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

The Air Force’s F-15E Strike Eagle jets have parts that wear down and break, causing unscheduled maintenance events that take away valuable time in the air for critical missions and training. Our team, Limitless Data, is working with Seymour Johnson Air Force Base to mine manually entered maintenance data to visualize and predict aircraft failures. We created a prototype data visualization product that will enable maintainers on the flight line and help them identify and repair critical failures before they happen, keeping jets ready to fly, fight and win.

 

Faculty Lead: Dr. Emma Rasiel

Client Lead: Lt. Devon Burger

Project Manger:  Vignesh Kumaresan

Neural Network-Based Self-Adjusting Computational Processors

This project aims to improve the computational efficiency of signal operations, e.g., sampling and multiplying signals. We design machine learning-based signal processing modules that use an adaptive sampling strategy and interpolation to generate a good approximation of the exact output. While ensuring a low error level, improvements in computational efficiency can be expected for digital signal processing systems using the implemented self-adjusting modules.

Project Leads: Yi Feng, Vahid Tarokh

 

Click here to view the project team's poster

 

Watch the team's final presentation (on Zoom) here:

 

Computational Approaches to the History of Cartography

Mapping History has focused on the categorizing, labelling, digitization, and 3D reconstruction of 16th & 17th century maps & atlases of London and Lisbon. Over the course of the summer, the Mapping History team has developed its own unique analytical dataset by painstakingly labelling every element contained within these maps, used python to digitize this dataset, and, now in the projects final stage, has begun the process of reconstructing these historical perspectives in a 3D game engine.

Project Lead: Philip Stern, Ed Triplett

Project Manager: Sam Horewood

 

View the team's final poster here

Watch the team's final presentation (on Zoom) below:

See all Research