The burden of head and neck squamous cell carcinoma (HNSCC) is greater for African Americans (AA) than for Whites with both incidence and disease-specific mortality higher in AA. Early detection tests are needed because the majority of patients present in late stage when cure rates reach only 40%. Our group has developed a prototype early detection test based on soluble CD44 (solCD44) and total protein levels that is simple, inexpensive and noninvasive. Subjects included 21 HNSCC AA patients and 68 AA controls recruited from the University of Miami Hospital and Clinics, Jackson Memorial Hospital and Liberty City Community in Miami-Dade County. The majority of cases and controls come from underserved populations of low SES. A one-way ANOVA of the two cancer types and controls followed by pairwise comparisons using the Student t-test, and comparison of all cases vs. controls using the contrast method were computed. The effects of solCD44 and total protein on the risk of being a case was evaluated using univariate and multivariate logistic regression analyses. We report odds ratio (OR) estimates with corresponding 95% confidence interval (95% CI) and area under the curve (AUC) of the operating characteristic curve (ROC) for all fitted models. The groups did not differ in regards to gender (p=0.684), but did differ significantly in age (cases were older, mean 60 vs. 50, p<.0001) and ethnicity (19% Hispanic in cases vs. none in controls, p=.003). Log2solCD44 and total protein showed significantly elevated levels in cases compared to controls (2.12 vs. 0.76, p<.0001 and 1.19 vs. 0.72, p=.001, respectively). Lip/oral cavity and oropharynx cancer cases and controls were significantly different with respect to log2solCD44 and total protein at p<0.05. Among cancer patients, difference in means by disease site was only statistically significant for log2solCD44 (lower for oropharynx than lip/oral cavity (1.86 vs. 3.21) at p<0.05). Univariately, both markers seemed to have good predictive ability (log2solCD44 AUC=0.814, OR=4.52, p=.0001 and total protein AUC=0.714, OR=5.66, p=.0015). The best models were log2solCD44 (OR=3.66) adjusted for age and gender (AUC=0.930), total protein (OR=5.25) adjusted for age and gender (AUC=0.899), and log2solCD44 (OR=3.17) and protein (OR=1.80) together adjusted for age and gender (AUC=0.930). In this last model, log2solCD44 maintained its effect in the presence of protein; however log2solCD44 appears to be a confounder for protein, reducing considerably its effect. Tests of effect modification (interactions) of ethnicity and age on log2 solCD44, on total protein, and between markers were carried out, and there were no interactions found. Our preliminary data on AA underserved populations show great promise for detecting HNSCC. In addition, our findings strongly suggest that race, gender and age are very important components for HNSCC early detection studies.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 3627. doi:1538-7445.AM2012-3627