Summer School - Medical Informatics Experiences in Undergraduate Research 2018, USA

University of Chicago


March 06, 2018

Event Date:

June 11, 2018 - August 17, 2018

Opportunity Cover Image - Summer School - Medical Informatics Experiences in Undergraduate Research 2018, USA

MedIX: MEDical Informatics eXperiences

The Medical Informatics (MedIX) program’s main objectives are to encourage talented undergraduates to pursue graduate education and to expose students to interdisciplinary research, especially at the border of information technology and medicine. 

All of the projects on which students will work are inspired by state-of-the-art research questions in imaging informatics.  Students will work as part of faculty-undergraduate teams on new problems ranging from traditional image processing (e.g. liver segmentation and computer-aided diagnosis, breast density assessment for cancer detection) to structured reporting and natural language processing of radiology reports, to workflow and process re-engineering to the application of data mining and ontology-based means for image annotation and markup (e.g. lung nodule detection and interpretation). Ultimately, each project has the long-term potential to increase the quality of healthcare available to people everywhere.

Faculty mentors will conduct tutorials on imaging informatics, conduct biweekly meetings of students and mentors, and create an environment that will expose students to all phases of research and graduate school.  Students will participate in defining the direction of their research, give presentations to the group as well as other audiences, write up  and publish research results in the format of a conference or journal article, and participate in relevant conferences.

The MedIX REU site will be hosted by two interdisciplinary laboratories: the Medical Informatics Laboratory at DePaul University and the Imaging Research Institute at the University of Chicago; the research environment will offer the students the opportunity to interact with computer scientists, medical physicists, and medical doctors.

Important Dates

January 11, 2018:  Application Submission Opens
March 6, 2018:  Application Submission Deadline
March 30, 2018:  Notification of Decision
April 6, 2018:  Confirmation of Participation
June 11, 2018: MedIX Program Orientation
August 17, 2018: MedIX Program Last Day

 Summer Support for Each Student:

  • $5,000 stipend (paid a weekly stipend of $500 for each of the 10 weeks)
  • $500 travel support to/from the REU site at the beginning/end of the program
  • $2,600 subsistence allowance (includes full expected housing costs and food allowance)
  • Housing on campus will be arranged for all students 

Program Requirements:

  • US citizenship or permanent residency
  • Full-time undergraduate student with at least 1 year of college/university course work completed as of September 1, 2018
  • Students are not allowed to have other jobs/internships during the REU program or to be registered for any courses


  • Classifying Lung Nodules using Convolutional Neural Networks

Early detection of lung nodules in CT scans is key to improving lung cancer treatment, but poses a challenge for radiologists due to the high throughput required of them. Computer-Aided Detection system aim to improve diagnosis rates by automatically detecting nodules. Most current systems divide the task into two steps:Identifying candidates which resemble nodules and classifying the candidates as nodules or not. MedIX proposes a simpler system using one combined step using 3D convolutional neural networks.

  • Constructing phylogenies of cichlids using normalized compression distance

A phylogenetic tree on a set of species is a hypothesis about how genetically close the various species are to each other.  The data used for this are DNA/RNA sequences.  Biologists have their methods, that are almost always statistically based and computationally expensive.  MedIX work determines the distance using normalized compression distance, a technique based on the theoretical notion of Kolmogorov complexity.  Because the Kolmogorov complexity of a string is not computable, MedIX uses instead a compression algorithm (e.g., Huffman, Lempel-Ziv).  One outcome is that this approach is much less expensive computationally.  One goal of the project at this point is to investigate how well this approach works with some of the large collection of very recent compression algorithms specifically designed for DNA/RNA sequences. MedIX does this by generating trees with the method from a set of sequences for which there are published trees generated using the methods traditionally used in biology.  Another goal is to scale up the technique. MedIX has applied it to sets of 20 to 30 species; it would like to apply it to sets of at least an order of magnitude higher. 

  • Texture Analysis in CT Scans Demonstrating Interstitial Disease

Interstitial lung disease is often difficult to diagnose, causing discrepancies among radiologists. However an early and accurate diagnosis is necessary in order to provide proper treatment and prognosis to patients. Computer-aided diagnosis (CAD) systems can use texture analysis to provide quantitative measures of scans to detect subtle infiltrates. This project focuses on using texture analysis on various patient CT scans, specifically those with nonspecific interstitial pneumonia (NSIP), in order to determine the characteristics that will help in distinguishing between healthy lungs and lungs with NSIP.

  • Projection Matrix Modeling for Limited Angle Tomography

The project is interested in exploring the application of iterative image reconstruction techniques to limited angle tomography. Test phantoms will be used to simulate the process of breast tomographic image reconstruction in a clinical setting.  Image quality parameters will be used to determine the precision of the final image. With the goal of ultimately determining optimal pixel size for image reconstruction based on image and data error.

