Leucine-rich alpha-2 glycoprotein reflects transmural inflammation assessed by through-the-scope endoscopic ultrasonography: A prospective study in small-bowel Crohn’s disease

Abstract

BACKGROUND

Assessment of transmural inflammation in small-bowel Crohn’s disease (CD) remains challenging. Although serum leucine-rich alpha-2 glycoprotein (LRG) has shown potential for evaluating transmural disease activity in CD, its association with small-bowel transmural inflammation remains unclear.

AIM

To evaluate the association between serum LRG levels and transmural inflammation assessed by through-the-scope endoscopic ultrasonography during double-balloon enteroscopy (DBE-EUS).

METHODS

This prospective single-center diagnostic study enrolled 132 consecutive patients with small-bowel CD between January 2022 and March 2025. All participants underwent baseline assessment, including LRG measurement and DBE-EUS examination, and were subsequently followed longitudinally for exploratory evaluation of clinical outcomes. Serum LRG levels were measured by enzyme-linked immunosorbent assay. DBE-EUS was performed using a through-the-scope miniprobe ultrasound system, and EUS parameters were assessed. Transmural inflammation was defined according to two separate criteria, namely a total wall thickness (TWT) ≥ 2.65 mm and a submucosal thickness (SMT) ≥ 0.95 mm, based on previously established cutoff values.

RESULTS

Serum LRG levels were significantly higher in patients with transmural inflammation than in those with transmural healing and showed strong positive correlations with DBE-EUS parameters, particularly TWT (r = 0.664) and SMT (r = 0.642). For detecting transmural inflammation defined by TWT ≥ 2.65 mm, LRG demonstrated an area under the curve (AUC) of 0.875 [95% confidence interval (CI): 0.816-0.935], which was higher than that of C-reactive protein (CRP) (AUC = 0.732; 95%CI: 0.641-0.822) and comparable to fecal calprotectin (FC) (AUC = 0.847; 95%CI: 0.780-0.913). For SMT ≥ 0.95 mm, LRG achieved an AUC of 0.833 (95%CI: 0.764-0.902), comparable to FC (AUC = 0.837; 95%CI: 0.766-0.908) and higher than that of CRP (AUC = 0.720; 95%CI: 0.631-0.810). Similar diagnostic performance was observed among patients in clinical remission, with LRG showing AUCs of 0.871 (95%CI: 0.794-0.948) for TWT ≥ 2.65 mm and 0.875 (95%CI: 0.790-0.960) for SMT ≥ 0.95 mm. In the exploratory prognostic analyses, elevated serum LRG levels and DBE-EUS evidence of persistent transmural inflammation were associated with an increased risk of hospitalization, whereas neither parameter significantly predicted clinical relapse in patients who had achieved clinical remission.

CONCLUSION

Serum LRG levels are significantly associated with transmural inflammation in small-bowel CD as assessed by DBE-EUS. LRG may serve as a reliable and noninvasive biomarker for evaluating transmural disease activity in small-bowel CD.

Key Words: Crohn’s disease; Leucine-rich alpha-2 glycoprotein; Endosonography; Transmural inflammation; Diagnostic accuracy

Core Tip: This prospective study investigated the association between serum leucine-rich alpha-2 glycoprotein (LRG) levels and transmural inflammatory activity in patients with isolated small-bowel Crohn’s disease (CD) using endoscopic ultrasonography during double-balloon enteroscopy (DBE-EUS). Serum LRG levels were strongly correlated with DBE-EUS-derived measures of transmural inflammation and demonstrated diagnostic performance comparable to fecal calprotectin and superior to C-reactive protein. Elevated LRG levels were also associated with an increased risk of CD-related rehospitalization, supporting LRG as a promising noninvasive biomarker for evaluating transmural disease activity in small-bowel CD.

INTRODUCTION

Crohn’s disease (CD) is a chronic, relapsing inflammatory disorder characterized by transmural inflammation of the gastrointestinal tract[1]. In recent years, therapeutic objectives for CD have shifted from mere symptom control toward the achievement of transmural healing (TH), which is associated with improved long-term outcomes[2,3]. However, the accurate assessment of transmural inflammation in small-bowel CD remains challenging. Compared with colonic disease, the evaluation of TH in the small intestine is inherently more difficult, largely due to its anatomical inaccessibility, greater procedural invasiveness, and the limited performance of conventional assessment tools. Consequently, the identification of reliable and noninvasive biomarkers for evaluating small-bowel TH is of considerable clinical importance.

Conventional ileocolonoscopy primarily evaluates the mucosal surface and is unable to adequately assess the deeper layers of the bowel wall, particularly in the proximal small intestine. Cross-sectional imaging modalities, such as intestinal ultrasound (IUS) and magnetic resonance enterography (MRE), have emerged as essential tools for evaluating TH in CD[4-6]. Among these, IUS is widely accepted due to its accessibility and noninvasiveness. However, it is susceptible to bowel gas interference, and predominantly offers cross-sectional imaging, which may compromise the accuracy of detailed bowel wall stratification. In addition, current definitions of TH are largely derived from colonic data, and physiological differences in bowel wall thickness between the small intestine and colon may lead to inconsistencies in its assessment.

