To the Editor: We read with great interest the many abstracts in your January 2000 issue regarding Her-2/neu and its value in the evaluation of breast carcinoma (1, 2, 3, 4, 5, 6, 7, 8, 9). More than one of these abstracted studies compared the immunohistochemical and fluorescence in situ hybridization (FISH) techniques for semiquantitation of Her-2/neu (5, 6). These studies reflect the substantial debate now occurring as to which technique is the superior methodology for evaluating Her-2/neu status in breast carcinomas. The answer seems unclear, in part because the question is two questions, neither of which we believe has been precisely framed. Her-2/neu status seems to have both prognostic and predictive value. The answer as to which assay methodology is superior may depend on whether one is interested in Her-2/neu as a prognostic or as a predictive marker. We believe that the preponderance of evidence indicates the superiority of direct assessment of the genetic material for prognostic stratification (10). Current data support a significant association between amplification of the Her-2/neu oncogene and a poor prognosis. This link has been most reproducibly demonstrated by the polymerase chain reaction and FISH techniques rather than by immunohistochemistry (IHC) (10).

The picture is less clear when Her-2/neu analysis is used as a predictive marker for response to Herceptin therapy (6). Herceptin therapy uses an antibody directed against the protein product of the Her-2/neu oncogene as a specific delivery system for a cytotoxic agent. The Her-2/neu protein product is expressed on the cell membrane surface in increased amounts in a subset of breast carcinomas. Theoretically, the greater the amount of Her-2/neu protein on the cell surface, the greater the amount of cytotoxic agent that can be delivered to the neoplastic cell. Hence, Herceptin therapy would be most effective against cancers that express high levels of cell membrane Her-2/neu protein. In the majority of Her-2/neu–positive cases, increased protein is secondary to amplification of Her-2/neu oncogene, but other pathways for overexpression exist, including posttranscriptional and posttranslational events. It is also possible that amplification of the Her-2/neu oncogene may not result in overexpression of a protein product recognizable by antibodies used in an IHC assay. Thus, one would not expect perfect correlation between FISH and IHC assay results.

From these observations, two conclusions can be drawn. First, assay of the Her-2/neu protein on the cell surface by IHC is a potentially superior technique for the prediction of Herceptin binding and response to Herceptin therapy than is measurement of DNA amplification by FISH techniques. Second, the IHC methodology would be most predictive of response if the assay antibody had in vitro binding characteristics identical to the Herceptin-carrying antibody’s in vivo binding characteristics. Although similar, the antibody used in the HercepTest kit is a different clone than the carrying antibody used in the therapeutic agent Herceptin (11, 12). We are unaware of any published data comparing the binding properties of these two antibodies or the superiority of the clone used in the HercepTest over other commercially available clones for the prediction of response to Herceptin therapy.

Large long-term clinical studies are needed to document the superiority of FISH or IHC in the prediction of response to Herceptin therapy. The abstract by Kaptain et al. (6) addressed this issue but did not include sufficient patients with long-term follow-up for definitive conclusions to be drawn. Large studies with long-term clinical correlation are necessary to identify which antibody clones are best able to predict the in vivo binding characteristics of the Herceptin-carrying antibody so that decisions regarding Herceptin therapy can be most accurately made based on Her-2/neu analysis of tissue specimens.