{
  "id": 264073,
  "title": "hounsfield units ranges for brain scans",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/264073",
  "author_name": "",
  "post_date": "2021-08-11T02:34:06.874112400Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Any resource for mapping brain sections to HU units?</p>\n<p>Thanks</p>",
  "messages": [
    {
      "id": "1465339",
      "postDate": "08/11/2021 02:34:06",
      "content": "<p>Any resource for mapping brain sections to HU units?</p>\n<p>Thanks</p>",
      "rawMarkdown": "Any resource for mapping brain sections to HU units?\n\nThanks",
      "votes": null
    },
    {
      "id": "1466008",
      "postDate": "08/11/2021 09:11:39",
      "content": "<p>+1 on this</p>\n<p>It's not very clear to me what the units are in the MRI scans. It doesn't look as clear as the CT scans that have HU. I am wondering whether there even is a standard. I was not able to find an answer to this, so if anyone knows, that would be super helpful.</p>",
      "rawMarkdown": "1 on this\n\nIt's not very clear to me what the units are in the MRI scans. It doesn't look as clear as the CT scans that have HU. I am wondering whether there even is a standard. I was not able to find an answer to this, so if anyone knows, that would be super helpful.",
      "votes": null
    },
    {
      "id": "1466330",
      "postDate": "08/11/2021 12:21:07",
      "content": "<p>H.U. is not used in MR. It's a CT concept only.</p>",
      "rawMarkdown": "H.U. is not used in MR. It's a CT concept only.",
      "votes": null
    },
    {
      "id": "1466367",
      "postDate": "08/11/2021 12:38:12",
      "content": "<p>I haven't been able to find any information about units used in MRI scans.</p>\n<p>My basis for using HU has been this paper <a href=\"https://pubmed.ncbi.nlm.nih.gov/24387496/\" target=\"_blank\">https://pubmed.ncbi.nlm.nih.gov/24387496/</a> <br>\nAlthough it does conclusively provide a basis for using HU in MRI, this gave me the idea that CT and MRI units might be correlated.</p>\n<p>Would really appreciate any advice on quantifying the scans.</p>\n<p>Thanks</p>",
      "rawMarkdown": "I haven't been able to find any information about units used in MRI scans.\n\nMy basis for using HU has been this paper https://pubmed.ncbi.nlm.nih.gov/24387496/ \nAlthough it does conclusively provide a basis for using HU in MRI, this gave me the idea that CT and MRI units might be correlated.\n\nWould really appreciate any advice on quantifying the scans.\n\nThanks",
      "votes": null
    },
    {
      "id": "1466505",
      "postDate": "08/11/2021 13:46:32",
      "content": "<p>There isn't a concept in MR similar to HU because of the physics involved. HU is a mapping of densities on a CT to densities in the real world. MR does not provide density information in the way CT does.</p>\n<p>I.E. in CT, the x-ray beam passes through some tissue and creates a shadow on a detector. By comparing the attenuation (how much radiation was blocked by the tissue) to the attenuation of a known density (water) .. we get Hounsfield Units. Then, we can say Tissue A is X times as dense as water. </p>\n<p>In MR, the signals come from magnetizing the tissue (thus aligning the magnetic poles in the same direction), and then releasing them to relax back to their normal state .. radio waves are then given off from the patient's body and are detected by the machine. Different tissues relax at different rates. So when you see an MR, you're not seeing density information like in CT. (this is a basic overview of a simple MR sequence .. it's a little more complex than this description).</p>\n<p>In the paper you referenced, they combined CT and MR images to produce a pseudo HU .. that's not practical (or even possible in some cases) in the real world.</p>",
      "rawMarkdown": "There isn't a concept in MR similar to HU because of the physics involved. HU is a mapping of densities on a CT to densities in the real world. MR does not provide density information in the way CT does.\n\nI.E. in CT, the x-ray beam passes through some tissue and creates a shadow on a detector. By comparing the attenuation (how much radiation was blocked by the tissue) to the attenuation of a known density (water) .. we get Hounsfield Units. Then, we can say Tissue A is X times as dense as water. \n\nIn MR, the signals come from magnetizing the tissue (thus aligning the magnetic poles in the same direction), and then releasing them to relax back to their normal state .. radio waves are then given off from the patient's body and are detected by the machine. Different tissues relax at different rates. So when you see an MR, you're not seeing density information like in CT. (this is a basic overview of a simple MR sequence .. it's a little more complex than this description).\n\nIn the paper you referenced, they combined CT and MR images to produce a pseudo HU .. that's not practical (or even possible in some cases) in the real world.",
      "votes": null
    },
    {
      "id": "1469389",
      "postDate": "08/12/2021 22:40:35",
      "content": "<p>+1 ; <a href=\"https://www.kaggle.com/davidbroberts\" target=\"_blank\">@davidbroberts</a> is absolutely right. there is no concept like HU in MR. you can ignore the paper you referenced :) </p>",
