To the editor: We read with interest the article by Esposito et al1 that appeared in the January 2007 edition of your journal. It essentially corroborates our 2004 study on 27 examples of the same lesion that established the morphologically hybrid nature of tubulolobular carcinomas with an immunophenotype more similar to tubular carcinoma than lobular carcinoma (Wheeler et al).2 When a tumor is being evaluated for expression of an antibody and the results are compared with those of another study, it is important to apply the same methodologies. As pointed out in one of our previous studies (Bratthauer et al),3 the high-molecular weight cytokeratin clone 34βE12 is not by itself suitable for the distinction of classic lobular neoplasia from ductal or hybrid neoplastic lesions and that it is of value only when used in tandem with E-cadherin immunostains. This earlier report also pointed out that for immunohistochemical staining of 34βE12, a heat retrieval form of antigen recovery was used—not the proteolytic enzyme digestion performed in the study by Esposito et al.1 Although it is possible that the subset of lobular, tubulolobular, and tubular carcinomas used in Esposito et al's1 study may have reacted differently with 34βE12 compared to the cases we reported, comparisons are valid only when exactly the same methodologies are used. It would be of interest to repeat the high-molecular weight cytokeratin (34βE12) assay on Esposito et al's1 tumor samples with a pH 6.0 heat retrieval pretreatment to see if the results are any different than those obtained using enzyme digestion.