Variation in both nuclear shape and size ("pleomorphism"), coupled with changes in chromatin amount and distribution, remains the basic criteria for microscopy in a cytologic diagnosis of cancer. The biological determinants of nuclear shape irregularities are not clarified, so, rather than on the genesis of nuclear irregularities, we here focus our attention on a descriptive analysis of nuclear pleomorphism. We keep in mind that evaluation of nuclear shape as currently practiced in routine preparations is improper because it is indirectly based on the distribution of DNA as revealed by the affinity for basic dyes. Therefore, over the last years we have been using as criteria morphological features of nuclei of thyroid and breast carcinomas as determined by immunofluorescence, in situ hybridization, and 3D reconstruction. We have translated this approach to routine diagnostic pathology on tissue sections by employing immunoperoxidase staining for emerin. Direct detection of nuclear envelope irregularities by tagging nuclear membrane proteins such as lamin B and emerin has resulted in a more objective definition of the shape of the nucleus. In this review we discuss in detail methodological issues as well as diagnostic and prognostic implications provided by decoration/staining of the nuclear envelope in both thyroid and breast cancer, thus demonstrating how much it matters "to be in the right shape" when dealing with pathological diagnosis of cancer.