Enrico Martin

121 Diagnostic accuracy of non-invasive tests Seven studies reported on cystic degeneration or necrosis. 22,30,32,37–40 Pooled sensitivity and specificity were 0.48 (95% CI: 0.23-0.71) and 0.86 (95% CI: 0.61-0.98) respectively. Pooled pLR was 5.75 (1.27-23.69) and nLR was 0.61 (0.34-0.91). There was moderate heterogeneity between studies. Sensitivity was higher in studies with smaller sample of lesions, smaller proportion of MPNST, and when only histologically proven lesions were included. Specificity was higher among studies with larger sample of lesions and lower in studies including NF1 patients only or histologically proven lesions only. Three studies reported on signal heterogeneity on T1 sequences. 31,32,35 Pooled sensitivity and specificity were 0.85 (95% CI: 0.56-1.00) and 0.48 (95% CI: 0.03-0.96) respectively. Pooled pLR was 9.23 (0.81-31.82) and nLR was 1.60 (0.01-5.42). There was substantial heterogeneity between studies. Sensitivity was lower in studies including NF1 patients only. Specificity was higher in studies including NF1 patients only and those with a higher proportion of MPNST. Five studies reported on signal heterogeneity on T2 sequences. 29–32,38 Pooled sensitivity and specificity were 0.78 (95% CI: 0.64-0.90) and 0.52 (95% CI: 0.23-0.80) respectively. Pooled pLR was 1.94 (0.90-4.82) and nLR was 0.49 (0.15-1.37). There was substantial heterogeneity between studies. Sensitivity was lower in studies with a smaller sample of lesions and in those that included histologically proven lesions only. Six studies reported on irregular or peripheral tumor enhancement after contrast administration. 16,29,32,34,39,40 Pooled sensitivity and specificity were 0.63 (95% CI: 0.50-0.76) and 0.81 (95% CI: 0.60-0.95) respectively. Pooled pLR was 4.81 (1.44-16.60) and nLR was 0.46 (0.28-0.72). There was moderate heterogeneity between studies. Sensitivity was lower in studies including histologically proven lesions only. Specificity was higher in studies with a smaller sample of lesions and higher prevalence of MPNST. Five studies reported on intratumoral lobulation. 32,35,36,38,39 Pooled sensitivity and specificity were 0.57 (95% CI: 0.41-0.72) and 0.89 (95% CI: 0.83-0.93) respectively. Pooled pLR was 5.38 (2.87-9.31) and nLR was 0.49 (0.30-0.68). There was limited heterogeneity between studies. Heterogeneity in sensitivity may be caused by studies with higher total number of lesions and including NF1 patients only. No sources were found explaining heterogeneity in specificity. Three studies reported on absence of split-fat sign. 30,34,39 The split-fat sign represents fat deposition around the lesion and is usually seen as a tapered rim of fat signal near the proximal and distal ends of the lesion. Pooled sensitivity and specificity were 0.76 (95% CI: 0.57-0.91) and 0.44 (95% CI: 0.16-0.78) respectively. Pooled pLR was 1.67 (0.82- 4.56) and nLR was 0.68 (0.15-1.94). There was limited heterogeneity between studies. Sensitivity was higher in studies with smaller proportion of MPNSTs. Specificity was higher in studies including NF1 patients only. 6