Swati Biswas

Bayesian Clinical Trials at the University of Texas M. D. Anderson Cancer Center

Background The Bayesian approach is being used increasingly in medical research. In particular, it has become a standard in designing clinical trials at the University of Texas M. D. Anderson Cancer Center. Purpose/Methods To address the extent and nature of Bayesian trials conducted at M. D. Anderson, we reviewed the protocols registered in the Protocol Document Online System between 2000 and early 2005. We summarize our findings and give details for three innovative trials that typify those in which a Bayesian approach has played a major role at the center. Results Of 964 protocols reviewed, 59% were conducted solely at M. D. Anderson and the rest were multicenter trials. Bayesian designs and analyses were used in about 20% (195/964) of the protocols that we reviewed. Of the 520 protocols identified as phase I or II drug trials, about 34% were Bayesian. Most of the 195 Bayesian trials were designed by M. D. Anderson statisticians. The Bayesian design features most commonly used were the continuous reassessment method in phase I (toxicity) trials, adaptive randomization in phase II trials, and designs to monitor efficacy and toxicity simultaneously. We also provide an insiders view regarding some practical considerations that have made the design and implementation of so many Bayesian trials possible. Limitations We reviewed only a subset of all M. D. Anderson protocols, but did not exclude any available in electronic form. Conclusions The large number of Bayesian trials conducted at M. D. Anderson testifies to the receptivity to the Bayesian approach within the center, including principal investigators, regulatory review committees, and patients. Statisticians who take a Bayesian perspective can successfully work to establish a culture of innovation in clinical trial design.

Factors associated with attention deficit/hyperactivity disorder among US children: Results from a national survey

BackgroundThe purpose of this study was to investigate the association between Attention Deficit/Hyperactivity Disorder (ADHD) and various factors using a representative sample of US children in a comprehensive manner. This includes variables that have not been previously studied such as watching TV/playing video games, computer usage, family member’s smoking, and participation in sports.MethodsThis was a cross-sectional study of 68,634 children, 5–17 years old, from the National Survey of Children’s Health (NSCH, 2007–2008). We performed bivariate and multivariate logistic regression analyses with ADHD classification as the response variable and the following explanatory variables: sex, race, depression, anxiety, body mass index, healthcare coverage, family structure, socio-economic status, family members’ smoking status, education, computer usage, watching television (TV)/playing video games, participation in sports, and participation in clubs/organizations.ResultsApproximately 10% of the sample was classified as having ADHD. We found depression, anxiety, healthcare coverage, and male sex of child to have increased odds of being diagnosed with ADHD. One of the salient features of this study was observing a significant association between ADHD and variables such as TV usage, participation in sports, two-parent family structure, and family members’ smoking status. Obesity was not found to be significantly associated with ADHD, contrary to some previous studies.ConclusionsThe current study uncovered several factors associated with ADHD at the national level, including some that have not been studied earlier in such a setting. However, we caution that due to the cross-sectional and observational nature of the data, a cause and effect relationship between ADHD and the associated factors can not be deduced from this study. Future research on ADHD should take into consideration these factors, preferably through a longitudinal study design.

Association between Obesity and asthma in US children and adolescents

Background. To explore the association between obesity and asthma in US children and adolescents with adjustment of other structural and behavioral factors. Method. Prevalence and associated risk factors of asthma were explored in 102,273 children and adolescents in the National Survey of Childrens Health (2003–2004). Subgroup analysis was performed for subjects of 0-6 year-old, 7–12 year-old, and 13–17 year-old. Crude and adjusted odds ratios for the potential risk factors were examined in univariate and multivariate logistic regressions. Results. The overall prevalence of obesity was 24.5% and that of asthma was 12.5%. The adjusted odds ratio of asthma with obesity remains significantly bigger than 1 for children in the 7–12 and the 13–17 year-old age-groups. Gender and race were significantly associated with asthma in all age groups. The two parent family structure showed significant protectiveness against asthma with children in the 0–6 year-old age group. Poverty was positively associated with asthma in the 7–12 years old age group. Having a smoker in the household increased the odds of asthma by 29% and 23.5% in the 0–6 and 13–17 year-old age-groups, respectively. Higher education level of the parents and access to healthcare showed positive association with asthma in the 13–17 year-old age group. Conclusion. Gender and race were significantly associated with asthma. In the 13–17 year-old age-groups, obesity, household education level, healthcare coverage, and household smoking were positively associated with asthma. Further studies should characterize how the family structure and household education level influence childhood asthma in 0–6 and 13–17 year-old age-groups respectively.

