Skip to main content
. 2009 Apr 14;7:25. doi: 10.1186/1479-5876-7-25

Table 2.

Diagnostic performance of 4 scoring methods for measuring p16INK4a expression in correctively distinguishing 14 ECA from 24 EMA

Scoring method C N Mean of C plus N
Sensitivity 35.7% 78.6% 71.4%
(95% CI) (20.5%,50.9%) (65.5%,91.6%) (57.1%,85.8%)
Specificity 83.3% 70.8% 87.5%
(95% CI) (71.5%,95.2%) (56.4%,85.3%) (77.0%,98.0%)
PPV 55.6% 61.1% 76.9%
(95% CI) (39.8%,71.4%) (45.6%,76.6%) (63.5%,90.3%)
NPV 69.0% 85.0% 84.0%
(95% CI) (54.3%,83.7%) (73.6%,96.4%) (72.3%,95.7%)
Accuracy 65.8% 73.7% 81.6%
(95% CI) (50.7%,80.9%) (59.7%,87.7%) (69.3%,93.9%)

PS:

1. C: Method of scoring based on independent cytoplasmic staining alone, irrespective of nucleic staining.

2. N: Method of scoring based on independent nucleic staining alone, irrespective cytoplasmic staining.

3. Mean of C plus N: Method of scoring based on mean of cytoplasmic score plus nucleic score.

Note:

Negative p16INK4a expression (score 0–3) tends to be EMA, whereas positive p16INK4a expression (scores 4–12) tends to be ECA.