Methodology Development

Methodology Workshops

BERD provides leadership and resources on development of new methodologies related to clinical and translational research. The BERD KCA has identified promising areas for methodology development and has sponsored methodology workshops on relevant areas including:

  • Sequential multiple assignment randomized treatment(SMART) trials
  • Big data methodology
  • Support vector machine(SVM) classifier and regression

The methodology workshops sponsored by BERD KCA can be accessed here.

Methodology Pilot Projects

Lynette Smith, PhD

Assistant Professor
Department of Biostatistics
University of Nebraska Medical Center

Project: Time to Event Prediction Based on Risk Calculation from Longitudinal Biomarkers using Bayesian Hierarchical Changepoint Mixture Models (Poster Download)

Pancreatic cancer (PC) is an extremely aggressive malignancy with 5-year overall survival rate of <8%. While identification of patients at a resectable stage results in increased patient survival, PC patients are often diagnosed at late stages due to the asymptomatic nature of PC. Biomarkers are needed that can be used to screen individuals for PC during the asymptomatic period in order to detect disease and predict survival rate. Longitudinally measured biomarkers can show a variety of trends over time. Most control subjects show little to no change in biomarkers, but cancer patients with progressive disease could reflect either no change or rapid changes in biomarkers prior to diagnosis of cancer or prior to death from the disease. Hence, the rate of change in biomarker levels, in combination with the absolute level of the biomarker, could better predict presence/progression of disease and patient survival. The aim of this project is to develop methodology for a Bayesian changepoint mixture model could be used to effectively predict patient survival or time to recurrence.

Christopher Wichman, PhD

Assistant Professor
Department of Biostatistics
University of Nebraska Medical Center

Project: Are Linear Models Sufficient for Analyzing Adolescent BMI z-scores: a comparison of methods for right skewed data

Childhood obesity is currently considered one of the most serious public health problems. The states included in the Great Plains IDeA-CTR Network (NE, ND, SD and KS) are estimated to have one of the highest national percentiles of 2-19 year olds classified as overweight or obese. My research looks at the most appropriate ways of analyzing Body Mass Index (BMI) z-score data. BMI z-scores are measures of relative weight adjusted for child age and sex. Due to inclusion criteria in a number of comparative studies that require participants to have a minimum BMI percentile of 85%, the data tends to be skewed right. This calls into question whether the appropriate models are being used for these measurements. Success of this pilot will provide a more precise representation of BMI z-score data.

Chi Zhang, PhD

Associate Professor
Biological Sciences
University of Nebraska-Lincoln

Project: Predictive modeling and visual analytics of radiotherapy on pancreatic cancer treatment, diagnosis, and prognosis

Ductal pancreatic adenocarcinoma is the fourth most common cause of death due to cancer in the United States. Although some new front-line chemotherapy/radiotherapy regimens have modified the landscape of therapy for pancreatic cancer, such as Stereotactic Body RadioTherapy (SBRT), there remain several problems, such as not every patient responds to all therapies and treatment. In this project, by applying the latest technology and methods in artificial intelligence, visualization, and computational systems biology to clinical images and data, we propose a mathematic model on pancreatic tumor growth and shrinkage to predict tumor response underwent SBRT and utilize deep-learning method for auto-segmentation. If successful, these models and methods can have direct clinical impacts on pancreatic cancer radiotherapy, such as to facilitate clinical decision making, so that the patients could receive the most effective treatments with the least toxicity.