Machine learning for prediction of concurrent endometrial carcinoma in patients diagnosed with endometrial intraepithelial neoplasia
Bibliographic record
Abstract
OBJECTIVE: To identify predictive clinico-pathologic factors for concurrent endometrial carcinoma (EC) among patients with endometrial intraepithelial neoplasia (EIN) using machine learning. METHODS: a retrospective analysis of 160 patients with a biopsy proven EIN. We analyzed the performance of multiple machine learning models (n = 48) with different parameters to predict the diagnosis of postoperative EC. The prediction variables included: parity, gestations, sampling method, endometrial thickness, age, body mass index, diabetes, hypertension, serum CA-125, preoperative histology and preoperative hormonal therapy. Python 'sklearn' library was used to train and test the models. The model performance was evaluated by sensitivity, specificity, PPV, NPV and AUC. Five iterations of internal cross-validation were performed, and the mean values were used to compare between the models. RESULTS: Of the 160 women with a preoperative diagnosis of EIN, 37.5% (60) had a post-op diagnosis of EC. In univariable analysis, there were no significant predictors of EIN. For the five best machine learning models, all the models had a high specificity (71%-88%) and a low sensitivity (23%-51%). Logistic regression model had the highest specificity 88%, XG Boost had the highest sensitivity 51%, and the highest positive predictive value 62% and negative predictive value 73%. The highest area under the curve was achieved by the random forest model 0.646. CONCLUSIONS: Even using the most elaborate AI algorithms, it is not possible currently to predict concurrent EC in women with a preoperative diagnosis of EIN. As women with EIN have a high risk of concurrent EC, there may be a value of surgical staging including sentinel lymph node evaluation, to more precisely direct adjuvant treatment in the event EC is identified on final pathology.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 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; a candidate call from one teacher head, 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".