30th Annual Meeting • American Society of Preventive Oncology, Bethesda, Maryland • February 26–28, 2006

Purpose: To examine the association between recreational physical activity and invasive colon adenocarcinoma among women enrolled in a prospective cohort study.

Methods: 120,147 CTS participants residing in California and ages 20-84 years with no prior history of colon cancer were included in the analyses. Three hundred nine-five were diagnosed with invasive colon cancer between 1996 and 2002. The relative risks associated with lifetime recreational physical activity were estimated using multivariable Cox proportional hazards regression models.

Results: Recreational physical activity was not associated with colon cancer risk in the cohort overall. However, physical activity reduced colon cancer risk among post-menopausal women who had never taken estrogen or combined hormone therapy. Women who reported an average lifetime moderate or strenuous recreational physical activity (from high school through age 54 years) of at least 4 hrs/wk/yr had 48% lower colon cancer risk (relative risk 0.52, 95% confidence interval 0.31-0.85) than women with a lifetime average of less than 0.5hr/wk/yr. Risk was not reduced among postmenopausal women with a history of hormone therapy use. We observed no effect modification by age, smoking status, level of folate intake, or body mass index.

Summary: These data suggest that lifetime recreational physical activity protects against colon cancer in post-menopausal women who have not taken hormone therapy. Hormone therapy users benefit from a lower colon cancer risk associated with hormone therapy use, but recreational physical activity does not appear to reduce risk further among these women.

Purpose: Chinese women have among the lowest breast cancer screening rates in the US. We developed and evaluated a culturally-tailored educational video guided by the Health Belief Model to promote Chinese women's use of mammography.

Method: This study included three phases: 1) focus-group discussions and an advisory board meeting including Chinese community leaders and cancer survivors to guide the video development, 2) producing the video with community actors, and 3) conducting a pre-post test pilot to evaluate the efficacy of the video in changing knowledge, beliefs, and screening intentions among Chinese women (age>39) who were not adherent to current NCI mammography guidelines (n = 50).

Results: A 17-minute video was produced in Mandarin and dubbed with Cantonese voices. The video included a soap-opera addressing barriers to screening and a segment with a physician recommending screening. Our preliminary evaluation of the video showed that compared to 37% at baseline, 88% of the participants intended to obtain a mammogram after viewing the video (P < .0001). There were significant increases in knowledge about breast cancer and mammography (P = .001) and decreases in Eastern cultural views of cancer (P < .0001). More than 84% of the women liked the video and said it was understandable, persuasive, and clear.

Conclusion: Our video was successfully created based on an intensive collaboration within our local Chinese community. This culturally-tailored video has the potential to motivate Chinese women to adhere to mammography screening. We will be testing the efficacy in future trials with broader community populations.

We conducted a pilot study to examine associations between polymorphisms in genes related to estrogen biosynthesis (CYP17 T→C, CYP19 TTTA repeats) and metabolism (CYP1B1 codon 432 G→C and codon 453 A→G, COMT codon 158 G→A) and urinary estrogen metabolites (2-hydroxyestrogens (2-OHE), 16-alphahydroxyestrone (16-alpha-OHE1), and their ratio) in 64 pre- and postmenopausal women with a family history of breast cancer. Women were participants in the Metropolitan New York Registry, one of six NCI Breast Cancer Family Registries. We used linear regression to examine the associations between genetic polymorphisms and log-transformed urinary metabolite levels. After adjusting for menstrual status, BMI and age, we found that carriers of the CYP1B1 codon 453 G allele had 31.0% lower levels of 2-OHE (P-values = 0.053) and 40.2% lower levels of 16-alpha- OHE1 (P = 0.005). When we restricted the analyses to premenopausal women (n = 41), we found similar results. Consistent with other studies, among premenopausal women the COMT codon 158 A allele was associated with increased 2-OHE levels (P = 0.031) and an increased 2-OHE/16-alpha-OHE1 ratio (P = 0.035); the CYP17 C allele was associated with increased 2- OHE levels (P = 0.082). To our knowledge this is the first report showing the effect of the CYP1B1 codon 453 G allele on urinary 2-OHE and 16-alpha-OHE1 metabolites, despite the small sample size. Further larger studies should be done to confirm these results.