In recent years, our center has pioneered a novel technique, through-the-scope endoscopic ultrasonography during double-balloon enteroscopy (DBE-EUS), which, to the best of our knowledge, is currently performed exclusively at our institution. By introducing a miniprobe ultrasound catheter through the working channel of the enteroscope, DBE-EUS enables real-time ultrasonographic evaluation under direct endoscopic visualization and provides a novel approach for assessing transmural disease activity in small-bowel CD. DBE-EUS provides circumferential imaging and enables detailed visualization of the layered structure of the bowel wall under direct endoscopic guidance, which may facilitate a comprehensive assessment of transmural inflammatory changes in the small bowel. In our previous study, DBE-EUS was demonstrated to be feasible, and preliminary cutoff values for defining deep remission in small-bowel CD were established separately for total wall thickness (TWT) and submucosal thickness (SMT), at 2.65 mm and 0.95 mm, respectively[7]. Notably, the TWT threshold identified for the small bowel was lower than the commonly used 3-mm cutoff for defining TH in colonic disease, which is biologically plausible given the inherently thinner wall of the small intestine. However, owing to the technical complexity and invasive nature of double-balloon enteroscopy, DBE-EUS is not well suited for long-term dynamic monitoring of TH. Therefore, it is essential to identify a reliable and noninvasive biomarker that reflects transmural inflammatory activity in the small bowel for therapeutic evaluation and prognostic monitoring.

Leucine-rich alpha-2 glycoprotein (LRG) has recently emerged as a promising biomarker in inflammatory bowel disease. Previous studies have demonstrated significant differences in serum LRG levels between the active and remission phases of CD, suggesting that LRG is a reliable indicator of disease activity[8-10]. Notably, LRG expression is induced by multiple proinflammatory cytokines and is independent of the interleukin-6 signaling pathway[11]. Previous studies have suggested that, compared with conventional biomarkers such as C-reactive protein (CRP), LRG may better reflect transmural intestinal inflammation[12,13]. However, most existing studies have focused on colonic transmural inflammation, and whether serum LRG levels are associated with transmural inflammation in small-bowel CD remains unclear.

Therefore, we hypothesized that serum LRG levels could reflect DBE-EUS-defined transmural inflammation in patients with small-bowel CD. The primary objective of this prospective diagnostic study was to evaluate the diagnostic performance of serum LRG for identifying transmural inflammation assessed by DBE-EUS. Secondary objectives were to compare the diagnostic performance of LRG with conventional biomarkers and disease activity indices, and to explore the prognostic associations of serum LRG levels and DBE-EUS parameters with subsequent hospitalization and clinical relapse.

MATERIALS AND METHODS

Study design and participants

This was a prospective, single-center diagnostic study with exploratory longitudinal follow-up. Between January 2022 and March 2025, consecutive patients with small-bowel CD at the First Affiliated Hospital of Anhui Medical University were enrolled. All patients with a clinical diagnosis of small-bowel CD underwent DBE-EUS as part of routine clinical care. Potentially eligible patients were screened by the study team before undergoing DBE-EUS. Written informed consent was obtained from all participants prior to enrollment. Baseline data, including medical history, demographic characteristics, Montreal classification, serum CRP levels, and fecal calprotectin (FC) levels, were collected. CRP and FC values obtained within 1 week before or after the DBE-EUS procedure were included for analysis. Exclusion criteria included age < 18 years, colonic CD, acute infection, internal fistula, concomitant autoimmune disease, and malignancy.

Sample size calculation

Sample size was calculated based on the primary diagnostic objective of evaluating the diagnostic performance of serum LRG for identifying DBE-EUS-defined transmural inflammation. Assuming an expected area under the curve (AUC) of 0.70, a null hypothesis AUC of 0.50, a two-sided α level of 0.05, a power of 90%, and an estimated transmural inflammation to non-inflammation ratio of approximately 1.5:1, a minimum sample size of approximately 100 patients was required.

DBE-EUS procedure and assessment

All patients underwent antegrade or retrograde double-balloon enteroscopy after an overnight fast of at least 8 hours. Procedures were performed using the EN-580T enteroscope system (Fujifilm, Tokyo, Japan) by three experienced endoscopists, each of whom had performed more than 200 DBE procedures. Before the procedure, the insertion route and the approximate location and distribution of small-bowel lesions were determined based on prior imaging findings, including computed tomography enterography (CTE), MRE, or capsule endoscopy. The segment exhibiting the most severe involvement, as identified during DBE, was selected for subsequent EUS evaluation. A miniprobe ultrasound system (P2620-L, 2.6 mm in diameter; Fujifilm) was introduced through the working channel. To optimize image quality, the balloons were inflated and sterile water was instilled into the intestinal lumen. The probe was then advanced to the target lesion for assessment. During DBE, all accessible involved small-bowel lesions were evaluated under white-light endoscopy. The lesion exhibiting the most severe inflammatory activity, as determined by endoscopic findings, was selected as the target site for EUS assessment, and EUS measurements were subsequently performed at the corresponding location. EUS parameters included TWT, mucosal thickness (MT), SMT, and muscularis propria thickness (MPT). TWT was defined as the distance from the mucosal interface to the serosal layer, and SMT was defined as the thickness of the submucosal layer. Additional features, including bowel wall stratification (classified as clear or unclear) and serosal integrity (smooth or irregular), were recorded according to a standardized reporting protocol. Representative DBE-EUS images of TH and transmural inflammation were shown in Supplementary Figure 1.

