Table 2. Development of the microRNA scoring system using the training set.
# | Selected for panel |
MicroRNA | AUC-ROC | p value | Youden J Associated Criterion Value |
95% Confidence interval |
Sensitivity | Specificity | Sample size |
Positive group | Negative group |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | x | 148a-3p | 1 | 0.0001 | ≤28.04 | ≤27.93 to ≤28.04 | 100 | 100 | 11 | 6 (54.55%) | 5 (45.45%) |
2 | x | 301a-3p | 0.9 | 0.0001 | ≤25.08 | ≤24.64to ≤25.08 | 100 | 69.23 | 20 | 7 (35.00%) | 13 (65.00%) |
3 | x | 671-3p | 0.98 | 0.0001 | ≤28.80 | ≤28.03 to ≤28.75 | 100 | 88.89 | 14 | 5 (35.71%) | 9 (64.29%) |
4 | x | 181a | 0.84 | 0.0005 | ≤25.84 | ≤24.61 to ≤26.91 | 85.71 | 72.73 | 18 | 7 (38.89%) | 11 (61.11%) |
5 | x | 210 | 0.81 | 0.002 | ≤21.04 | ≤19.70 to ≤22.19 | 71.43 | 84.62 | 20 | 7 (35.00%) | 13 (65.00%) |
6 | x | 1267 | 0.79 | 0.006 | ≤14.31 | ≤14.15to ≤15.47 | 85.71 | 76.92 | 20 | 7 (35.00%) | 13 (65.00%) |
7 | x | 223 | 0.77 | 0.02 | ≤12.33 | ≤9.95 to ≤12.61 | 85.71 | 61.54 | 20 | 7 (35.00%) | 13 (65.00%) |
8 | x | 340-5p | 0.79 | 0.02 | ≤21.31 | ≤20.55to ≤23.01 | 57.14 | 100 | 20 | 7 (35.00%) | 13 (65.00%) |
9 | 16 | 0.76 | 0.06 | ≤13.49 | ≤13.12 to ≤15.31 | 85.71 | 76.92 | 20 | 7 (35.00%) | 13 (65.00%) | |
10 | 193a-3p | 0.75 | 0.08 | ≤26.67 | ≤25.00 to ≤29.78 | 71.43 | 75 | 15 | 7 (46.67%) | 8 (53.33%) | |
11 | 132 | 0.73 | 0.1 | ≤24.74 | ≤23.25 to ≤29.33 | 71.43 | 76.92 | 20 | 7 (35.00%) | 13 (65.00%) | |
12 | 133b | 0.73 | 0.1 | ≤23.88 | ≤22.85to ≤27.48 | 71.43 | 76.92 | 20 | 7 (35.00%) | 13 (65.00%) | |
13 | 30e-3p | 0.67 | 0.19 | ≤16.80 | ≤16.51 to ≤16.80 | 100 | 46.15 | 20 | 7 (35.00%) | 13 (65.00%) | |
14 | 146a | 0.68 | 0.2 | ≤17.20 | ≤16.37 to ≤18.82 | 71.43 | 69.23 | 20 | 7 (35.00%) | 13 (65.00%) | |
15 | 1244 | 0.66 | 0.28 | ≤23.56 | ≤22.84 to ≤26.64 | 71.43 | 75 | 19 | 7 (36.84%) | 12 (63.16%) | |
16 | 7-5p | 0.64 | 0.36 | >28.47 | >25.4 to >30.79 | 71.43 | 62.5 | 15 | 7 (46.67%) | 8 (53.33%) | |
17 | 582-5p | 0.7 | 0.4 | ≤29.13 | ≤27.77 to ≤29.70 | 80 | 75 | 9 | 5 (55.56%) | 4 (44.44%) | |
18 | 575 | 0.67 | 0.43 | ≤27.34 | ≤23.94 to ≤27.34 | 100 | 66.67 | 12 | 6 (50.00%) | 6 (50.00%) | |
19 | 199b-5p | 0.67 | 0.53 | >26.50 | >26.06 to >29.70 | 83.33 | 66.67 | 9 | 6 (66.67%) | 3 (33.33%) | |
20 | 1229 | 0.67 | 0.62 | ≤26.75 | ≤25.50 to ≤26.75 | 100 | 66.67 | 7 | 4 (57.14%) | 3 (42.86%) | |
21 | 155 | 0.51 | 0.97 | ≤14.13 | ≤12.11 to ≤14.62 | 28.57 | 46.15 | 20 | 7 (35.00%) | 13 (65.00%) | |
22 | 196a | 0.5 | 1 | >23.79 | >22.49 to >26.72 | 100 | 50 | 9 | 5 (55.56%) | 4 (44.44%) | |
23 | 1 | NA | NA | NA | NA | NA | NA | NA | NA | NA | |
24 | 144-3p | NA | NA | NA | NA | NA | NA | NA | NA | NA | |
25 | 199a-5p | NA | NA | NA | NA | NA | NA | NA | NA | NA | |
26 | 219-5p | NA | NA | NA | NA | NA | NA | NA | NA | NA | |
27 | 221-5p | NA | NA | NA | NA | NA | NA | NA | NA | NA | |
28 | 33a-5p | NA | NA | NA | NA | NA | NA | NA | NA | NA | |
29 | 424-5p | NA | NA | NA | NA | NA | NA | NA | NA | NA | |
30 | 513-5p | NA | NA | NA | NA | NA | NA | NA | NA | NA |
Using the 20 sample training set, an AUC-ROC was calculated for each of the 30 microRNAs used in the study.
MicroRNA where the p value was less than 0.05 were included in the panel.
“NA” represents microRNA for which an insufficient number of samples signaled to generate a ROC curve.
Eight microRNAs were included in the scoring panel: microRNA 148a, -301a, -671, -181a, -210, -1267, -223, and microRNA -340.
A cutoff value for each of these microRNAs was concurrently determined as the Youden Index J Associated Criterion Value.
Cutoff values for all eight microRNAs were then used to verify the eight-microRNA risk score for individual patients in the validation set.