Evaluation of computational tools to determine prognostic significance of TP53 mutation in head and neck squamous cell carcinoma (HNSCC).

Head and Neck Cancer
Session Type and Session Title: 
Poster Highlights Session, Head and Neck Cancer
Abstract Number: 
J Clin Oncol 32:5s, 2014 (suppl; abstr 6035)
David Masica, Shuli Li, Chris Douville, Judith Manola, Robert L. Ferris, Barbara Burtness, Arlene A. Forastiere, Wayne Koch, Rachel Karchin, Christine H. Chung; The Johns Hopkins University School of Medicine, Baltimore, MD; Dana-Farber Cancer Institute, Boston, MA; University of Pittsburgh School of Medicine, Pittsburgh, PA; Fox Chase Cancer Center, Philadelphia, PA; eviti, Inc., Philadelphia, PA

Abstracts that were granted an exception in accordance with ASCO's Conflict of Interest Policy are designated with a caret symbol (^).

Abstract Disclosures


Background: TP53 is the most commonly mutated gene in HNSCC. The specific mutations in TP53 can be prognostic of patient survival. Thus, a framework to predict patient survival from previously uncharacterized mutation in TP53 is valuable. There are many computational tools for predicting the phenotypic impact of genetic variation, but the overall clinical value of these algorithms remains unclear. Methods: Sixteen different models to predict HNSCC patient survival based on TP53 mutations were assessed using the TP53 mutation and clinical data from ECOG 4393 [Poeta, M. L., et al. NEJM (2007) 357(25) 2552-2561]. These models include: server-based computational tools SIFT, PolyPhen-2, and Align-GVGD; our in-house POSE and VEST algorithms; the rules devised in Poeta et al.1 with and without considerations for splice-site mutations; location of mutation in the DNA-bound TP53 protein structure; and a functional assay measuring WAF1 transactivation in TP53-mutated yeast. Results: We assessed model performance using overall survival (OS) and progression-free survival (PFS) from 420 HNSCC patients, of whom 224 had TP53 mutations. Each mutation was categorized as “disruptive” or “non-disruptive”. For each model, we compared the outcome between the predicted disruptive group vs. the non-disruptive group. The rules devised by Poeta et al. 1 (disruptive mutations: non-conservative mutations in the key DNA-binding domain, or stop codons vs. non-disruptive mutations: all mutations excluding the disruptive mutations) with or without our modification were observed to be superior to others. While the differences in OS (disruptive vs. non-disruptive) appear to be marginally significant (Poeta rule+modification, p=0.089; Poeta rule: p=0.053), both algorithms identified a disruptive group that has significantly worse PFS outcome (Poeta rule+modification, p=0.011; Poeta rule, p=0.027). Conclusions: In general, prognostic performance was low among the computational methods assessed here. Further studies are required to develop and validate computational models that can predict functional and clinical significance of TP53mutations in HNSCC patients.