Interview with Dr. Erin Gillespie
Dr. Erin Gillespie is currently an attending physician at Memorial Sloan Kettering Cancer Center and focuses on breast and lung cancers. She completed her residency in 2017 from the University of California San Diego where the concept of eContour first began. She is here to give us an inside look into the creation, challenges, and future direction of this valuable resource for radiation oncologists!
SWRO: What initially inspired the creation of eContour?
Dr. Gillespie: As a resident learning how to contour, I found it difficult to keep track of all the different protocols and publications that helped describe what should be covered. Often, especially in the absence of a resource, I was advised by my senior resident to look at a prior case contoured by our attending. I then learned that one residency program had put a set of H&N cases onto a disc for residents to take with them – I realized this issue of access to information might become even more challenging after residency, especially if leaving the institution and/or going to a practice with fewer colleagues. So I proposed the idea of eContour as a case viewer with links to evidence to my research mentor, Dr. Jim Murphy. At the same time, our research group was looking at outcomes related to H&N cancer volume in the era of IMRT (Boero et al JCO 2016), and got to thinking this could be an opportunity to try implementing a solution. In January 2015, I wrote our first grant for this project in about 10 days and, needless to say, it was rejected, but provided us the first set of peer reviewed feedback which was valuable.
SWRO: Who was instrumental in helping bring it to fruition?
Dr. Gillespie: The most critical person in turning this idea into a reality was a then third-year medical student, Neil Panjwani, who was rotating with me (and my attending) in clinic during the spring of 2015. He was planning to go into radiology but became intrigued by rad onc, so decided to take an extra research year in our department. Neil had both programming and imaging research experience, so Dr. Murphy proposed to him our eContour idea and within 2 weeks he came back to us with a prototype of a web-based viewer of 3D images and contours! We spent the next couple months fixing up the website with about 8 cases to show it in a focus group of 12 residents from different residency programs we hosted at ASTRO 2015. This group (which included several friends from the interview trail!) was critical in expressing enthusiasm that gave us the motivation (and constructive feedback) to take the project to the next level. With a prototype, we were able to start securing external grant funding, first from ACRO then RSNA (who had given our first rejection/feedback). We decided to first build out a comprehensive set of H&N cases, given data that this would be the highest yield disease site. It was at this time that Dr. Parag Sanghvi became heavily involved in selecting cases and reviewing in depth every contour and written description presented, which was critical to initial case development.
SWRO: What unforeseen obstacles did you face with this project?
Dr. Gillespie: When we started reviewing cases, we quickly realized that contours adequate for clinical practice were not always good for teaching/reference because minor edits were needed to precisely represent guidelines. The process of reviewing and modifying contours has proven to be somewhat labor-intensive. Of note, this issue of contour QA has also limited the development of autosegmentation algorithms from large clinical datasets. Second, we underestimated how difficult it would be to receive cases from outside institutions, as the process of exporting, anonymizing, and sharing DICOMs is often cumbersome. To remedy this, we are considering a partnership with MIM to use MIMcloud as an anonymized cloud-based DICOM repository.
SWRO: What has been the most rewarding part of this endeavor for you?
Dr. Gillespie:The most rewarding part is honestly the emails and comments we get from people who tell us that eContour has been useful in their practice. We email these around to the whole team every so often as it seems to be a good motivator for all of us.
SWRO: What are the future goals of eContour?
Dr. Gillespie: We are currently developing a web-based contouring platform based on user feedback — practicing physician users even more than residents wanted to be able to draw contours then overlay expert contours. A significant component of our research now involves simulation, in addition to implementation. There is of course an increasing need to understand how physicians are accessing evidence-based information in their routine practice to keep current with increasingly conformal radiation planning. Clinical pathways have demonstrated the most promising evidence for reducing variation in care, so we entered an agreement with Via Oncology to incorporate our content into their pathways
SWRO: What advice would you give to current radiation oncology residents who may ideas for their own innovative projects but do not know where to start?
Dr. Gillespie: In general, understanding the target audience – both the number of potential users and their enthusiasm for your product —is critical for determining feasibility and ultimately securing funding. For a business model, you have to consider willingness to pay, which is usually more successful if you can identify a way for the user to increase reimbursement. We have so far taken the academic/research route, which involves identifying data that supports the product’s potential to improve patient outcomes and quality of care, which we then have shaped into research questions and grant applications.
As a resident, you usually do need to identify a faculty member that will champion your project. Jim Murphy was my co-founder who helped shape my raw idea into something fundable, but he also secured initial funding from the department to support our programmer’s research year.
I have also been reminded multiple times to do one thing and to do it well. People will try to convince you to tackle related problems, but you have to be careful to not stray from your primary focus.