Combining traditional analysis and machine learning to predict early, middle, and long-term recurrence of intrahepatic cholangiocarcinoma
Bibliographic record
Abstract
INTRODUCTION: Intrahepatic cholangiocarcinoma (ICC) is a rare and highly aggressive cancer. Few patients are eligible for radical surgery, and most face the high risk of recurrence. METHODS: We developed early-, middle- and long-term (1-, 2-, and 3-year) ICC disease-free survival (DFS) prediction models using traditional Logistic analysis combined with machine learning (ML) and systematically compared the performance of traditional analysis and MLs. RESULTS: 275, 256, and 238 ICC patients under radical surgery were included in the 1-, 2-, and 3-year DFS groups respectively. Five-fold cross-validation results demonstrated that both traditional Logistics and ML models exhibited remarkable robustness. MLs outperformed traditional Logistic models for DFS prediction across the AUC, accuracy and F1-scores. Specifically, the average AUC of training cohorts for the ML models were 0.878, 0.897 and 0.917 in 3 groups, compared to 0.657 (P < 0.001), 0.817 (P = 0.05), and 0.798 (P = 0.005) in traditional models. The average AUCs of testing cohorts for ML models were 0.831, 0.768, 0.803 in ML models in 3 groups, compared to 0.619 (P < 0.001), 0.719 (P = 0.008), 0.698 (P < 0.001) in traditional models. SHAP analysis identified lymph node metastasis played significant role in all-round recurrence, T stage and neural invasion had strong correction with middle and long-term recurrence in ICC patients. CONCLUSION: Models with high predictive efficiency across early, middle, and long-term recurrence have been successfully built. ML models outperformed Logistic models for DFS prediction in ICC patients. This study suggests new possibilities for advancing statistical analysis software, such as SPSS and Stata, through ML integration.
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 | 0.004 |
| 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 | 0.000 | 0.001 |
| Research integrity | 0.000 | 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".