Lung Lobe Segmentation and Quantification Dataset
Bibliographic record
Abstract
Quantitative lobar lung function provides valuable information for tailoring treatment plans, including surgical planning. However, its clinical application has been limited due to the need for tedious, manual lung lobe segmentation by expert operators. To address this, we present a new thoracic computed tomography (CT)/lung perfusion single-photon emission computed tomography (SPECT) dataset, featuring accurate lung lobe and trachea segmentations. Database Creation This database was developed with approval from the Ottawa Hospital Research Ethics Board (Protocol ID: 20220303-01H), using secondary use of clinical data. To create a clinically representative dataset, we retrospectively collected chest CT scans from patients who had undergone lung perfusion SPECT for pre-operative lung function quantification at The Ottawa Hospital. In our local practice, prior diagnostic or low-dose CT images are used for lung lobe segmentation to reduce patient radiation exposure and avoid suboptimal CT image quality. Consequently, the CTs in this dataset encompass a wide range of chest diagnostic and low-dose scans acquired on various CT devices, using diverse image acquisition and reconstruction parameters, with or without contrast, and representing a broad spectrum of anatomical and pathological variations. Segmentation Methods The majority of lung lobe segmentations were initially performed semi-automatically using the Hybrid3D™ software (Hermes Medical Solutions, Stockholm, Sweden). Remaining inaccuracies, if any, were corrected using either an in-house trained nnUNet model or manually using the open-source 3DSlicer software. All segmentations were reviewed and validated by a medical physicist and a physician, both with over 10 years of experience. Trachea segmentations were generated using a 3D region-growing algorithm, encompassing the trachea and a few initial generations of bronchi, depending on CT image quality. The trachea and lobar segmentations are mutually exclusive. Data Characteristics The dataset includes diagnostic and low-dose CT scans of the thorax, paired with lung perfusion SPECT acquired within an average interval of 61 days. CT scans were obtained using various imaging protocols, with slice thicknesses ranging from 0.8 to 3 mm. Lung perfusion SPECT acquisition followed standardized local clinical practice (74-185 MBq 99mTc-MAA, 128 steps, 8 seconds per step, OSEM reconstruction, 128×128 pixels, 4.83 mm pixel size). SPECT data includes rigidly co-registered reconstructed volumes. In this first release of the dataset, we have prepared 100 studies, each including the CT scan, the co-registered SPECT scan, and trachea and lung lobe segmentations in NIFTI format. We will continue to refine the segmentation masks and provide additional studies for the community in the near future. Please cite our paper if you use the dataset.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.003 | 0.001 |
| Insufficient payload (model declined to judge) | 0.010 | 0.012 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".