Table 4.

Predictive accuracy of logistic regression models for steroid-free clinical remission at week 22 comprising the top 10 contributing clinical features.

Clinical features employed for the logistic regression model	Prediction accuracy in Cohort 1 (%)	Prediction accuracy in Cohort 2 (%)
Partial Mayo score	82.4	42.9
Partial Mayo score, MCH	76.5	60.0
Partial Mayo score, MCH, BMI	82.4	62.9
Partial Mayo score, MCH, BMI, BUN	76.5	60.0
Partial Mayo score, MCH, BMI, BUN, Concomitant AZA (+ /−)	82.4	60.0
Partial Mayo score, MCH, BMI, BUN, Concomitant AZA (+ /−), Lymphocyte	82.4	60.0
Partial Mayo score, MCH, BMI, BUN, Concomitant AZA (+ /−), Lymphocyte, Height	88.2	62.9
Partial Mayo score, MCH, BMI, BUN, Concomitant AZA (+ /−), Lymphocyte, Height, CRP	100.0	68.6
Partial Mayo score, MCH, BMI, BUN, Concomitant AZA (+ /−), Lymphocyte, Height, CRP, Total Cholesterol	100.0	68.6
Partial Mayo score, MCH, BMI, BUN, Concomitant AZA (+ /−), Lymphocyte, Height, CRP, Total Cholesterol, Neutrophil	100.0	65.7

MCH mean corpuscular hemoglobin, BMI body mass index, BUN blood urea nitrogen, AZA azathioprine, CRP C-reactive protein.