Skip to main content
. 2020 Aug 26;15(3):391–400. doi: 10.1093/ecco-jcc/jjaa173

Table 6.

Results from PLS discriminant analysis for prediction of quiescent and severe patterns.

Quiescent vs most severe Quiescent vs all other patients Most severe vs all other patients
Sens Spec AUROC p-value Sens Spec AUROC p-value Sens Spec AUROC p-value
Montreal classificationa 0.678 0.638 0.715 <0.001 0.678 0.594 0.697 <0.001 0.610 0.656 0.683 <0.001
All baseline variablesb 0.617 0.766 0.770 <0.001 0.593 0.722 0.722 <0.001 0.553 0.710 0.687 <0.001
Baseline variables and 6 month follow-up combinedc 0.677 0.755 0.793 <0.001 0.634 0.754 0.779 <0.001 0.589 0.679 0.701 <0.001

PLS, partial least square; Sens, sensitivity; Spec, specificity; AUROC, area under the receiver operating characteristic curve.

aAge, disease localisation, disease behaviour, perianal disease, upper gastrointestinal involvement.

bMontreal classification and gender, smoking status, and surgery at diagnosis.

cAll baseline variables and disease progression, need for resective surgery, need for prednisone, need for immunomodulators, need for biologics, and need for hospitalisation within the first 6 months after diagnosis.