LNW2023 Dataset¶
Collection summary¶
| Attribute | Value |
|---|---|
| Full name | Mediastinal Lymph Node Quantification (LNQ): Segmentation of Heterogeneous CT Data |
| Patients | 513 |
| Series | 1,026 (513 CT + 513 SEG) |
| Images | 66,870 |
| Total size | 35.27 GB |
| License | CC BY 4.0 |
| Last updated | August 2024 |
Overview¶
This page provides a quick high-level overview of the LNQ2023 data in graphical form. Each line represents a patient scan and hovering over a line provides a thumbnail view of the case. Clicking the scan opens an interactive viewer hosted by the NCI Imaging Data Commons.
Access¶
- CT Lymph Nodes on IDC (grayscale) — Related NIH lymph node dataset
- Mediastinal Lymph Node SEG on IDC (segmentation) — NCI Imaging Data Commons portal
Data contents¶
| Modality | Series | Total Size | Avg per Series |
|---|---|---|---|
| CT | 513 | 34.96 GB | 68 MB |
| SEG | 513 | 314 MB | 0.6 MB |
Annotation types:
- Fully annotated: 120 cases (all mediastinal lymph nodes segmented)
- Partially annotated: 393 cases (one or more lymph nodes segmented)
Data source¶
The dataset originates from the Tumor Imaging Metrics Core (TIMC), a multi-institutional imaging core lab comprising:
- Massachusetts General Hospital
- Dana Farber Cancer Institute
- Brigham and Women's Hospital
TIMC provides multimodality imaging measurements to evaluate treatment response in patients enrolled in oncology clinical trials, having processed data from over 1,400 clinical trials conducted between 2007–2020.
Patient population¶
All scans were captured at baseline timepoints during clinical trials. The top six primary cancers (60% of patients):
- Breast cancer
- Chronic lymphocytic leukemia (CLL)
- Non-small cell lung cancer
- Hodgkin's lymphoma
- Small cell lung cancer
- Renal cell cancer
Additional cancer types (40% of patients): thyroid, adenocarcinoma, endometrial adenocarcinoma, melanoma, head and neck, non-Hodgkin's lymphoma, prostate, mesothelioma, esophageal, ovarian, colon.
Methodology¶
Annotation process¶
-
Clinical trial localization Initial lymph node localization performed during clinical trial reads by TIMC staff. Each case reviewed by US board-certified radiologists.
-
3D Slicer segmentation CT scans and bidimensional measurements loaded into 3D Slicer (v4.11). Segment Editor module used for manual delineation at native CT resolution.
-
Freehand boundary drawing Draw tool used to create freehand boundaries on axial cross-sections while referencing sagittal and coronal planes. Large vessels, artifacts, and non-nodal components excluded from lesion boundaries.
-
De-identification & quality review Data fully de-identified per MGB IRB approval (2020P000211 and 2020P003754). All header data manually reviewed; image series checked for quality and de-identification integrity.
Annotator qualifications¶
All annotators were trained radiologists or radiology domain experts with over 10 years of experience.