Measurements and definitions

In this study, the primary assessment parameters included inflammatory biomarkers, disease activity indices, and DBE-EUS parameters. Blood samples were collected on the morning of DBE-EUS examination before the procedure. Samples were centrifuged at 3000 rpm for 10 minutes within 2 hours after collection, and serum aliquots were stored at -80 °C until analysis. Serum LRG levels were measured using an enzyme-linked immunosorbent assay kit (Thermo Fisher Scientific, United States; catalog number: EH305RB). Serum samples were diluted 1:2000 before analysis. The assay was performed according to the manufacturer’s protocol, with a standard curve generated in each assay run. The calibration range was 0.307-75 ng/mL, and absorbance was measured at 450 nm. The minimum detectable dose of the assay was 0.3 ng/mL. According to the manufacturer’s validation data, intra-assay and inter-assay coefficients of variation were < 10% and < 12%, respectively. Laboratory personnel performing LRG measurements were blinded to DBE-EUS findings and clinical outcomes.

TWT, MT, SMT, and MPT were independently measured three times by two experienced endoscopists who were blinded to each other’s measurements and all patient information. The mean of the three measurements was calculated for each observer. Intraobserver reliability was assessed based on the three repeated measurements performed by each observer, whereas interobserver reliability was assessed based on the mean measurements obtained by the two observers. Reliability was quantified using intraclass correlation coefficients (ICCs). All interobserver and intraobserver ICCs were ≥ 0.75 (range: 0.80-0.93), indicating good-to-excellent reliability of DBE-EUS measurements. Based on previously established cutoff values, transmural inflammation was assessed separately using the TWT criterion of ≥ 2.65 mm and the SMT criterion of ≥ 0.95 mm. The present cohort was entirely independent of the threshold-development cohort, with no overlap in participants between the two studies.

Endoscopic activity was evaluated using the simple endoscopic score for CD (SES-CD). DBE was performed to systematically evaluate all accessible small-bowel segments, including the jejunum, proximal ileum, and terminal ileum, in every participant. SES-CD was assessed according to the established criteria using a standardized protocol, with scores calculated for each evaluated segment based on four parameters: Ulcer size, ulcerated surface, affected surface, and stenosis. The total SES-CD score was calculated by summing the scores from the assessed segments[14]. Clinical disease activity was assessed using the CD activity index (CDAI), with clinical remission defined as a CDAI score ≤ 150.

CD-related hospitalization was defined as any hospital admission resulting from CD activity or disease-related complications. Clinical relapse was defined as an increase of ≥ 70 points in the CDAI from baseline among patients who had achieved clinical remission. During follow-up, CD-related hospitalization and clinical relapse were identified through review of medical records, outpatient visits, and telephone interviews. Investigators responsible for outcome assessment were blinded to baseline DBE-EUS findings and biomarker measurements. Follow-up was conducted until May 2026. Patients without CD-related hospitalization or clinical relapse during follow-up were censored at the end of the study period. Patients lost to follow-up were censored at the date of their last contact.

Statistical analysis

Continuous variables were expressed as mean ± SD or median interquartile range (IQR), as appropriate. Categorical variables were presented as frequencies and percentages. Between-group comparisons were performed using the independent-samples t-test or Mann-Whitney U test. Pearson’s or Spearman’s correlation analyses were used to assess the relationships between serum LRG levels and bowel wall parameters. Detailed procedures for the multiple imputation of missing FC and CRP values were provided in Supplementary material, and detailed descriptions of the logistic regression model construction were provided in Supplementary material. Receiver operating characteristic (ROC) curve analysis was performed to assess the diagnostic performance of LRG, CRP, FC, SES-CD, and CDAI, and the area under the ROC curve (AUC) was calculated. Comparisons between AUCs were performed using the DeLong test. The optimal cutoff values were determined using the Youden index. Internal validation of the ROC-derived cutoff values was performed using bootstrap resampling with 1000 repetitions. ROC-derived LRG cutoff values for DBE-EUS-defined transmural inflammation were used to stratify patients for Kaplan-Meier survival analyses. Differences in hospitalization-free survival and relapse-free survival were assessed using the log-rank test, and univariate Cox regression analyses were performed to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). A two-sided P value < 0.05 was considered statistically significant.

Ethics

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Clinical Medical Research Ethics Committee of the First Affiliated Hospital of Anhui Medical University (Approval No. 5101116, dated 1 March 2021). Written informed consent was obtained from all participants.

