Markers of prognosis in breast cancer--the relationship between binding of the lectin HPA and histological grade, SPF, and ploidy.

Abnormal cellular glycosylation as demonstrated by the binding of a lectin from Helix pomatia (HPA) to paraffin-embedded sections has been shown in several studies to be associated with aggressive biological behaviour and poor long-term patient prognosis in breast cancer. This study aims to address the possibility that expression of the HPA binding ligand may be of prognostic significance through an association with increased cellular proliferation (as measured by S-phase fraction and histological grade), anaplasia (reflected in histological grade), or ploidy (DNA index). In a 24 year retrospective study, paraffin-embedded sections of 366 primary breast cancers were stained for binding of HPA. All tumours were assessed for histological grade. Flow cytometry was performed on all cases for which sufficient tumour tissue was available (358/366 cases) and S-phase fraction (SPF) and ploidy calculated. Data regarding patient age at diagnosis, nodal status, and tumour size were also recorded. Life table analyses revealed survival advantage for HPA 'non stainers' in comparison to 'stainers' (p < 0.001); for patients with tumours of low grade vs. high grade (p < 0.001); and for those with tumours of low SPF vs. high SPF (p < 0.001). No survival advantage was shown for those with diploid vs. aneuploid tumours (p = 0.17). No association was apparent between HPA binding and grade, SPF, or ploidy (Chi squared values not significant). This was confirmed by multivariate analysis in which nodal status, tumour size, and SPF were independently predictive of survival. There was no confounding effect of grade, SPF, or ploidy upon the correlation between survival and HPA binding. HPA was, however, not independently predictive owing to its strong association with nodal status. The results of this study suggest that the prognostic significance of altered glycosylation, as detected by HPA binding, is unlikely to be through an association with proliferative rate, degree of anaplasia, or cellular ploidy, but may rather be through a direct association with the presence of nodal metastases.