Comparative Validation of Breast Cancer Risk Prediction Models and Projections for Future Risk Stratification

BACKGROUND

External validation of risk models is critical for risk-stratified breast cancer prevention. We used the Individualized Coherent Absolute Risk Estimation (iCARE) as a flexible tool for risk model development and comparative model validation and to make projections for population risk stratification.

METHODS

Performance of two recently developed models, one based on the Breast and Prostate Cancer Cohort Consortium analysis (iCARE-BPC3) and another based on a literature review (iCARE-Lit), were compared with two established models (Breast Cancer Risk Assessment Tool and International Breast Cancer Intervention Study Model) based on classical risk factors in a UK-based cohort of 64 874 white non-Hispanic women (863 patients) age 35-74 years. Risk projections in a target population of US white non-Hispanic women age 50-70 years assessed potential improvements in risk stratification by adding mammographic breast density (MD) and polygenic risk score (PRS).

RESULTS

The best calibrated models were iCARE-Lit (expected to observed number of cases [E/O] = 0.98, 95% confidence interval [CI] = 0.87 to 1.11) for women younger than 50 years, and iCARE-BPC3 (E/O = 1.00, 95% CI = 0.93 to 1.09) for women 50 years or older. Risk projections using iCARE-BPC3 indicated classical risk factors can identify approximately 500 000 women at moderate to high risk (>3% 5-year risk) in the target population. Addition of MD and a 313-variant PRS is expected to increase this number to approximately 3.5 million women, and among them, approximately 153 000 are expected to develop invasive breast cancer within 5 years.

CONCLUSIONS

iCARE models based on classical risk factors perform similarly to or better than BCRAT or IBIS in white non-Hispanic women. Addition of MD and PRS can lead to substantial improvements in risk stratification. However, these integrated models require independent prospective validation before broad clinical applications.

Robustness of risk-based allocation of resources for disease prevention

Risk models for disease incidence can be useful for allocating resources for disease prevention if risk assessment is not too expensive. Assume there is a preventive intervention that should be given to everyone, but preventive resources are limited. We optimize risk-based prevention strategies and investigate robustness to modeling assumptions. The optimal strategy defines the proportion of the population to be given risk assessment and who should be offered intervention. The optimal strategy depends on the ratio of available resources to resources needed to intervene on everyone, and on the ratio of the costs of risk assessment to intervention. Risk assessment is not recommended if it is too expensive. Preventive efficiency decreases with decreasing compliance to risk assessment or intervention. Risk measurement error has little effect nor does misspecification of the risk distribution. Ignoring population substructure has small effects on optimal prevention strategy but can lead to modest over- or under-spending. We give conditions under which ignoring population substructure has no effect on optimal strategy. Thus, a simple one-population model offers robust guidance on prevention strategy but requires data on available resources, costs of risk assessment and intervention, population risk distribution, and probabilities of acceptance of risk assessment and intervention.

Confidence intervals for the difference between two relative risks

We provide methods to estimate the confidence interval for the difference between two relative risks. Letting p₀, p₁, and p₂ be the probabilities of an event in three groups (i.e. control, treatment 1, treatment 2), our methods estimate a confidence interval for r = p₁/p₀ - p₂/p₀. We highlight that our methods can handle small sample sizes, covariates, and study populations from multiple strata. We specifically developed these methods for vaccine trials to estimate the difference between two vaccine efficacies, where VE₁ = 1 - p₁/p₀, VE₂ = 1 - p₂/p₀ and r = VE₂ - VE₁. We showcase our methods by using interim data from one of these trials to suggest that one dose of the human papillomavirus vaccine may be as efficacious as two doses of the vaccine.

Abstract A39: The human oral microbiota and risk of lung cancer: An analysis of three prospective cohort studies

Background: The oral microbiota may be associated with lung cancer risk through direct mechanisms, including infection, immune responses, and periodontal disease, and through indirect mechanisms such as the modification of the oral microbiota by tobacco. We conducted a case-cohort study nested within three US prospective cohort studies to evaluate the association between oral microbiota ascertained years before a cancer diagnosis and risk of lung cancer.

