Problem: A current shortcoming in the prognosis and treatment of head and neck squamous cell carcinoma (HNSCC) is a lack of methods that adequately address the complexity and diversity of the disease. Molecular modeling to assess the predictive ability of an evidence-based panel of gene loci implicated in head and neck cancer to differentiate early- and late-stage HNSCC disease was tested in a pilot cohort of 15 HNSCC cases in an exploratory analysis. Methods: Patient cohort DNA was interrogated for gene loss and gain at 122 gene loci. An interpretation of gene loss or gain was measured as the number of copies in a range of 0 to more than 2 copies, respectively, where 2 (copies) was normal. Tumor stages were classified into 3 categories. Category 1 included stage 0 (preneoplastic), stages 1 and 2 (40%; 6 cases). Category 2 included stages 3 and 4 (40%; 6 cases). Category 3 included stage unknown (20%; 3 cases). The difference in mean copy number among stage categories was tested using ANOVA. A gene probe with overall P-value <0.1 (candidate genes) was further tested for pair-wise difference between stage categories. To test for correlation (Spearmen) among gene probes, an absolute correlation coefficient (ACC) was set in the range of 0 to 1. Results: Nine gene probes were selected as candidate genes: ABCG2(04q22), IFNG(12q14), IL4(05q31.1), ING1(13q34), MLH1 (03p21.3), MME(03q21-q27), RB1(13q14), SCYA31(17q11-q21), and TINF2(14q12-q21.3). The pairs ING1 and TINF2, MME and SCYA3 /TINF2, and SCYA3 and TINF2 are correlated with an ACC of 70%. Conclusion: Our exploratory analysis indicates that molecular modeling strategies can be informative in delineating candidate genes associated with early- and late-stage diesase. Significance: Additional prognostic markers are needed to supplement the TNM staging system. The genetic information obtained in our study through molecular modeling strategies has the potential to provide a multivariable instrument to aid in the diagnosis and prognosis of HNSCC. Support: NIH CA 70923, DAMD DAMD17-00-1-0288 and DAMD17-02-1-0406