Hi all: excited for the competition! I noticed that the non-brain tissue has been removed from the dicom images (which is certainly a good thing from the perspective of data privacy) but I don't see any mention of that in the data documentation. Could you let us know
\na) what tool was used
\nand
\nb) what other pre-processing, if any, was applied to the data?
\nThanks
\nRichard
", "messages": [ { "id": "1387501", "postDate": "07/14/2021 08:35:04", "content": "Hi all: excited for the competition! I noticed that the non-brain tissue has been removed from the dicom images (which is certainly a good thing from the perspective of data privacy) but I don't see any mention of that in the data documentation. Could you let us know
\na) what tool was used
\nand
\nb) what other pre-processing, if any, was applied to the data?
\nThanks
\nRichard
", "rawMarkdown": "Hi all: excited for the competition! I noticed that the non-brain tissue has been removed from the dicom images (which is certainly a good thing from the perspective of data privacy) but I don't see any mention of that in the data documentation. Could you let us know \n\na) what tool was used\n\nand \n\nb) what other pre-processing, if any, was applied to the data?\n\nThanks\n\nRichard", "votes": null }, { "id": "1388803", "postDate": "07/15/2021 08:35:59", "content": "Hey Richard,
\nCheck https://arxiv.org/abs/2107.02314,
\nThey used the CaPTk preprocessing pipeline with this skull-stripping tool:
\nhttps://www.sciencedirect.com/science/article/pii/S105381192030567X
\nFan of you work here! Good luck!
", "rawMarkdown": "Hey Richard, \n\nCheck https://arxiv.org/abs/2107.02314, \n\nThey used the CaPTk preprocessing pipeline with this skull-stripping tool: \n\nhttps://www.sciencedirect.com/science/article/pii/S105381192030567X\n\nFan of you work here! Good luck!", "votes": null }, { "id": "1485335", "postDate": "08/22/2021 02:43:37", "content": "I'm still not clear about the pre-processing that was done to produce the contest data. The paper at https://arxiv.org/abs/2107.02314 says:
\nStandardized pre-processing has been applied to all the BraTS mpMRI scans.Specifically, the applied pre-processing routines include conversion of the DI-COM files to the NIFTI file format [16], re-orientation to a common orientation system (i.e., RAI), co-registration to the same anatomical template (SRI24) [17], resampling to a uniform isotropic resolution (1mm3), and finally skull-stripping.
\nNote that the above says the files were converted from DICOM to NIFTI, but the data for this contest is provided in DICOM form. So my question is: what pre-processing steps were actually applied?
", "rawMarkdown": "I'm still not clear about the pre-processing that was done to produce the contest data. The paper at https://arxiv.org/abs/2107.02314 says:\n\nStandardized pre-processing has been applied to all the BraTS mpMRI scans.Specifically, the applied pre-processing routines include conversion of the DI-COM files to the NIFTI file format [16], re-orientation to a common orientation system (i.e., RAI), co-registration to the same anatomical template (SRI24) [17], resampling to a uniform isotropic resolution (1mm3), and finally skull-stripping.\n\nNote that the above says the files were converted from DICOM to NIFTI, but the data for this contest is provided in DICOM form. So my question is: what pre-processing steps were actually applied?", "votes": null }, { "id": "1485363", "postDate": "08/22/2021 03:28:15", "content": "I just noticed that later the article says:
\nFor Task 2 (Radiogenomic Classification), all the imaging volumes were converted from NIFTI to DICOM files, while ensuring that the original patient space is preserved
\nSo that answers my question.
