Skip to main content

Table 2 Performance of the artificial neural network classifier in differential diagnosis of colorectal adenocarcinoma in patients and controls

From: Early detection of colorectal adenocarcinoma: a clinical decision support tool based on plasma porphyrin accumulation and risk factors

 

Training set

Validation set

Test set

Total population

146

33

33

Sensitivity

88.9%

84.6%

86.7%

Specificity

91.9%

85.0%

88.9%

NPV

89.5%

89.5%

88.9%

PPV

91.4%

78.6%

86.7%

  1. NPV Negative Predictive Value, PPV Positive Predictive Value