, it has repeatedly been shown that MRI is associated with something one could call a “reverse length time bias”. Cancer detection in MRI is biased towards prognostically important disease. Cancers only detected by MRI tend to exhibit high nuclear grade, high Ki-67 values, i.e. hallmarks of rapid growth. In turn, malignant lesions that went undetected by MRI screening, and picked up by mammography alone, typically constitute of low grade DCIS. Trials that compared the added value of mammography in women undergoing MRI for screening concordantly found that the additional cancers diagnosed through mammographic screening mainly represent disease with limited, if any, prognostic importance.