Purpose: Using data from a prospective cohort study, we investigated associations of interviewer-measured body mass index (BMI) and postload plasma glucose (PLG) levels with risk of mortality from non-Hodgkin lymphoma (nHL) among persons without diabetes at baseline and to explore associations with leukemia and multiple myeloma.

Methods: Employees of 84 Chicago-area organizations, with an average age of 40 years at baseline, were screened from 1967 to 1973. Height and weight were measured by study nurses. A 50g oral glucose load was administered to nondiabetic participants. Of the at-risk cohort of 35,420 men and women, 129 died of NHL, 151 died of leukemia, and 66 died of multiple myeloma during an average of 31 years follow-up.

Results: Among men, there were positive dose-response relations of BMI with NHL (Hazard Ratio (HR) = 2.57, 95% confidence interval (CI) = 1.24-5.34 for the highest vs. lowest quartile, P-trend = 0.01) and leukemia (HR = 1.98; 1.07-3.69, P-trend = 0.02), after adjustment for age, education, smoking status, and race. PLG also was positively related to NHL (HR = 2.86, 1.35-6.06 for the highest vs. lowest category, P-trend = 0.004). For women, a higher BMI was positively associated with leukemia (HR = 2.47; 0.96-6.36; P-trend = 0.02) and the highest level of PLG was associated with a three-fold higher risk of mortality from multiple myeloma (HR = 3.06; 1.05-8.93). The risk estimates for obesity and PLG remained essentially unchanged after adjusting for each other.

Conclusion: Our data suggest that factors associated with BMI and/or abnormal PLG might play an important role in the mortality from NHL and possibly, leukemia, and from myeloma in women.

Background: Several papers have examined the relationship between treatment delay and survival among patients who are diagnosed with cancer. None has yet relied on a large, population- based dataset to systematically examine 5-year survival among women within different ethnic/racial groups who delay treatment.

Method: Subjects were 49,865 female Medicare enrollees age 65 and older who were diagnosed with breast cancer between 1992 and 1999 and identified by the Surveillance, Epidemiology, and End Results (SEER) program. Dates of their health care visits were identified through the linkage of SEER with Medicare claims data. Mortality from Breast Cancer was Assessed through Linkage with Death Certificates.

Results: Using the log-rank test for comparing survival curves, non-Hispanic whites (P <.001), blacks (P <.01), and Hispanics (P <.05) with treatment delays 6 months or more had diminished 5-year survival relative to those with less delay. Logistic regression analyses of 5-year survival (with adjustments) revealed that subjects with 6-month delay in treatment had a reduced odds of 5-year survival (adjusted OR 2.52, 95% CI, 1.70-3.73) than those with less delay. Blacks had significantly lessened 5-year survival than women in other races/ethnicities (OR = 1.63, 95% CI, 1.38-1.94), but the interaction between race/ethnicity and delay was not statistically significant. Women who were diagnosed at stage 4 (OR = 298.01, 95% CI, 197.2-450.4), were older (OR = 3.51, 95% CI, 2.92-4.23), unmarried (OR = 1.48, 95% CI, 1.34-1.64), and had 3 or more comorbidities (OR = 1.53, 95% CI, .13-2.08) predicted reduced 5 year survival.

Conclusions: Delay in accessing breast cancer treatment has a clear relationship to survival, among all racial/ethnic subgroups. Rapid access to treatment is recommended for all women with breast cancer.

Purpose: Increasing incidence of hepatocellular carcinoma (HCC) in the US has been associated with hepatitis C (HCV) infections. We report a study of HCC in Egypt, a country with an epidemic of HCV and HCC. The goal of our study is to identify serum peptides associated with HCC for early detection and improved classification of the disease.

