An Intelligent Clinical Decision Support System for Patient-Specific Predictions to Improve Cervical Intraepithelial Neoplasia Detection
Table 20
Performance of several type “AND” combinations between more than two tests in detecting CIN2+.
Combinations of medical tests
Cytology cutoff
Sensitivity (%)
Specificity (%)
PPV (%)
NPV (%)
Youden’s index
Pap test and HPV DNA and NASBA
ASCUS+
74.5
91.5
71.0
92.8
0.66
Pap test and HPV DNA and NASBA
LSIL+
70.2
92.9
73.4
91.8
0.63
Pap test and HPV DNA and NASBA
HSIL+
57.8
98.1
89.4
89.3
0.56
Pap test and HPV DNA and FLOW
ASCUS+
86.3
87.7
66.2
95.8
0.74
Pap test and HPV DNA and FLOW
LSIL+
80.7
88.9
67.0
94.3
0.70
Pap test and HPV DNA and FLOW
HSIL+
63.4
97.4
87.2
90.5
0.61
Pap test and HPV DNA and p16
ASCUS+
55.3
94.8
74.8
88.4
0.50
Pap test and HPV DNA and p16
LSIL+
53.4
95.0
74.8
88.0
0.48
Pap test and HPV DNA and p16
HSIL+
46.6
98.6
90.4
86.9
0.45
Pap test and HPV DNA and NASBA and FLOW
ASCUS+
71.4
93.1
74.2
92.1
0.65
Pap test and HPV DNA and NASBA and FLOW
LSIL+
67.1
94.0
75.5
91.1
0.61
Pap test and HPV DNA and NASBA and FLOW
HSIL+
55.3
98.4
90.8
88.8
0.54
Pap test and HPV DNA and NASBA and FLOW and p16
ASCUS+
44.7
98.8
91.1
86.5
0.44
Pap test and HPV DNA and NASBA and FLOW and p16
LSIL+
43.5
98.8
90.9
86.3
0.42
Pap test and HPV DNA and NASBA and FLOW and p16
HSIL+
38.5
99.5
95.4
85.3
0.38
HPV DNA and NASBA and FLOW
72.0
93.1
74.4
92.3
0.65
HPV DNA and NASBA and p16
46.6
97.6
84.3
86.8
0.44
HPV DNA and NASBA and FLOW and p16
45.3
98.8
91.3
86.7
0.44
NASBA and FLOW and p16
47.2
98.8
91.6
87.1
0.46
Statistical measures have been calculated using all cases of the dataset (Table 1). Histology endpoint is CIN2+ for all cases. Definition of positivity of each medical test is presented in Table 15. CIN2+: cervical intraepithelial neoplasia grade 2 or worse, HR-HPV: high-risk human papillomavirus, NASBA: nucleic acid sequence based amplification for the identification of E6/E7 mRNA of the HPV types 16, 18, 31, 33, and 45, FLOW: flow cytometric E6/E7 HPV mRNA assay, PPV: positive predictive value, and NPV: negative predictive value.