Metabolomics Analysis-Based Machine Learning for Endometrial Cancer Diagnosis: Integration of Biomarker Discovery and Explainable Artificial Intelligence
Bibliographic record
Abstract
Objective: Endometrial cancer (EC) is the most frequent gynecological malignancy in women worldwide. This study aims to develop a predictive model integrating machine learning (ML) approaches with explainable artificial intelligence (XAI) using metabolomics panel data for significant biomarker discovery in EC. Materials and Methods: This study applied metabolomics and XAI to uncover diagnostic biomarkers for EC, the most common gynecologic malignancy. A total of 191 EC cases and 204 controls were analyzed using mass spectrometry. ML and XAI techniques were incorporated, including SHapley Additive exPlanation, Random Forest, BaggedCART, LightGBM, Adaptive Boosting, and Extreme Gradient Boosting. Results: Statistically significant differences (adjusted p<0.05) were found in 25 metabolites. Effect sizes (ES) of m/z=219.125 (ES=1.516), m/z=672.6961 (ES=0.913), and m/z=203.1564 (ES=0.839) were notably large, suggesting strong discriminatory ability. These metabolites are involved in lipid dysregulation, steroid hormone pathways, and oxidative stress, reflecting cancer-specific metabolic reprogramming. The ML models, particularly LightGBM, demonstrated high accuracy and good calibration. After training with the final feature dataset, SHapley Additive exPlanations (SHAP) analysis identified m/z=219.125, m/z=672.6961, and m/z=127.0769 as the top contributing features, aligning with their biological impact on EC pathogenesis. Conclusion: This study suggests non-invasive biomarkers for early detection of EC screening, highlighting the heterogeneity of metabolic adaptation in EC and the need for multi-omics approaches to understand disease mechanisms. Limitations include diverse cohorts and reliance on tandem mass spectrometry. Nonetheless, these findings represent a step forward in precision oncology.
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 | 0.003 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| 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".