GraftIQ: A Hybrid Multi-Class Neural Network Integrating Clinical Insight for Multi-Outcome Prediction in Liver Transplant Recipients
Bibliographic record
Abstract
ABSTRACT Background and Aims Liver transplant recipients (LTRs) are at risk of developing graft injury, leading to cirrhosis and reduced survival. Liver biopsy remains the gold standard method for the diagnosis of graft pathology but is invasive and risky. Our study aimed to develop a novel hybrid multi-class neural network (NN) model ‘GraftIQ’ integrating clinician expertise for non-invasive diagnosis of graft pathology. Methods Graft injury diagnosis was based on liver biopsies from LTRs (1992-2020). Demographic, clinical, and laboratory data from the 30 days before biopsy were used to train a multi-class NN model to classify biopsies into six categories. The dataset was split into 70% training and 30% test sets, with external validation on additional biopsies from 2020-2024. To enhance predictive capabilities, clinician expertise was integrated with neural network predictions using Bayesian fusion to combine clinician-provided probabilities with data-driven outcomes. Results Our dataset comprises 5,217 biopsies categorized into six graft etiology groups. In response to findings from expert versus machine implementation analysis, Bayesian fusion of clinical expertise and NN predictions enhanced predictive performance. GraftIQ (MulticlassNN + clinical insight) achieved an overall AUC of 0.902 (95% CI: 0.884, 0.919), improving from an AUC of 0.885 using the NN alone. Robustness validated through 10-fold internal cross-validation and external validation, showed AUC improvements of 10-16% compared to conventional machine learning approaches. Conclusion Our multi-class neural network model demonstrates high accuracy in predicting common causes of graft pathology. Through the integration of clinician expertise, we observed an improvement in its performance, affirming the effectiveness of GraftIQ as a valuable clinical decision support tool. Availability of code https://github.com/divya031090/multiclassNN
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.002 | 0.009 |
| 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".