Computed Tomography Enterography-Based Lémann Density Index as an Independent Predictor of Surgical Resection in Crohn’s Disease: A Firth’s Penalized Regression Analysis

Abstract Objectives: Background and Aims: The conventional Lémann Index (LI) measures cumulative damage but often "dilutes" the predictive power of critical localized lesions. We introduced the Lémann Density Index (LDI) and evaluated its efficacy for predicting 1-year resection risk in Crohn’s disease (CD). Materials and methods: This retrospective study included 110 CD patients assessed via computed tomography enterography (CTE). LDI was defined as the standard small-bowel LI divided by the number of affected segments. To ensure reliability, 25 cases were re-evaluated by a multidisciplinary consensus team (radiologist and gastroenterologist) after a 2-month interval. Independent predictors were identified using Firth’s penalized logistic regression. Results: Results: Ten patients (9.1%) underwent resection. ROC analysis showed LDI (AUC=0.902) significantly outperformed LI (AUC=0.740). An LDI cutoff ≥0.200 yielded 90% sensitivity and 82% specificity. Reliability analysis showed excellent reproducibility for LDI (ICC=0.864; Mean Bias=0.001). In multivariable Firth’s regression, LDI remained a strong independent predictor of surgery (adjusted odds ratio (aOR)=11.27 per 0.1-unit increase, p < 0.001), while penetrating behavior (B3) became non-significant ( p = 0.084). Decision curve analysis confirmed LDI provided higher net benefit than the standard LI. Conclusion: LDI is a robust, independent predictor of 1-year surgical risk in CD. By quantifying lesion intensity and mitigating the dilution effect, LDI offers a more precise objective tool for surgical decision-making and risk stratification than the conventional cumulative LI.