Jatoi comments on the greater risk of breast cancer death after unilateral mastectomy (ULM) versus breast-conserving therapy (BCT) that we recently reported,1 and he concludes that this finding likely reflects residual confounding. We drew the same conclusion about unmeasured confounders in the selection of surgical treatment both in our recent article1 and in our earlier study of survival after BCT versus ULM and bilateral mastectomy (BLM).2 In both studies, we found no significant difference in survival between BLM and BCT, but we did find worse survival associated with ULM. The characteristics of patients who underwent ULM were different from the characteristics of those who underwent BCT or BLM in both studies on the basis of measurable patient sociodemographic and clinical characteristics, and it is very likely that unmeasured confounders were also present. We certainly agree with Jatoi that the treatment of primary breast cancer should always be predicated on the results of large randomized trials. However, only a small and select percentage of US patients with cancer are treated in clinical trials, and many patients currently treated with ULM or BCT might not have met the entry criteria for the relevant trials. Population-based registry data are thus an essential partner to clinical trial results because they capture long-term health outcomes that represent and are relevant to the underlying population of US patients with cancer.3 Alison J. Canchola reports grants from the National Institutes of Health and the California Department of Public Health during the conduct of the study. Allison W. Kurian reports research funding to her institution from Myriad Genetics outside the submitted work. The other authors made no disclosures.
Abstract Background/Objective: Many Hispanic or Latino/a individuals (hereafter Hispanic) reside in ethnic enclaves. Hispanic enclaves are places characterized by high prevalence of Hispanic individuals, immigrants, individuals who speak Spanish, and/or ethnic-specific businesses. This population-based study of Hispanic women with breast cancer examines how residence in Hispanic enclaves is associated with late stage diagnosis. Methods: We used data from population-based cancer registries in four states (CA, NJ, NY, TX) to identify Hispanic women with breast cancer diagnosed between 2000-2017. Hispanic enclaves were defined using principal components analysis of four census tract-level variables: percent Hispanic, percent foreign-born Hispanic residents, percent with limited English and percent linguistically isolated and speaking Spanish. All census tracts in the four states were classified into quintiles of ethnic enclave score (pooled across all four states). Using log binomial regression with clustering by census tract, we examined associations of enclave residence (in quintiles) on late stage (regional or distant) compared to early stage (in-situ or localized) diagnosis; we fit unadjusted and adjusted models with age, year of diagnosis, insurance type, state, metropolitan/non-metropolitan census tract residence, and census tract percent poverty as covariates. Results: Among 165,226 Hispanic women, 35.1% were diagnosed at late stage. Two-thirds of women resided in census tracts with the highest ethnic enclave scores: 25.3% in Q4 and 43.6% in Q5. The percent late stage varied by ethnic enclave quintile (29.5% in Q1 to 37.9% in Q5; Chi-square across all quintiles p<0.01. Residence in the highest (Q5) compared to lowest (Q1) enclave quintile was associated with increased risk of late-stage diagnosis in the unadjusted model (RR=1.28 95% CI: 1.24-1.33) and in the fully adjusted model (RR=1.11; 95% CI: 1.07-1.15). Conclusions: Enhanced understanding of the features of ethnic enclaves that may contribute to later stage at breast cancer diagnosis among Hispanic women is important to inform future outreach and intervention. Additional analyses will assess potential for moderation by poverty and will examine whether results vary by enclave characteristics such as the predominant ethnic group (e.g., Mexican, Puerto Rican) or enclave transitions between 2000-2010 (never, former, persistent, emerging enclave). Citation Format: Sandi L. Pruitt, Aniruddha B. Rathod, Kathryn L. Shahan, Alison J. Canchola, Francis P. Boscoe, Kevin A. Henry, Robert A. Hiatt, Amy E. Hughes, Katherine Lin, Dan Meltzer, Paulo S. Pinheiro, Antoinette M. Stroup, Hong Zhu, Scarlett L. Gomez, Salma Shariff-Marco. Residence within Hispanic ethnic enclaves is associated with later stage breast cancer diagnosis among Hispanic women in four U.S. states [abstract]. In: Proceedings of the 16th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2023 Sep 29-Oct 2;Orlando, FL. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2023;32(12 Suppl):Abstract nr B093.
