Abstract T P191: Using Clinical Trial Data to Generate Causative Classification System (CCS) Ischemic Stroke Phenotypes for the NINDS Stroke Genetics Network (SiGN)
Bibliographic record
Abstract
Background: Recruitment and phenotyping constitute major costs in stroke genetics research. Clinical trial datasets that include biological samples provide an opportunity to leverage richly phenotyped cases for genomic research. However, the heterogeneity of stroke and methodological differences across studies present challenges. We developed and validated a systematic method of mapping data abstracted from a multicenter randomized clinical trial to the inputs of the CCS stroke subtyping system for the SiGN study. Methods: The SiGN Phenotype Committee and the Secondary Prevention of Small Subcortical Strokes (SPS3) research team systematically reviewed SPS3 case report forms (CRF) to identify key elements required to generate CCS subtyping, and drafted and refined mapping rules using a derivation set of 30 charts. The resultant algorithm was compared to manual entry of clinical information into the CCS in a test set of 30 randomly selected charts. This identified problems due to multiple versions of CRFs (up to 7 versions) and prompted revision of the mapping rules. We assessed the revised algorithm in an independent test set. Results: None of the subtype classifications agreed in the initial testing prompting revision to allow capture information from different versions of the CRFs. The revised mapping algorithm applied to the same test set performed well; 29/30 agreed - 20 small artery occlusion (SAO) evident, 4 SAO possible, 2 SAO probable and 3 supra-aortic large artery atherosclerosis evident). The one chart that disagreed was classified as SAO evident by manual entry but as undetermined by the data mapping rules. In the 2nd independent test set, 30/30 agreed: 22 SAO evident, 4 SAO possible, 3 supra-aortic large artery atherosclerosis evident and 1 undetermined unknown - incomplete evaluation. Conclusions: We created mapping rules using CRFs from a clinical trial to allow reliable subtype classification using the CCS. Issues related to multiple iterations of CRFs presented challenges. Highly phenotyped stroke cases from clinical trials represent a cost-effective opportunity for genomic research.
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 | 0.063 | 0.301 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.005 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.004 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.006 | 0.002 |
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".