Figure 3
From: A novel autoantibody test for the detection of pre-neoplastic lung lesions
Diagnostic accuracy for predicting the pre-neoplastic samples in validation test. Classifiers generated in the training tests were used to predict the samples statures in the blinded validation test. Corresponding receiver operating characteristics (ROC) curves and the value of the area under the curve (AUC) were calculated. Panel A was generated in the validation test based on the values from 100 AAH samples and 100 control samples. Panel B was generated based on the values from 100 SCD samples and 100 control samples.