Reciprocal Regulation of Annexin A2 and EGFR with Her-2 in Her-2 Negative and Herceptin-Resistant Breast Cancer

Alternative survival pathways are commonly seen to be upregulated upon inhibition of receptor tyrosine kinases (RTK), including Her-2. It is established that treatment with Herceptin leads to selective overexpression and activation of epidermal growth factor receptor (EGFR) and Src which further contributes to oncogenesis in Herceptin resistant and triple negative breast cancer (TNBC) patients. Here, we show a co-regulated upregulation in the expression of Annexin A2 (AnxA2), a known substrate of Src and one of the regulators of EGFR receptor endocytosis, in Herceptin resistant and Her-2 negative breast cancer. Immunohistochemical expression analysis revealed a reciprocal regulation between Her-2 and AnxA2 in breast cancer clinical samples as well as in cell lines as confirmed by protein and RNA analysis. The siRNA and Herceptin mediated downregulation/inhibition of Her-2 in Her-2 amplified cells induced AnxA2 expression and membrane translocation. In this study we report a possible involvement of AnxA2 in maintaining constitutively activated EGFR downstream signaling intermediates and hence in cell proliferation, migration and viability. This effect was consistent in Herceptin resistant JIMT-1 cells as well as in Her-2 negative breast cancer. The siRNA mediated AnxA2 downregulation leads to increased apoptosis, decreased cell viability and migration. Our studies further indicate the role of AnxA2 in EGFR-Src membrane bound signaling complex and ligand induced activation of downstream signaling pathways. Targeting this AnxA2 dependent positive regulation of EGFR signaling cascade may be of therapeutic value in Her-2 negative breast cancer.

Assessing the added value of breast tumor markers in genetic risk prediction model BRCAPRO

Abstract The BRCAPRO model estimates carrier probabilities for the BRCA1 and BRCA2 genes, and was recently enhanced to use estrogen receptor (ER) and progesterone receptor (PR) status of breast cancer. No independent assessment of the added value of these markers exists. Moreover, earlier versions of BRCAPRO did not use human epidermal growth factor receptor 2 (Her-2/neu) status of breast cancer. Here, we incorporate Her-2/neu in BRCAPRO and validate all the markers. We trained the enhanced model on 406 germline tested individuals, and validated on a separate clinical cohort of 796 individuals for whom test results and family history are available. For model-building, we estimated joint probabilities of ER, PR, and Her-2/neu status for carriers and non-carriers of BRCA1/2 mutations. For validation, we obtained BRCAPRO predictions with and without markers. We calculated area under the receiver operating characteristic curve (AUC), sensitivity, specificity, predictive values, and correct reclassification rates. The AUC for predicting BRCA1 status among individuals who are carriers of at least one mutation improved when ER and PR were used. The AUC for predicting the presence of either mutation improved when Her-2/neu was added. Use of markers also produced highly significant correct reclassification improvements in both cases. Breast tumor markers are useful for prediction of BRCA1/2 mutation status. ER and PR improve discrimination between BRCA1 and BRCA2 mutation carriers while Her-2/neu helps discriminate between carriers and non-carriers, particularly among women who are ER positive and Her-2/neu negative. These results support the use of the enhanced version of BRCAPRO in clinical settings.

Impact of a Community Based Implementation of Reach II Program for Caregivers of Alzheimer's Patients

Background In 2009 an estimated 5.3 million people in the United States were afflicted with Alzheimers disease, a degenerative form of dementia. The impact of this disease is not limited to the patient but also has significant impact on the lives and health of their family caregivers. The Resources for Enhancing Alzheimers Caregiver Health (REACH II) program was developed and tested in clinical studies. The REACH II program is now being delivered by community agencies in several locations. This study examines the impact of the REACH II program on caregiver lives and health in a city in north Texas. Study design Family caregivers of Alzheimers patients were assessed using an instrument covering the multi-item domains of Caregiver Burden, Depression, Self-Care, and Social Support upon enrollment in the program and at the completion of the 6 month intervention. The domain scores were analyzed using a multivariate paired t-test and Bonferroni confidence interval for the differences in pre- and post-service domain scores. Results A total of 494 families were enrolled in the program during the period January 1, 2011 through June 30, 2012. Of these families 177 completed the 6 month program and have pre – and post service domain scores. The median age for the caregivers was 62 years. The domain scores for Depression and Caregiver Burden demonstrated statistically significant improvements upon program completion. Conclusion The REACH II intervention was successfully implemented by a community agency with comparable impacts to those of the clinical trial warranting wider scale implementation.

