{"success":true,"data":{"author":"Pandit et al. (National Hospital for Neurology and Neurosurgery, Queen Square / UCL); mirrored by MedOtter","bucketAccessType":"public","bucketName":"MedOtter/Pituitary_tumor","bucketPrefix":"","category":"Health & Biomedical","cloudProvider":"huggingface","createdAt":"2026-11-21T22:16:41.121Z","description":"Annotated brain MRI of 136 patients with pituitary neuroendocrine tumors (pituitary adenomas) scanned at the National Hospital for Neurology and Neurosurgery, Queen Square, London, between 2021 and 2023, with manual masks of the tumor-gland complex and both intracranial internal carotid arteries. The carotid labels support measuring cavernous sinus invasion (Knosp grade) and the distance between tumor and artery before transsphenoidal surgery. Every patient has a contrast-enhanced T1-weighted volume; 100 also have a co-registered T2-weighted volume. Volumes were bias-corrected with FSL and resampled to 0.5 mm isotropic voxels. The mirror keeps the authors' original binary masks under original_masks/ and adds a merged label map per patient (0 background, 1 tumor, 2 carotids) under dataset/<pid>/, plus train.jsonl and metadata.csv with Knosp grade, tumor volume, age and sex. About 47 GB of NIfTI (.nii.gz, measured 2026-11-21); read with nibabel or SimpleITK. Mirrored by MedOtter from the authors' figshare record (DOI 10.6084/m9.figshare.27894084), released under CC BY 4.0. Cite Pandit et al., Scientific Data 12:80 (2025), doi 10.1038/s41597-024-04218-8.","displayTitle":"Pituitary neuroendocrine tumor (PitNET) MRI: 136 patients with tumor and internal carotid artery segmentations (UCL Queen Square)","doi":"10.6084/m9.figshare.27894084.v1","huggingFaceUrl":"https://huggingface.co/datasets/MedOtter/Pituitary_tumor","id":"9c8c808f-fb43-52d2-9522-710612f35c5c","license":"CC-BY-4.0","licenseEvidenceUrl":"https://api.figshare.com/v2/articles/27894084","objectCountEstimate":651,"probeStatus":"ok","referenceUrl":"https://doi.org/10.1038/s41597-024-04218-8","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"MedOtter/Pituitary_tumor\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"MedOtter/Pituitary_tumor\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=MedOtter, license:cc-by-4.0, size_categories:n<1K, format:parquet, downloads=1070"],"format":"mixed","generatedAt":"2026-11-21T22:32:00.524Z","layout":{"example":"metadata.csv","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"case_id,patient_id,image,image_t2,mask,mask_tumor_original,mask_carotids_original,shape,spacing_mm,orientation,isotropic,has_t2,has_carotids,carotid_absent_reason,resampled_to_image_grid,t2_resampled_to_image_grid,tumor_voxels,carotid_voxels,tumor_volume_mm3,t1_sequence,t1_voxel_native,t1_matrix_native,t2_sequence,t2_voxel_native,t2_matrix_native,te_ms,age,sex,micro_macro,functional_status,knosp,ki67,cell_type,lineage,published_tumor_volume_mm3\r\n1,1,dataset/1/1_T1.nii.gz,dataset/1/1_T2.nii.gz,dataset/1/1_ma","encoding":"utf8","key":"metadata.csv","truncatedAt":512}],"sampleObjects":["dataset/1/1_T1.nii.gz","dataset/1/1_T2.nii.gz","dataset/1/1_mask.nii.gz","dataset/10/10_T1.nii.gz","dataset/10/10_mask.nii.gz"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\ntask_categories:\n- image-segmentation\ntags:\n- medical\n- mri\n- brain\n- pituitary\n- pituitary-adenoma\n- pitnet\n- carotid\n- skull-base\n- 3d\npretty_name: Pituitary_tumor - pituitary neuroendocrine tumour and carotid MRI\nsize_categories:\n- n<1K\ndataset_info:\n  config_name: preview\n  features:\n  - name: image\n    dtype: image\n  - name: mask\n    dtype: image\n  - name: overlay\n    dtype: image\n  - name: case_id\n    dtype: string\n  - name: patient_id\n    dtype: string\n  - name: slice_index\n    dtype: int32\n  - name: slice_selection\n    dtype: string\n  - name: n_slices\n    dtype: int32\n  - name: has_t2\n    dtype: bool\n  - name: has_carotids\n    dtype: bool\n  - name: carotid_absent_reason\n    dtype: string\n  - name: tumor_voxels\n    dtype: int64\n  - name: carotid_voxels\n    dtype: int64\n  - name: tumor_volume_mm3\n    dtype: float64\n  - name: shape\n    dtype: string\n  - name: spacing_mm\n    dtype: string\n  - name: isotropic\n    dtype: bool\n  - name: resampled_to_image_grid\n    dtype: bool\n  - name: age\n    dtype: int32\n  - name: sex\n    dtype: string\n  - name: micro_macro\n    dtype: string\n  - name: functional_status\n    dtype: string\n  - name: knosp\n    dtype: string\n  - name: ki67\n    dtype: string\n  - name: lineage\n    dtype: string\n  splits:\n  - name: train\n    num_bytes: 31143697\n    num_examples: 136\n  download_size: 31145695\n  dataset_size: 31143697\nconfigs:\n- config_name: preview\n  data_files:\n  - split: train\n    path: preview/train-*\n---\n\n# Pituitary_tumor\n\nMedOtter mirror of **\"Mapping pituitary neuroendocrine tumors: an annotated MRI\ndataset profiling tumor and carotid characteristics\"** (Pandit et al.,\n*Scientific Data* **12**:80, 2025).\n\n- Upstream: [figshare 27894084](https://doi.org/10.6084/m9.figshare.27894084.v1)\n- Paper: [10.1038/s41597-024-04218-8](https://doi.org/10.1038/s41597-024-04218-8)\n- License: **CC BY 4.0** - redistribution and commercial use permitted with attribution.\n\n136 patients, one scan each, from the Victor Horsley Department of Neurosurgery,\nNational Hospital for Neurology and Neurosurgery (Queen Square) / UCL, recruited\nJuly 2021 - December 2023.\n\n## Contents\n\n```\ndataset/<pid>/<pid>_T1.nii.gz        T1w contrast-enhanced, 136/136\ndataset/<pid>/<pid>_T2.nii.gz        T2w, co-registered, 100/136\ndataset/<pid>/<pid>_mask.nii.gz      merged uint8 {0=bg, 1=tumour, 2=carotids}\noriginal_masks/<pid>/                untouched upstream binaries\nupstream/                            the authors' FSL scripts + supplementary sheet\ntrain.jsonl                          136 rows, one per patient\nmetadata.csv                         same fields, flat\n```\n\n## Classes\n\n| id | name | present |\n|----|------|---------|\n| 0 | background | - |\n| 1 | pituitary tumour-gland complex | 136/136 |\n| 2 | intracranial internal carotid arteries (both sides, one label) | 133/136 |\n\n**The two structures are disjoint.** Measured across all 133 patients that carry\nboth: maximum overlap **0 voxels** - including the 9 Knosp-4 cases where the\ntumour encases the carotid and the naive expectation is overlap. Annotators\ncarve the carotid out of the tumour label. So `mask == k` equality is exact and\nno fan-out is needed.\n\nCarotids are absent for **PID 73** and **PID 75** (never published) and **PID\n132** (upstream file is zero bytes). `has_carotids` and `carotid_absent_reason`\nrecord this; an absent carotid is *not* an empty mask.\n\n## Geometry\n\n0.5 mm isotropic, **RAS**, `qform = sform = 1`, float32 images. Per-patient FOV\ndiffers (24 distinct shapes). **PID 80 escaped the upstream isotropic step** and\nis native 0.4883 x 0.4883 x 1.5 mm; `isotropic` flags it.\n\nUpstream preprocessing (scripts under `upstream/`): `fsl_anat` bias correction\nwith `--noreorient --nocrop --noreg`, then `flirt -applyisoxfm 0.5`, then\n`fslmaths -thr 0.9 -bin` for masks. No raw DICOM is published, so this is a\n**preprocessed** release, not native acquisition data.\n\n## What this mirror changes\n\nEvery change is recorded per case in `train.jsonl`; `original_masks/` keeps the\nuntouched upstream binaries.\n\n1. **Merged label map.** Upstream ships","encoding":"utf8","key":"README.md","name":"README.md","size":7058,"truncated":true}],"sourceObjectKey":"metadata.csv","starterCell":"import pandas as pd\ndf = pd.read_csv(\"hf://datasets/MedOtter/Pituitary_tumor/metadata.csv\")\ndf.head()","starterSamplePath":"dataset/1/1_T1.nii.gz","totalObjects":651},"summary":"Annotated brain MRI of 136 patients with pituitary neuroendocrine tumors (pituitary adenomas) scanned at the National Hospital for Neurology and Neurosurgery, Queen Square,…","title":"Pituitary neuroendocrine tumor (PitNET) MRI: 136 patients with tumor and internal carotid artery segmentations (UCL Queen Square)","updatedAt":"2026-11-22T00:54:12.719Z","visibility":"public","websiteUrl":"https://doi.org/10.6084/m9.figshare.27894084.v1"}}