", "rawMarkdown": "I just noticed that later the article says:\n\nFor Task 2 (Radiogenomic Classification), all the imaging volumes were converted from NIFTI to DICOM files, while ensuring that the original patient space is preserved\n\nSo that answers my question.", "votes": null }, { "id": "1485624", "postDate": "08/22/2021 09:26:05", "content": "\"while ensuring that the original patient space is preserved\", this line make me feel unclear about the orientation of the images. Does it mean Task 2 data is RAI base (NIFTY format -> DICOM) or is it converted back to whatever the orientation of the original DICOM files ? If it's the latter, we still need to normalize Task 2 images to the same base (whether it's RAI or LPS)
", "rawMarkdown": "\"while ensuring that the original patient space is preserved\", this line make me feel unclear about the orientation of the images. Does it mean Task 2 data is RAI base (NIFTY format -> DICOM) or is it converted back to whatever the orientation of the original DICOM files ? If it's the latter, we still need to normalize Task 2 images to the same base (whether it's RAI or LPS)", "votes": null }, { "id": "1486201", "postDate": "08/22/2021 18:14:26", "content": "Yes, I think you're correct, the details of the pre-processing are still not clear.
", "rawMarkdown": "Yes, I think you're correct, the details of the pre-processing are still not clear.", "votes": null }, { "id": "1486804", "postDate": "08/23/2021 08:21:23", "content": "I think this description is not correct, because the data is not resampled to a uniform isotropic resolution (1mm3).
\nI would appreciate it if the host team ( @cdcarr, @ujjwalbaid, @sbakas, or @juliaelliott) could answer the questions raised in this topic.
processing for 1 task was done using the CaPTk program https://github.com/CBICA/CaPTk - as I understand it. This program has a separate item, by selecting which, you will do all the preprocessing + segmentation. one problem - it will take about 25 minutes per patient.
", "rawMarkdown": "processing for 1 task was done using the CaPTk program https://github.com/CBICA/CaPTk - as I understand it. This program has a separate item, by selecting which, you will do all the preprocessing + segmentation. one problem - it will take about 25 minutes per patient.", "votes": null }, { "id": "1489268", "postDate": "08/24/2021 20:52:12", "content": "The images were converted back to their original orientation and resolution (original scanner acquisition). Further details will be provided in the Github page (https://github.com/CBICA/CaPTk) and manuscript (https://arxiv.org/abs/2107.02314) soon.
", "rawMarkdown": "The images were converted back to their original orientation and resolution (original scanner acquisition). Further details will be provided in the Github page (https://github.com/CBICA/CaPTk) and manuscript (https://arxiv.org/abs/2107.02314) soon.", "votes": null } ], "comments": [ { "id": 1388803, "author_name": "ranafago", "author_url": "", "post_date": "07/15/2021 08:35:59", "content": "Hey Richard,
\nCheck https://arxiv.org/abs/2107.02314,
\nThey used the CaPTk preprocessing pipeline with this skull-stripping tool:
\nhttps://www.sciencedirect.com/science/article/pii/S105381192030567X
\nFan of you work here! Good luck!
", "votes": null, "replies": [ { "id": 1485335, "author_name": "dslate", "author_url": "", "post_date": "08/22/2021 02:43:37", "content": "I'm still not clear about the pre-processing that was done to produce the contest data. The paper at https://arxiv.org/abs/2107.02314 says:
\nStandardized pre-processing has been applied to all the BraTS mpMRI scans.Specifically, the applied pre-processing routines include conversion of the DI-COM files to the NIFTI file format [16], re-orientation to a common orientation system (i.e., RAI), co-registration to the same anatomical template (SRI24) [17], resampling to a uniform isotropic resolution (1mm3), and finally skull-stripping.
\nNote that the above says the files were converted from DICOM to NIFTI, but the data for this contest is provided in DICOM form. So my question is: what pre-processing steps were actually applied?
", "votes": null, "replies": [] }, { "id": 1485363, "author_name": "dslate", "author_url": "", "post_date": "08/22/2021 03:28:15", "content": "I just noticed that later the article says:
\nFor Task 2 (Radiogenomic Classification), all the imaging volumes were converted from NIFTI to DICOM files, while ensuring that the original patient space is preserved
\nSo that answers my question.
", "votes": null, "replies": [] }, { "id": 1485624, "author_name": "tungvs", "author_url": "", "post_date": "08/22/2021 09:26:05", "content": "\"while ensuring that the original patient space is preserved\", this line make me feel unclear about the orientation of the images. Does it mean Task 2 data is RAI base (NIFTY format -> DICOM) or is it converted back to whatever the orientation of the original DICOM files ? If it's the latter, we still need to normalize Task 2 images to the same base (whether it's RAI or LPS)
", "votes": null, "replies": [] }, { "id": 1486201, "author_name": "dslate", "author_url": "", "post_date": "08/22/2021 18:14:26", "content": "Yes, I think you're correct, the details of the pre-processing are still not clear.