Methods: Incident lung cancer cases within the Agricultural Health Study (AHS; N=244), NIH-AARP Diet and Health Study (N=376), and the Prostate, Lung, Colorectum, and Ovarian Cancer Screening Trial (PLCO; N=700) who provided an oral wash sample were identified. The median time between oral sample collection and diagnosis was approximately 6.6 years, 3.4 years, and 4.5 years for AHS, NIH-AARP, and PLCO, respectively. A referent subcohort was randomly selected by strata of age, sex, and cigarette smoking history. We extracted DNA using the DSP DNA Virus Pathogen kit, and the V4 region of the 16S rRNA gene was PCR amplified and sequenced using the MiSeq. The sequencing data were processed using QIIME2 with the DADA2 plugin and we generated alpha and beta diversity metrics. Cox proportional hazards models were used to evaluate the hazard ratios (HR) and 95% confidence intervals (CI) for the association between the oral microbial measures and the risk of lung cancer with adjustment for known lung cancer risk factors, and estimates from the three cohorts were meta-analyzed.

Results: Increased alpha diversity was associated with decreased lung cancer risk, although only the association with the Shannon Index reached statistical significance (HR for 5th quintile versus 1st quintile 0.74; 95% CI 0.60, 0.92) with no evidence of between-study heterogeneity (p = 0.5968). Specific principal coordinate vectors from the beta diversity matrices were also significantly associated with lung cancer risk, suggesting differing bacterial communities between future lung cancer cases. When stratified by histologic subtypes, the inverse association with alpha diversity was restricted to squamous cell carcinoma, with all alpha diversity metrics reaching statistical significance (e.g., Faith’s phylogenetic diversity HR for 5th quintile versus 1st quintile 0.57; 95% CI: 0.37, 0.87). Similarly, when stratified by smoking history, the inverse association with alpha diversity was restricted to former smokers (e.g., observed species HR for 5th quintile versus 1st quintile 0.63; 95% CI: 0.44, 0.89).

Conclusions: In oral wash samples collected years before diagnosis, we found significant associations between both alpha and beta diversity metrics of the oral microbial communities and risk of lung cancer. Additional work is required to understand the associations by histologic subtype and smoking history.

Citation Format: Emily Vogtmann, Xing Hua, Guoqin Yu, Autumn Hullings, Yunhu Wan, Casey L Dagnall, Kristine Jones, Belynda D. Hicks, Amy Hutchinson, Shalabh Suman, Bin Zhu, Barry Graubard, Mitchell H. Gail, J. Gregory Caporaso, William Wheeler, Dale Sandler, Laura E. Beane Freeman, Linda Liao, Neal D. Freedman, Neil Caporaso, Rashmi Sinha, Jianxin Shi, Christian C Abnet. The human oral microbiota and risk of lung cancer: An analysis of three prospective cohort studies [abstract]. In: Proceedings of the AACR Special Conference on the Microbiome, Viruses, and Cancer; 2020 Feb 21-24; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2020;80(8 Suppl):Abstract nr A39.

The Need for Combined Assessment of Multiple Outcomes in Noninferiority Trials in Oncology

Noninferiority trials in oncology assess novel therapies with the potential for slightly worse recurrence or death outcomes (ie, the margin of noninferiority) than standard therapies. This poses a dilemma because, in the absence of potential health outcome advantages, these trials may not provide the treatment equipoise required for an ethical study. Any new treatment with the potential for slightly worse recurrence or death outcomes should have countervailing health outcome advantages, but these are rarely taken into account in the design of noninferiority trials. This article presents the argument that not only the potentially worse health outcomes but also the potential benefits of the novel therapy should be considered when designing, analyzing, and reporting noninferiority trials. Some approaches to study design and analysis that consider both primary and secondary end points are discussed, and reporting the joint distributions of end points for the novel and standard treatments is recommended.

Design choices for observational studies of the effect of exposure on disease incidence

The purpose of this paper is to help readers choose an appropriate observational study design for measuring an association between an exposure and disease incidence. We discuss cohort studies, sub-samples from cohorts (case-cohort and nested case-control designs), and population-based or hospital-based case-control studies. Appropriate study design is the foundation of a scientifically valid observational study. Mistakes in design are often irremediable. Key steps are understanding the scientific aims of the study and what is required to achieve them. Some designs will not yield the information required to realise the aims. The choice of design also depends on the availability of source populations and resources. Choosing an appropriate design requires balancing the pros and cons of various designs in view of study aims and practical constraints. We compare various cohort and case-control designs to estimate the effect of an exposure on disease incidence and mention how certain design features can reduce threats to study validity.