Methods: Serum samples were obtained in collaboration with NCI, Cairo, Egypt. Controls were matched to cases on gender, age, and residence. We developed MALDI-TOF/TOF methods for analysis of serum peptides enriched by denaturing ultrafiltration. Analysis of TOF-MS spectra of 78 HCC cases and 72 controls in the 0.8-5 kDa mass range identified 264 peptides, a subset of which was identified by TOF/TOF sequencing. The abundance of 45 peptides was increased (34) or decreased (11) in patients with HCC. Using newly developed computational methods, we selected 6 peptides that classify the disease with 100% sensitivity and 92% specificity in an independent set of 50 samples. Logistic regression analysis showed that the association of biomarker-candidates with HCC is not substantially altered by age, gender, viral infections, and date of sample collection.

Conclusion: Using novel analytical methods, we identified six peptides that classify HCC with high prediction accuracy. These peptides may be useful in examining progression of chronic hepatitis C viral infection to malignancy.

Purpose: Carcinogens produced in the course of cooking meat at high temperatures can lead to DNA damage which can be repaired by the nucleotide excision repair (nER) pathway. We tested whether non-synonymous single nucleotide polymorphisms (nsSNPs) in NER genes may modify the association between meat intake and colon cancer risk.

Methods: Colon cancer cases (n = 643) and a random selection of controls (n = 1049) were selected from 33 counties in North Carolina from 1996 to 2000. Information on meat intake and preparation methods was collected by dietary interview. Genomic DNA from whole blood was used for genotyping 7 NER nsSNPs: XPC A499V and K939Q, ERCC2 D312N and K751Q, ERCC4 R415Q, ERCC5 D1104H, and RAD23B A249V. Adjusted odds ratios (OR) and 95% confidence intervals (CI) were calculated using logistic regression.

Results: No association was observed for any of the NER nsSNPs and risk of colon cancer. Increased risk of colon cancer was observed for high consumption of total meat as compared to low consumption in individuals homozygous variant for the ERCC2 D312N or the K751Q SNP (OR = 2.7, 95%CI = 1.2, 6.3, and OR = 2.4, 95%CI = 1.2, 4.8, respectively) and in carriers with at least one V allele for the RAD23B A249V SNP (OR = 1.6, 95%CI = 1.0, 2.6), but not in homozygous wildtype-carriers for these SNPs.

Conclusions: Specific NER nsSNPs may modify the association between meat intake and colon cancer risk. This work was supported by R01CA90898 and an ASPO/CRPF Cancer Prevention Research Fellowship.

Purpose: We studied the association of the Gail model variables with education, insurance status, and race/ethnicity among women who completed a risk assessment form (RAF) for the Study of Tamoxifen and Raloxifene (STAR), a breast cancer chemoprevention trial.

Methods: We analyzed the association of Gail model risk factors, education, and insurance with race/ethnicity using chi-square tests and two-sided P-values. We developed logistic regression models of trial eligibility, controlling for the Gail model risk factors, education, and insurance status.

Results: Among 823 women who completed an RAF, white women were 10 times as likely as Hispanic women and 45 times as likely as black women to be eligible for STAR. Age at first birth (P = 0.04), having an affected first degree relative (P < 0.0001), having had a biopsy (P < 0.0001), education (P < 0.0001), and insurance status (P < 0.0001) varied by race/ethnicity; all except insurance status were associated with eligibility when race was excluded from the model. In a model that included race/ethnicity, the same factors remained statistically significant.

Summary: Both race/ethnicity and socioeconomic factors were barriers to eligibility for and contributed to low minority participation in a breast cancer prevention trial. The same factors are likely to function as barriers to preventive treatment. Given the relatively high breast cancer mortality among minority women, the race/ethnicity variable should be eliminated from the Gail model formula, and other efforts should be made to improve minority and low-income women's access to breast cancer prevention.