RESULTS

Characteristics of patients

A total of 218 patients with small-bowel CD who underwent DBE-EUS between January 2022 and March 2025 were initially screened. Of these, 188 patients underwent concurrent serum LRG measurement. Subsequently, 56 patients were excluded due to colonic involvement (n = 48), active infection (n = 4), internal fistula (n = 3), or concomitant autoimmune diseases (n = 1). Ultimately, 132 patients with small-bowel CD were included in the final analysis. Serum LRG measurements were available for all included patients, whereas serum CRP and FC measurements were available in 129 and 121 patients, respectively. The patient selection process is illustrated in Figure 1. Baseline patient characteristics are summarized in Table 1. Most patients were male (74.2%), with a median age of 32 years (IQR: 22-40) and a median disease duration of 36 months (IQR: 12-60). All patients had isolated ileal disease (L1), and three patients (2.3%) also had upper gastrointestinal involvement (L4). According to the Montreal classification, 59.8% of patients had non-stricturing, non-penetrating disease (B1), 38.6% had stricturing disease (B2), and 1.5% had penetrating disease (B3). Regarding current treatment, the majority of patients were receiving biologic therapy (79.5%), while smaller proportions were treated with azathioprine (15.2%) or corticosteroids (9.1%); 5.3% of patients were not receiving any treatment. A history of prior intestinal resection was reported in 4.5% of patients. At baseline, the median CDAI score was 129 (IQR: 89-198), and the median SES-CD was 5 (IQR: 0-10). Seventy-eight patients (59.1%) were in clinical remission at baseline. The median CRP, serum LRG, and FC levels were 1.58 mg/L (IQR: 0.60-6.92, n = 129), 14.64 μg/mL (IQR: 10.58-19.09) and 172.97 μg/g (IQR: 44.75-602.55, n = 121), respectively. Regarding bowel wall parameters, the median TWT, MT, SMT, and MPT was 2.90 mm (IQR: 2.32-3.70), 0.80 mm (IQR: 0.60-0.90), 0.90 mm (IQR: 0.72-1.20), and 0.70 mm (IQR: 0.60-0.80), respectively. Among the 132 enrolled patients, transmural inflammation was identified in 81 patients (61.3%) according to the TWT criterion (TWT ≥ 2.65 mm) and in 64 patients (48.4%) according to the SMT criterion (SMT ≥ 0.95 mm). Clear bowel wall stratification was observed in 69.6% of patients, and smooth serosal integrity was noted in 95.4%.

Figure 1 Flowchart of patient selection. CD: Crohn’s disease; DBE-EUS: Endoscopic ultrasonography during double-balloon enteroscopy; LRG: Leucine-rich alpha-2 glycoprotein.

Table 1 Baseline characteristics of the cohort, n (%)/median interquartile range.

Characteristics	n = 132
Male	98 (74.2)
Median age, year	32 (22-40)
Median disease duration, months	36 (12-60)
Location
L1 (ileal)	132 (100)
L2 (colonic)	0 (0)
L3 (ileocolonic)	0 (0)
L4 (upper gastrointestinal)	3 (2.3)
Behavior
B1 (nonstricturing, nonpenetrating)	79 (59.8)
B2 (stricturing)	51 (38.6)
B3 (penetrating)	2 (1.5)
Concomitant treatment
No therapy	7 (5.3)
Corticosteroid	12 (9.1)
Azathioprine	20 (15.2)
Biologic agents	105 (79.5)
Previous intestinal resection	6 (4.5)
Median CDAI score	129 (89-198)
Clinical remission (CDAI ≤ 150)	78 (59.1)
Median SES-CD score	5 (0-10)
Median CRP (mg/L), n = 129	1.58 (0.60-6.92)
Median LRG (μg/mL)	14.64 (10.58-19.09)
Median FC (μg/g), n = 121	172.97 (44.75-602.55)
Transmural activity by DBE-EUS
TWT, mm	2.90 (2.32-3.70)
MT, mm	0.80 (0.60-0.90)
SMT, mm	0.90 (0.72-1.20)
MPT, mm	0.70 (0.60-0.80)
TWT ≥ 2.65 mm	81 (61.3)
SMT ≥ 0.95 mm	64 (48.4)
Bowel-wall stratification (clear)	92 (69.6)
Serosal integrity (smooth)	126 (95.4)

Unless otherwise indicated, percentages were calculated using the overall cohort (n = 132) as the denominator. Variables with missing data were presented using variable-specific denominators. CDAI: Crohn’s disease activity index; SES-CD: Simple endoscopic score for Crohn’s disease; CRP: C-reactive protein; LRG: Leucine-rich alpha-2 glycoprotein; FC: Fecal calprotectin; DBE-EUS: Endoscopic ultrasonography during double-balloon enteroscopy; TWT: Total wall thickness; SMT: Submucosal thickness; MT: Mucosal thickness; MPT: Muscularis propria thickness.

Serum LRG levels were elevated in small-bowel CD with transmural inflammation

In the overall cohort, serum LRG levels were significantly higher in patients with TWT ≥ 2.65 mm than in those with TWT < 2.65 mm [17.86 μg/mL (IQR: 14.65-22.74) vs 10.61 μg/mL (IQR: 4.96-12.73), P < 0.0001]. Similar findings were observed for SMT, with higher serum LRG levels in patients with SMT ≥ 0.95 mm than in those with SMT < 0.95 mm [18.85 μg/mL (IQR: 15.14-24.37) vs 11.32 μg/mL (IQR: 6.93-14.57), P < 0.0001] (Figure 2A). In the clinical remission subgroup, serum LRG levels were higher in patients with TWT ≥ 2.65 mm than in those with TWT < 2.65 mm [15.84 μg/mL (IQR: 12.65-19.74) vs 9.32 μg/mL (IQR: 5.96-11.73), P < 0.0001]. Serum LRG levels were higher in patients with SMT ≥ 0.95 mm than in those with SMT < 0.95 mm [16.26 μg/mL (IQR: 13.97-21.34) vs 10.52 μg/mL (IQR: 5.79-13.68), P < 0.0001] (Figure 2B).