On Mendelian randomization analysis of case-control study

Mendelian randomization (MR) analysis uses genotypes as instruments to estimate the causal effect of an exposure in the presence of unobserved confounders. The existing MR methods focus on the data generated from prospective cohort studies. We develop a procedure for studying binary outcomes under a case-control design. The proposed procedure is built upon two working models commonly used for MR analyses and adopts a quasi-empirical likelihood framework to address the ascertainment bias from case-control sampling. We derive various approaches for estimating the causal effect and hypothesis testing under the empirical likelihood framework. We conduct extensive simulation studies to evaluate the proposed methods. We find that the proposed empirical likelihood estimate is less biased than the existing estimates. Among all the approaches considered, the Lagrange multiplier (LM) test has the highest power, and the confidence intervals derived from the LM test have the most accurate coverage. We illustrate the use of our method in MR analysis of prostate cancer case-control data with vitamin D level as exposure and three single nucleotide polymorphisms as instruments.

Using standard microbiome reference groups to simplify beta-diversity analyses and facilitate independent validation

Motivation

Comparisons of microbiome communities across populations are often based on pairwise distance measures (beta-diversity). Standard analyses (principal coordinate plots, permutation tests, kernel methods) require access to primary data if another investigator wants to add or compare independent data. We propose using standard reference measurements to simplify microbiome beta-diversity analyses, to make them more transparent, and to facilitate independent validation and comparisons across studies.

Results

Using stool and nasal reference sets from the Human Microbiome Project (HMP), we computed mean distances (actually Bray-Curtis or Pearson correlation dissimilarities) to each reference set for each new sample. Thus, each new sample has two mean distances that can be plotted and analyzed with classical statistical methods. To test the approach, we studied independent (not reference) HMP subjects. Simple Hotelling tests demonstrated statistically significant differences in mean distances to reference sets between all pairs of body sites (stool, skin, nasal, saliva and vagina) at the phylum, class, order, family and genus levels. Using the distance to a single reference set was usually sufficient, but using both reference sets always worked well. The use of reference sets simplifies standard analyses of beta-diversity and facilitates the independent validation and combining of such data because others can compute distances to the same reference sets. Moreover, standard statistical methods for survival analysis, logistic regression and other procedures can be applied to vectors of mean distances to reference sets, thereby greatly expanding the potential uses of beta-diversity information. More work is needed to identify the best reference sets for particular applications.

Availability and implementation

https://github.com/NCI-biostats/microbiome-fixed-reference.

Supplementary information

Supplementary data are available at Bioinformatics online.

Risk-Reducing Salpingo-Oophorectomy and Breast Cancer Risk Reduction in the Gynecologic Oncology Group Protocol-0199 (GOG-0199)

Background

Risk-reducing salpingo-oophorectomy (RRSO) has been associated with approximately 50% breast cancer risk reduction among women with a pathogenic variant in BRCA1 or BRCA2 (BRCA1/2), a finding that has recently been questioned.

Methods

We estimated incidence rates of breast cancer and all cancers combined during 5 years of follow-up among participants selecting RRSO or ovarian cancer screening (OCS) among women with a BRCA1/2 pathogenic variant or strong breast and/or ovarian cancer family history. Ovarian or fallopian tube or peritoneal cancer incidence rates were estimated for the OCS group. Breast cancer hazard ratios (HRs) for time-dependent RRSO were estimated using Cox regression with age time-scale (4943 and 4990 women-years in RRSO and OCS cohorts, respectively). All statistical tests were two-sided.

Results

The RRSO cohort included 925 participants, and 1453 participants were in the OCS cohort (381 underwent RRSO during follow-up), with 88 incident breast cancers diagnosed. Among BRCA1/2 pathogenic variant carriers, a non-statistically significant lower breast cancer incidence was observed in the RRSO compared with the OCS cohort (HR = 0.86, 95% confidence interval  = 0.45 to 1.67; P = .67). No difference was observed in the overall population or among subgroups stratified by prior breast cancer history or menopausal status. Seven fallopian tube and four ovarian cancers were prospectively diagnosed in the OCS cohort, and one primary peritoneal carcinoma occurred in the RRSO cohort.

Conclusions

These data suggest that RRSO might be associated with reduced breast cancer incidence among women with a BRCA1/2 pathogenic variant, although the effect, if present, is small. This evolving evidence warrants a thorough discussion regarding the impact of RRSO on breast cancer risk with women considering this intervention.