The equivocal or mildly abnormal Pap smear is the cytologic manifestation of HPV infection. With time, persistent carcinogenic infection can lead to cervical precancer/cancer. Limited studies have adequately examined type-specific effects of HPV infection, viral load and age on cytology. To examine the relationships of these factors with cytologic diagnoses, we conducted an analysis of 1,454 women infected with a single HPV type using a 10,000 women population-based prospective study in Guanacaste, Costa Rica. Enrollment cervical specimens were tested for >40 HPV types by MY09/MY11 L1 consensus primer PCR. Stratifying by age, we calculated the frequency of type-specific cytologic abnormality and examined viral load using PCR signal-strength. Our analysis reveals that overall, 21.6% of single HPV infections resulted in equivocal or worse cytologic diagnoses ranging from 0%-80% based on HPV type. Having HPV16, the cause of >50% of cancers, resulted in only 38.5% abnormality. Nonetheless, infection with HPV16 or another carcinogenic type resulted in more abnormalities as viral load increased (P-trend = <0.0001 and <0.001, respectively). Specifically, HPV16 positive women aged 35-55 years with high viral loads had the most cytologic abnormalities (77%). Our results suggest that cytologic abnormality varies by HPV type and age and that in women with HPV16 and other carcinogenic types, viral load may be a predictor of abnormal cytology.

Most breast cancer risk factors are modest (< 2-fold) making it difficult to rule out bias. One explanation for the modest associations is that breast tumors are heterogeneous. Laboratory and clinical data point to different phenotypes characterized by tumor markers such as p53, HER2neu, estrogen receptor and progesterone receptor status, and cell cycle control genes such as Cyclin E and Cyclin D1. We previously found the relationship between oral contraceptive use and breast cancer risk to be limited to tumors overexpressing Cyclin D1. Using the same study population — a population-based study of young women under age 45 years in New Jersey — we analyzed whether oral contraceptives and other suspected and known risk factors for breast cancer were associated with protein overexpression of Cyclin E in breast cancer tissue measured by immunohistochemistry. Unordered polytomous logistic regression was used to estimate the odds ratios (OR) for two case groups — 1) breast cancer with Cyclin E overexpression (n = 156), and 2) breast cancer without overexpression (n = 179) — compared with 462 population-based controls. Multivariate-adjusted odds ratio (OR) for ever use of oral contraceptives (OCs) was 1.5 (95% CI 1.0-2.4) for cases that overexpressed Cyclin E and 1.0 (0.7-1.4) for those with no overexpression. Among, women who started using OCs at least 10 years prior to the reference date, the OR was increased two-fold for breast cancer with Cyclin E overexpression (2.2, 95%CI 1.1-4.3) but not for breast cancer without Cyclin E overexpression (OR = 1.1, 95% CI 0.6-2.3). These findings were independent of Cyclin D1 protein overexpression. If replicated, these findings suggest that early OC use may be associated with the subset of mammary tumors that overexpress Cyclin E.

Objective: Many women lack knowledge about cervical cancer screening and colposcopy and can experience worry and anxiety during and following these exams. We analyzed the content of the information discussed during consultations in which women were informed of their colposcopy exam results. We also examined the womens perceptions of the discussion and their condition.

Method: Forty-seven results visits were taped, transcribed, and analyzed for content. Following the visit, women were asked to rate their perceived severity and worry about their condition and their rating of the providers explanation.

Results: Nearly all (n = 46) of the providers gave a name or description of the womans diagnosis and all discussed some aspect of the follow-up plan for screening or treatment. Risk factors were rarely discussed: human papillomavirus (HPV) was discussed in 6 visits, smoking was discussed in 3 visits and sexual activity was discussed in 1 visit. Thirty-one women asked questions and 6 brought a friend or family member to the visit. Forty of the clinicians asked if the woman understood their explanation or if she had any questions. Using 10-point Likert scales, women rated the perceived seriousness of their condition (4.3), their perceived worry (4.4), and the quality of the clinicians explanation (9.6).

Conclusion: Overall, the content of these results visits was focused on follow-up screening and treatment plans whereas prevention messages were rarely discussed.

Purpose: To estimate the potential risk of radiation-induced lung cancer mortality from annual CT screening for current male smokers, compared to the potential reduction in lung cancer mortality.

