Abstract WP100: Automatic Large Vessel Occlusion Detection On Computed Tomography Angiography Using A 3D Convolutional Neural Network
Bibliographic record
Abstract
Purpose: To validate the performance of a 3D convolutional neural network (CNN) based algorithm i.e. Stroke SENS LVO, in automatically detecting the presence of large vessel occlusions (LVO) on computed tomography angiography (CTA) images of the head. Method: A total of 400 studies (217 LVO, 183 non-LVO) were used in the analysis. The LVO group includes internal carotid artery (ICA) and m1 segment of the middle cerebral artery (M1-MCA) occlusions; and the non-LVO group includes more distal or posterior cerebral artery occlusions, no occlusions, and hemorrhagic stroke cases. Expert consensus reads were used as reference standard. Performance was evaluated using sensitivity and specificity and corresponding 95% confidence intervals (CI). Additional analysis was performed on several subgroups of interest. Results: For detecting LVO, the algorithm achieved a sensitivity of 0.894 [0.853, 0.935] and specificity of 0.874 [0.826, 0.922]. Furthermore, sensitivities of 0.857 [0.779, 0.935] on ICA cases (N=77) and 0.914 [0.868, 0.961] on M1-MCA cases (N=140) were noted; similarly, specificities of 0.891 [0.833, 0.949] on hemorrhagic stroke cases (N=110) and 0.849 [0.767, 0.931] on non-LVO-non-hemorrhage cases (N=73) were noted. Similar performances were observed across stratified datasets based on age, sex, scanner manufacturer and slice thickness when compared to the full cohort. Conclusion: Stroke SENS LVO demonstrated high accuracy in automatic detection of LVO on a large heterogeneous 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 | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 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".