Trials import / upload

[dkfzsinn]

Dear mindminer-curate Dev’s,

We wanted to evaluate mindminer-cruate for a project, but having trouble in getting it properly running.

I followed carefully all of the “Preparation” instructions from the git page.

The “yarn install” command runs, except some warnings, successfully.
The “yarn start” command runs, except one warning, successfully.
I can navigate to the WEB-UI at IP_ADRESS:9000 as well.

Trials-tab looks like this:

Data-tab looks like this:

The problem I have now is to get data/trials from clinicaltrials.gov (all in json format) into the application.
The upload function from the “Data”-tab does not work for me at all.
Adding some data with console.firebase.google.com manually or via the json upload function to the firebase realtime database did not helped.

More details:
OS: CentOS Linux release 7.8.2003 (Core)
yarn install logfile: yarn_install.log
yarn start logfile: yarn_start.log

Do you have any idea what the problem could be or how to debug further?
Thanks in advance for your time and help.
Best regards
Marcel

[zhx828]

@dkfzsinn thanks for using the tool. Did you try to import a trial with NCT id in the Trials tab using the search box on the top?

[dkfzsinn]

@dkfzsinn thanks for using the tool. Did you try to import a trial with NCT id in the Trials tab using the search box on the top?

Thanks a lot for your quick reply and offering your tool to the public!

I already tried to import a NCT id e.g. “NCT02057146” which leads to the message:
“NCT02057146 not found”

Just out of curiosity, I tried “NCT02057133” which is displayed as an example for the trial format, which surprisingly worked.

I downloaded the trial as json from the Data-tab, deleted it from the database and tried this time to import it via the upload functionality from the Data-tab with no success.

Is this the expected behavior because it’s not build for production and was just started with “yarn start”?

Another question: I looked into the json of “NCT02057133” after I downloaded it as described above and saw it has a different structure compared to trials downloaded from clinicaltrials.gov
Does matchminer-curate understand the clinicaltrials.gov jsons aswell or do they need to be formatted exactly as the previous downloaded “NCT02057133” json ?

Thanks in advance for your time and help.

Best regards
Marcel

[zhx828]

I already tried to import a NCT id e.g. “NCT02057146” which leads to the message:
“NCT02057146 not found”

Just out of curiosity, I tried “NCT02057133” which is displayed as an example for the trial format, which surprisingly worked.

We use the API from cancer.gov instead. There may be minor data difference comparing with clinicaltrails.gov https://clinicaltrialsapi.cancer.gov/

I downloaded the trial as json from the Data-tab, deleted it from the database and tried this time to import it via the upload functionality from the Data-tab with no success.

Is this the expected behavior because it’s not build for production and was just started with “yarn start”?

This is a good question. The feature is mainly for internal usage. I haven't used it for awhile but happy to investigate if you cannot find out the issue.

Another question: I looked into the json of “NCT02057133” after I downloaded it as described above and saw it has a different structure compared to trials downloaded from clinicaltrials.gov
Does matchminer-curate understand the clinicaltrials.gov jsons aswell or do they need to be formatted exactly as the previous downloaded “NCT02057133” json ?

We has specific data structure for matchminor-curate, the “NCT02057133” json you downloaded. We do not use data from clinicaltrials.gov

[zhx828]

@dkfzsinn BTW, if you have a gmail account, I can give you access to one of our instances which includes more data. That may help you get more familiar with the app.

[dkfzsinn]

@dkfzsinn BTW, if you have a gmail account, I can give you access to one of our instances which includes more data. That may help you get more familiar with the app.

Hello @zhx828 , thanks again for your quick response.
The information about using clinicaltrialsapi.cancer.gov helped a lot already, thanks for that !

From the trials I've imported so far, I see the "Treatment List" for curation.
Is it possible to do the same with "Eligibility Criteria", or is the tool meant for "Treatment List" curation only ?

As always thanks for your time, answer and help in advance.

Best regards
Marcel

[zhx828]

@dkfzsinn the section is the core of this tool. The information could come from Eligibility Criteria or other parts of the trial determines on how you interpret the trial. Once you import the trial from cancer.gov, you need to curate the information under Treatment List.

Worth to mention, this file structure(CTML) eventually is feeding into https://matchminer.org/ . Not sure whether it's the use case for you.

[dkfzsinn]

@zhx828 I just had a meeting regarding these "Eligibility Criteria".

What we are still unsure about is the following

Example Trial: NCT02057133
Request URL: https://clinicaltrialsapi.cancer.gov/v1/clinical-trial/NCT02057133
curl: curl -X GET --header 'Accept: application/json' 'https://clinicaltrialsapi.cancer.gov/v1/clinical-trial/NCT02057133'

I guess this is the API call matchminer-curate does when I try to import via the search box.
If I look manually into the response, I have e.g. this part :

{ "assay_purpose": "Eligibility Criterion - Inclusion", "long_name": "HER2/Neu Negative", "name": "HER2/Neu Negative", "synonyms": [ "HER-2 Negative", "HER2 Negative", "ERBB2 Negative" ], "hugo_biomarker_code": null, "nci_thesaurus_concept_id": "C68749", "eligibility_criterion": "inclusion" }

So we would like to curate (further) this "Inclusion Eligibility Criterion" but it's not shown within the "Treatment List".
Is it because it's "already curated" ?
If so, I thought the data would be included if I download it via the Data-Tab, but that's not the case.

The question would be how not lose the data and get the "Inclusion Eligibility Criterion" shown in matchminer-curate ?

The CTML part is important for us as well, since we've planned to move the curated trial data afterwards to matchminer.

[zhx828]

@dkfzsinn we do not import Inclusion/Exclusion criteria from the API since often time it's not well structured. Rather, you should curate them by manually adding sections under Treatment List.