Identify Non-mutational p53 Functional Deficiency in Human Cancers.

Area of Science:

• Oncology
• Genomics
• Bioinformatics

Background:

• Accurate p53 functional status assessment is crucial for cancer genomic medicine.
• Non-mutational p53 inactivation is often missed by DNA sequencing, leading to misclassification of TP53 wild-type tumors.
• This misclassification impacts cancer prognosis and downstream analyses.

Purpose of the Study:

• To develop and validate machine learning models for reassessing p53 functional status in TP53 wild-type tumors.
• To identify TP53 wild-type tumors with functionally compromised p53 that are undetectable by sequencing.

Main Methods:

• Support vector machine (SVM) models were constructed to evaluate p53 functional status.
• Models were trained and cross-validated on The Cancer Genome Atlas (TCGA) data from multiple cohorts.
• Performance metrics included AUROC, precision, and recall.

Main Results:

• SVM models demonstrated high performance (AUROC 0.9822, precision 0.9747, recall 0.9784).
• A significant proportion (87%-99%) of TP53 wild-type tumors exhibited functionally compromised p53 (TP53WT-pRF).
• TP53WT-pRF tumors shared genomic (instability) and pathophysiologic (hypoxia) features with TP53-mutant tumors.

Conclusions:

• A substantial number of TP53 wild-type tumors possess functionally impaired p53, necessitating advanced assessment methods.
• TP53WT-pRF tumors are clinically distinct, showing poorer survival and altered treatment sensitivity compared to TP53WT-pN tumors.
• These findings have implications for personalized cancer medicine and therapeutic strategies.