Methods: Radiation risk models from the Japanese atomic bomb survivors were used to estimate the lifetime risk of radiation-induced lung cancer mortality from a decade of annual CT screening for male current smokers starting at age 55 or 65 years. Lung cancer mortality rates for male current smokers (from the Cancer Prevention Study-II) were used to estimate the number of deaths that could be prevented by each decade of screening, assuming a 0-25% reduction in lung cancer mortality.

Results: A decade of screening starting at age 55 or 65 years was estimated to increase lung cancer mortality by 0.4 (range = 0.2-0.8) and 0.2 (range = 0.1-0.3) deaths per 1000 men screened, respectively. The estimated number of lung cancer deaths that could be prevented by a decade of screening starting at age 55 or 65 years was 0-7, and 0-14 per 1000 men screened.

Summary: Despite uncertainty surrounding the mortality benefits from lung CT screening a number of facilities are already offering this service in the US. These estimates suggest that the benefits will outweigh the radiation risks if the mortality reduction from screening is 3% or more, and that the net benefit would be greater for a decade of screening starting at age 65 than at age 55 years.

Purpose: We examined the roles of antihistamine and anti- inflammatory drug use in the development of gliomas considering infections and other inflammatory risk factors.

Methods: Data from The University of Texas M. D. Anderson Cancer Center Harris County Adult Glioma Study were used for this analysis. Included in this analysis are 429 glioma cases and 429 population-based controls frequency-matched on age, race, and gender. Controls were selected using random-digit dialing. Logistic regression models were used to examine the effects of antihistamine and antiinflammatory drug use, as well as other inflammation-associated risk factors among glioma cases and controls. Briefly, each potential covariate was analyzed in univariate models with the outcome. Those significant at the 0.20 level were considered for inclusion in a multivariate model.

Results: The use of antiinflammatory drugs (OR = 0.84, 95%CI: 0.61-1.17) and history of shingles (OR = 0.55, 95%CI: 0.31-0.96) or asthma/allergies (OR = 0.53, 95%CI: 0.37-0.76) appear protective against glioma formation. However, antihistamine use shows a higher risk among cases (OR = 1.92, 95%CI: 1.30-2.82).

Conclusions: This analysis confirms a protective effect of a history of shingles, and is the first, to our knowledge, to report increased risk with the use of antihistamines. There is also modification of the effect of antihistamines by anti-inflammatory use. More detailed analyses of the different antihistamine drugs will be needed to confirm these findings. It would also be of interest to ascertain any differences in risk according to brain tumor histology.

Purpose: New genomic technology has improved the accuracy of estimates of breast cancer recurrence risk. This test has considerable implications for chemotherapy decisions yet little is known about how well women understand the test and its results.

Methods: We interviewed post-treatment female breast cancer patients (n = 144), assessing their health literacy (using REALM) as well as knowledge of and attitudes toward the test. The women had a median age of 58 (range,36-85). Most were Caucasian (80%) and relatively well-educated (50% had a college degree).

Results: After a presentation of information on the recurrence risk test, women with lower health literacy retained less information than women with higher health literacy (75% vs. 90% accurate), t (140) = 4.73,P <.001. Overall, women found verbal comparisons (your risk is higher than the average person) hardest to understand and mixed verbal-percentage formats (7% risk and that is a low risk) easiest to understand. However, there were health literacy-related differences in understandability. Women with lower health literacy stated equal understanding of all risk formats, F(5,135) = 1.43,P >.20. In contrast, women with higher health literacy expressed better understanding of the mixed verbal-percentage format while finding verbal comparisons least understandable, F (5,135) = 17.95,P <.001. Self-efficacy partially mediated this effect (P <.05).

Summary: Differences in perceived understandability of risk communication format were driven entirely by higher health literacy women and their self-efficacy in interpreting risk information. The optimal risk format for women with higher health literacy is not problematic for women with lower health literacy.

