TABLE 4.
ROC analysis and diagnostic performance of MMP‐8 analysis methods Biosensor, IFMA and ELISA in classifying between (A) periodontitis and gingivitis versus health; and (B) periodontitis versus health and gingivitis
| Periodontitis + Gingivitis vs Health | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| AUC | Cutoff | OR | Se(%) | Sp(%) | FN(%) | FP(%) | Acc(%) | MCC | |
| Biosensor | 0.808 (0.736–0.881) | 1.6 | 9.8 | 71.6 | 79.6 | 42.6 | 12.0 | 74.2 | 0.48 |
| IFMA | 0.782 (0.696–0.867) | 347.7 | 14.2 | 89.2 | 63.3 | 26.2 | 16.5 | 80.8 | 0.55 |
| ELISA | 0.857 (0.792–0.922) | 187.6 | 19.5 | 83.3 | 79.6 | 30.4 | 10.5 | 82.1 | 0.61 |
| Periodontitis vs Health + Gingivitis | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| AUC | Cutoff | OR | Se(%) | Sp(%) | FN(%) | FP(%) | Acc(%) | MCC | |
| Biosensor | 0.857 (0.793–0.920) | 2.5 | 21.4 | 74.1 | 88.2 | 15.5 | 20.4 | 82.8 | 0.63 |
| IFMA | 0.720 (0.641–0.800) | 378.3 | 27.4 | 96.6 | 49.5 | 4.2 | 45.6 | 67.5 | 0.48 |
| ELISA | 0.832 (0.768–0.896) | 386.5 | 16.2 | 56.9 | 92.5 | 22.5 | 17.5 | 78.8 | 0.54 |
Cutoff calculated by Youden's index.
Abbreviations: Acc, accuracy; FN, false negatives; FP, false positives; MCC, Matthews correlation coefficient; OR, odds ratio; Se, sensitivity; Sp, specificity.