Skip to main content

Table 7 Diagnostic algorithms developed for autistic spectrum disorder from plasma and urinary analytes

From: Advanced glycation endproducts, dityrosine and arginine transporter dysfunction in autism - a source of biomarkers for clinical diagnosis

Algorithm no 1 2 3 4
Compartment and analyte Plasma protein adduct residues Plasma amino acids Plasma protein adduct residues and amino acids Urinary amino acids
Features CML, 3DG-H, CMA, and DT CML and CMA CML, 3DG-H, CMA, and DT residues with G-H1 and GSA free adducts GSA and pyrraline free adducts
Accuracy (%) 88.3 (85.5–91.2) 74.8 (71.7–77.9) 89.0 (87.0–91.0) 76.8 (74.6–79.0)
Sensitivity (%) 91.9 (89.1–94.6) 80.5 (75.1–86.0) 90.4 (87.7–93.1) 77.1 (73.4–80.8)
Specificity (%) 83.9 (79.3–88.4) 67.1 (58.9–75.4) 87.3 (84.1–90.5) 76.4 (72.0–80.8)
AUROC 0.94 (0.91–0.96) 0.80 (0.77–0.83) 0.95 (0.94–0.96) 0.79 (0.76–0.81)
Positive likelihood ratio 5.69 (4.49–6.89) 2.85 (2.16–3.55) 7.23 (6.09–8.38) 4.16 (2.88–5.44)
Negative likelihood ratio 0.10 (0.07–0.13) 0.28 (0.21–0.35) 0.11 (0.08–0.14) 0.30 (0.25–0.34)
Positive predictive value (%) 88.2 (85.0–91.4) 77.1 (72.9–81.4) 90.2 (87.9–92.5) 80.6 (77.6–83.5)
Negative predictive value (%) 89.1 (85.5–92.6) 75.0 (70.6–79.4) 88.0 (85.1–91.0) 73.7 (71.0–76.5)
F score 0.90 (0.87–0.92) 0.78 (0.75–0.81) 0.90 (0.88–0.92) 0.78 (0.76–0.81)
  1. Algorithm outcomes for twofold cross-validation (10 randomized repeat trials for robustness) using SVMs (95% CI given in brackets)