Lémman Index in Paediatric Crohn’s disease—Evidence Is Accumulating

We thank Torres et al.¹ for allowing us to further elaborate on our study exploring the performance of the Lémann Index [LI]² in paediatrics. A damage score in children needs to be sensitive to reflect subtle changes, since paediatric cohorts have much shorter disease duration and thus less bowel damage than adult cohorts. Moreover, children have higher inflammatory burden than adults, which may falsely increase the damage score more than in adults. We found that the LI’s overall psychometric performance was insufficient to recommend its use in children.³

The development of an index may require sterile conditions, but external validation requires a real-world setting. Since clinical trials using the LI as an outcome do not typically stratify sites by disease location or provide scores per bowel segments, these should also be the validation conditions. Nonetheless, similar to the original LI study, all sites recruited >30% with small bowel and >30% with colonic involvement. Likewise, rather than per segment/organ, it is the total LI score for the entire bowel that should be validated as it mimics the way the LI is being used in studies. The radiologist scoring the LI cannot be the same one providing the assessment for construct validation; otherwise, the two may be falsely correlated. To ensure independence of assessment, the site radiologist scored global damage as a construct for validation and the central radiologists completed the LI.

Another concern involved the lack of systematic endoscopic data. However, all 240 ImageKids children underwent ileocolonoscopy in addition to the magnetic resonance enterography [MRE] at baseline, and these data were used in all validation and inter-observer reliability analyses. It was only in the repeated visit of 116 children, used for responsiveness exploration, that colonoscopy was missing as repeated ileocolonoscopic assessment in children is less acceptable. The requirement of endoscopic assessment is, therefore, another shortcoming of the LI in pediatrics. Although the SES-CD, used to capture the endoscopic data does not directly assess ulceration depth, its ulceration scoring does differentiate deep from superficial ulcerations.⁴ Otherwise, we calculated the LI exactly as instructed by the LI user guide, including using the most severe finding from MRE/colonoscopy/pelvic magnetic resonance imaging/gastroenterologist assessment.⁵

The ImageKids study has notable strengths. It was a prospective, multicentre study involving 22 sites and a large sample size. It also included standardisation of the imaging and calibration of the readings provided by a site radiologist and three independent central radiologists. Like pieces being added to a jigsaw puzzle slowly revealing the bigger picture, our study has started to provide answers to the question of how bowel damage should be quantified in paediatric CD. As internal penetrating lesions are rare in children (found in ~5% in the ImageKids cohort despite systematic imaging on all children), a paediatric damage score might focus on stenosis only, with a sensitive range of scores highlighting subtle changes. Such an index could be particularly useful when exploring the effectiveness of emerging antifibrotic medications.

Funding

None.

Conflict of Interest

GF: consultation fee from Abbvie and Lilly; DT: consultation fee, research grant, royalties, or honoraria from Janssen, Pfizer, Hospital for Sick Children, Ferring, Abbvie, Takeda, Atlantic Health, Shire, Celgene, Lilly, Roche, ThermoFisher, BMS.

Author Contributions

GF and DT contributed to the manuscript concept, writing, and revision.

References