Evaluation of Glial and Neuronal Blood Biomarkers Compared With Clinical Decision Rules in Assessing the Need for Computed Tomography in Patients With Mild Traumatic Brain Injury
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Importance: In 2018, the combination of glial fibrillary acidic protein (GFAP) and ubiquitin C-terminal hydrolase (UCH-L1) levels became the first US Food and Drug Administration-approved blood test to detect intracranial lesions after mild to moderate traumatic brain injury (MTBI). How this blood test compares with validated clinical decision rules remains unknown. Objectives: To compare the performance of GFAP and UCH-L1 levels vs 3 validated clinical decision rules for detecting traumatic intracranial lesions on computed tomography (CT) in patients with MTBI and to evaluate combining biomarkers with clinical decision rules. Design, Setting, and Participants: This prospective cohort study from a level I trauma center enrolled adults with suspected MTBI presenting within 4 hours of injury. The clinical decision rules included the Canadian CT Head Rule (CCHR), New Orleans Criteria (NOC), and National Emergency X-Radiography Utilization Study II (NEXUS II) criteria. Emergency physicians prospectively completed data forms for each clinical decision rule before the patients' CT scans. Blood samples for measuring GFAP and UCH-L1 levels were drawn, but laboratory personnel were blinded to clinical results. Of 2274 potential patients screened, 697 met eligibility criteria, 320 declined to participate, and 377 were enrolled. Data were collected from March 16, 2010, to March 5, 2014, and analyzed on August 11, 2021. Main Outcomes and Measures: The presence of acute traumatic intracranial lesions on head CT scan (positive CT finding). Results: Among enrolled patients, 349 (93%) had a CT scan performed and were included in the analysis. The mean (SD) age was 40 (16) years; 230 patients (66%) were men, 314 (90%) had a Glasgow Coma Scale score of 15, and 23 (7%) had positive CT findings. For the CCHR, sensitivity was 100% (95% CI, 82%-100%), specificity was 33% (95% CI, 28%-39%), and negative predictive value (NPV) was 100% (95% CI, 96%-100%). For the NOC, sensitivity was 100% (95% CI, 82%-100%), specificity was 16% (95% CI, 12%-20%), and NPV was 100% (95% CI, 91%-100%). For NEXUS II, sensitivity was 83% (95% CI, 60%-94%), specificity was 52% (95% CI, 47%-58%), and NPV was 98% (95% CI, 94%-99%). For GFAP and UCH-L1 levels combined with cutoffs at 67 and 189 pg/mL, respectively, sensitivity was 100% (95% CI, 82%-100%), specificity was 25% (95% CI, 20%-30%), and NPV was 100%; with cutoffs at 30 and 327 pg/mL, respectively, sensitivity was 91% (95% CI, 70%-98%), specificity was 20% (95% CI, 16%-24%), and NPV was 97%. The area under the receiver operating characteristic curve (AUROC) for GFAP alone was 0.83; for GFAP plus NEXUS II, 0.83; for GFAP plus NOC, 0.85; and for GFAP plus CCHR, 0.88. The AUROC for UCH-L1 alone was 0.72; for UCH-L1 plus NEXUS II, 0.77; for UCH-L1 plus NOC, 0.77; and for UCH-L1 plus CCHR, 0.79. The GFAP biomarker alone (without UCH-L1) contributed the most improvement to the clinical decision rules. Conclusions and Relevance: In this cohort study, the CCHR, the NOC, and GFAP plus UCH-L1 biomarkers had equally high sensitivities, and the CCHR had the highest specificity. However, using different cutoff values reduced both sensitivity and specificity of GFAP plus UCH-L1. Use of GFAP significantly improved the performance of the clinical decision rules, independently of UCH-L1. Together, the CCHR and GFAP had the highest diagnostic performance.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 it