Modern classifications of ductal in situ carcinoma (DCIS) of the breast suffer from unsatisfactory reproducibility in inter-observer circulation analyses. Ducts in DCIS are markedly enlarged in the range of 360 microm in diameter. Since the diffusion of oxygen from peri-ductal vessels is limited to 100 microm, cells in the center of DCIS are poorly oxygenated and become either necrotic or remain hypoxic but viable. There is evidence that such alternative fate is dictated by the biological characteristics of the neoplastic cells. Therefore, determination of presence or absence of necrosis in ducts up to 360 microm in diameter might represent a simple, reproducible, and biologically sound criterion to classify DCIS. In the present work, following this criterion, we classified 32 cases of intra-ductal lesions as either "necrotic" or "hypoxic" and tested the reproducibility of such classification using K statistics. These cases had already been circulated among a group of European pathologists, who classified the lesions using five different classifications. The K statistics value obtained with the presently proposed system was extremely high (0.91). It remains to be established whether the classification "necrotic/hypoxic" withstands large inter-observer circulation analyses, whether it is predictive of the clinical evolution of DCIS, and whether it might constitute a reproducible basis for selecting appropriate treatments.