Predicting Pancreatic Cancer

Project Summary

Albert Antar(Biology), and Zidi Xiu (Biostatistics) spent ten weeks leveraging Duke Electronic Medical Record (EMR) data to build predictive models of Pancreatic ductal adenocarcinoma (PDAC). PDAC is the 4th leading cause of cancer deaths in the US, and is most often is diagnosed in stage IV, with a survival rate of only 1% and life expectancy measured in months. Diagnosis of PDAC is very challenging due of deep anatomical placement, and significant risk imposed by traditional biopsy. The goal of this project is to utilize EMR data to identify potential avenues for diagnosing PDAC in the early treatable stages of disease.

Themes and Categories
Health
Data+
Year
2016

Project Results

The team first constructed a patient timeline leading up to PDAC using diagnostic codes in the EMR data. They then applied a supervised topic model to the diagnosis code data, resulting in highly interpretable groups of diagnoses, and a promising predictive model for pancreatic cancer. The study team is following up with clinical colleagues in the Duke Department of Medicine to initiate further studies based on the team's work.

Download the Executive Summary (PDF)

Faculty Sponsors

Project Manager

  • Shaobo Han, post-doc, Electrical and Computer Engineering

Participants

  • Albert Antar, Duke University Biology
  • Zidi Xiu, Student Mentor, Master's student Biostatistics, Duke University

Disciplines Involved

  • Biostatistics
  • Pre-med
  • All quantitative STEM

Related People

Related Projects

Predictive Maintenance of F15s

A team of students that worked together for a semester in the Mission Driven Startups class will obtain and analyze data to create a predictive maintenance model for F15-E Fighter Jets from Seymour Johnson Air Base. Using data provided by the Base, the Data+ team will evaluate the relationship between unscheduled maintenance and external factors such as weather, sortie hours between repairs, and failure frequency of aircraft components. These findings will then feed into a predictive maintenance model to enhance the Air Force Crew’s ability to anticipate maintenance needs, helping to minimize unscheduled aircraft downtime. 

 

Faculty Lead: Dr. Emma Rasiel

Client Lead: Lt. Devon Burger

Project Manger:  Vignesh Kumaresan

Neural Network-Based Self-Adjusting Computational Processors

A team of students, led by Electrical and Computer Engineering professor Vahid Tarokh, will develop methods to improve the efficiency of information processing with adaptive decisions according to the structure of new incoming data. Students will have the opportunity to explore data-driven adaptive strategies based on neural networks and statistical learning models, investigate trade-offs between error threshold and computational complexity for various fundamental operations, and implement software prototypes. The outcome of this project can potentially speed up many systems and networks involving data sensing, acquisition, and computation.

Project Leads: Yi Feng, Vahid Tarokh

Computational Approaches to the History of Cartography

A team of students will explore new ways of reading pre-modern maps and perspectival views through image tagging, annotation and 3D modeling. Each student will build a typology of icons found in these early maps (for example, houses, churches, roads, rivers, etc.). By extracting, modeling, and cataloging these features, the team will create a library of 2D and 3D objects that will be used to (a) identify patterns in how space and power are represented across these maps, and (b) to create a model for “experiencing” these maps in 3D, using the Unity game engine platform. This is a combined Data+ / Bass Connections project that will instruct students in qualitative and quantitative mapping techniques, basic 3D modeling and the history of cartography.

Project Lead: Philip Stern, Ed Triplett

Project Manager: Sam Horewood

See all Research