", "votes": null, "replies": [] }, { "id": 1486804, "author_name": "tomooinubushi", "author_url": "", "post_date": "08/23/2021 08:21:23", "content": "I think this description is not correct, because the data is not resampled to a uniform isotropic resolution (1mm3).
\nI would appreciate it if the host team ( @cdcarr, @ujjwalbaid, @sbakas, or @juliaelliott) could answer the questions raised in this topic.
processing for 1 task was done using the CaPTk program https://github.com/CBICA/CaPTk - as I understand it. This program has a separate item, by selecting which, you will do all the preprocessing + segmentation. one problem - it will take about 25 minutes per patient.
", "votes": null, "replies": [] } ] }, { "id": 1489268, "author_name": "lechuck0", "author_url": "", "post_date": "08/24/2021 20:52:12", "content": "The images were converted back to their original orientation and resolution (original scanner acquisition). Further details will be provided in the Github page (https://github.com/CBICA/CaPTk) and manuscript (https://arxiv.org/abs/2107.02314) soon.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1387501": "Hi all: excited for the competition! I noticed that the non-brain tissue has been removed from the dicom images (which is certainly a good thing from the perspective of data privacy) but I don't see any mention of that in the data documentation. Could you let us know \n\na) what tool was used\n\nand \n\nb) what other pre-processing, if any, was applied to the data?\n\nThanks\n\nRichard", "1388803": "Hey Richard, \n\nCheck https://arxiv.org/abs/2107.02314, \n\nThey used the CaPTk preprocessing pipeline with this skull-stripping tool: \n\nhttps://www.sciencedirect.com/science/article/pii/S105381192030567X\n\nFan of you work here! Good luck!", "1485335": "I'm still not clear about the pre-processing that was done to produce the contest data. The paper at https://arxiv.org/abs/2107.02314 says:\n\nStandardized pre-processing has been applied to all the BraTS mpMRI scans.Specifically, the applied pre-processing routines include conversion of the DI-COM files to the NIFTI file format [16], re-orientation to a common orientation system (i.e., RAI), co-registration to the same anatomical template (SRI24) [17], resampling to a uniform isotropic resolution (1mm3), and finally skull-stripping.\n\nNote that the above says the files were converted from DICOM to NIFTI, but the data for this contest is provided in DICOM form. So my question is: what pre-processing steps were actually applied?", "1485363": "I just noticed that later the article says:\n\nFor Task 2 (Radiogenomic Classification), all the imaging volumes were converted from NIFTI to DICOM files, while ensuring that the original patient space is preserved\n\nSo that answers my question.", "1485624": "\"while ensuring that the original patient space is preserved\", this line make me feel unclear about the orientation of the images. Does it mean Task 2 data is RAI base (NIFTY format -> DICOM) or is it converted back to whatever the orientation of the original DICOM files ? If it's the latter, we still need to normalize Task 2 images to the same base (whether it's RAI or LPS)", "1486201": "Yes, I think you're correct, the details of the pre-processing are still not clear.", "1486804": "I think this description is not correct, because the data is not resampled to a uniform isotropic resolution (1mm3).\nI would appreciate it if the host team ( @cdcarr, @ujjwalbaid, @sbakas, or @juliaelliott) could answer the questions raised in this topic.", "1487817": "processing for 1 task was done using the CaPTk program https://github.com/CBICA/CaPTk - as I understand it. This program has a separate item, by selecting which, you will do all the preprocessing + segmentation. one problem - it will take about 25 minutes per patient.", "1489268": "The images were converted back to their original orientation and resolution (original scanner acquisition). Further details will be provided in the Github page (https://github.com/CBICA/CaPTk) and manuscript (https://arxiv.org/abs/2107.02314) soon." }, "source": "meta" }