Maarten Cozijnsen

44 Table 3. Accuracy of cutoffs for the MINI index, FC, PCDAI, CRP and ESR Cohort Measure MINI index FC (μg/g) PCDAI/ wPCDAI CRP (mg/L) ESR (mm/ hr) Cutoff <6 ≥8 >11 <300 <10 / <12,5 <5 <20 Detect MH DH MI* MSI MH MH MH MH Derivation (n=154) Sensitivity 75% 82% 85% 89% 83% 67% 75% 79% Specificity 90% 88% 88% 61% 79% 73% 64% 47% PPV 58% 45% 97% 68% 43% 31% 28% 22% NPV 95% 98% 51% 86% 96% 92% 93% 92% Validation (n=168) Sensitivity 79% NA 87% 93% 80% 93% 93% 79% Specificity 93% NA 84% 73% 87% 69% 69% 64% PPV 86% NA 90% 66% 78% 63% 63% 56% NPV 88% NA 78% 95% 89% 95% 95% 84% *For MINI≥8, we set SESCD≥3 as positive result (thus NPV and PPV are switched). CRP=C- reactive protein, ESR=erythrocyte sedimentation rate, FC=fecal calprotectin; MH=mucosal healing (SESCD<3), DH=deep healing (i.e. SESCD<3 and remission by MRE as defined by radiologic global assessment<20 mm), MI=mucosal inflammation (SESCD≥3), MSI=moderate-to-severe inflammation (SESCD>9), PCDAI=pediatric Crohn’s disease activity index; wPCDAI=mathematically weighted PCDAI; PPV=positive predictive value, NPV=negative predictive value, NA=not applicable. In the derivation cohort, the MINI index had good accuracy to detect the 17 children with deep healing (Table 3); a cut-off value of < 6 points detected deep healing with 82% sensitivity, 88% specificity, 45% PPV and 98% NPV (AUROC 0.82). The low PPV is a result of the low deep healing rate in the derivation cohort (17/154 [11%]). Fourteen out of 17 children with deep healing (82%) had a MINI < 6, versus 17 out of 137 (12%) without deep healing (p < 0.001).