While exploring DICOM files, there are different numbers of .dcm files for each patient. CT Scan of the patient is taken by the first week of their treatment. one patient is having almost more than 1000 .dcm files. A number of slices portrait a full CT Scan image. but what is the reason behind the higher number of .dcm files?
", "messages": [ { "id": "930603", "postDate": "07/15/2020 15:30:53", "content": "While exploring DICOM files, there are different numbers of .dcm files for each patient. CT Scan of the patient is taken by the first week of their treatment. one patient is having almost more than 1000 .dcm files. A number of slices portrait a full CT Scan image. but what is the reason behind the higher number of .dcm files?
", "rawMarkdown": "While exploring DICOM files, there are different numbers of .dcm files for each patient. CT Scan of the patient is taken by the first week of their treatment. one patient is having almost more than 1000 .dcm files. A number of slices portrait a full CT Scan image. but what is the reason behind the higher number of .dcm files?", "votes": null }, { "id": "930682", "postDate": "07/15/2020 16:37:30", "content": "That's the nature of CT scans. Depends on what type of machine the scans were taken on, the parameters used (e.g. slice thickness, etc.). Also, when they take CT scans, where the \"chest\" portion of a CT starts and ends is typically at the discretion of a technician who will manually perform the cropping. Another factor is what kind of postprocessing was done after the DICOM images were collected. In short, there are hundreds of reasons why some scans have so few slices, and others have so many, but unfortunately this is the nature of working with real-life medical data.
", "rawMarkdown": "That's the nature of CT scans. Depends on what type of machine the scans were taken on, the parameters used (e.g. slice thickness, etc.). Also, when they take CT scans, where the \"chest\" portion of a CT starts and ends is typically at the discretion of a technician who will manually perform the cropping. Another factor is what kind of postprocessing was done after the DICOM images were collected. In short, there are hundreds of reasons why some scans have so few slices, and others have so many, but unfortunately this is the nature of working with real-life medical data.", "votes": null }, { "id": "931125", "postDate": "07/16/2020 02:14:02", "content": "Thanks a lot JJ. In that case, what kind of information, we can get from the each slice? or a combination of slices? The information like patient details, or FVC value like that..
", "rawMarkdown": "Thanks a lot JJ. In that case, what kind of information, we can get from the each slice? or a combination of slices? The information like patient details, or FVC value like that..", "votes": null }, { "id": "931174", "postDate": "07/16/2020 03:28:12", "content": "I think that's the $55,000 dollar question! @sandorkonya and @jryoungw explain some of the underlying concepts here and here respectively.
\n\nIf I had to venture a guess it would be that features such as fibrosis, consolidation (lung that appears squished), etc. that whatever model you're using can pick up, will provide baseline features of the patient to help you predict future FVC. It's possible that focusing on other things besides just the lungs can also give you information. For example, we don't have the patient's height, or weight, but perhaps the height of the ribcage and the amount of fat can provide an approximation of body mass index. I think there's a lot of room to get creative!
", "rawMarkdown": "I think that's the $55,000 dollar question! @sandorkonya and @jryoungw explain some of the underlying concepts [here ](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/165727)and [here ](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/165253)respectively. \n\nIf I had to venture a guess it would be that features such as fibrosis, consolidation (lung that appears squished), etc. that whatever model you're using can pick up, will provide baseline features of the patient to help you predict future FVC. It's possible that focusing on other things besides just the lungs can also give you information. For example, we don't have the patient's height, or weight, but perhaps the height of the ribcage and the amount of fat can provide an approximation of body mass index. I think there's a lot of room to get creative!", "votes": null } ], "comments": [ { "id": 930682, "author_name": "jjinho", "author_url": "", "post_date": "07/15/2020 16:37:30", "content": "That's the nature of CT scans. Depends on what type of machine the scans were taken on, the parameters used (e.g. slice thickness, etc.). Also, when they take CT scans, where the \"chest\" portion of a CT starts and ends is typically at the discretion of a technician who will manually perform the cropping. Another factor is what kind of postprocessing was done after the DICOM images were collected. In short, there are hundreds of reasons why some scans have so few slices, and others have so many, but unfortunately this is the nature of working with real-life medical data.
", "votes": null, "replies": [ { "id": 931125, "author_name": "subbuvolvosekar", "author_url": "", "post_date": "07/16/2020 02:14:02", "content": "Thanks a lot JJ. In that case, what kind of information, we can get from the each slice? or a combination of slices? The information like patient details, or FVC value like that..
", "votes": null, "replies": [] }, { "id": 931174, "author_name": "jjinho", "author_url": "", "post_date": "07/16/2020 03:28:12", "content": "I think that's the $55,000 dollar question! @sandorkonya and @jryoungw explain some of the underlying concepts here and here respectively.
\n\nIf I had to venture a guess it would be that features such as fibrosis, consolidation (lung that appears squished), etc. that whatever model you're using can pick up, will provide baseline features of the patient to help you predict future FVC. It's possible that focusing on other things besides just the lungs can also give you information. For example, we don't have the patient's height, or weight, but perhaps the height of the ribcage and the amount of fat can provide an approximation of body mass index. I think there's a lot of room to get creative!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "930603": "While exploring DICOM files, there are different numbers of .dcm files for each patient. CT Scan of the patient is taken by the first week of their treatment. one patient is having almost more than 1000 .dcm files. A number of slices portrait a full CT Scan image. but what is the reason behind the higher number of .dcm files?", "930682": "That's the nature of CT scans. Depends on what type of machine the scans were taken on, the parameters used (e.g. slice thickness, etc.). Also, when they take CT scans, where the \"chest\" portion of a CT starts and ends is typically at the discretion of a technician who will manually perform the cropping. Another factor is what kind of postprocessing was done after the DICOM images were collected. In short, there are hundreds of reasons why some scans have so few slices, and others have so many, but unfortunately this is the nature of working with real-life medical data.", "931125": "Thanks a lot JJ. In that case, what kind of information, we can get from the each slice? or a combination of slices? The information like patient details, or FVC value like that..", "931174": "I think that's the $55,000 dollar question! @sandorkonya and @jryoungw explain some of the underlying concepts [here ](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/165727)and [here ](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/165253)respectively. \n\nIf I had to venture a guess it would be that features such as fibrosis, consolidation (lung that appears squished), etc. that whatever model you're using can pick up, will provide baseline features of the patient to help you predict future FVC. It's possible that focusing on other things besides just the lungs can also give you information. For example, we don't have the patient's height, or weight, but perhaps the height of the ribcage and the amount of fat can provide an approximation of body mass index. I think there's a lot of room to get creative!" }, "source": "meta" }