A colorectal liver metastasis prediction model based on the combination of lipoprotein-associated phospholipase A2 and serum biomarker levels.

OBJECTIVE

This study aims to assess the predictive value of serum lipoprotein-associated phospholipase A2 (Lp-PLA2) in colorectal liver metastasis (CRLM) patients.

METHODS

A total of 507 participants were recruited for this study, comprising 162 healthy controls (HCs), 186 non-CRLM patients, and 159 CRLM patients. Serum Lp-PLA2 levels were measured across these three groups, and a CRLM prediction model was developed using machine learning (ML) algorithms in conjunction with traditional serological markers. The performance of each model was assessed using the area under the receiver operating characteristic (ROC) curve (AUC), sensitivity, specificity, and other relevant metrics.

RESULTS

The serum Lp-PLA2 levels in CRLM patients were significantly elevated compared to those in HCs group and the non-CRLM group (P < 0.0001). The CRLM prediction model developed using the Random forest algorithm demonstrated superior performance, incorporating six features: Lp-PLA2, ALB, GLB, ALT, LDH, and TC. This model achieved an AUC of 0.918, with a sensitivity of 0.823, specificity of 0.889, positive predictive value (PPV) of 0.861, and negative predictive value (NPV) of 0.857.

CONCLUSION

The Random forest model, incorporating serum Lp-PLA2 level and conventional laboratory parameters, demonstrates robust predictive capability for CRLM and holds promise for enhancing early detection in CRLM patients.