Datasets:

ibrahimhamamci

/

CT-RATE

like 48

Hi @farrell236 , The automatic annotation algorithm is trained on another set of reports from the same hospital randomly. They might and might not be in the CT-RATE. All classification labels provided in CT-RATE are from the inference of the label classifier model. We provide training and validation labels (not inferred) for the classifier in the GitHub repository separately: https://github.com/ibrahimethemhamamci/CT-CLIP/tree/main/text_classifier/data

I can ask the person who worked on the label classifier to provide me the accession numbers of the ones that are also in CT-RATE, but again not all of them are in the CT-RATE.

Hi @sezginer , if you can identify the overlap between that dataset and CT-RATE via the accession numbers, that would be super helpful! thanks! 🙂

Hi @farrell236 , sorry for the delay, I added the mappings to the Github repo: https://github.com/ibrahimethemhamamci/CT-CLIP/tree/main/text_classifier/data

NameinCTRATE column shows the directory path (e.g. valid_128_a). This is not the case in reports.csv files (e.g. valid_128_a_1.nii.gz, and valid_128_a_2.nii.gz) but reports are same for both of them so you can use any. Hope this helps!

Thank you so much @sezginer !! 🙂 This definitely helps

Hi @sezginer , thanks very much for all the help with mapping the accessions in CT-RATE. Since there are multiple studies for some patients, @farrell236 and I were wondering if the time difference between the studies for the same patient could be calculated? We want to assess longitudinal changes in patients who underwent serial imaging within a short timespan (e.g., 6 months - 1 year) and we can't properly assess the reports for changes without the time difference between consecutive studies. May we request your help with that?

Hi @tsm91 , time of the study is added as metadata column (StudyDate) in metadata files. You can calculate time difference between studies with that.

Thank you @sezginer , appreciate the help!