Andrew Louis Wentland

Company: GE Healthcare
Dates: May 2004 - August 2004
Position Title: Software Visualization Engineer
Fields: Radiology, BME & Computer Science
Internship or Co-op: Internship
Location: Waukesha, WI
Began: After junior year

How I earned the position
I earned this internship through UW's Engineering Career Services around the time of November. In late spring GE contacts its interns and gives them a list of project descriptions. Interns are told to select their top five choices and they'll try their best to accommodate the interns' interests. I was very interested in an MRI-based project since my research deals with a similar subject matter. Therefore, all of the projects I chose dealt with MRI. It just so happened, however, that the advisers of those projects wanted an intern with a master's degree. I actually had to fight for a few weeks to earn such a position. I was only able to convince them to grant me the position because of my experience with MRI research and programming.

Project Description
GE Healthcare sells a program called ReportCARD, which provides radiologists with useful tools for analyzing cardiovascular MRI images. When I began my position the program had been issued for a year and needed to be updated. Therefore, the project required extensive programming knowledge (in Java) and also the ability to design analysis tools from a radiologist's perspective.

During my internship I added three major features to the program. The first project involved bull's eye plots. Bull's eye plots are segmented circles with each segment corresponding to a region of the heart. Radiologists color code these segments to implicate various diseased states of the hearts' regions. The bull's eye plots already included in ReportCARD contained 16 segments, but should have contained 17 segments according to the standards of the American Heart Association (AHA). After adding the seventeenth segment I added a "painting" feature that allowed radiologists to quickly color-code the segments (previously, a radiologist had to right-click on each segment and choose the color).

My second feature was by far my biggest project. This feature may be a little difficult to describe without some knowledge of MRI's capabilities, but hopefully it will make sense. The goal of this feature was to provide the ability to analyze a patent foramen ovale (PFO) [a partial gap or flap in the septum between the right and left atria]. This PFO causes a bit of blood to leak from the right atrium to the left atrium. A PFO can be detected in MRI images if signal (blood) can be detected in the left atrium when there should otherwise be no signal (blood). So my feature graphed the signal in the left atrium throughout the cardiac cycle. A spike in signal thus indicates the presence of a PFO.

My final project was very basic. This feature allowed for the export of .gif images and .mpeg movies.

How I liked the job
I'm not a very big fan of industry, so if you are, you may want to take my opinions lightly.

As mentioned, my project required a lot of programming. An entire software program contains a lot of code, and this is where difficulty arises. Programming new features wasn't difficult because of the concept of programming itself, but because I didn't have a working knowledge of the existing code. Rather, I was thrown into a million lines of code that had a structure clear only in the minds of the program's creators. Oftentimes I needed help with locating something in the code, but never received any help.

The people I worked with all had degrees in computer science. Frequently I needed to explain to them MRI-related matters and why features should be designed a certain way for radiologists. GE often hires outside consultants to evaluate their products; for ReportCARD they fly in a medical scientist (MD/PhD). I got to work with him a number of times. Working with him was one of my favorite parts of the internship.

When I was all done and back in Madison working on my research, something exciting happened as a result of my internship. Since I work in the Department of Radiology at the UW Hospital, I actually got to see ReportCARD in action with the very features I added to the program! Therefore, radiologists all around the world are using the tools I had devoted my summer to designing.

Why I sought the position
I'm much more of a research-oriented person than an industry person. My primary reason for wanting to work at GE was because all of my MRI research deals with GE MRI equipment. My second reason was to confirm that I don't want to pursue industry for a career. GE actually offered me an internship for the summer of 2005 and a permanent position in the group with whom I worked. As for the former, I turned the offer down so that I could work on my research. I also rejected the latter in pursuit of a career in research.