Automated Quantification of Neuropad Improves Its Diagnostic Ability in Patients with Diabetic Neuropathy

Figure 1

ROC curve analysis was used to compare the diagnostic accuracy of Neuropad in detecting large with small nerve fibre damage (black line). The grey line is the null value of the ROC curve. The AUC for large nerve fibre damage was (a) 75% for Vibration Perception Threshold (VPT) (SE: 0.049, 95% CI 0.53–0.97) and (b) 86% for sural sensory nerve action potential (SNAP) (SE: 0.04, 95% CI 0.42–1). The AUC for small nerve fibre damage was (c) 82% for Corneal Nerve Fibre Density (CNFD) (SE: 0.053, 95% CI 0.48–1), (d) 91% for Deep Breathing-Heart Rate Variability (DB-HRV) (SE: 0.031, 95% CI 0.49–1), (e) 63% for intraepidermal nerve fibre density (IENFD) (SE: 0.068, 95% CI 0.45–0.81), and (f) 69% for warm perception thresholds (WPT) (SE: 0.05, 95% CI 0.49–0.88).

(a)

(b)

(c)

(d)

(e)

(f)