Coronary artery stenosis and vulnerable plaque quantification on CCTA by deep learning methods
Bibliographic record
Abstract
Abstract Introduction Coronary computed tomography angiography (CCTA) has emerged as a reliable non-invasive modality to assess coronary artery stenosis (CAS) severity and vulnerable plaque (VP). However, comprehensive CCTA assessment, especially VP, is time-consuming and dependent on reader expertise, limiting CCTA's true potential. Purpose In this study, we aim to develop and validate a deep learning (DL) based system capable of evaluating CAS severity and characterising VP on CCTA. Methods A DL system was trained to assess CAS severity on 3909 expert annotated vessels. A subset of 824 vessels was used to train the model to assess for the presence of VP. The model was based on a 2D U-Net and 3D convolutional neural network architecture. The system automatically performed vessel tracking and segmentation to quantify stenosis severity and characterise the presence of VP. CAS severity was categorised as 0%, 1–49% and ≥50%. VP was defined as: low attenuation plaque (LAP; ≤30 Hounsfield units), positive remodelling (PR; ≥10% diameter) and spotty calcification (SC; <3mm). The model was then tested on 1435 vessels for CAS (mean calcium score 197±502) and a subset of 365 vessels for VP (mean calcium score 419±551), and its diagnostic performance compared with expert readers. Results The CAS testing data had a prevalence of 75% (1080/1435), 18% (257/1435) and 6.8% (98/1435) for 0%, 1–49% and ≥50% stenosis, respectively. VP was present in 20% (72/365) in the respective dataset with 44% (32/72) LAP, 19% (14/72) PR and 36% (26/72) SC. Average analysis time for CAS severity and VP was 3.7±2.0s and 3.5±1.8s, respectively. Diagnostic performance of our system is summarised in Tables 1 (CAS severity) and 2 (VP characteristics). Conclusions We developed an DL based system capable of rapidly evaluating CAS severity and characterising VP on CCTA. Our system demonstrated high specificity and accuracy for both CAS severity and VP quantification when compared with expert readers. Funding Acknowledgement Type of funding sources: None.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".