systematic_review·oncology, obstetrics and gynecology, systematic review, clinical trial, research methods·PMC10103471
Cervical cancer survival prediction by machine learning algorithms: a systematic review
BMC Cancer · 4 authors, 4 centres
AI SUMMARY
FIDELITY 100%
This summary was generated by AI from a single paper. It has not been reviewed by a clinician and is not clinical advice. Verify against the source before acting on it.
This systematic review analyzed 13 studies that used machine learning algorithms to predict survival outcomes in cervical cancer patients. The central finding is that machine learning techniques, often combined with heterogeneous data types, show promise for improving survival prediction, but challenges in interpretability and generalization remain.
Full summary
816 CHARS
This systematic review examined 13 original articles from a pool of 229, all retrospective studies using machine learning to predict cervical cancer survival. The included studies used datasets ranging from 85 to 14,946 patients, incorporating clinical, molecular, and PET/CT image data. Common models included random forest, logistic regression, support vector machines, ensemble/hybrid learning, and deep learning. Performance metrics like AUC and C-index varied widely. The review identifies 15 influential predictive variables and notes that data preprocessing is critical. Most models were internally validated, with only two having external validation. The authors conclude that combining multidimensional data with machine learning is influential for prediction but requires further study for standardization.
