Development and Validation of a Bronchoalveolar Lavage Genomic Classifier for Acute Cellular Rejection
Bibliographic record
Abstract
Background: Acute cellular rejection (ACR) is the main risk factor for chronic lung allograft dysfunction (CLAD), but diagnosis requires invasive transbronchial biopsy (TBB). We previously demonstrated the feasibility of bronchoalveolar lavage cell pellet (BAL-cp) gene expression for ACR diagnosis. We sought to develop and validate a genomic classifier for ACR in a multicenter cohort. Methods: We performed RNA-seq on 806 BAL-cp from 181 lung transplant recipients enrolled in CTOT-20. Differential expression was based on fold difference >2.0 and False Discovery adjusted p-value <0.05. Samples were randomly split 80:20 into training and testing sets. A Random Forest model was optimized for area under the curve (AUC), and the threshold for genomic ACR was selected for classification accuracy. We validated performance in an independent single-center cohort. Cox models evaluated risk for CLAD. Findings: From 37 cases and 151 controls, we identified 62 ACR genes, indicating upregulation of T-cell receptor signaling, and downregulation of CTLA4 signaling in cytotoxic lymphocytes, among other enriched pathways. A 31-gene Random Forest model’s AUC was 0.99 (SE 0.0053) in the training set, and 0.72 (SE 0.0874) in the test set. At a probability threshold of 0.396, accuracy for distinguishing clinically significant ACR cases from stable controls was 93.1% (specificity 95.4%, sensitivity 83.8%). In the independent validation cohort, accuracy was 82.1% (specificity 87.5%, sensitivity 73.3%). The model classified 138 (17.1%) CTOT-20 samples as genomic ACR. Late genomic ACR (≥90 days posttransplant) associated with increased CLAD risk (HR 2.52, 95% CI 1.47 - 4.34, p<0.001). Interpretation: A BAL-cp genomic classifier can identify ACR, predict CLAD risk, and may be a less invasive alternative to TBB after lung transplant.
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 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 | 0.005 | 0.001 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.002 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.004 | 0.006 |
| Research integrity | 0.001 | 0.001 |
| 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 itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
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".