Redefining aberrant P53 expression of gastric cancer and its distinct clinical significance among molecular-histologic subtypes

Shih-Chiang Huang,Ian Yi-Feng Chang,Tse-Ching Chen,Hsiao-Ching Lin,Chun-Yi Tsai,Jun-Te Hsu,Chun-Nan Yeh,Shih-Cheng Chang,Ta-Sen Yeh
DOI: https://doi.org/10.1016/j.asjsur.2024.05.121
Abstract:Background: Numerous studies have demonstrated a correlation between p53 overexpression and diminished survival in gastric cancer patients. However, conflicting findings exist, and we hypothesize that these discrepancies arise from the cancer's complexity and heterogeneity, coupled with a lack of consensus on aberrant p53 expression. Methods: We enrolled a cohort of 187 patients with surgically resected gastric cancer. Patient categorization was based on Epstein-Barr virus (EBV), microsatellite instability (MSI), and Lauren classification (intestinal, diffuse and mixed). Utilizing an incremental algorithm, we evaluated p53 immunohistochemical (IHC) patterns in all 187 cases, while next-generation sequencing was successfully performed on 152 cases to identify TP53 mutations (mutTP53). Results: MutTP53 was identified in 32 % of the 152 cases, comprising 36 missense, 5 nonsense, and 7 frameshift alterations. Missense mutations predominantly correlated with p53 overexpression, while nonsense and frameshifting alterations related to null expression. Trial calculations indicated that null expression and a p53 IHC cutoff at >40 % offered the best prediction of mutTP53 (kappa coefficient, 0.427), with the highest agreement (0.524) observed in diffuse type and the lowest (0.269) in intestinal type. Null expression and a p53 IHC cutoff at >10 %, but not mutTP53 per se, provided the optimal prediction of survival outcome (p = 0.043), particularly in diffuse type (p = 0.044). Multivariate analysis showed that aberrant p53 IHC expression was not an independent prognostic factor. Conclusions: P53 IHC patterns are predictive biomarkers for mutTP53 and gastric cancer outcomes, where a prerequisite involves a nuanced approach considering cutoff values and molecular-histologic subtyping.
What problem does this paper attempt to address?