Abstract Objectives: We investigated racial and ethnic differences in colorectal cancer (CRC) mortality in adults with early onset colorectal cancer (ages <50; EO-CRC) in California and quantified the contribution of sociodemographic, health system, and clinical factors to racial and ethnic disparities in mortality. Methods: California Cancer Registry data were used to estimate CRC-specific mortality for adults ages 18-49 with EO-CRC between 2000-2019 for each racial/ethnic group (non-Hispanic Black [NHB], Hispanic, Asian American, American Indian/Alaska Native [AIAN], Native Hawaiian/Pacific Islander [NHPI]) compared with non-Hispanic White (NHW) adults. Additional disaggregated comparisons were conducted in Asian American ethnic groups: Chinese, Filipino, Japanese, Korean, South Asian, Southeast Asian, Other Asian and NHPI, compared to NHW individuals and separately using Chinese adults as a referent group. Cox proportional hazards models were used to measure association between race/ethnicity and CRC mortality risk, yielding unadjusted hazard ratios (HR) and adjusted hazard ratios (aHR) accounting for sociodemographic, clinical and health system factors, with corresponding 95% confidence intervals (95% CI). Mediation analyses were conducted to measure the percentage of contribution of factors to overall racial/ethnic CRC mortality rates using a sequence of multivariable Cox models. Results: There were 22,997 adults ages 18-49 with an EO-CRC diagnosis between 2000-2019, with 3,544 Asian (1,039 deaths), 6,889 Hispanic (1,998 deaths), 125 AIAN (36 deaths), 1,668 NHB (670 deaths), 10,473 NHW (3,089 deaths), 135 NHPI (51 deaths) and 163 (<5 deaths) unknown race/ethnicity adults. Compared to NHW adults, higher CRC mortality was shown for NHPI adults (HR=1.69; 95% CI: 1.27-2.23; aHR=1.36; 95% CI: 1.03-1.79) and NHB adults (HR=1.53; 95% CI: 1.41-1.66; aHR=1.17; 95% CI: 1.07-1.29). Hispanic adults had an increased risk of CRC mortality in the unadjusted model (HR=1.15, 95% CI: 1.09-1.22) that was attenuated in the adjusted model (aHR=0.99, 95% CI: 0.92-1.04), compared to NHW adults. After disaggregating the Asian American group, Southeast Asian adults had increased CRC mortality risk in the unadjusted (HR=1.29, 95% CI: 1.13-1.46) but not adjusted model (aHR=1.10; 95% CI: 0.97-1.26), compared to NHW adults. In mediation analyses that examined contributors to the association between race/ethnicity and CRC mortality risk, neighborhood socioeconomic status and insurance status were among the top five most influential factors for Hispanic, NHB, and NHPI adults. However, the magnitude of influence differed by racial and ethnic group. Conclusions: Compared to NHWs, NHPI and NHB adults with EO-CRC had increased CRC mortality risk, even after adjusting for clinical and sociodemographic factors. Mediation analyses highlighted variation in the importance of various sociodemographic, clinical, health system and neighborhood factors, underscoring the need to better disentangle factors within racial and ethnic groups that contribute to disparities in CRC-related mortality. Citation Format: Joshua Demb, Scarlett Gomez, Alison Canchola, Alexander Qian, James D. Murphy, Samir Gupta, Maria Elena Martinez. Contribution of neighborhood and clinical factors to differences in early-onset colorectal cancer mortality across different racial and ethnic groups [abstract]. In: Proceedings of the 15th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2022 Sep 16-19; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2022;31(1 Suppl):Abstract nr A074.
The development of endometrial cancer is largely related to prolonged exposure to unopposed estrogens. Phytoestrogens (i.e., weak estrogens found in plant foods) may have antiestrogenic effects. We evaluated the associations between dietary intake of seven specific compounds representing three classes of phytoestrogens (isoflavones, coumestans, and lignans) and the risk of endometrial cancer.In a case-control study from the greater San Francisco Bay Area, we collected dietary information from 500 African American, Latina, and white women aged 35-79 years who were diagnosed with endometrial cancer between 1996 and 1999 and from 470 age- and ethnicity-matched control women identified through random-digit dialing. Unconditional logistic regression analyses were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs).Isoflavone (OR = 0.59, 95% CI = 0.37 to 0.93 for the highest versus lowest quartile of exposure) and lignan (OR = 0.68, 95% CI = 0.44 to 1.1) consumptions were inversely related to the risk of endometrial cancer. These associations were slightly stronger in postmenopausal women (OR = 0.44, 95% CI = 0.26 to 0.77 and OR = 0.57, 95% CI = 0.34 to 0.97 for isoflavones and lignans, respectively). Obese postmenopausal women consuming relatively low amounts of phytoestrogens had the highest risk of endometrial cancer (OR = 6.9, 95% CI = 3.3 to 14.5 compared with non-obese postmenopausal women consuming relatively high amounts of isoflavones); however, the interaction between obesity and phytoestrogen intake was not statistically significant.Some phytoestrogenic compounds, at the levels consumed in the typical American-style diet, are associated with reduced risk of endometrial cancer.
