A machine learning model for colorectal liver metastasis post-hepatectomy prognostications

Area of Science:

• Oncology
• Surgical Oncology
• Machine Learning in Medicine

Background:

• Surgical resection is the primary curative treatment for colorectal liver metastases (CRLM).
• Accurate prognostication is crucial for selecting patients for surgery to optimize outcomes.
• Existing prognostic tools require enhancement for improved patient stratification.

Purpose of the Study:

• To develop and validate a machine learning-based survival prediction model for CRLM patients.
• To compare the model's performance against established prognostic scoring systems.
• To enhance individualized prognostication for patients undergoing CRLM resection.

Main Methods:

• A multicenter cohort of 572 patients who underwent hepatectomy for CRLM was analyzed.
• Cox proportional hazards and LASSO regression were used to build the prediction model.
• The model's discriminative ability was assessed using the concordance index (C-index) and compared to the Fong Clinical Risk Score.

Main Results:

• The developed CRLM Machine-learning Algorithm Prognostication (CMAP) model identified 8 key predictive variables.
• CMAP demonstrated superior prediction for overall survival (OS) (C-index=0.651) compared to Fong CRS (C-index=0.571 for 1-year, 0.574 for 5-year).
• CMAP also showed strong performance for recurrence-free survival (RFS) (C-index=0.651).

Conclusions:

• A novel machine learning algorithm (CMAP) shows significant promise for individualizing prognostication in CRLM.
• The CMAP model offers improved discriminative ability for predicting survival outcomes post-resection.
• This tool can aid in refining patient selection for surgical management of CRLM.