Effects of Nutrition Intervention on Total and Cancer Mortality: 25-Year Post-trial Follow-up of the 5.25-Year Linxian Nutrition Intervention Trial

Background

A beneficial effect of supplementation with selenium, vitamin E, and beta-carotene was observed on total and cancer mortality in a Chinese population, and it endured for 10 years postintervention, but longer durability is unknown.

Methods

A randomized, double-blind, placebo-controlled trial was conducted in Linxian, China, from 1986 to 1991; 29 584 residents age 40 to 69 years received daily supplementations based on a factorial design: Factors A (retinol/zinc), B (riboflavin/niacin), C (vitamin C/molybdenum), and/or D (selenium/vitamin E/beta-carotene), or placebo for 5.25 years, and followed for up 25 years. Hazard ratios (HRs) and 95% confidence intervals (CIs) for the intervention effects on mortalities were estimated using Cox proportional hazards models.

Results

Through 2016, the interventions showed no effect on total mortality. The previously reported protective effect of Factor D against total mortality was lost 10 years postintervention. The protective effect of Factor D for gastric cancer was attenuated (HR = 0.93, 95% CI = 0.85 to 1.01), but a newly apparent protective effect against esophageal cancer was found for Factor B (HR = 0.92, 95% CI = 0.85 to 1.00, two-sided P = .04). Other protective/adverse associations were observed for cause-specific mortalities. Protective effects were found in people younger than age 55 years at baseline against non-upper gastrointestinal cancer death for Factor A (HR = 0.80, 95% CI = 0.69 to 0.92) and against death from stroke for Factor C (HR = 0.89, 95% CI = 0.82 to 0.96). In contrast, increased risk of esophageal cancer was found when the intervention began after age 55 years for Factors C (HR = 1.16, 95% CI = 1.04 to 1.30) and D (HR = 1.20, 95% CI = 1.07 to 1.34).

Conclusions

Multiyear nutrition intervention is unlikely to have a meaningful effect on mortality more than a decade after supplementation ends, even in a nutritionally deprived population. Whether sustained or repeat intervention would provide longer effects needs further investigation.

Breast Cancer Risk Model Requirements for Counseling, Prevention, and Screening

Background

Incorporation of polygenic risk scores and mammographic density into models to predict breast cancer incidence can increase discriminatory accuracy (area under the receiver operating characteristic curve [AUC]) from 0.6 for models based only on epidemiologic factors to 0.7. It is timely to assess what impact these improvements will have on individual counseling and on public health prevention and screening strategies, and to determine what further improvements are needed.

Methods

We studied various clinical and public health applications using a log-normal distribution of risk.

Results

Provided they are well calibrated, even risk models with AUCs of 0.6 to 0.7 provide useful perspective for individual counseling and for weighing the harms and benefits of preventive interventions in the clinic. At the population level, they are helpful for designing preventive intervention trials, for assessing reductions in absolute risk from reducing exposure to modifiable risk factors, and for resource allocation (although a higher AUC would be desirable for risk-based allocation). Other public health applications require higher AUCs that can only be achieved with risk predictors 1.6 to 8.8 times as strong as all those yet discovered combined. Such applications are preventing an appreciable proportion of population disease when employing a high-risk prevention strategy and deciding who should be screened for subclinical disease.

Conclusions

Current and foreseeable risk models are useful for counseling and some prevention activities, but given the daunting challenge of achieving, for example, an AUC of 0.8, considerable effort should be put into finding effective preventive interventions and screening strategies with fewer adverse effects.

Perspectives for Consideration in the Development of Microbial Cell Reference Materials

Microbiome measurement and analyses benefit greatly from incorporation of reference materials as controls. However, there are many points to consider in defining an ideal whole-cell reference material standard. Such a standard would embody all the diversity and measurement challenges present in real samples, would be completely characterized to provide "ground truth" data, and would be inexpensive and widely available. This ideal is, unfortunately, not readily attainable because of the diverse nature of different sequencing projects. Some applications may benefit most from highly complex reference materials, while others will value characterization or low expense more highly. The selection of appropriate microbial whole-cell reference materials to benchmark and validate microbial measurements should be considered carefully and may vary among specific applications. In this article, we describe a perspective on the development of whole-cell microbial reference materials for use in metagenomics analyses.

Effects of Helicobacter pylori treatment and vitamin and garlic supplementation on gastric cancer incidence and mortality: follow-up of a randomized intervention trial

OBJECTIVE

To assess the effects of Helicobacter pylori treatment, vitamin supplementation, and garlic supplementation in the prevention of gastric cancer.

