Let's do an eeDAP reader study

This is the first email to this group.

Thanks for your interest in conducting a reader study with eeDAP. This email is for connecting everyone, stating goals, and getting started.

 

Connecting everyone (introductions):

  • Brandon Gallas, Qi Gong, and Marios Gavrielides: FDA/CDRH/OSEL/Division of Imaging, Diagnostics, and Software review, Silver Spring, MD
  • Mark Simpson, Chuck Halsey, Stephen Hewitt: NIH/NCI/CCR/Laboratory of Cancer Biology and Genetics, Rockville, MD
  • Darren Treanor: Leeds Teaching Hospitals NHS Trust and University of Leeds, Leeds, UK
  • Yukako Yagi, Jo Sirintrapun, Tom Fuchs: Memorial Sloan Kettering Cancer Center, New York, NY
  • James Monaco, Richard Morroney, Santosh Bhargava, Mark Lloyd: Inspirata, Inc., Tampa, FL
  • Esther Abels: Philips Digital Pathology Solutions, Best, Netherlands
  • “Mohan” Mohanasankar Sivaprakasan, Jayaraj: IIT, Indian Institutes of Technology, Chennai, India
  • Mitra Partha: Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

 

Stating goals:

  • My main goal is to design, execute, and analyze an eeDAP study to provide evidence demonstrating that task-based pathologist evaluations of tissue features taken with eeDAP are more reproducible than pathologist evaluations taken following a standard clinical method. This evidence is needed for qualifying eeDAP as an FDA/CDRH Medical Device Development Tool (Summary of eeDAP and MDDT) (Draft MDDT Proposal Announcement) (Draft MDDT proposal). I am currently planning that the study task will be mitotic counting and the results will be compared to a recently completed study that followed standard clinical methods for counting mitoses.
  • My next goal is to help 1-3 other groups design, execute, and analyze an eeDAP study of their own. While these studies will provide additional evidence to the utility of eeDAP, the purpose (and the feature-based task) of these studies will be defined by the group leading the study.
  • My immediate plans are to quickly prepare a pilot study and collect data before Pathology Informatics Summit in May. Then at PI I plan to install an eeDAP system, demo the pilot study, and share preliminary results (demonstrating a fast and precise study). The purpose is to generate interest for recruiting pathologists and sites to participate in a larger study, mine or yours.
  • I don’t have any funding beyond our division’s basic lab budget. I will be recruiting volunteer pathologists.
  • My level of commitment is high. I am willing to share equipment, help install equipment, and modify the eeDAP code (MATLAB) for different hardware and different tasks.

 

Getting  started:

  • I have created another NCIPhub group to coordinate our discussions. Please join https://nciphub.org (if you haven’t already) and then join this new group https://nciphub.org/groups/eedapstudies . This is your full opt-in with the ability to opt-out later. Please invite other colleagues that will be involved to join too. I will assume the first person listed for each group is the group’s lead point of contact.
  • I have stated my goals. Will you please share your goals with the group? Just one from each group by the end of next week please.
  • Once we finish this basic introduction, we will likely need a T-con.

 

Finally, I think you all should understand my approach to working together. As a federal regulator and scientist, everything I hear and say related to this work generally should be considered public information. As a federal scientist, I can freely talk and do science. Of course we can have a private conversation, but then I am a federal regulator and there are limits to what I can say and do. This doesn’t rule out a formal research agreement, but I’m hoping that won’t be necessary as it is a serious impediment to work.

 

I hope to hear back from you soon. All the best.

 

Brandon D. Gallas, PhD
Mathematician     

Center For Devices and Regulatory Health
Office of Science and Engineering Laboratories
U.S. Food and Drug Administration

Tel: 301-796-2531
Brandon.Gallas@fda.hhs.gov

Comments on this entry

There are no comments at this time.

Post a comment

You must log in to post comments.

Please keep comments relevant to this entry.

Line breaks and paragraphs are automatically converted. URLs (starting with http://) or email addresses will automatically be linked.