To the Editor Esserman and Yau1 make important points about the therapeutic approach to the lesions currently labeled by pathologists as ductal carcinoma in situ (DCIS), but even these experts might be confused about the nature of the actual pathologic abnormalities shrouded behind the DCIS terminology. As the authors know, the term cancer has traditionally been applied to breast lesions that invade, metastasize, or have the high-grade cytological features associated with clinically overt cancers. Pathologists do encounter a type of DCIS that shows strikingly malignant cytological features, but increased identification of these lesions (also referred to as comedo intraductal carcinoma) is not the only explanation for the DCIS epidemic. In fact, a substantial portion of the epidemic is due to a completely different lesion that lacks all 3 of these classic malignant features. In these latter cases, pathologists abandoned the classic morphological cancer definition that was tightly linked to behavior familiar to patients and clinicians and began diagnosing large numbers of low-grade mammographically detected intraductal breast abnormalities as cancer. These latter diagnoses were based on a long list of difficult-to-agree-on cytological and architectural features.2