DESIGN

Blinded randomized placebo controlled trial.

SETTING

Linqu County, Shandong province, China.

PARTICIPANTS

3365 residents of a high risk region for gastric cancer. 2258 participants seropositive for antibodies to H pylori were randomly assigned to H pylori treatment, vitamin supplementation, garlic supplementation, or their placebos in a 2×2×2 factorial design, and 1107 H pylori seronegative participants were randomly assigned to vitamin supplementation, garlic supplementation, or their placebos in a 2×2 factorial design.

INTERVENTIONS

H pylori treatment with amoxicillin and omeprazole for two weeks; vitamin (C, E, and selenium) and garlic (extract and oil) supplementation for 7.3 years (1995-2003).

MAIN OUTCOME MEASURES

Primary outcomes were cumulative incidence of gastric cancer identified through scheduled gastroscopies and active clinical follow-up through 2017, and deaths due to gastric cancer ascertained from death certificates and hospital records. Secondary outcomes were associations with other cause specific deaths, including cancers or cardiovascular disease.

RESULTS

151 incident cases of gastric cancer and 94 deaths from gastric cancer were identified during 1995-2017. A protective effect of H pylori treatment on gastric cancer incidence persisted 22 years post-intervention (odds ratio 0.48, 95% confidence interval 0.32 to 0.71). Incidence decreased significantly with vitamin supplementation but not with garlic supplementation (0.64, 0.46 to 0.91 and 0.81, 0.57 to 1.13, respectively). All three interventions showed significant reductions in gastric cancer mortality: fully adjusted hazard ratio for H pylori treatment was 0.62 (95% confidence interval 0.39 to 0.99), for vitamin supplementation was 0.48 (0.31 to 0.75), and for garlic supplementation was 0.66 (0.43 to 1.00). Effects of H pylori treatment on both gastric cancer incidence and mortality and of vitamin supplementation on gastric cancer mortality appeared early, but the effects of vitamin supplementation on gastric cancer incidence and of garlic supplementation only appeared later. No statistically significant associations were found between interventions and other cancers or cardiovascular disease.

CONCLUSIONS

H pylori treatment for two weeks and vitamin or garlic supplementation for seven years were associated with a statistically significant reduced risk of death due to gastric cancer for more than 22 years. H pylori treatment and vitamin supplementation were also associated with a statistically significantly reduced incidence of gastric cancer.

TRIAL REGISTRATION

ClinicalTrials.gov NCT00339768.

Estimating the decision curve and its precision from three study designs

The decision curve plots the net benefit ( N B ) of a risk model for making decisions over a range of risk thresholds, corresponding to different ratios of misclassification costs. We discuss three methods to estimate the decision curve, together with corresponding methods of inference and methods to compare two risk models at a given risk threshold. One method uses risks (R) and a binary event indicator (Y) on the entire validation cohort. This method makes no assumptions on how well-calibrated the risk model is nor on the incidence of disease in the population and is comparatively robust to model miscalibration. If one assumes that the model is well-calibrated, one can compute a much more precise estimate of N B based on risks R alone. However, if the risk model is miscalibrated, serious bias can result. Case-control data can also be used to estimate N B if the incidence (or prevalence) of the event ( Y = 1 ) is known. This strategy has comparable efficiency to using the full ( R , Y ) data, and its efficiency is only modestly less than that for the full ( R , Y ) data if the incidence is estimated from the mean of Y. We estimate variances using influence functions and propose a bootstrap procedure to obtain simultaneous confidence bands around the decision curve for a range of thresholds. The influence function approach to estimate variances can also be applied to cohorts derived from complex survey samples instead of simple random samples.

