Table 5 Details on sensitivity, specificity, positive predictive values (NPV), and negative predictive values (NPV) at different risk percentiles based on model 1.

From: Identification of patients at risk for pancreatic cancer in a 3-year timeframe based on machine learning algorithms

Risk percentiles

Sensitivity (%)

Specificity (%)

PPV (%)

NPV (%)

1%

6.78 [6.13, 7.43]

99.01 [99.01, 99.01]

0.65 [0.58, 0.72]

99.93 [99.92, 99.93]

5%

18.36 [16.82, 19.89]

95.01 [95.01, 95.01]

0.34 [0.31, 0.37]

99.92 [99.92, 99.92]

10%

29.14 [27.70, 30.58]

90.02 [90.02, 90.02]

0.28 [0.26, 0.29]

99.93 [99.92, 99.93]

20%

42.71 [40.66, 44.76]

80.02 [80.02, 80.03]

0.19 [0.19, 0.20]

99.93 [99.92, 99.94]

Back to article page