Pre-analytical processing protocol of breast biopsies affects multigene panel results

Abstract

The creation of multigene panels for prognostic and predictive purposes allows a more accurate indication of adjuvant chemotherapy for patients with breast cancer. In a previous study, we reproduced a multigene panel of 21 genes based on the commercial Oncotype-DX method. We submitted 183 embedded specimens obtained from breast surgery on patients with locoregional disease (stages I to III) between 2005 and 2010 performed at the Hospitals of the Medical School of the ABC Foundation. When we analysed the correlations between the score of the multigene panel and the progression-free interval (PFI) in all patients, we did not find a statistically significant association. However, when we selected only the 71 samples that had amplification of at least eight non-housekeeping genes, we observed that those with scores above the 75th percentile had a significantly lower PFI (p = .0054). Samples processed with nonbuffered formaldehyde were associated with a worse quality of extracted RNA (p = .004) and a significantly higher multigene panel score (p = .021). We conclude that variations in the pre-analytical processing of specimens destined for multigene panel amplification can significantly affect the results, with a potential impact on clinical management.

Keywords: multigene panel; neoplasia; prognosis; recurrence score.

FIGURE 1

Flowchart for inclusion and exclusion of tumour samples in this study

FIGURE 2

Kaplan–Meier curve of set 1 when we separated patients with an X2_8 score above the 75th percentile (53.31) and those below this value. The log‐rank test showed a value of p = .0054. PFS stands for progression‐free survival

FIGURE 3

Kaplan–Meier relapse‐free survival curve for patients with positivity for hormone receptors and negativity for HER2/neu. There was a trend toward better progression with X2_8 values below the 75th percentile than with values above the 75th percentile (Gehan–Wilcoxon log‐rank, p = .0717). We found no relationship between X2_8 and survival in this group of patients (data not shown)

FIGURE 4

X2_8 levels in relation to set 1 (batch 1) and set 2 (batch 2). The ANOVA test showed a value of p = .02667

FIGURE 5

Scores of X2_8 according to the year of collection (A) and fixation protocol (B). In (A), the median scores for the years 2009 and 2010 are significantly higher than those of previous years (p = .02141). In (B), the ANOVA test revealed a value of p = .02314