Predicting EGFR mutation, ALK rearrangement, and uncommon EGFR mutation in NSCLC patients by driverless artificial intelligence: a cohort study

Table 2 Performance measures of the stacked ensemble model for prediction two classifications of EGFR mutation

Cohort	Sensitivity	Specificity	Accuracy	PPV	NPV
The training cohort	0.835	0.677	0.578	0.886	0.732
The testing cohort	0.856	0.680	0.638	0.877	0.750

EGFR epidermal growth factor receptor; PPV positive predictive value; NPV negative predictive value

ISSN: 1465-993X