Lack of independent prognostic and predictive value of centromere 17 copy number changes in breast cancer patients with known HER2 and TOP2A status

The clinical benefit of anthracyclines has been connected to HER2 status, TOP2A status and centromere 17 copy numbers (CEN-17). Data from a clinical trial randomizing patients to anthracyclines was used to assess whether the number of CEN-17 in breast cancers may predict incremental responsiveness to anthracyclines besides what is obtained when used relatively to TOP2A and HER2. As cut sections of paraffin-embedded tissue are prone to truncation of nuclei, strict definition of ploidy levels is lacking. We therefore used normal breast tissue to assist define ploidy levels in cut sections. Fluorescence in situ hybridization (FISH) with centromere 17 (CEN-17) and TOP2A was performed on 120 normal breast specimens. The diploid CEN-17 copy number was reduced from the expected two signals in whole nuclei to an average of 1.68 signals per nucleus in cut sections of normal breast. Ploidy levels determined in normal breast were applied to data on 767 patients with known HER2 and TOP2A status randomized to anthracyclines in the DBCG 89D trial. CEN-17 ploidy levels were in cut sections from the 767 breast cancer patients established as: Haploid: ≤1.25 (10%), diploid: 1.26-2.09 (60%), triploid: 2.10-2.93 (21%), tetraploid: 2.94-3.77 (5%) or higher ploidy: ≥3.78 (4%). Amplification of HER2 and deletion of TOP2A were frequently observed in tumors with a high ploidy level. In univariate analyses increasing ploidy was associated with decreased disease-free survival (DFS) (P=0.0001) and overall survival (OS) (P