While many reports describe quality of life (QOL) among breast cancer survivors, few compare QOL before and after diagnosis. QOL was examined using data from a cohort that included all women (n = 2,762; 83% of eligible) who were residents of Beaver Dam, WI and were aged 43-86 years at the time of a baseline examination during 1988-1990. Participants were re-contacted up to 5 times through 2002 to ascertain QOL using a 4-level overall health question (rated as excellent, good, fair, or poor) and the SF-36. Data on medical and lifestyle factors and demographics were also collected. Of the 130 incident breast cancer cases identified by data linkage with the statewide cancer registry, 43% (n = 56) contributed exam data both prior and subsequent to the diagnosis. QOL scores for cases were compared to scores for women without breast cancer matched on age and exam year. The 4-level health question was not sufficiently discriminating; cases were similarly likely as controls to decline in self-reported health (odds ratio, OR, 1.18, 95% CI 0.59-2.35). However, using the 5level general health question from the SF-36, breast cancer cases were substantially more likely to report declines in self-reported health following their diagnosis (OR 3.64, 95% CI 1.45-9.14). This relation strengthened after adjustment for self-rated health prior to diagnosis (OR 6.62, 95% CI 2.17-20.2). Greater declines in QOL were observed for cases than controls in both mental and physical subscales of the SF-36. This study will allow us to identify breast cancer survivors who have the greatest risk of having lower quality of life after diagnosis. These women might benefit most from interventions to prevent or delay additional morbidity.

Purpose: The aim of the study was to examine the association of BMI and weight change with the risk of adenomas.

Methods: The Insulin Resistance Atherosclerosis Study was a multi-center prospective cohort study conducted among participants with normal or impaired glucose tolerance or non-insulin requiring type-2 diabetes. Weight and height as well as other risk factors were measured at three time points with colonoscopies conducted at the last time point. Participants were stratified by sex and univariate and multivariate analyses were conducted to assess differences between participants with and without an adenoma.

Results: Among the 610 participants who underwent a colonoscopy, the mean age was 64 years and the majority were women (54.6%). Participants with adenomas were more likely to be males (P = 0.017). After adjustment for age, clinic, ethnicity, diabetes status, and smoking status, BMI at time of colonoscopy was associated with adenomas in women (highest vs. lowest tertile BMI OR = 4.51, 95% CI 1.61-12.62) but not in men (OR = 1.23, 95% CI 0.49-3.10). After adjustment for the same variables, women participants who had gained >5 lbs versus those who had lost or maintained weight over ten years had an increased risk of adenomas (OR = 3.88, 95%CI 1.24-6.72). This association was not found in men (OR = 1.47, 95% CI 0.59-3.70). No independent association was detected between ethnicity and risk of adenomas.

Summary: These data indicate that there is an increased risk of colorectal adenomas with increased weight gain and higher BMI in women but not in men.

Purpose: To investigate the impact of FGFR4 G388R polymorphism in the development or progression of bladder cancer.

Methods: 219 bladder cancer patients and 151 controls were recruited from Memorial Sloan-Kettering Cancer Center between October 1993 and June 1997. The PCR-RFLP method was used to assess the genotype of FGFR4, and status of TP53 mutation were detected by a genechip-based method.

Results: Prevalence of Gly/Gly, Gly/Arg, and Arg/Arg genotypes were similar among patients and controls, suggesting that FGFR4 G388R may not be involved in the early development of bladder cancer. We detected no clear correlation between the FGFR4 G388R polymorphism and pathological parameters such as age at diagnosis, tumor stage, grade, or TP53 mutations. Neither TP53 mutations nor FGFR4 G388R showed any evidence of an association with time to disease-specific death. However, the combination of Gly/Gly genotype and TP53 mutations was strongly associated with reduced disease-specific survival time within a 120 months follow-up period (HR = 1.94; P = 0.037).

Conclusion: Our findings suggested that the FGFR4 gene might have a limited role in bladder cancer development. However, combined information of FGFR4 G388R polymorphism and TP53 mutations may serve as a prognostic factor in predicting disease-specific survival for patients with bladder cancer and a potential target for therapeutic strategy.