Table 3.
Performance of NASHmap in classifying patients in the NIDDK cohort stratified by T2DM status.
| Predicted class (NASHmap) | Metrics | ||||
|---|---|---|---|---|---|
| Positive | Negative | ||||
| T2DM (N = 112) | True class (biopsy confirmed NASH) | Positive | TP = 72 | FN = 12 | Sensitivity/TPR = 86% |
| Negative | FP = 10 | TN = 18 | Specificity/TNR = 64% | ||
| Metrics | Precision/PPV = 88% | NPV = 60% | Accuracy = 80% | ||
| Non-T2DM (N = 169) | True class (biopsy confirmed NASH) | Positive | TP = 75 | FN = 22 | Sensitivity/TPR = 77% |
| Negative | FP = 27 | TN = 45 | Specificity/TNR = 63% | ||
| Metrics | Precision/PPV = 74% | NPV = 67% | Accuracy = 71% | ||
Accuracy: TP + TN/(TP + FP + FN + TN); FN: false negative; FP: false positive; NIDDK = National Institute of Diabetes and Digestive Kidney Diseases; NASH = nonalcoholic steatohepatitis; NPV: negative predictive value (TN/[FN + TN]); PPV: positive predictive value (TP/[TP + FP]); TN: true negative; TNR: true negative rate (TN/[FP + TN]); TP: true positive; TPR: true positive rate (TP/[TP + FN]).