OHDSI Home | Forums | Wiki | Github

Hematologic Malignancy Response Criteria


(JD Liddil) #1

Another malignancy I trying to figure out coding


(Christian Reich) #2

@jliddil1:

Here is the deal:

The overall disease states like Complete Remission, Partial Remission, Stable Disease, Progressive Disease, Relapsed Disease are going to be stored in the new EPISODE table. It will be filled either from pre-abstracted (mostly registry) data, or from abstraction.

The detailed attributes are tumor attributes need to be developed. The bad news is they differ for each tumor type. For lymphomas, they are topology of lesions, radiographic characterization of lesions (LDi, SDi, SPD, PPD), extralymphatic manifestations, PET results and scores, morphology, etc. Hopefully, we can steal a big chunk of these from available sources. But we will need community help to nail them.

We are also interested in engaging with folks who are developing automatic abstraction methods, which take the tumor markers and create the disease status.

Makes sense?


(JD Liddil) #3

Makes perfect sense. I need to get on the calls to help out. I have two CTRs in house I can go to for stuff. I know most of the codes etc. I know the differ for each form of cancer since I’ve been in the field all my working life. BTW how do I find info on the WG? Can’t seem to find it


(Michael Gurley) #4

Here is the information on the Oncology Workgroup meeting:

https://www.ohdsi.org/web/wiki/doku.php?id=projects:workgroups:oncology-sg


(Seng Chan You) #5

@Christian_Reich Ajou university is working on developing an automatic abstract method for identification of regimens, which will be presented in the OHDSI symposium 2019 as a poster. In addition, we’re developing an NLP tool, which can be used to annotate tumor markers and disease status.This was presented in the OHDSI call recently, and will be presented in the OHDSI symposium as a software demo.


(JD Liddil) #6

Less than straight forward to find the info on the group. Thanks. In our case we are abstracting a data set of oncology data including histology and pathology data that we then combine with NGS, WES and RNA data to predict disease response for primarily immuno-oncology medications.


(Christian Reich) #7

@jliddil1:

Yeah. Our fault. Will put them on the WG list properly.

Folks in the WG are converting our respective data to the standard we are developing and testing them with use cases. Want to join in that? It will make a splash I believe.


(JD Liddil) #8

I’ve add the standing meeting to my calendar


t