Machine Learning to Predict the Need for Postmastectomy Radiotherapy after Immediate Breast Reconstruction
Bibliographic record
Abstract
Background: Post mastectomy radiotherapy (PMRT) is an independent predictor of reconstructive complications. PMRT may alter the timing and type of reconstruction recommended. This study aimed to create a machine learning model to predict the probability of requiring PMRT after immediate breast reconstruction (IBR). Methods: In this retrospective study, breast cancer patients who underwent IBR from January 2017 to December 2020 were reviewed and data were collected on 81 preoperative characteristics. Primary outcome was recommendation for PMRT. Four algorithms were compared to maximize performance and clinical utility: logistic regression, elastic net (EN), logistic lasso, and random forest (RF). The cohort was split into a development dataset (75% of cohort for training-validation) and 25% used for the test set. Model performance was evaluated using area under the receiver operating characteristic curve (AUC), precision-recall curves, and calibration plots. Results: In a total of 800 patients, 325 (40.6%) patients were recommended to undergo PMRT. With the training-validation dataset (n = 600), model performance was logistic regression 0.73 AUC [95% confidence interval (CI) 0.65-0.80]; RF 0.77 AUC (95% CI, 0.74-0.81); EN 0.77 AUC (95% CI, 0.73-0.81); logistic lasso 0.76 AUC (95% CI, 0.72-0.80). Without significantly sacrificing performance, 81 predictive factors were reduced to 12 for prediction with the EN method. With the test dataset (n = 200), performance of the EN prediction model was confirmed [0.794 AUC (95% CI, 0.730-0.858)]. Conclusion: A parsimonious accurate machine learning model for predicting PMRT after IBR was developed, tested, and translated into a clinically applicable online calculator for providers and patients.
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 | 0.009 |
| 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 | 0.000 |
| Open science | 0.000 | 0.000 |
| 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".