Deep-learning triage of 3D pathology datasets for comprehensive and efficient pathologist assessments
Bibliographic record
Abstract
Standard-of-care slide-based 2D histopathology severely undersamples spatially heterogeneous tissue specimens, with each thin 2D section representing <1% of the entire tissue volume (in the case of a biopsy). Recent advances in non-destructive 3D pathology, such as open-top light-sheet microscopy (OTLS), enable comprehensive high-resolution imaging of large clinical specimens. While fully automated computational analyses of such 3D pathology datasets are being explored, a potential low-risk route for accelerated clinical adoption would be to continue to rely upon pathologists to provide final diagnoses. Since manual review of these massive and complex 3D datasets is infeasible for routine clinical practice, we present CARP3D, a deep learning triage framework that identifies high-risk 2D cross sections within large 3D pathology datasets to enable time-efficient pathologist evaluation. CARP3D assigns risk scores to all 2D levels within a tissue volume by leveraging context from a subset of neighboring depth levels, outperforming models in which predictions are based on isolated 2D levels. In two use cases - risk stratification based on prostate cancer biopsies and screening for dysplasia/cancer in endoscopic biopsies of Barrett's esophagus - AI-triaged 3D pathology, enabled by CARP3D, demonstrates the potential to improve the detection of high-risk diseases in comparison to slide-based 2D histopathology while optimizing pathologist workloads.
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 | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.003 |
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".