Hi fellow kagglers,
\nafter wondering about the aortic_hu meta data value and its description, I realized I didn't quite understand how a multiphasic CT scan works. I found this nice explanation including example images though: https://www.youtube.com/watch?v=4ggULZ7trGI
\nDoes someone know if the data contains different phases for the same patient?
Some of the patients have more than one scan, some do not.
", "rawMarkdown": "Some of the patients have more than one scan, some do not.", "votes": 2, "replies": [ { "id": 2395630, "postDate": "2023-08-17T17:32:08.587Z", "content": "I'm not only referring to more than one scan for the same patient, but scans with different phases. Based on this summary https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/431475, the different series could correspond to different phases for the same patient, but it could also just be repeated scans (maybe for healthy patients who are scanned not because of an immediate injury, not sure if injured patients would just have repeated scans).
", "rawMarkdown": "I'm not only referring to more than one scan for the same patient, but scans with different phases. Based on this summary https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/431475, the different series could correspond to different phases for the same patient, but it could also just be repeated scans (maybe for healthy patients who are scanned not because of an immediate injury, not sure if injured patients would just have repeated scans).", "votes": 1, "replies": [ { "id": 2396041, "postDate": "2023-08-18T03:03:16.990Z", "content": "I used the term \"scan\" loosely here. Some of the patient directories contain more than one \"phase\" of the same study. Some are just a single non-contrast series. We can verify if the series are from the same study by comparing their StudyInstanceUID tags.
\nIf the series share the same StudyInstanceUID tag, there's a good chance the patient is positioned in the same place for each series and did not move between them. This means we can \"register\" the series together in 3D space and the anatomy should align. Cool things like digital subtraction, MIPing and 3D rendering can happen with these contrasted scans, given enough slices.
\nI agree though, some series might not be part of the same study and could be repeats or, rechecks. I haven't looked through the tags to verify.
", "rawMarkdown": "I used the term \"scan\" loosely here. Some of the patient directories contain more than one \"phase\" of the same study. Some are just a single non-contrast series. We can verify if the series are from the same study by comparing their StudyInstanceUID tags. \n\nIf the series share the same StudyInstanceUID tag, there's a good chance the patient is positioned in the same place for each series and did not move between them. This means we can \"register\" the series together in 3D space and the anatomy should align. Cool things like digital subtraction, MIPing and 3D rendering can happen with these contrasted scans, given enough slices.\n\nI agree though, some series might not be part of the same study and could be repeats or, rechecks. I haven't looked through the tags to verify.", "votes": 1 }, { "id": 2396267, "postDate": "2023-08-18T06:54:25.650Z", "content": "Thanks a lot for your detailed answer and the explanation of the StudyInstance UID tag!
", "rawMarkdown": "Thanks a lot for your detailed answer and the explanation of the StudyInstance UID tag!" } ] } ] }, { "id": 2395186, "postDate": "2023-08-17T10:54:21.437Z", "content": "Hi fellow kagglers,
\nafter wondering about the aortic_hu meta data value and its description, I realized I didn't quite understand how a multiphasic CT scan works. I found this nice explanation including example images though: https://www.youtube.com/watch?v=4ggULZ7trGI
\nDoes someone know if the data contains different phases for the same patient?
In case some else has the same question, https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/427226 answers it very nicely!
", "rawMarkdown": "In case some else has the same question, https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/427226 answers it very nicely!" } ], "comments": [ { "id": 2395232, "author_name": "David Roberts", "author_url": "", "post_date": "2023-08-17T11:46:45.360000", "content": "Some of the patients have more than one scan, some do not.
", "votes": 2, "replies": [ { "id": 2395630, "author_name": "Michael Wolff", "author_url": "", "post_date": "2023-08-17T17:32:08.587000", "content": "I'm not only referring to more than one scan for the same patient, but scans with different phases. Based on this summary https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/431475, the different series could correspond to different phases for the same patient, but it could also just be repeated scans (maybe for healthy patients who are scanned not because of an immediate injury, not sure if injured patients would just have repeated scans).
", "votes": 1, "replies": [ { "id": 2396041, "author_name": "David Roberts", "author_url": "", "post_date": "2023-08-18T03:03:16.990000", "content": "I used the term \"scan\" loosely here. Some of the patient directories contain more than one \"phase\" of the same study. Some are just a single non-contrast series. We can verify if the series are from the same study by comparing their StudyInstanceUID tags.
\nIf the series share the same StudyInstanceUID tag, there's a good chance the patient is positioned in the same place for each series and did not move between them. This means we can \"register\" the series together in 3D space and the anatomy should align. Cool things like digital subtraction, MIPing and 3D rendering can happen with these contrasted scans, given enough slices.
\nI agree though, some series might not be part of the same study and could be repeats or, rechecks. I haven't looked through the tags to verify.
", "votes": 1, "replies": [] }, { "id": 2396267, "author_name": "Michael Wolff", "author_url": "", "post_date": "2023-08-18T06:54:25.650000", "content": "Thanks a lot for your detailed answer and the explanation of the StudyInstance UID tag!
", "votes": 0, "replies": [] } ] } ] }, { "id": 2395775, "author_name": "Michael Wolff", "author_url": "", "post_date": "2023-08-17T20:04:24.003000", "content": "In case some else has the same question, https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/427226 answers it very nicely!
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "2395232": "Some of the patients have more than one scan, some do not.", "2395186": "Hi fellow kagglers, \nafter wondering about the aortic_hu meta data value and its description, I realized I didn't quite understand how a multiphasic CT scan works. I found this nice explanation including example images though: https://www.youtube.com/watch?v=4ggULZ7trGI\nDoes someone know if the data contains different phases for the same patient?", "2395775": "In case some else has the same question, https://www.kaggle.com/competitions/rsna-2023-abdominal-trauma-detection/discussion/427226 answers it very nicely!" } }