p53 immunostaining cannot be used to predict TP53 mutations in gastric cancer: results from a large Central European cohort.

Four molecular subgroups of gastric cancer (GC) have been proposed, ie, Epstein-Barr virus (EBV)-positive, microsatellite instable, chromosomal instable (CIN), and genomically stable GC. Based on the complex relationship between chromosomal instability and TP53 mutational status, we hypothesized that the typical clinicopathological characteristics caused by chromosomal instability are correlated with the p53 expression that is detected by immunohistochemistry. Four hundred sixty-seven whole-tissue sections of patients with therapy-naive GC were stained with anti-p53 antibody. The histoscore and staining pattern were analyzed for each slide. Different algorithms of immunohistochemistry evaluation were formed and correlated with clinicopathological characteristics. The algorithms were validated by assessing the mutational status of TP53 in 111 cases. Four hundred forty-two GCs were p53 positive, and 25 were negative, including 414 GCs with a homogeneous pattern and 53 GCs with a heterogeneous staining pattern. There was no correlation with overall or tumor-specific survival. In comparison with clinicopathological characteristics, the algorithm high versus low showed correlations with microsatellite instability, hepatocyte growth factor receptor (MET), and TP53 mutational status. The algorithm Q1/Q4 versus Q2/Q3 appeared to be correlated with the phenotype as per the Laurén classification, microsatellite instability, EBV status, and p53 expression pattern. The algorithm <90% = 0 and <50% = 3+ versus ≥90% = 0 or ≥50% = 3+ showed correlations with the EBV status, microsatellite instability, grading, and p53 expression pattern. The algorithm homogeneous versus heterogeneous did not correlate with any clinicopathological characteristic. Our results showed that the immunohistochemistry of p53, TP53 mutational status, and CIN subtype were connected. However, different algorithms for p53 immunohistochemical evaluation cannot be used to predict TP53 mutations in CIN GCs in individual cases.