Figure 2 Serum leucine-rich alpha-2 glycoprotein levels in patients with transmural healing and transmural inflammation. A: In the overall cohort; B: Among patients in clinical remission. TWT: Total wall thickness; SMT: Submucosal thickness.

Performance of serum LRG in detecting transmural inflammation in small-bowel CD

In the overall cohort, logistic regression analyses were conducted to identify factors independently associated with transmural inflammation. In the multivariable analysis, both LRG and SES-CD remained independently associated with transmural inflammation. For transmural inflammation defined by TWT ≥ 2.65 mm, LRG [odds ratio (OR) = 1.23; 95%CI: 1.09-1.40; P = 0.001] and SES-CD (OR = 1.24; 95%CI: 1.07-1.43; P = 0.003) were independently associated with transmural inflammation. Similar findings were observed when transmural inflammation was defined by SMT ≥ 0.95 mm, with independent associations for LRG (OR = 1.14; 95%CI: 1.03-1.25; P = 0.007) and SES-CD (OR = 1.22; 95%CI: 1.08-1.38; P = 0.001) (Supplementary Tables 1 and 2). The performance of serum LRG for identifying transmural inflammation was evaluated using ROC curve analysis. In the overall cohort, using TWT ≥ 2.65 mm as the definition of transmural inflammation, serum LRG demonstrated good diagnostic performance, with an AUC of 0.875 (95%CI: 0.816-0.935), which was significantly superior to that of CRP (AUC = 0.732, 95%CI: 0.641-0.822; P = 0.003) and showed diagnostic performance comparable to FC (AUC = 0.847, 95%CI: 0.780-0.913; P = 0.511) (Figure 3A). The optimal cutoff value for LRG was 14.68 μg/mL, yielding a sensitivity of 75.3% and a specificity of 92.2%, indicating the highest specificity among the evaluated biomarkers (Table 2). Similarly, when SMT ≥ 0.95 mm was used to define transmural inflammation, LRG also showed good diagnostic performance, with an AUC of 0.833 (95%CI: 0.764-0.902), outperforming CRP (AUC = 0.720, 95%CI: 0.631-0.810; P = 0.029) and demonstrating comparable accuracy to FC (AUC = 0.837, 95%CI: 0.766-0.908; P = 0.932) (Figure 3B). The optimal cutoff value was 15.8 μg/mL, with a sensitivity of 75.0% and a specificity of 83.8% (Table 2). When compared with disease activity indices, LRG showed diagnostic performance comparable to SES-CD, with no significant differences observed for either TWT ≥ 2.65 mm (AUC: 0.875 vs 0.870; P = 0.880) or SMT ≥ 0.95 mm (AUC: 0.833 vs 0.847; P = 0.682). In contrast, LRG demonstrated significantly better performance than CDAI for TWT ≥ 2.65 mm (AUC: 0.875 vs 0.769; P = 0.016), whereas no significant difference was observed for SMT ≥ 0.95 mm (AUC: 0.833 vs 0.756; P = 0.110) (Figure 3C and D).

Figure 3 Receiver operating characteristic curves for assessing transmural inflammation in the overall cohort and clinical remission. A-D: In the overall cohort: Biomarkers for total wall thickness (TWT) ≥ 2.65 mm (A); Biomarkers for submucosal thickness (SMT) ≥ 0.95 mm (B); Disease activity indices for TWT ≥ 2.65 mm (C); Disease activity indices for SMT ≥ 0.95 mm (D); E and F: In clinical remission: Biomarkers for TWT ≥ 2.65 mm (E); Biomarkers for SMT ≥ 0.95 mm (F). TWT: Total wall thickness; SMT: Submucosal thickness; CRP: C-reactive protein; LRG: Leucine-rich alpha-2 glycoprotein; FC: Fecal calprotectin; AUC: Area under the curve; CDAI: Crohn’s disease activity index; SES-CD: Simple endoscopic score for Crohn’s disease.

Table 2 Area under the curve, cutoff values, sensitivity, specificity in the overall cohort, n (%).

Outcome	n	Variable	AUC	95%CI	P value	Cutoff	Sn	Sp
TWT ≥ 2.65 mm	81 (61.3)	LRG	0.875	0.816-0.935	< 0.0001	14.68 μg/mL	0.753	0.922
		CRP	0.732	0.641-0.822	< 0.0001	1.12 mg/L	0.753	0.686
		FC	0.847	0.780-0.913	< 0.0001	124.89 μg/g	0.778	0.824
		SES-CD	0.870	0.808-0.933	< 0.0001	2	0.914	0.784
		CDAI	0.769	0.687-0.851	< 0.0001	133.5	0.642	0.784
SMT ≥ 0.95 mm	64 (48.4)	LRG	0.833	0.764-0.902	< 0.0001	15.80 μg/mL	0.750	0.838
		CRP	0.720	0.631-0.810	< 0.0001	1.17 mg/L	0.781	0.647
		FC	0.837	0.766-0.908	< 0.0001	224.54 μg/g	0.766	0.853
		SES-CD	0.847	0.780-0.913	< 0.0001	4.5	0.813	0.779
		CDAI	0.756	0.675-0.837	< 0.0001	136.5	0.688	0.735

