Back

Peritoneal Metastasis Prediction in Gastric Cancer: A Machine Learning Approach

Jalali, S.; Dadkhah, K.; Ghazi, M. M.

2025-04-13 gastroenterology
10.1101/2025.04.11.25325702 medRxiv
Show abstract

ObjectiveEvaluate the predictive efficacy of six machine learning (ML) algorithms in identifying peritoneal metastasis in gastric cancer (GC) patients. MethodsData from 809 GC patients (712 non-metastasis, 97 metastasis) were split into training and test sets (80:20). Six ML models--Decision Trees (DT), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Naive Bayes (NB), Random Forest (RF), and Logistic Regression (LR)--were assessed for feature importance and predictive performance. ResultsLymph node positivity, lymph nodes cleaned, invasion depth, lymphatic invasion, and node dissection extent were key predictors. Among inflammatory markers, PLR was significant (p = 0.018), while NLR was not (p = 0.121). RF achieved the highest accuracy (97%), followed by SVM and LR. ConclusionML enables robust prediction of peritoneal metastasis, with RF demonstrating the best performance. These findings highlight MLs role in risk stratification, though multi-center validation is required for clinical application.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.