      "rawMarkdown": "1 ; @davidbroberts is absolutely right. there is no concept like HU in MR. you can ignore the paper you referenced :)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1466008,
      "author_name": "smoschou55",
      "author_url": "",
      "post_date": "08/11/2021 09:11:39",
      "content": "<p>+1 on this</p>\n<p>It's not very clear to me what the units are in the MRI scans. It doesn't look as clear as the CT scans that have HU. I am wondering whether there even is a standard. I was not able to find an answer to this, so if anyone knows, that would be super helpful.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1466330,
      "author_name": "davidbroberts",
      "author_url": "",
      "post_date": "08/11/2021 12:21:07",
      "content": "<p>H.U. is not used in MR. It's a CT concept only.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1466367,
          "author_name": "aryamansharma47",
          "author_url": "",
          "post_date": "08/11/2021 12:38:12",
          "content": "<p>I haven't been able to find any information about units used in MRI scans.</p>\n<p>My basis for using HU has been this paper <a href=\"https://pubmed.ncbi.nlm.nih.gov/24387496/\" target=\"_blank\">https://pubmed.ncbi.nlm.nih.gov/24387496/</a> <br>\nAlthough it does conclusively provide a basis for using HU in MRI, this gave me the idea that CT and MRI units might be correlated.</p>\n<p>Would really appreciate any advice on quantifying the scans.</p>\n<p>Thanks</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1466505,
          "author_name": "davidbroberts",
          "author_url": "",
          "post_date": "08/11/2021 13:46:32",
          "content": "<p>There isn't a concept in MR similar to HU because of the physics involved. HU is a mapping of densities on a CT to densities in the real world. MR does not provide density information in the way CT does.</p>\n<p>I.E. in CT, the x-ray beam passes through some tissue and creates a shadow on a detector. By comparing the attenuation (how much radiation was blocked by the tissue) to the attenuation of a known density (water) .. we get Hounsfield Units. Then, we can say Tissue A is X times as dense as water. </p>\n<p>In MR, the signals come from magnetizing the tissue (thus aligning the magnetic poles in the same direction), and then releasing them to relax back to their normal state .. radio waves are then given off from the patient's body and are detected by the machine. Different tissues relax at different rates. So when you see an MR, you're not seeing density information like in CT. (this is a basic overview of a simple MR sequence .. it's a little more complex than this description).</p>\n<p>In the paper you referenced, they combined CT and MR images to produce a pseudo HU .. that's not practical (or even possible in some cases) in the real world.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1469389,
          "author_name": "mpsampat",
          "author_url": "",
          "post_date": "08/12/2021 22:40:35",
          "content": "<p>+1 ; <a href=\"https://www.kaggle.com/davidbroberts\" target=\"_blank\">@davidbroberts</a> is absolutely right. there is no concept like HU in MR. you can ignore the paper you referenced :) </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1465339": "Any resource for mapping brain sections to HU units?\n\nThanks",
    "1466008": "1 on this\n\nIt's not very clear to me what the units are in the MRI scans. It doesn't look as clear as the CT scans that have HU. I am wondering whether there even is a standard. I was not able to find an answer to this, so if anyone knows, that would be super helpful.",
    "1466330": "H.U. is not used in MR. It's a CT concept only.",
    "1466367": "I haven't been able to find any information about units used in MRI scans.\n\nMy basis for using HU has been this paper https://pubmed.ncbi.nlm.nih.gov/24387496/ \nAlthough it does conclusively provide a basis for using HU in MRI, this gave me the idea that CT and MRI units might be correlated.\n\nWould really appreciate any advice on quantifying the scans.\n\nThanks",
    "1466505": "There isn't a concept in MR similar to HU because of the physics involved. HU is a mapping of densities on a CT to densities in the real world. MR does not provide density information in the way CT does.\n\nI.E. in CT, the x-ray beam passes through some tissue and creates a shadow on a detector. By comparing the attenuation (how much radiation was blocked by the tissue) to the attenuation of a known density (water) .. we get Hounsfield Units. Then, we can say Tissue A is X times as dense as water. \n\nIn MR, the signals come from magnetizing the tissue (thus aligning the magnetic poles in the same direction), and then releasing them to relax back to their normal state .. radio waves are then given off from the patient's body and are detected by the machine. Different tissues relax at different rates. So when you see an MR, you're not seeing density information like in CT. (this is a basic overview of a simple MR sequence .. it's a little more complex than this description).\n\nIn the paper you referenced, they combined CT and MR images to produce a pseudo HU .. that's not practical (or even possible in some cases) in the real world.",
    "1469389": "1 ; @davidbroberts is absolutely right. there is no concept like HU in MR. you can ignore the paper you referenced :)"
  },
  "source": "meta"
}