Clinical Data for Predicting Symptomatic Pelvic Lymphocele After Cervical Cancer Surgery

19.9% of 221 cervical cancer patients developed symptomatic pelvic lymphocele after pelvic lymphadenectomy at a Chinese hospital between January 2023 and September 2024. Yiyue Wang created this dataset to develop interpretable machine learning models, with the K-Nearest Neighbors model achieving an AUC of 0.952 on the training set. The data includes clinical characteristics and laboratory results used to identify key predictive features like diabetes and tumor size.

Use Cases

• Train interpretable ML models for surgical risk prediction based on clinical characteristics mentioned in the description.
• Analyze feature importance for post-lymphadenectomy complications based on SHAP-identified factors like diabetes and surgical approach.
• Benchmark predictive performance of models like Logistic Regression and XGBoost based on metrics such as AUC and Brier Score.
• Investigate the relationship between laboratory markers like preoperative monocyte-to-lymphocyte ratio and surgical outcomes.

Strengths

• Includes data from 221 patients with a defined outcome (44 with SPL, 177 without).
• Model performance is quantitatively reported, with a training set AUC of 0.952 for the KNN model.
• Key predictive features were identified via SHAP analysis, providing interpretability.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data is from a single hospital (Affiliated Hospital of North Sichuan Medical College) over a limited 21-month period, which may limit generalizability.

Provenance

Source

Affiliated Hospital of North Sichuan Medical College.

Collection Method

Retrospective analysis of clinical data from patients who underwent pelvic lymphadenectomy.

Time Range

January 2023 to September 2024.

Freshness

Last updated 2026-04-15 05:44:15; freshness should be verified.

Geography

China (hospital location not further specified).