Prenatal diethylstilbestrol exposure and cancer risk in women

In the Diethylstilbestrol [DES] Combined Cohort Follow-up, the age- and calendar-year specific standardized incidence ratio [SIR] for clear cell adenocarcinoma [CCA] was 27.6 (95% confidence interval [CI] 7.51-70.6) for the exposed women. The SIR for breast cancer was 1.17 (95% CI 1.01-1.36) and the hazard ratio [HR] adjusted for birth year and cohort for comparison with the unexposed was 1.05 (95% CI 0.79-1.41). The SIR for pancreatic cancer was 2.43 (95% CI 1.21-4.34) and the adjusted HR for comparison with unexposed women was 7.16 (95% CI 0.84-61.5). There was little evidence of excess risk for other sites. There appeared to be a deficit in risk for endometrial cancer among the exposed (SIR 0.61; 95% CI 0.35-0.98), and an excess in the unexposed (SIR 1.55; 95% CI 0.95-2.40); the adjusted HR was 0.45 (95% CI 0.22-0.93) for the internal comparison. There was no overall excess cancer risk in exposed women compared with general population rates (1.06; 95% CI 0.95-1.17) or with unexposed participants (adjusted HR 1.03; 95% CI 0.84-1.25). These data do not support the suggestion that there is a diathesis of cancers in DES exposed female offspring The excess risk of breast and pancreatic cancers that we observed is concerning and warrants continued follow-up and mechanistic investigation. Environ. Mol. Mutagen. 60:395-403, 2019. Published 2017. This article is a US Government work and is in the public domain in the USA.

Design and analysis considerations for combining data from multiple biomarker studies

Pooling data from multiple studies improves estimation of exposure-disease associations through increased sample size. However, biomarker exposure measurements can vary substantially across laboratories and often require calibration to a reference assay prior to pooling. We develop two statistical methods for aggregating biomarker data from multiple studies: the full calibration method and the internalized method. The full calibration method calibrates all biomarker measurements regardless of the availability of reference laboratory measurements while the internalized method calibrates only non-reference laboratory measurements. We compare the performance of these two aggregation methods to two-stage methods. Furthermore, we compare the aggregated and two-stage methods when estimating the calibration curve from controls only or from a random sample of individuals from the study cohort. Our findings include the following: (1) Under random sampling for calibration, exposure effect estimates from the internalized method have a smaller mean squared error than those from the full calibration method. (2) Under the controls-only calibration design, the full calibration method yields effect estimates with the least bias. (3) The two-stage approaches produce average effect estimates that are similar to the full calibration method under a controls only calibration design and the internalized method under a random sample calibration design. We illustrate the methods in an application evaluating the relationship between circulating vitamin D levels and stroke risk in a pooling project of cohort studies.

Next steps in studying the human microbiome and health in prospective studies, Bethesda, MD, May 16-17, 2017

The National Cancer Institute (NCI) sponsored a 2-day workshop, "Next Steps in Studying the Human Microbiome and Health in Prospective Studies," in Bethesda, Maryland, May 16-17, 2017. The workshop brought together researchers in the field to discuss the challenges of conducting microbiome studies, including study design, collection and processing of samples, bioinformatics and statistical methods, publishing results, and ensuring reproducibility of published results. The presenters emphasized the great potential of microbiome research in understanding the etiology of cancer. This report summarizes the workshop and presents practical suggestions for conducting microbiome studies, from workshop presenters, moderators, and participants.

Circulating 25-hydroxyvitamin D up to 3 decades prior to diagnosis in relation to overall and organ-specific cancer survival

While vitamin D has been associated with improved overall cancer survival in some investigations, few have prospectively evaluated organ-specific survival. We examined the accepted biomarker of vitamin D status, serum 25-hydroxyvitamin D [25(OH)D], and cancer survival in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study. Of 4616 cancer cases with measured serum 25(OH)D, 2884 died of their cancer during 28 years of follow-up and 1732 survived or died of other causes. Proportional hazards regression estimated hazard ratios (HR) and 95% confidence intervals (CI) for the association between pre-diagnostic 25(OH)D and overall and site-specific survival. Serum 25(OH)D was significantly lower among cases who subsequently died from their malignancy compared with those who did not (medians 34.7 vs. 36.5 nmol/L, respectively; p = 0.01). Higher 25(OH)D was associated with lower overall cancer mortality (HR = 0.76, 95% CI 0.67-0.85 for highest vs. lowest quintile, p-trend < 0.0001). Higher 25(OH)D was related to lower mortality from the following site-specific malignancies: prostate (HR = 0.74, 95% CI 0.55-1.01, p-trend = 0.005), kidney (HR = 0.59, 95% CI 0.35-0.98, p-trend = 0.28), and melanoma (HR = 0.39, 95% CI 0.20-0.78, p-trend = 0.01), but increased mortality from lung cancer (HR = 1.28, 95% CI 1.02-1.61, p-trend = 0.19). Improved survival was also suggested for head and neck, gastric, pancreatic, and liver cancers, though not statistically significantly, and case numbers for the latter two organ sites were small. Higher 25(OH)D status years prior to diagnosis was related to improved survival for overall and some site-specific cancers, associations that should be examined in other prospective populations that include women and other racial-ethnic groups.