  • Computer Aided Diagnosis: Navigating Uncertainty with Confidence  

Computer-aided diagnosis (CAD) systems provide radiologists with a second opinion when analyzing medical images. These systems typically aim to reduce the work required for a radiologist to assess an image through their ability to analyze and predict malignancy characteristics of the case in an efficient and effective way. In this project, you will investigate strategies to better analyze CAD systems’ decision making when handling complex diagnostic cases with uncertainty. MedIX first examines Belief Decision Tree classifiers and Iterative Classification approaches to quantify the uncertainty through a vector of belief assignments corresponding to the malignancy class distribution. Second, MedIX proposes to apply a measure of a system’s diagnostic reliability and identify the typicality of a specific case to augment the CAD outcome. Combining these approaches have the potential to produce smarter CAD systems by providing more detailed and descriptive CAD output for radiologists in a clinical setting.


  • DePaul University

The Intelligent Multimedia Processing (IMP) Laboratory will host the project activities at the School of Computing and Digital Media (CDM). The CDM research laboratory has seven faculty members and a mission focused on medical imaging, image processing, computer vision, content-based multimedia retrieval, data analysis and data mining.  The goal is to develop both the theory and the tools for real world applications from various domains.   There are around 10 students performing research in the lab every year. The IMP lab facilities include six high-end workstations boasting Intel i7 processors at 2.8 GHz per core and having large amounts of memory (~ 8GB). 

The Medical Imaging Informatics (MedIX) laboratory houses ten workstations providing workspace for ten-full-time students. The MedIX workstations have the latest Intel i7 processors at 2.8 GHz per core 8GB of DDR each.

Both labs are equipped with image processing and statistical software to enable the proposed studies.  Prototyping of image processing algorithms will be done on MatLab; statistical analyses will be conducted in MatLab and SPSS.  Machine learning and data mining algorithms will be implemented in MatLab and AnswerTree.  The open source will be developed in C#. The labs are also equipped with a small library of relevant texts.                                            

The IMP and MedIX labs are located at DePaul University and thus benefit from the Universitiy’s extensive internal networking support, including wireless ports throughout the university. DePaul has an Internet2 high-speed connection for conducting research, and the library has considerable holdings online.   Faculty and staff also have access to any of the University’s many public labs.

  • University of Chicago 

Laboratory: The Radiology Imaging Research Institute located in the Department of Radiology at the University of Chicago will be the facility for the proposed research.  The laboratories for x-ray imaging have over 30 faculty and staff members with 14,000 square-feet of recently renovated laboratory space.  The facility includes two dedicated computer rooms, seven laboratories (one psychophysics with two rooms dedicated for observer performance studies, one film digitization, one scientific computing and visualization, three digital image processing), three x-ray rooms (containing five x-ray generators and seven x-ray tubes, including a Fischer digital stereotactic system and a Faxitron DX20), one darkroom, and three conference rooms. It has in the laboratory a R2 Technology, Inc, ImageChecker 1000 system modified so that students can archive the digitized images and have access to a file that contains the x-y location and lesion type of all computer detections.

 Clinical: The Radiology Department has 47 attending radiologists, 7 fellows, and 24 residents.  There are 6 radiologists and 1 fellow performing breast imaging.  Over 18,000 mammographic examinations are performed each year in the department.  The department has three dedicated mammographic imaging units, one digital stereotactic breast needle biopsy unit, one breast ultrasound unit, two GE full-field digital mammography (FFDM) system, and a FujiFilm CR system for mammography. 

Computer: Computer workstations include: 16 SGI workstations, 5 SUN Ultrasparc workstations, and 20 Dell or equivalent workstations running Linux operating system. Each computer has at least 40 Gbytes of disk space on average with data backup done centrally.  There are 3 Apple Macintosh computers in the laboratories and one Apple Macintosh computer in each investigator's office.  All workstations are connected via high-speed Internet.  The Department of Radiology facilities include a Silicon Graphics Power Onyx (R10,000 6 processors) with an Immersadesk and 500 Gbytes of disk space.

The Rossmann Laboratories have a Silicon Graphics Power Onyx (R10,000 multi-processors) that contains four 200 MHz processors with 1.5 gigabytes of RAM and 140 gigabytes of hard disk space.  The graphics hardware on the Onyx operates an Immersadesk for display of volume-rendered, virtual reality representations of CT data sets and computer detection output.  For image display, the Rossmann Laboratories have three 1k x 1k Imlogix CRT monitors, four 2k x 2k Megascan CRT monitors, six Seikosha gray-scale paper printers, two Konica laser film printers (2k x 2K), one Konica high-quality laser film printer (4k x 5K), and one Kodak high-quality laser film printer. 

For more information click "LINK TO ORIGINAL" below.

Eligible Countries
Host Country
Study Levels
Publish Date
February 26, 2018
Link To Original