Abstract – Objective: To determine if medical clinicians are as accurate as dental clinicians in recognizing diagnostic characteristics of HIV‐related oral lesions. Methods: In 355 HIV‐infected participants at five Women’s Interagency HIV Study sites, we paired oral examinations conducted within 7 days of each other by dental and medical clinicians. We used the former as a gold standard against which to evaluate the accuracy of the latter. We assessed the accuracy of the medical clinicians’ findings based both on their observations of abnormalities and on their descriptions of these abnormalities. Results: Dental clinicians diagnosed some oral abnormality in 38% of participants. When “abnormality” was used as the medical clinicians’ outcome, sensitivities were 75% for pseudomembranous candidiasis and 58% for erythematous candidiasis, but only 40% for hairy leukoplakia. When a precise description of the abnormality was used as their outcome, sensitivities were 19%, 12% and 20%, respectively. Conclusions: Medical clinicians recognize that HIV‐related oral abnormalities are present in 40–75% of cases, but less often describe them accurately. Low sensitivity implies that the true associations of specific oral lesions with other HIV phenomena, such as time until AIDS, must be stronger than the literature suggests.
Data are limited on effective methods for recruiting persons, especially from ethnically diverse populations, into population-based studies. The goal of this study was to evaluate the variation among and representativeness of controls identified using multiple methods for a population-based case-control study of breast cancer among Asian Americans, Native Hawaiians and Pacific Islanders (AANHPIs) in the San Francisco Bay Area.We used a unique combination of targeted recruitment strategies, including address-based sampling, community-based methods, and internet-based and media-based approaches for recruiting controls, frequency matched on age and ethnicity to a population-based sample of cases. To characterise the participating controls, we compared the distribution of sociodemographic characteristics and cancer risk factors between recruitment sources using χ(2) tests. To ensure that the controls we recruited were representative of the underlying at-risk population, we compared characteristics of the controls, by ethnicity and in aggregate, to data from the California Health Interview Survey (CHIS), and adjusted the relative mix of recruitment strategies throughout the study as needed to achieve representativeness.As expected, controls (n=483) recruited by any single method were not representative. However, when aggregated across methods, controls were largely representative of the underlying source population, as characterised by CHIS, with regard to the characteristics under study, including nativity, education, marital status and body mass index.A multimode approach using targeted recruitment strategies is an effective and feasible alternative to using a single recruitment method in identifying a representative, diverse control sample for population-based studies.
Abstract Background: Cancer patients who are married at diagnosis have lower mortality than the unmarried. Although this effect has been attributed to increased social support among married patients, whether economic resources influence this association remains unclear. Purpose: We assessed whether overall mortality differences between married and unmarried cancer patients is modified by neighborhood socioeconomic status (nSES) and mediated by health insurance status. Methods: We studied patients newly diagnosed (first invasive primary) with one of the 10 most common causes of cancer deaths from 2000 through 2009 in California. Information on patient nSES (based on block-group- level Census 2000 or 2007-2011 American Community Survey data), insurance (primary and secondary payer source), demographic and tumor characteristics, and follow-up through 2012 were obtained from the California Cancer Registry. Using Cox proportional hazards regression, we estimated overall mortality [hazard ratio (HR)] associated with marital status among N = 377,932 males (194,216 deaths) and N = 378,447 females (175,414 deaths), stratified on stage and adjusting for age, race/ethnicity, cancer site, nSES, insurance status, and treatment. Results: Prior to adjustment for insurance status, unmarried patients had higher overall mortality than married patients [HR (males) = 1.28 (1.27-1.29), HR (females) = 1.20 (1.19-1.21)]. This association was marginally stronger among patients from higher SES neighborhoods [HR (males) = 1.30 (1.28-1.32), HR (females) = 1.21 (1.20-1.23)] than from lower SES neighborhoods [HR (males) = 1.27 (1.26-1.29), HR (females) = 1.19 (1.18-1.21)] and only slightly lower after adjustment for insurance status. The magnitude of the associations varied by race/ethnicity and cancer site, with the largest marital status effect sizes seen for: high SES non-Hispanic White males with non-Hodgkin lymphoma [1.62 (1.52-1.73)], low SES Black males with pancreatic cancer [1.40 (1.20-1.63)], high SES Hispanic males with prostate cancer [1.50 (1.33-1.69)], high SES Asian males with pancreatic cancer [1.57 (1.28-1.92)], high SES non-Hispanic White females with breast cancer [1.34 (1.29-1.38)] or NHL [1.34 (1.24-1.44)], low SES Black females with NHL [1.43 (1.13-1.81)], and high SES Hispanic [1.44 (1.12-1.86)] and Asian [1.59 (1.24-2.05)] females with leukemia. The largest attenuation of HRs after adjustment for insurance was seen among Blacks, regardless of nSES [males: from 1.26 (1.22-1.30) to 1.20 (1.16-1.24), females: from 1.15 (1.11-1.20) to 1.10 (1.06-1.15)]. Conclusions: Neighborhood SES and insurance status had no considerable impact on the association between marital status and mortality after cancer diagnosis. Citation Format: Scarlett L. Gomez, Alison Canchola, Susan Hurley, Christina A. Clarke, Iona Cheng, Theresa H. M. Keegan, Sally L. Glaser, Maria Elena Martinez. Lower mortality among married cancer patients: How much of the effect is explained by socioeconomic and health insurance status. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 885. doi:10.1158/1538-7445.AM2015-885