Proportion of U.S. Trends in Breast Cancer Incidence Attributable to Long-term Changes in Risk Factor Distributions

Background: U.S. breast cancer incidence has been changing, as have distributions of risk factors, including body mass index (BMI), age at menarche, age at first live birth, and number of live births.Methods: Using data for U.S. women from large nationally representative surveys, we estimated risk factor distributions from 1980 to 2008. To estimate ecologic associations with breast cancer incidence, we fitted Poisson models to age- and calendar year-specific incidence data from the NCI's Surveillance, Epidemiology and End Results registries from 1980 to 2011. We then assessed the proportion of incidence attributable to specific risk factors by comparing incidence from models that only included age and calendar period as predictors with models that additionally included age- and cohort-specific categorized mean risk factors. Analyses were stratified by age and race.Results: Ecologic associations usually agreed with previous findings from analytic epidemiology. From 1980 to 2011, compared with the risk factor reference level, increased BMI was associated with 7.6% decreased incidence in women ages 40 to 44 and 2.6% increased incidence for women ages 55 to 59. Fewer births were associated with 22.2% and 3.99% increased incidence in women ages 40 to 44 and 55 to 59 years, respectively. Changes in age at menarche and age at first live birth in parous women did not significantly impact population incidence from 1980 to 2011.Conclusions: Changes in BMI and number of births since 1980 significantly impacted U.S. breast cancer incidence.Impact: Quantifying long-term impact of risk factor trends on incidence is important to understand the future breast cancer burden and inform prevention efforts. Cancer Epidemiol Biomarkers Prev; 27(10); 1214-22. ©2018 AACR.

Evidence for single-dose protection by the bivalent HPV vaccine-Review of the Costa Rica HPV vaccine trial and future research studies

The Costa Rica Vaccine Trial (CVT), a phase III randomized clinical trial, provided the initial data that one dose of the HPV vaccine could provide durable protection against HPV infection. Although the study design was to administer all participants three doses of HPV or control vaccine, 20% of women did not receive the three-dose regimens, mostly due to involuntary reasons unrelated to vaccination. In 2011, we reported that a single dose of the bivalent HPV vaccine could be as efficacious as three doses of the vaccine using the endpoint of persistent HPV infection accumulated over the first four years of the trial; findings independently confirmed in the GSK-sponsored PATRICIA trial. Antibody levels after one dose, although lower than levels elicited by three doses, were 9-times higher than levels elicited by natural infection. Importantly, levels remained essentially constant over at least seven years, suggesting that the observed protection provided by a single dose might be durable. Much work has been done to assure these non-randomized findings are valid. Yet, the group of recipients who received one dose of the bivalent HPV vaccine in the CVT and PATRICIA trials was small and not randomly selected nor blinded to the number of doses received. The next phase of research is to conduct a formal randomized, controlled trial to evaluate the protection afforded by a single dose of HPV vaccine. Complementary studies are in progress to bridge our findings to other populations, and to further document the long-term durability of antibody response following a single dose.

Excess Deaths Associated With Underweight, Overweight, and Obesity: An Evaluation of Potential Bias

As the prevalence of obesity has increased over time in the United States (1,2), concern over the association between body weight and excess mortality also increased. In 2005, an analysis of estimated excess deaths, relative to the normal weight category (body mass index [BMI] 18.5-24.9), that were associated with underweight (BMI less than 18.5), overweight (BMI 25.0-29.9), and obesity (BMI greater than or equal to 30) in U.S. adults in 2000 was published (3). Both underweight and obesity, particularly higher levels of obesity, were associated with increased mortality relative to the normal weight category. Obesity was estimated to be associated with 111,909 excess deaths (95% confidence interval [CI]: 53,754 to 170,064) in 2000 relative to the normal weight category, and underweight with 33,746 excess deaths (95% CI: 15,726 to 51,766). Overweight was associated with reduced mortality (-86,094 deaths; 95% CI: -161,223 to -10,966). This report evaluates several potential sources of bias in that analysis.