Missing C-reactive protein and fecal calprotectin values were imputed using predictive mean matching before receiver operating characteristic analysis. AUC: Area under the curve; CI: Confidence interval; Sn: Sensitivity; Sp: Specificity; TWT: Total wall thickness; SMT: Submucosal thickness; CRP: C-reactive protein; LRG: Leucine-rich alpha-2 glycoprotein; FC: Fecal calprotectin; CDAI: Crohn’s disease activity index; SES-CD: Simple endoscopic score for Crohn’s disease.

Internal validation of the ROC-derived cutoff values was performed in the overall cohort using bootstrap resampling with 1000 repetitions. For TWT-defined transmural inflammation, the bootstrap-corrected AUC was 0.873 (95%CI: 0.810-0.932), which was consistent with the original ROC analysis (AUC = 0.875). The median optimal cutoff value across bootstrap samples was 14.7 μg/mL. For SMT-defined transmural inflammation, the bootstrap-corrected AUC was 0.832 (95%CI: 0.759-0.899), comparable to the original estimate (AUC = 0.833), with a median optimal cutoff value of 16.1 μg/mL. Sensitivity analyses based on complete-case datasets yielded comparable ROC results to those obtained after multiple imputation. For TWT ≥ 2.65 mm and SMT ≥ 0.95 mm definitions of transmural inflammation, the relative differences in AUC values between the two approaches were all < 10%, supporting the robustness of the diagnostic performance estimates.

Among patients in clinical remission, 37 patients (47.4%) had TWT ≥ 2.65 mm-defined transmural inflammation, and 26 patients (33.3%) had SMT ≥ 0.95 mm-defined transmural inflammation. ROC curve analysis for TWT ≥ 2.65 mm showed that LRG demonstrated robust diagnostic accuracy (AUC = 0.871, 95%CI: 0.794-0.948). LRG exhibited higher diagnostic performance than CRP (AUC = 0.732, 95%CI: 0.614-0.849; P = 0.038), while no significant difference was observed compared with FC (AUC = 0.827, 95%CI: 0.732-0.922; P = 0.469) (Figure 3E). Similarly, for SMT ≥ 0.95 mm, LRG maintained good diagnostic performance (AUC = 0.875, 95%CI: 0.790-0.960). LRG showed significantly better performance than CRP (AUC = 0.699, 95%CI: 0.576-0.822; P = 0.021), whereas its performance was comparable to that of FC (AUC = 0.837, 95%CI: 0.733-0.940; P = 0.587) (Figure 3F). The corresponding optimal cutoff values, sensitivity, and specificity were presented in Tables 2 and 3.

Table 3 Area under the curve, cutoff values, sensitivity, specificity among patients in clinical remission, n (%).

Outcome	n	Variable	AUC	95%CI	P value	Cutoff	Sn	Sp
TWT ≥ 2.65 mm	37 (47.4)	LRG	0.871	0.794-0.948	< 0.0001	13.99 μg/mL	0.703	0.951
		CRP	0.732	0.614-0.849	< 0.0001	1.02 mg/L	0.784	0.659
		FC	0.827	0.732-0.922	< 0.0001	118.24 μg/g	0.703	0.854
SMT ≥ 0.95 mm	26 (33.3)	LRG	0.875	0.790-0.960	< 0.0001	15.72 μg/mL	0.692	0.942
		CRP	0.699	0.576-0.822	0.001	1.04 mg/L	0.808	0.596
		FC	0.837	0.733-0.940	< 0.0001	162.39 μg/g	0.769	0.846

Missing C-reactive protein and fecal calprotectin values were imputed using predictive mean matching before receiver operating characteristic analysis. AUC: Area under the curve; CI: Confidence interval; Sn: Sensitivity; Sp: Specificity; TWT: Total wall thickness; SMT: Submucosal thickness; CRP: C-reactive protein; LRG: Leucine-rich alpha-2 glycoprotein; FC: Fecal calprotectin.

Correlation between serum LRG levels and DBE-EUS parameters

Correlation analysis demonstrated that serum LRG levels were positively correlated with all DBE-EUS parameters in both the overall cohort and the clinical remission subgroup. In the overall cohort, the strongest correlation was observed between LRG and TWT (r = 0.664, P < 0.0001), followed by SMT (r = 0.642, P < 0.0001). By contrast, the correlations between LRG and MT (r = 0.262, P = 0.002) and between LRG and MPT (r = 0.246, P = 0.004) were weaker, although they remained statistically significant (Figure 4A). Among patients in clinical remission, LRG showed strong positive correlations with TWT (r = 0.719, P < 0.0001) and SMT (r = 0.674, P < 0.0001). In contrast, the correlations between LRG and MT (r = 0.272, P = 0.015) and between LRG and MPT (r = 0.267, P = 0.018) were comparatively weaker but remained statistically significant (Figure 4B).

