{
  "id": 606892,
  "title": "Clarification on Single-Slice Annotations in RSNA Intracranial Aneurysm Detection Dataset",
  "url": "/competitions/rsna-intracranial-aneurysm-detection/discussion/606892",
  "author_name": "Jeslin Thomaskutty",
  "post_date": "2025-09-10T19:39:07.812000",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi, I’m trying to understand the RSNA Intracranial Aneurysm Detection dataset. The CSV provides x and y coordinates for a specific image ID, but there are no z-axis annotations. Does this mean the aneurysm is annotated only in that single slice? How should we handle detection across the rest of the volume(z-axis) when training a model — should we only use the annotated slice, or include neighboring slices as context?</p>",
  "messages": [
    {
      "id": 3287113,
      "postDate": "2025-09-10T22:12:02.457Z",
      "content": "<p>Thanks for your question. The aneurysm coordinates in the spreadsheet are just a single point approximating the center of the aneurysm. As such there is only a single point per aneurysm, which is a single x, y coordinate plus a slice instance, which is the z coordinate. For a subset of 3D DICOM data in the dataset, we also provide a frame number, which is the equivalent of a z coordinate. Hope this helps!</p>",
      "rawMarkdown": "Thanks for your question. The aneurysm coordinates in the spreadsheet are just a single point approximating the center of the aneurysm. As such there is only a single point per aneurysm, which is a single x, y coordinate plus a slice instance, which is the z coordinate. For a subset of 3D DICOM data in the dataset, we also provide a frame number, which is the equivalent of a z coordinate. Hope this helps!",
      "votes": 1,
      "replies": [
        {
          "id": 3287196,
          "postDate": "2025-09-11T04:25:47.467Z",
          "content": "<p>Thanks a lot for the clarification</p>",
          "rawMarkdown": "Thanks a lot for the clarification"
        }
      ]
    },
    {
      "id": 3287025,
      "postDate": "2025-09-10T19:39:07.813Z",
      "content": "<p>Hi, I’m trying to understand the RSNA Intracranial Aneurysm Detection dataset. The CSV provides x and y coordinates for a specific image ID, but there are no z-axis annotations. Does this mean the aneurysm is annotated only in that single slice? How should we handle detection across the rest of the volume(z-axis) when training a model — should we only use the annotated slice, or include neighboring slices as context?</p>",
      "rawMarkdown": "Hi, I’m trying to understand the RSNA Intracranial Aneurysm Detection dataset. The CSV provides x and y coordinates for a specific image ID, but there are no z-axis annotations. Does this mean the aneurysm is annotated only in that single slice? How should we handle detection across the rest of the volume(z-axis) when training a model — should we only use the annotated slice, or include neighboring slices as context?",
      "votes": 2
    }
  ],
  "comments": [
    {
      "id": 3287113,
      "author_name": "Evan Calabrese",
      "author_url": "",
      "post_date": "2025-09-10T22:12:02.457000",
      "content": "<p>Thanks for your question. The aneurysm coordinates in the spreadsheet are just a single point approximating the center of the aneurysm. As such there is only a single point per aneurysm, which is a single x, y coordinate plus a slice instance, which is the z coordinate. For a subset of 3D DICOM data in the dataset, we also provide a frame number, which is the equivalent of a z coordinate. Hope this helps!</p>",
      "votes": 1,
      "replies": [
        {
          "id": 3287196,
          "author_name": "Jeslin Thomaskutty",
          "author_url": "",
          "post_date": "2025-09-11T04:25:47.467000",
          "content": "<p>Thanks a lot for the clarification</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3287113": "Thanks for your question. The aneurysm coordinates in the spreadsheet are just a single point approximating the center of the aneurysm. As such there is only a single point per aneurysm, which is a single x, y coordinate plus a slice instance, which is the z coordinate. For a subset of 3D DICOM data in the dataset, we also provide a frame number, which is the equivalent of a z coordinate. Hope this helps!",
    "3287025": "Hi, I’m trying to understand the RSNA Intracranial Aneurysm Detection dataset. The CSV provides x and y coordinates for a specific image ID, but there are no z-axis annotations. Does this mean the aneurysm is annotated only in that single slice? How should we handle detection across the rest of the volume(z-axis) when training a model — should we only use the annotated slice, or include neighboring slices as context?"
  }
}