Chu Chen, PhD, NRCC, DABCC

Selected Research Findings

Prognostic Gene Signatures for Oral Cancer

Oral squamous cell carcinoma (OSCC), comprising malignancies of the oral cavity and oropharynx, is the 6^th most common cancer world-wide, with an estimated 529,000 new cases in 2012. This disease is often fatal; the five-year survival has remained at about 50% for the past four decades. For those who survive, the disease and treatment often result in the loss of orofacial function, pain, and disfigurement.

Recent investigations all show HPV status to be an important and independent prognostic factor for oropharyngeal cancer (OPC). The 5-year survival under current standard of care regardless of the specific therapy given is about 80-85% for HPV-positive OPC patients, but was only about 30-35% for HPV-negative OPC patients.

For oral cavity cancers (OCC), in which HPV is relatively uncommon (10 to15%), the overall 5-year survival is about 70% for stage I or II disease, 45% for stage III disease and about 35% for stage IV disease (http://www.cancer.org). Unfortunately, studies to date indicate that HPV status is not likely to be useful as a prognostic marker in patients with OCC, and there are no other prognostic biomarkers that can reliably predict the survival of these patients to aid clinical management. At present, the classification of OCC to inform treatment or prognosis is heavily dependent on AJCC stage. Yet, the ability of staging to predict prognosis in OCC is limited; patients with tumors of the same clinical and pathologic staging have heterogeneous response to clinical treatment, and different probability of recurrence and survival. The treatment for OCC varies from uni-modality treatment (surgery or radiation) for early stage disease to multimodality treatments (some combination of surgery, radiation, and chemotherapy) for late stage disease. For those patients who present with early stage I/II disease treated with uni-modality therapy (i.e., surgery) and for whom open neck surgery to fully assess high-risk features (multiple positive lymph nodes, or extracapsular spread in positive lymph nodes) may not be warranted, more precise knowledge of whether a patient’s tumor is associated with a poor prognosis might justify treatment intensification with a second modality (i.e., radiation +/- chemotherapy). Although randomized clinical trials show that adding chemotherapy to radiation treatment improves prognosis in OSCC patients, we lack biomarkers to identify which patients with high risk features will respond to more aggressive treatment, and thus which patients could be spared significant treatment toxicity. Thus, an assay to improve the prediction of survival of OCC patients under the current standard of care could help physicians and patients to make better personalized treatment choices.

Building on our work using Affymetrix gene expression arrays to subclassify prospectively-collected fresh OSCC from the Oralchip study, we have identified a gene expression profile of 131 probesets (representing 108 unique known genes) which not only differentiate invasive OSCC from normal oral epithelium but also predict OSCC survival. Furthermore, we have identified a subset of 13 genes from this 131 probeset list for which the gene expression is strongly associated with OSCC-specific survival for patients with HPV-negative oral cavity and oropharyngeal cancer irrespective of treatment modalities, and validated the performance of this 13-gene signature in predicting survival of OCC patients using an external independent dataset This is the first demonstration of a gene set that provides prognostic information beyond AJCC stage for OCC patients.  We are currently conducting an international multicenter study to convert the testing of this gene signature to a potentially useful clinical test to aid the management of patients with HPV-negative OCC using funds from a recent NIH award.

Etiology of Head and Neck Squamous Cell Cancer (HNSCC)

We have been investigating etiologic factors for HNSCC by pooling our results from our case-control study with those from investigators in the International Head and Neck Cancer Consortium (INHANCE), These efforts led to the following observations: 1) Sexual behaviors, such as higher number of sexual partners, higher number of  oral sex partners, early age of sexual debute, are associated with cancer risk at the head and neck cancer subsites that have previously been associated with HPV infection. 2) The risk for HNSCC increases with greater cigarette and alcohol use and lower body mass index (BMI). Adult height is inversely associated with HNSCC risk. 3) A genetic variant at chromosome 12q24 and three ADH variants were associated with HNSCC susceptibility. 4) High fruit/vegetable and low red meat intake were associated with reduced HNSCC risk. 5) A history of diabetes was weakly associated with HNSCC overall, with a positive association among those who never smoked cigarettes and no association among ever smokers. 6) Cigar- and pipe-smoking are independently associated with increased risk of HNSCC. Marijuana use was associated with the risk of oral tongue cancer but not oropharyngeal cancer.

Etiology of Lung Cancer

In a nested case-control study comparing 746 lung cancer cases and 1,477 controls, all of whom were non-Hispanic white smokers in the β-Carotene and Retinol Efficacy Trial (CARET), we examined whether lung cancer risk is associated with single nucleotide polymorphisms (SNPs) in base-excision repair, nucleotide excision repair, mismatch repair, and homologous end joining DNA repair pathways and whether diet modifies the association, if any. We found that inherited variants in LIG1 and possibly other nucleotide excision repair genes may predispose to smoking-related lung cancer. We also found MSH5/BAT3 locus to be associated with increased lung cancer risk among smokers. We confirmed the reported relation between chromosome 15q24-25.1 variation and lung cancer risk and found diet has little influence on this relation. Our collaboration with investigators in the International Lung Cancer Consortium (ILCCO) and the Women’s Health Initiative (WHI) led to the following observations: 1) Several genetic variants on 5p15, 6p21 and 15q25 were associated with lung cancer risk. Histology-specific effects for 5p15, 6p21, 9p21 and 12p13 loci but not for 15q25 were observed. 2) Associations between genetic variants in TERT (rs2736100), HLA, CHEK2, BRCA2 and lung cancer risk were observed. 3) Unlike use of estrogen plus progestin, which was associated with an increased death rate from lung cancer, use of conjugated equine estrogen alone did not increase either incidence or mortality from lung cancer.

Etiology of Endometrial Cancer

We examined single-nucleotide polymorphisms (SNPs) in the NER genes in a population-based case-control study in Washington State, with 783 endometrial cancer cases and 795 controls. We found that certain ERCC5, LIG1, XPA, and XPC genotypes might influence endometrial cancer risk. A history of incomplete pregnancies was not associated with endometrial cancer risk to any appreciable degree. Older age at first birth was more strongly associated with decreased endometrial cancer risk than was older age at last birth. A long duration of use of continuous-combined estrogen-progestin therapy was associated with a reduced risk of endometrial cancer. Our collaboration with investigators in the Etiology of Endometrial Cancer Consortium (E2C2) and WHI has led to the following observations: 1) Late age at last birth was independently associated with a reduced risk of endometrial cancer, and the reduced risk persisted for many years. 2) Type I and Type II endometrial cancer share many common etiologic factors. The etiology of type II tumors may not be completely estrogen independent, as previously suggested. 3) A large two-phase genome-wide association study did not find genetic variants that are associated with endometrial cancer with genome-wide statistical significance. 3) Among WHI women, after adjusting for BMI, the associations between diabetes, diabetes treatment, diabetes duration and the risk of endometrial cancer became non-significant. These results suggest that the relationship observed in previous research between diabetes and endometrial cancer incidence may be largely confounded by body weight, although some modest independent elevated risk remains.4) Ever use of IUDs was inversely related to endometrial cancer risk. 5) The major known risk factors for endometrial cancer, e.g., BMI, parity, smoking, oral contraceptives use, exert similar effects on black and white women.