LNTransformer: Lung Nodule Transformer for Sparse CT Segmentation
Bibliographic record
Abstract
Accurate segmentation of lung nodules in computed tomography (CT) scans is challenging due to extreme class imbalance, where nodules appear sparsely among healthy tissue. We introduce a novel two-stage approach for lung nodule segmentation, framing it as an anomaly detection problem. The method consists of two stages: Stage 1 employs a custom Detection Transformer architecture with deformable attention and focal loss to generate region proposals, addressing class imbalance and localizing sparse nodules. In Stage 2, the predicted bounding boxes are refined into segmentation masks using a fine-tuned variant of the Segment Anything Model (SAM). To address sparsity and enhance spatial context, a 5mm Maximum Intensity Projection is applied to improve differentiation between nodules, bronchioles, and vascular structures. The model achieves a stage-2 DiceC of 91.4%, with stage-1 yielding an F1 score of$94.2 \%, 95.2 \%$sensitivity, and 93.3% precision on the LUNA16 dataset despite extreme sparsity, where only 5% of slices contain a nodule, outperforming existing state-of-the-art methods. The model was additionally validated on a privately procured test dataset of 30 patients with significantly different characteristics, achieving a Dice coefficient of 78.3% despite significant distribution drift, demonstrating strong generalization to clinical variability and establishing our approach as the new state-of-the-art for lung nodule segmentation.
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) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.005 | 0.004 |
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".