Method for grading breast cancer.

labelling.' PCNA immunohistochemical expression (evaluated with the PCI0 mono-clonal antibody) seems to be related to cellular proliferation in many normal tissues and in some neoplasms,2 such as gastro-intestinal lymphomas,3 central nervous system tumours,4 lung neuroendocrine neoplasms,' and prostatic carcinomas.' However, in other tumours, like breast and gastric cancer, PCNA (PC 10) expression seems aberrant and not strictly related to proliferative activity.' 7 8 Various factors unrelated to cell proliferation may influence the immunohistochemical expression of PCNA, including post-transcriptional regulation (and deregulation) of the PCNA gene,8" long half-life of the PCNA protein,'0 involvement of PCNA protein in DNA repair synthesis," and tissue and section processing-type and ionic strength of the fixatives, fixation time, section heating, immunohistochemical techni-qus8. 2 1 3 ques. Further problems in PCNA immunohis-tochemical staining, as in other kinetic quantitative immunohistochemical studies, concern evaluation and scoring methods. '4 15 Should we use quantitative or semiquantita-tive methods? How many cells should be counted? Which tumour areas should be evaluated (the most positive or random selected areas)? Which immunoreactive cells should be evaluated (all positive cells or only the most intensely stained)? Particular attention should be also drawn to the kind of antibody used to localise PCNA. Different staining patterns may be seen with different antibodies, and this may add to conflicting and confusing results.'4 In our opinion PCNA immunostaining should be evaluated with great caution and in some fields even with scepticism. More work is needed to assess the extent and range of PCNA staining in different tissues and lesions (neoplastic and non-neoplastic). PCNA counts should be evaluated concurrently with the different anti-PCNA available antibodies and the results should be compared with other "proliferation markers" and especially with clinical data. The possibility that PCNA immunostaining may have diagnostic5 or prognostic value7 is intriguing and carefully performed clinicopathological studies are needed to assess this possibility further. This will be the only way to know if we are faced with an interesting but clinically worthless tool or with an important test to be added to the routine evaluation of neo-plasms. and Ki67 expression in breast carcinoma. Correlations with clinical and biological parameters.