Mucosal healing is the mainstream goal of modern treat-to-target strategy as it is associated with a significantly more favorable disease course in IBD patients with either ulcerative colitis or Crohn's disease. Recent advances in endoscopic imaging technologies have overcome the traditional concept of mucosal healing assessed with conventional white light imaging, allowing for multiple levels of endoscopic healing up to the boundaries of molecular and functional evaluation. In this review, we focused on conventional and emerging strategies to assess endoscopic healing in ulcerative colitis and ileocolonic Crohn's disease, examining their pros and cons in real life practice.

Some patients with eosinophilic esophagitis (EoE) remain symptomatic despite histological remission. In absence of strictures at endoscopy, empirical dilation has been suggested, however no data on usefulness are available. To investigate effectiveness of empirical dilation. Retrospective analysis of all EoE patients who achieved histological remission at our Institution between July 2018 and February 2024. Patients with persistent dysphagia and/or bolus impaction underwent dilation irrespective of stricture at endoscopy. Follow-up of ≥6 months was required. Forty-six patients (median age 42 years, 25 men) were included. Thirty-seven had no stricture and 9 had a stricture (i.e., S1 or R3 according to EREFS score). In the former group, 23/37 patients became asymptomatic whereas 14 were still symptomatic and underwent empirical dilation, clinically effective in 13 of them. A second dilation was needed in 2, after 25 and 31 months. In the latter group all patients underwent dilation and achieved clinical remission. All 45 patients remained in clinico-histological remission during a 14(8-46) months follow-up, with no change in medical treatment and a final I-SEE score of 1(0-1). Empirical esophageal dilation is highly effective when dysphagia/bolus impaction persists despite histological remission and should be considered before escalating treatment.

Background and aims: Liver transplantation is effective against hepatocellular carcinoma (HCC), but recurrence remains a challenge. Traditional criteria based on tumor size, nodule number, and AFP levels have had limited success in predicting aggressiveness. [18F]FDG PET/CT has shown promise in identifying high-risk tumor features, including microvascular invasion (MVI), which is a key predictor of recurrence. Methods: In this retrospective, single-center study, all consecutive patients who underwent LT for HCC between 2010 and 2019 were included. Prior to listing, the patients underwent [18F]FDG PET/CT, and explant pathology was analyzed for MVI and other histological features. The primary objective was to identify the predictors of early HCC recurrence (within 24 months after LT). Secondary objectives included identifying predictors of high-risk histological features of the explant, describing recurrence patterns, and assessing post-recurrence survival. Results: The study included 143 patients (median age 59 years [IQR 54–64], 85% males, median MELD 10 [IQR 8–14], median AFP value 8.5 [IQR 4–39] ng/mL) and 40 (28%) with intra-hepatic [18F]FDG PET/CT positivity. HCC recurred post-LT in 25 patients (17%) (median post-LT follow-up 49 months [IQR 28.5–77]) and within 24 months in 12 patients (48%). MVI at the explant stage was independently associated with early recurrence (HR: 7.20, 95% CI 1.82–28.45, p = 0.005), while intra-hepatic [18F]FDG PET/CT positivity before LT independently predicted MVI in explants (OR 3.90, 95% CI 1.30–11.71, p = 0.01). Conclusions: [18F]FDG PET/CT may offer a valuable tool for pre-transplant risk assessment by identifying MVI, which is an independent predictor of early cancer recurrence. Its incorporation into the selection criteria for LT may enhance patient stratification and post-transplant outcomes.

Guidelines recommend sampling 2-3 esophageal locations (4-6 biopsies) during EoE follow-up, but the optimal scheme is uncertain. We assessed whether two biopsies (distal, proximal) reliably classify activity versus three sites (distal, middle, proximal). Retrospective analysis of EoE follow-up endoscopies (2020-2024, tertiary centers) with three-site histology. Active disease was defined as ≥15 eos/0.3 mm² in ≥1 site; sub-analyses used ≥6 and ≥1 eos/0.3 mm². We compared active/inactive classification for two-site (distal+proximal) versus three-site sampling. Among 634 histologic evaluations, 306 three-site sets and 293 two-site sets were positive at ≥15 eos/0.3 mm²; all 328 three-site negatives remained negative with two sites. Thus, omitting the middle site would miss 2.0 % of active disease. Two-site performance: accuracy 98.0 %, sensitivity 96 %, specificity 100 %, PPV 100 %, NPV 96 %, AUC 0.98 (95 % CI, 0.97-0.99), Cohen's κ 0.98. At ≥6 eos/0.3 mm², sensitivity 97 %, accuracy 98.6 %, AUC 0.98 (95 % CI, 0.98-0.99). At ≥1 eos/0.3 mm², sensitivity 99 %, AUC 0.99 (95 % CI, 0.98-0.99). In EoE follow-up, two-site (distal+proximal) biopsies provide classification nearly equivalent to three-site sampling, with only a 2 % miss rate for active disease when the middle site is omitted. This approach may reduce procedure time, patient discomfort, and costs while maintaining excellent diagnostic performance.

Artificial intelligence (AI) is expanding in gastroenterology, particularly in endoscopy and imaging, where models support detection, classification, and risk stratification. However, translation into routine practice is hindered by inconsistent evaluation practices and limited clinical interpretability of reported metrics. We do not propose a generic reporting checklist; rather, we address a specific operational gap: insufficient support for clinicians in translating model outputs into use-case-specific decision thresholds, calibration assessment, and decision-analytic evaluation of clinical utility. These recommendations are based on narrative synthesis and expert opinion. We argue that evaluation should be anchored to pre-specified, clinically justified decision thresholds - or to a transparent interval of plausible thresholds elicited for the intended use case and prospective user community - rather than relying on global measures such as AUROC. We highlight how prevalence, calibration, and asymmetric error costs shape sensitivity, specificity, and predictive values. We recommend visualizations -ROC curves with marked operating points, calibration plots, and decision curve analysis - to communicate threshold-dependent behavior, probability reliability, and expected clinical utility. Beyond summary metrics, we emphasize the need for external validation, structured error analysis, and robustness assessment under distribution shift. Finally, we discuss transparency requirements under emerging European regulatory frameworks. These recommendations aim to support safer and more interpretable adoption of AI tools in gastroenterology.