Design and statistical considerations for studies evaluating the efficacy of a single dose of the human papillomavirus (HPV) vaccine

Cervical cancer is a leading cause of cancer mortality in women worldwide. Human papillomavirus (HPV) types 16 and 18 cause about 70% of all cervical cancers. Clinical trials have demonstrated that three doses of either commercially available HPV vaccine, Cervarix ® or Gardasil ®, prevent most new HPV 16/18 infections and associated precancerous lesions. Based on evidence of immunological non-inferiority, 2-dose regimens have been licensed for adolescents in the United States, European Union, and elsewhere. However, if a single dose were effective, vaccine costs would be reduced substantially and the logistics of vaccination would be greatly simplified, enabling vaccination programs in developing countries. The National Cancer Institute (NCI) and the Agencia Costarricense de Investigaciones Biomédicas (ACIB) are conducting, with support from the Bill & Melinda Gates Foundation and the International Agency for Research on Cancer (IARC), a large 24,000 girl study to evaluate the efficacy of a 1-dose regimen. The first component of the study is a four-year non-inferiority trial comparing 1- to 2-dose regimens of the two licensed vaccines. The second component is an observational study that estimates the vaccine efficacy (VE) of each regimen by comparing the HPV infection rates in the trial arms to those in a contemporaneous survey group of unvaccinated girls. In this paper, we describe the design and statistical analysis for this study. We explain the advantage of defining non-inferiority on the absolute risk scale when the expected event rate is near 0 and, given this definition, suggest an approach to account for missing clinic visits. We then describe the problem of estimating VE in the absence of a randomized placebo arm and offer our solution.

Durability of Protection Afforded by Fewer Doses of the HPV16/18 Vaccine: The CVT Trial

Background

Previously, we demonstrated similar human papillomavirus (HPV)16/18 vaccine efficacy estimates and stable HPV16/18 antibody levels four years postvaccination in a nonrandomized analysis of women who received a varying number of doses of the bivalent HPV16/18 vaccine. Here we extend data to seven years following initial vaccination.

Methods

We evaluated HPV16/18-vaccinated women who received one (n = 134), two (n 0/1 = 193, n 0/6 = 79), or three doses (n = 2043) to a median of 6.9 years postvaccination. Cervical HPV DNA was measured with the SPF10- DEIA-LiPA PCR system; HPV16/18-specific antibody levels were measured using enzyme-linked immunosorbent assays (n = 486). Infection and immunological measures were compared across vaccine dose groups. Prevalent HPV infection at year 7 was also compared with an unvaccinated control group (UCG). All statistical tests were two-sided.

Results

Among women in the three-dose, two-dose 0/6 , two-dose 0/1 , and one-dose groups, cumulative incident HPV16/18 infection rates (No. of events/No. of individuals) were 4.3% (88/2036, 95% confidence interval [CI] = 3.5% to 5.3%), 3.8% (3/78, 95% CI = 1.0% to 10.1%), 3.6% (7/192, 95% CI = 1.6% to 7.1%), and 1.5% (2/133, 95% CI = 0.3% to 4.9%; P = 1.00, .85, .17 comparing the two-dose 0/6 , two-dose 0/1 , and one-dose groups to the three-dose group, respectively). The prevalence of other carcinogenic and noncarcinogenic HPV types, excluding HPV16/18/31/33/45, were high and not statistically different among all dose groups, indicating that the low incidence of HPV16/18 in the one- and two-dose groups was not due to lack of exposure. At seven years, 100% of participants in all dose groups remained HPV16 and HPV18 seropositive. A non-statistically significant decrease in the geometric mean of the HPV16 antibody levels between years 4 and 7 was observed among women in the three-dose group: -10.8% (95% CI = -25.3% to 6.6%); two-dose (0/6 months) group: -17.3% (95% CI = -39.3% to 12.8%), two-dose (0/1 month) group: -6.9% (95% CI = -22.1% to 11.2%), and one-dose group: -5.5% (95% CI = -29.7% to 27.0%); results were similar for HPV18.

Conclusions

At an average of seven years of follow-up, we observed similar low rates of HPV16/18 infections and slight, if any, decreases in HPV16/18 antibody levels by dose group.

Power and sample size for multivariate logistic modeling of unmatched case-control studies

Sample size calculations are needed to design and assess the feasibility of case-control studies. Although such calculations are readily available for simple case-control designs and univariate analyses, there is limited theory and software for multivariate unconditional logistic analysis of case-control data. Here we outline the theory needed to detect scalar exposure effects or scalar interactions while controlling for other covariates in logistic regression. Both analytical and simulation methods are presented, together with links to the corresponding software.