Brandon started out with a bunch of announcements and reminders (see below). After, Ajay introduced some of the background for the study they are planning. The main concept is that they want to collect mulit-level information for individual cells. The multi-levels correlate with the confidence or level of agreement from pathologists that the cell is indeed a mitotic figure. This kind of data recognizes that the decisions given the slides and images are not binary. Pathologists do not agree 100%. Pathologists are not 100% sure about their decisions. We ran out of time but hope to discuss this at PI and hear more about how Inspirata might use this kind of data.
Reminder about Pathology Informatics program
- We have our meeting Sunday May 21 1:30pm-5pm, King’s Garden 3.
- Brandon is presenting an update on our project to the rest of the conference on Tuesday May 23 at 9am, Grand Ballroom 1.
- Qi Gong will be manning a table in the exhibitor hall, demonstrating eeDAP. We will also have posters.
- Do you want to present or provide a poster? Let Brandon and Marios know right away. We are finalizing the agenda this week.
Update on MDDT proposal
We received a preliminary response from FDA on the eeDAP MDDT proposal. I have put it on the wsi working group blog (LINK).
- The response explains the delay in the full response and it provides a limited response to our question about providing only technical performance results in the MDDT, no reader study data.
- “We are open to discussing the COU and the associated evidence as discussions about the MDDT proceed. The criteria for qualification do not necessitate clinical data, the evidence will primarily depend upon the MDDT and the context of use.“
- They say a full response will be coming within two weeks. This would be in time for us to discuss at Pathology Informatics.
- Will someone revise the MDDT proposal such that the supporting evidence would only be technical performance data (registration accuracy and color reproduction)?
We received a definitive response on the eeDAP MDDT proposal. eeDAP is accepted into the incubator phase of the MDDT program Here is a link to the acceptance letter.
Open submission or Mock submission
- Brandon and Marios have been given permission to consult on an open submission or a mock submission to the FDA.
- An open submission would be one that makes public (through a journal publication or other means) the discussion and content related to the consult.
- More details coming.
Some project members have run eeDAP
- IIT was the first. Philips was the second. Have you tried? Let Brandon know. It is the first step to running an eeDAP study.
- We have archived the questions and responses that helped make this possible eeDAPregistrationquestions, eeDAPwithImageScope. We have also updated the user manual and provided fresh resources (sample WSI and corresponding eeDAP input file). You may want to scan these discussions if you struggle getting eeDAP running.
- We have uploaded two videos demonstrating eeDAP, including registration. Check them out here!
- eeDAP successfully opened BIGtif!
- Qi Gong at the FDA has conducted preliminary registration accuracy studies. We have 1 failure out of 20 tasks/FOVs in “Fast Registation” mode. This failure is fixed when the “Best Registration” button is pressed. Except for the failure, we are within 2.5 microns.
- First draft of slides outlining study can be found here PI2017-Gong-eeDAPregistrationAccuracy-1.pdf (654 KB, uploaded by Brandon D. Gallas 4 years 8 months ago).
Funding for WSI WG meeting
I am working on the AV logistics of our WSI working group meeting on Sunday. The cost of the AV is less than $1000. Anyone willing to pay?
Please update. This list is derived from a screen shot of the WebEx participants.
- Darren Treanor at Leeds
- 516-455-8362, NY
- 646-888-2001, NY
- 813-570-8900, FL
- IIT Madras
- James Crawford, Northwell Health
- Marios Gavrielides, Qi Gong, Weijie Chen, FDA
- Partha Mitra, CSHL
- Peter Schueffler, MSKCC
- Prarthana Shrestha, Philips