Abstract
The tumor suppressor gene TP53 is inactivated by mutation in a large fraction of cancers. In head and neck squamous cell carcinoma (HNSCC) patients, a subset of TP53 mutations are nonsense mutations that lead to premature termination of non-functional p53 translation. These mutations have shown worse overall survival in HNSCC yet we have not been able to use this information to optimize treatment. We hypothesized that destructive-type TP53 mutations are associated with worse overall survival (OS), even after controlling for important clinical characteristics. In addition, we sought to identify clinical factors associated with destructive-type TP53 mutations.
We abstracted 461 patients from The Cancer Genome Atlas (TCGA) with stages I-IVB HPV-negative HNSCC, treated with surgery +/- adjuvant therapy or with radiotherapy +/- chemotherapy. We identified TP53 mutations and categorized them as “destructive-type” versus all other mutations (e.g., missense, silent, splice site). We collected patient, tumor, and treatment information. We performed univariable (not shown) and multivariable logistic regressions (LR) to determine which factors are associated with destructive-type TP53 mutations. Then, we performed univariable (not shown) and multivariable Cox proportional hazards (CPH) regressions on OS as a function of TP53 mutations and clinical factors. Backward elimination was used for multivariable feature selection, and variables were kept in the final models if they met the P-value threshold of less than 0.05.
51 patients had destructive-type TP53 nonsense mutations and 211 had non-destructive type mutations. On multivariable LR, the presence of destructive-type TP53 mutations was associated with advanced age (OR:2.33, 95% CI:1.09-4.73, P = 0.022), ENE (OR:2.17, 95% CI:1.12-4.13, P = 0.019), and close (OR:2.72, 95% CI:1.15-6.00, P = 0.017) or positive margins (OR:2.36, 95% CI:1.02-5.13, P = 0.036). On multivariable CPH regression, destructive-type TP53 mutations were independently associated with worse OS (HR:1.56, 95% CI:1.02-2.37, P = 0.039), while controlling for significant clinical factors, including PNI (HR:1.43, 95% CI:1.03-1.98, P = 0.032), ≥4 lymph nodes (HR:1.62, 95% CI:1.13-2.34, P = 0.010), microscopic ENE (HR:2.10, 95% CI:1.41-3.14, P = 0.0003), gross ENE (HR:2.0, 95% CI 1.19-3.35, P = 0.008), positive margins (HR:1.52, 95% CI:1.01-2.30, P = 0.047), and advanced age (HR:1.02, 95% CI:1.01-1.04, P = 0.010).
We found that TP53 nonsense mutations are independently associated with worse OS after controlling for significant confounders. Patients with HPV-negative HNSCC with destructive-type TP53 mutations are more likely to be of advanced age and have adverse pathologic features such as ENE and close or positive margins. This study supports characterizing TP53 mutations in patients with HPV-negative HNSCC as it may further optimize patient stratification, treatment, and clinical trial designs.