e15001 Background: Overall survival (OS) is an important endpoint, with the advantage that there is minimal ambiguity in defining an OS event; the patient is either alive or dead. Some single gene mutations, such as TP53, were identified as important predictors of shorter OS in malignant tumors. However, the relationship between gene co-mutation and OS of malignant tumors is poorly defined. Methods: Genomic and OS data were derived from The Cancer Genome Atlas (TCGA) databases. We defined the double-mutant gene pair as “hit pair” in the specific cancer type if statistically significant OS difference was observed in at least one of three comparison (double-mutant samples vs gene A-only-mutant samples; double-mutant samples vs gene B-only-mutant samples; gene A-only-mutant samples vs gene B-only-mutant samples). Results: PIK3CA-TP53 co-mutation was evaluated as “hit pair” in seven cancer types: GBM (Glioblastoma multiforme), HNSC (Head and neck squamous cell carcinoma), LUSC (Lung squamous cell carcinoma), SARC (Sarcoma), SKCM (Skin cutaneous melanoma), UCEC (Uterine corpus endometrial carcinoma), and UCS (Uterine carcinosarcoma). In SARC and SKCM, the expected hazard ratio was higher in both-mutant group than TP53-only-mutant group, and neither genes were correlated with OS. Especially in SKCM, after we took “neither” group (both genes were wild-type) into comparison, hazard ratio of both-mutant group was significantly higher than neither group. Single-gene mutation didn’t make difference on hazard ratio, which indicated that the both-mutant interaction made the leading contribution. In GBM, LUSC, and UCS, the significant difference of hazard ratio between TP53-only-mutant group and PIK3CA-only-mutant group was neutralized in the both-mutant group, to some extent equivalent to “average effect (OS curve of double-mutant group was between that of only single gene mutant groups)". For PIK3CA-TP53 in HNSC, both-mutant group and TP53-only-mutant group carried with a higher hazard ratio than PIK3CA-only-mutant group. It suggested that the increasing hazard ratio in both-mutant group might be resulted from TP53 mutation. For the circumstance in UCEC, TP53 and PIK3CA were testified to be correlated with higher and lower hazard ratio, respectively and both-mutant group and TP53-only-mutant group both with significantly higher hazard ratio than PIK3CA-only-mutant group, which further verified the greater strength of TP53 than PIK3CA in both-mutant group. Conclusions: In summary, our study identified PIK3CA-TP53 co-mutation as a predictor of OS in seven cancer types: GBM, HNSC, LUSC, SARC, SKCM, UCEC, and UCS, especially in SARC and SKCM.
Read full abstract