Letter to the editor regarding ‘Seol H, Lee HJ, Choi Y, et al. Intratumoural heterogeneity of HER2 gene amplification in breast cancer: its clinicopathological significance’

To the Editor: In Seol et al,¹ the authors provide a clinicopathologic analysis showing that intratumoral heterogeneity of HER2 gene amplification is associated with short disease-free survival. They conclude that it is likely that intratumoral heterogeneity is a surrogate for chromosomal instability, and thus a poor prognosis. This result would appear directly to conflict with the study of Bartlett et al,² showing that patients with tumors that are uniformly HER2-amplified do worse than those with heterogeneity (eg, 30–50% of cells with a ratio >2.2). Seol et al¹ attribute this difference to a variation in study design—that they have selected their heterogeneous cases from tumors that were already classified as HER2-amplified on whole-tissue sections. To this reader, an alternative interpretation presents itself, which takes into account patient treatment, as well as one study³ not cited by Seol et al (See Table 1).

Table 1 Prognostic significance of ‘HER2 heterogeneity’ accounting for treatment

From Table 1, it appears that intratumoral heterogeneity, in and of itself, is not a poor prognostic marker at all.² Rather, high/unequivocal HER2 amplification is a favorable predictor of response to (antracycline-based) chemotherapy—a result that has been well documented.^{3, 4} Moreover, patients with low-HER2-amplification—and heterogeneity, perhaps—still benefit from trastuzumab in addition to chemotherapy.^{1, 3}

Seol et al¹ rightly highlight the importance of determining the HER2 amplification status accurately, both overall and taking into account intratumoral heterogeneity. Based on our own work, a fully satisfactory definition of heterogeneity has not been forthcoming. A persistent problem is how to distinguish bonafide heterogeneity from statistical artifact.⁵ Both Bartlett et al² and Seol et al¹ raise the possibility of examining ‘regional heterogeneity’. The current guidelines address this by recommending that distinct (clustered) subregions with differences in HER2 status be scored separately.⁶ Further work is needed to define the most revealing testing parameters with respect to prognosis, trastuzumab response, and chemotherapy response.