Prognostic Models for Nonmetastatic Triple-Negative Breast Cancer Based on the Pretreatment Serum Tumor Markers with Machine Learning

Table 3

Regression coefficients of the lasso Cox model.

Tumor marker	β^∗
CEA	0.270645
CA19-9	0.262472
CA125	0.619252
CA242	0
CA211	0.713767
CA15-3	0.004887

^∗Positive regression coefficients indicate that higher serum tumor marker levels contributed to higher recurrence risks, while negative coefficients indicate that higher tumor marker levels contributed to lower recurrence risks. CEA: carcinoembryonic antigen; CA: cancer antigen.