Figure 4 Correlation analysis between serum leucine-rich alpha-2 glycoprotein levels and endoscopic ultrasonography during double-balloon enteroscopy parameters. A: In the overall cohort; B: Among patients in clinical remission. LRG: Leucine-rich alpha-2 glycoprotein; TWT: Total wall thickness; SMT: Submucosal thickness; MT: Mucosal thickness; MPT: Muscularis propria thickness.

Predictive performance of serum LRG and DBE-EUS parameters for clinical outcomes

After a median follow-up of 26 months (IQR: 18-36), 25 patients experienced CD-related hospitalization. In the exploratory prognostic analyses, patients were stratified according to the LRG cutoff values derived from the ROC analyses based on TWT-defined and SMT-defined transmural inflammation, respectively. In the overall cohort, Kaplan-Meier analysis showed that TWT ≥ 2.65 mm was associated with an increased risk of CD-related hospitalization compared with TWT < 2.65 mm (Figure 5A; log-rank P < 0.05), with a HR of 2.578 (95%CI: 1.155-5.754, P < 0.05). Similarly, patients with LRG ≥ 14.68 μg/mL (Figure 5B), SMT ≥ 0.95 mm (Figure 5C), and LRG ≥ 15.80 μg/mL (Figure 5D) were associated with an increased risk of CD-related hospitalization, with HRs of 3.445 (95%CI: 1.572-7.551, P < 0.05), 2.934 (95%CI: 1.326-6.493, P < 0.05), and 2.357 (95%CI: 1.075-5.168, P < 0.05), respectively.

Figure 5 Kaplan-Meier analysis. A-D: Crohn’s disease-related hospitalization-free survival in the overall cohort stratified by total wall thickness (TWT) ≥ 2.65 mm (A), leucine-rich alpha-2 glycoprotein (LRG) ≥ 14.68 μg/mL (B), submucosal thickness (SMT) ≥ 0.95 mm (C), LRG ≥ 15.80 μg/mL (D); E-H: Relapse-free survival among patients in clinical remission stratified by TWT ≥ 2.65 mm (E), LRG ≥ 13.99 μg/mL (F), SMT ≥ 0.95 mm (G), LRG ≥ 15.72 μg/mL (H). LRG: Leucine-rich alpha-2 glycoprotein; TWT: Total wall thickness; SMT: Submucosal thickness.

Among the 78 patients in clinical remission, 12 patients experienced clinical relapse during follow-up. Kaplan-Meier analysis showed that TWT ≥ 2.65 mm was not associated with a significantly increased risk of clinical relapse (Figure 5E; log-rank P = 0.088), with a HR of 1.541 (95%CI: 0.496-4.785, P = 0.445). Similarly, LRG ≥ 13.99 μg/mL (Figure 5F), SMT ≥ 0.95 mm (Figure 5G), and LRG ≥ 15.72 μg/mL (Figure 5H) were not significantly associated with clinical relapse, with HRs of 2.603 (95%CI: 0.795-8.527, P = 0.124), 2.137 (95%CI: 0.635-7.194, P = 0.176), and 2.050 (95%CI: 0.563-7.468, P = 0.208), respectively.

Overall, in this exploratory analysis, elevated serum LRG levels showed associations with CD-related hospitalization that were comparable to those observed for DBE-EUS parameters. Both elevated serum LRG levels and abnormal DBE-EUS findings were associated with an increased risk of hospitalization in the overall cohort, whereas neither predictor was significantly associated with clinical relapse among patients in clinical remission.

DISCUSSION

In this study, we investigated the association between serum LRG levels and transmural inflammatory activity in small-bowel CD assessed by DBE-EUS. By restricting the study population to patients with isolated ileal disease (L1), we minimized the potential confounding effects of colonic inflammation. Our findings demonstrated that serum LRG levels were strongly associated with DBE-EUS-derived measures of transmural inflammation, supporting the potential utility of LRG as a reliable, noninvasive biomarker for assessing TH in patients with small-bowel CD.

Previous studies have shown that achieving TH, compared with mucosal healing alone, is associated with improved long-term outcomes in CD[3,15]. Currently, cross-sectional imaging modalities such as CTE, MRE, and IUS are commonly used to assess transmural disease activity. However, each modality has inherent advantages and limitations. CTE is effective in detecting complications but exposes patients to ionizing radiation[16]. MRE provides excellent soft-tissue contrast and provides comprehensive assessment of bowel wall morphology and extraintestinal manifestations[17,18], but it is costly and time-consuming. IUS is a noninvasive and radiation-free modality that enables dynamic assessment of bowel wall thickness and vascularity; however, its image quality may be compromised in patients with deep bowel segments, obesity, or excessive bowel gas[19,20]. In addition, current criteria for assessing TH using MRE and IUS are largely based on colonic disease, with limited applicability to the small bowel.

