To the editor: We thank Mr Bratthauer, Dr Wheeler and Dr Tavassoli for their comments on our manuscript.

The methodology that we use for 34βE12 antibody is different from that used by Wheeler et al,1 beginning with section thickness. We do not use 6 μm sections for any immunostaining procedure in contrast to the study by Wheeler et al. Indeed, it is not uncommon for a multitude of protocols to exist for any given antibody across laboratories.

The methodology that we use in our laboratory has been optimized and validated against tissues and controls with the Benchmark XT (Ventana Medical Systems, Tuscon, AZ, USA), and the results have been quite satisfactory in our diagnostic experience. We have not optimized 34βE12 antibody in our laboratory for detecting lobular carcinomas, because 34βE12 is an antibody that detects a high-molecular-weight cytokeratin that ultrastructurally correlates with the presence of tonofilaments. Tonofilaments are present in carcinomas of any duct derivation in addition to squamous cell carcinomas.2 We detected 34βE12 in 50% of lobular carcinomas (10 cases examined) vs 80% in the paper by Wheeler et al (5 cases examined). The paper by Wheeler et al also clearly demonstrates that 34βE12 is not useful in the distinction of tubulolobular (93% positive) vs lobular carcinomas (80% positive). E-cadherin alone is a superior antibody for the distinction of ductal vs lobular phenotypes.

It was not the intent of our paper to discuss the usefulness of 34βE12 in making the diagnosis of lobular carcinoma. Rather, our goal was to demonstrate the ductal immunostaining patterns, specifically of E-cadherin and p120 catenin, both of which are biological markers for the E-cadherin/catenin complex. Ours is the first study to demonstrate the membrane-dominant ductal phenotype immunostaining pattern of p120 catenin in tubulolobular carcinomas. The cytoplasmic p120 catenin immunostaining pattern that we described in lobular neoplasia was absent in the tubulolobular group and present in the lobular carcinomas. P120 catenin and E-cadherin are equally sensitive and specific (100%)3 and far superior to 34βE12 for the distinction of ductal vs lobular phenotypes.