Bibliographic record
Abstract
Ex vivo lung perfusion (EVLP) is an advanced technology that reconditions donor lungs prior to transplantation. Lung monitoring during EVLP provides isolated lung data without confounding factors from other physiological systems. Herein, we used machine learning modelling to process EVLP diagnostic data and predict lung transplant outcomes. For functional data analysis, we first validated sampling methods for many biomarker assessments, by measuring biopsy mRNA and perfusate protein levels of inflammatory biomarkers from donor lungs declined for transplantation. Biopsy and perfusate samples across different locations were indeed representative of the whole lung, except for biopsies taken from the lingula or from lungs with gross focal injury. From there, we built an XGBoost algorithm and showed that lung functional data from clinical EVLP were highly predictive of transplant outcomes (transplanted lungs with <72h vs. ≥72h of recipient ventilation vs. declined lungs). We further investigated X-ray images acquired during clinical EVLP, which is another important assessment regularly performed in the Toronto protocol. We first established a standardized scoring method, analyzed findings in clinical ex vivo lung radiographs, and then developed a convolutional neural network (CNN) pipeline to simultaneously process temporal radiographs from different time points. We demonstrated the value of evaluating EVLP radiographs by showing that consolidation and infiltrate scores were indicative of lung injury. Moreover, automatically extracted radiographic features from our CNN strongly correlated with clinical consolidation and infiltrate findings, indicating that the trained CNN learned relevant information from clinical EVLP radiographs. The final, multi-modal model combining radiographic features and functional data significantly improved transplant outcome predictions. These foundational analyses and machine learning modelling of EVLP functional data and radiographs demonstrated the predictive value of isolated donor lung evaluations. In a high-intensity environment like lung transplantation where large amounts of data are constantly generated, these models can be readily deployed to support clinicians with accurate diagnostic information for more informed decisions.
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) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| 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".