EUS provides high-resolution visualization of bowel wall architecture and enables assessment of inflammatory changes extending beyond the mucosa. Although EUS has been primarily applied to perianal and colonic CD[21,22], our previous work demonstrated that EUS performed through the DBE channel enables detailed evaluation of transmural disease activity in the small bowel. In contrast to the cross-sectional imaging provided by IUS, EUS enables longitudinal visualization of bowel wall structure, allowing for a more sensitive assessment of bowel wall thickness. To our knowledge, our center is currently the only institution performing EUS through the DBE channel for the evaluation of small-bowel CD. In our previous work, we demonstrated that EUS performed through the DBE channel could be used to assess deep remission in small-bowel CD. Compared with conventional imaging modalities, DBE-EUS offers the unique advantage of real-time, simultaneous visualization of both mucosal and transmural abnormalities during enteroscopy, thereby eliminating the temporal discrepancy between mucosal healing assessed by DBE and TH assessed by EUS[7]. Nevertheless, it should be acknowledged that the definition of TH based on DBE-EUS has not yet been fully established, and the currently available cutoff values remain exploratory and require external validation. Furthermore, the invasive nature and technical complexity of DBE-EUS limit its feasibility for serial longitudinal assessments. Therefore, there is a compelling need to identify reliable, noninvasive biomarkers that accurately reflect transmural inflammation in patients with small-bowel CD.

Serum biomarkers are widely used in clinical practice due to their accessibility, convenience, and low risk. LRG has emerged as a promising biomarker in CD. However, most previous studies have primarily focused on its association with mucosal inflammatory activity. In a study involving 81 patients with CD, serum LRG levels decreased in parallel with improvements in clinical and endoscopic outcomes[23]. In a study by Kawamoto et al[8] focusing on small-bowel CD, LRG also demonstrated good diagnostic performance (AUC = 0.85). More recently, the role of LRG in assessing TH has gained attention. Takenaka et al[13] reported a significant correlation between serum LRG levels and MRE-based scores (sMaRIA score: r = 0.576; 5-point score: r = 0.633) in 227 patients with CD, and elevated LRG levels were associated with poorer prognosis[13]; however, the study population mainly included patients with colonic involvement. Komatsu et al[12] further demonstrated that LRG was associated with TH assessed by IUS, outperforming CRP; however, the criteria used to define TH were also largely derived from patients with colonic disease. Evidence regarding the association between LRG and transmural inflammation specifically in small-bowel CD remains limited. By focusing on a cohort of patients with disease confined to the small intestine and using DBE-EUS to assess transmural inflammation, our study provides further evidence that serum LRG levels are closely associated with transmural inflammation in small-bowel CD, further supporting the potential utility of LRG as a noninvasive biomarker for the evaluation of TH.

Our results demonstrated that serum LRG levels were significantly correlated with DBE-EUS-derived parameters reflecting bowel wall inflammation. Specifically, LRG demonstrated the strongest correlations with TWT and SMT (r = 0.664 and r = 0.642, respectively), whereas correlations with MT and MPT were weaker (r = 0.262 and r = 0.246, respectively). These findings suggest that LRG may better reflect the overall inflammatory burden of the bowel wall, particularly inflammation within the submucosal layer. In addition, ROC analysis demonstrated that LRG exhibited good performance in identifying transmural inflammatory activity, with diagnostic accuracy comparable to that of FC and superior to that of CRP. LRG is induced by multiple proinflammatory cytokines and produced not only by hepatocytes but also by intestinal epithelial cells and infiltrating inflammatory cells[24,25]. Compared with CRP, which is primarily liver-derived, and calprotectin, which originates predominantly from neutrophils, this multi-source production may explain why LRG more accurately reflects transmural intestinal inflammation[26,27]. Notably, LRG demonstrated a prognostic performance comparable to that of DBE-EUS parameters. Although no significant association between LRG or DBE-EUS parameters and clinical relapse was observed in patients who achieved clinical remission, this finding may be attributable to the limited sample size and the relatively small number of relapse events.

Several limitations of this study should be acknowledged. First, this was a single-center study with a relatively small sample size, which may limit the generalizability of our findings. Although all enrolled patients had small-bowel CD (L1), thereby minimizing the influence of colonic inflammation, validation in larger multicenter cohorts is warranted. Second, the cutoff values used to define transmural inflammation (TWT ≥ 2.65 mm and SMT ≥ 0.95 mm) were derived from our previous study and have not yet been externally validated, and thus should be interpreted with caution. Third, the limited availability of serial blood samples from the same patients at different disease stages prevented longitudinal within-patient comparisons between active and remission states. Finally, the relatively small sample size of the clinical-remission subgroup and the limited number of clinical events during follow-up restricted the statistical power for further multivariable analyses. Nevertheless, these exploratory findings provide preliminary evidence that serum LRG levels and DBE-EUS parameters may be associated with the risk of CD-related hospitalization, although their prognostic value for clinical relapse requires further validation. Future prospective, multicenter studies with larger sample sizes are needed to validate and extend these findings.

CONCLUSION

In conclusion, serum LRG levels are significantly associated with transmural inflammatory activity in small-bowel CD as assessed by DBE-EUS. Compared with CRP, LRG showed superior performance in reflecting DBE-EUS-defined transmural inflammation, while demonstrating comparable diagnostic performance to FC. LRG may serve as a potential noninvasive biomarker for assessing transmural disease activity in small-bowel CD, pending further validation in larger multicenter cohorts.

ACKNOWLEDGEMENTS

The authors sincerely extend their gratitude to all the physicians, nurses, and endoscopic nursing staff of the Department of Gastroenterology and the Endoscopy Center at The First Affiliated Hospital of Anhui Medical University for their invaluable support in patient care and data collection.