Simplifying clinical use of the genetic risk prediction model BRCAPRO

Health care providers need simple tools to identify patients at genetic risk of breast and ovarian cancers. Genetic risk prediction models such as BRCAPRO could fill this gap if incorporated into Electronic Medical Records or other Health Information Technology solutions. However, BRCAPRO requires potentially extensive information on the counselee and her family history. Thus, it may be useful to provide simplified version(s) of BRCAPRO for use in settings that do not require exhaustive genetic counseling. We explore four simplified versions of BRCAPRO, each using less complete information than the original model. BRCAPROLYTE uses information on affected relatives only up to second degree. It is in clinical use but has not been evaluated. BRCAPROLYTE-Plus extends BRCAPROLYTE by imputing the ages of unaffected relatives. BRCAPROLYTE-Simple reduces the data collection burden associated with BRCAPROLYTE and BRCAPROLYTE-Plus by not collecting the family structure. BRCAPRO-1Degree only uses first-degree affected relatives. We use data on 2,713 individuals from seven sites of the Cancer Genetics Network and MD Anderson Cancer Center to compare these simplified tools with the Family History Assessment Tool (FHAT) and BRCAPRO, with the latter serving as the benchmark. BRCAPROLYTE retains high discrimination; however, because it ignores information on unaffected relatives, it overestimates carrier probabilities. BRCAPROLYTE-Plus and BRCAPROLYTE-Simple provide better calibration than BRCAPROLYTE, so they have higher specificity for similar values of sensitivity. BRCAPROLYTE-Plus performs slightly better than BRCAPROLYTE-Simple. The Areas Under the ROC curve are 0.783 (BRCAPRO), 0.763 (BRCAPROLYTE), 0.772 (BRCAPROLYTE-Plus), 0.773 (BRCAPROLYTE-Simple), 0.728 (BRCAPRO-1Degree), and 0.745 (FHAT). The simpler versions, especially BRCAPROLYTE-Plus and BRCAPROLYTE-Simple, lead to only modest loss in overall discrimination compared to BRCAPRO in this dataset. Thus, we conclude that simplified implementations of BRCAPRO can be used for genetic risk prediction in settings where collection of complete pedigree information is impractical.

Recent Enhancements to the Genetic Risk Prediction Model BRCAPRO

BRCAPRO is a widely used model for genetic risk prediction of breast cancer. It is a function within the R package BayesMendel and is used to calculate the probabilities of being a carrier of a deleterious mutation in one or both of the BRCA genes, as well as the probability of being affected with breast and ovarian cancer within a defined time window. Both predictions are based on information contained in the counselees family history of cancer. During the last decade, BRCAPRO has undergone several rounds of successive refinements: the current version is part of release 2.1 of BayesMendel. In this review, we showcase some of the most notable features of the software resulting from these recent changes. We provide examples highlighting each feature, using artificial pedigrees motivated by complex clinical examples. We illustrate how BRCAPRO is a comprehensive software for genetic risk prediction with many useful features that allow users the flexibility to incorporate varying amounts of available information.

Detecting Rare Haplotype‐Environment Interaction With Logistic Bayesian LASSO

Two important contributors to missing heritability are believed to be rare variants and gene‐environment interaction (GXE). Thus, detecting GXE where G is a rare haplotype variant (rHTV) is a pressing problem. Haplotype analysis is usually the natural second step to follow up on a genomic region that is implicated to be associated through single nucleotide variants (SNV) analysis. Further, rHTV can tag associated rare SNV and provide greater power to detect them than popular collapsing methods. Recently we proposed Logistic Bayesian LASSO (LBL) for detecting rHTV association with case–control data. LBL shrinks the unassociated (especially common) haplotypes toward zero so that an associated rHTV can be identified with greater power. Here, we incorporate environmental factors and their interactions with haplotypes in LBL. As LBL is based on retrospective likelihood, this extension is not trivial. We model the joint distribution of haplotypes and covariates given the case–control status. We apply the approach (LBL‐GXE) to the Michigan, Mayo, AREDS, Pennsylvania Cohort Study on Age‐related Macular Degeneration (AMD). LBL‐GXE detects interaction of a specific rHTV in CFH gene with smoking. To the best of our knowledge, this is the first time in the AMD literature that an interaction of smoking with a specific (rather than pooled) rHTV has been implicated. We also carry out simulations and find that LBL‐GXE has reasonably good powers for detecting interactions with rHTV while keeping the type I error rates well controlled. Thus, we conclude that LBL‐GXE is a useful tool for uncovering missing heritability.

Linkage analysis of the simulated data – evaluations and comparisons of methods

The goal of this study is to evaluate, compare, and contrast several standard and new linkage analysis methods. First, we compare a recently proposed confidence set approach with MAPMAKER/SIBS. Then, we evaluate a new Bayesian approach that accounts for heterogeneity. Finally, the newly developed software SIMPLE is compared with GENEHUNTER. We apply these methods to several replicates of the Genetic Analysis Workshop 13 simulated data to assess their ability to detect the high blood pressure genes on chromosome 21, whose positions were known to us prior to the analyses. In contrast to the standard methods, most of the new approaches are able to identify at least one of the disease genes in all the replicates considered.