Early Intestinal Ultrasound Response to Biologic Therapy Predicts Endoscopic Remission in Children with Ileal Crohn’s Disease: Results from the Prospective Super Sonic Study

Abstract

Background and Aims

STRIDE-II recommends early biomarker targets for treatment optimization to achieve treat-to-target [T2T] endoscopic remission [ER] in Crohn’s disease [CD]. The predictive capabilities of intestinal ultrasound [IUS] for T2T ER remain unknown. We aimed to evaluate IUS response to predict ER in children with CD.

Methods

This was a prospective longitudinal cohort study of children with ileal [TI] CD initiating biologic therapy undergoing IUS, clinical disease activity, and C-reactive protein [CRP] assessments at baseline, week 8, 6 months, and T2T within 1 year. The primary outcome was the accuracy of optimal cut-points to predict TI ER [SES-CD ≤ 2] for change in bowel wall thickness [BWT] on IUS from baseline to week 8, and BWT at week 8. Area under the receiver operating curve [AUROC] analysis was performed and univariate analysis tested associations.

Results

In total, 44 children (median age 13 [IQR 12–17] years, 29 [66%] biologic naïve) were included, and 29 [66%] achieved ER. A ≥18% decrease in TI BWT at week 8 predicted ER with an AUROC of 0.99 [95% CI 0.98–1.00], 100% sensitivity, 93% specificity, 97% positive predictive value, and 100% negative predictive value, superior to a ≥46% decrease in PCDAI (AUROC 0.67 [95% CI 0.49–0.84]) and ≥84% decrease in CRP (AUROC 0.49 [95% CI 0.31–0.67]) at week 8.

Conclusions

Early change in TI BWT on IUS is highly predictive of ER in children with CD and superior to symptoms and CRP. Our findings suggest that IUS could be used for treatment optimization and tight control to guide T2T strategies.

Imaging, paediatrics, biomarkers

1. Introduction

The treat-to-target [T2T] STRIDE-II consensus guidelines for Crohn’s disease [CD], a progressive, transmural inflammatory bowel disease [IBD], recommend utilizing early objective non-invasive biomarker targets to guide therapeutic optimization towards a goal of achieving T2T endoscopic remission 1 year after treatment initiation.¹ Utilization of early biomarker targets, rather than relying on clinical symptoms alone, is supported by the CALM trial, in which a biomarker-dominant monitoring strategy with faecal calprotectin [FC] and C-reactive protein [CRP] was superior to clinical symptom monitoring to achieve endoscopic remission at week 48.² While the CALM trial was a positive study, use of FC and CRP targets to guide treatment optimization led to endoscopic remission in only 46% of adult patients. Additionally, analysis of the PAILOT trial, a prospective study of children with CD initiating treatment with adalimumab [ADA] comparing proactive vs reactive therapeutic drug monitoring, found that FC and CRP were not the primary drivers of treatment intensification.³ Lastly, in children with ileal CD, a treatment optimization strategy guided by FC targets may be even less ideal due to weak correlation with the Simple Endoscopic Score for Crohn’s Disease [SES-CD] in the terminal ileum [TI].⁴ As such, early FC, CRP, and clinical symptom targets may not be adequate to guide early treatment optimization in children with ileal CD in order to achieve T2T endoscopic remission.

Advancements in cross-sectional imaging, particularly intestinal ultrasound [IUS], introduce new potential early treatment targets to guide therapeutic optimization strategies for children with ileal CD. IUS is a non-invasive, point-of-care, cross-sectional imaging tool capable of directly monitoring transmural IBD activity without the need for bowel preparation, fasting, or oral or intravenous contrast agents.^5–7 Unlike FC, IUS has been shown to be accurate to detect endoscopic activity in the TI based on the SES-CD and clinical CD activity.^8–11

By directly assessing responsiveness of the TI bowel wall to treatment, IUS may be an ideal tool with which to guide treatment optimization towards achieving endoscopic remission. With this in mind, the international expert consensus definition of IUS treatment response in CD was developed, but without evidence from prospective studies targeting endoscopic remission, recommending targeting a ≥25% reduction in bowel wall thickness [BWT], or >2 mm reduction, or >1 mm reduction and reduction by 1 point in colour Doppler signal, 14 weeks after treatment initiation, with consideration of response assessment in 4–8 weeks in certain situations, which are not well defined.¹² Utilizing this IUS response definition of ≥25% decrease in BWT in the prospective STARDUST sub-study of adults with CD initiating treatment with ustekinumab [UST], only 65% of patients who achieved IUS response in the ileum at week 8 achieved endoscopic response at week 48.¹³ After the development of the consensus definition of early IUS response, a small prospective study of adults with ileal and colonic CD undergoing treatment with anti-tumour necrosis factor [TNF]-α therapy was performed and found that an 18% reduction in BWT 4–8 weeks after treatment initiation had a 77% accuracy to predict T2T endoscopic response at weeks 12–32.¹⁴

To date, there have been only two small studies in children with CD, neither of which included comparison to endoscopy, that have shown BWT and hyperaemia on IUS decreases early in response to biologic therapy induction.^15,16 In children with CD, there have been no prospective studies assessing the role of IUS to predict T2T endoscopic outcomes. Accordingly, we aimed to evaluate if early IUS response at week 8 predicts T2T endoscopic outcomes and compare IUS to early changes in clinical disease activity and CRP in children with TI CD initiating biologic therapy.

2. Materials and Methods

2.1. Study design

This was a prospective, single-centre, longitudinal, cohort study. Children <18 years old with active TI CD only [defined as ileocolonoscopy SES-CD ≥ 4 for the TI segment] and/or transmural inflammation on IUS as evidenced by increased TI BWT [defined as BWT > 3 mm] and the presence of any bowel wall hyperaemia detected by colour Doppler signal initiating biologic therapy treatment with either anti-TNF-α therapy (infliximab [IFX] or ADA) or UST were eligible for inclusion between September 2020 and September 2021 and enrolled in the PRospective Intestinal Ultrasound [PRIUS] Database. Children with a history of or current colonic disease detected on either ileocolonoscopy or IUS were excluded. Additional exclusion criteria included prior CD surgery, prior stricturing or internal penetrating complications, and failure to complete a T2T ileocolonoscopy 8–12 months after initiating biologic therapy.

2.2. Study procedures

Demographics, including age at the time of enrollment, age at diagnosis, disease duration, gender, CD location and behaviour [Paris Classification^17], and surgical and treatment history were collected at the initial visit. IUS was performed, the Pediatric Crohn’s Disease Activity Index [PCDAI] was calculated, and serum was collected for CRP at each study visit: within 2 weeks prior to the initiation of biologic therapy [baseline], 8 weeks and 6 months [±1 month] after treatment initiation, and at the time of T2T ileocolonoscopy performed within 12 months after treatment initiation. IUS was performed within 21 days of T2T ileocolonoscopy, but never on the same day, avoiding any potential effects from bowel preparation. Treatment intensification decisions [either increasing the dose and/or decreasing the treatment interval] were not protocolized and were made by the treating gastroenterologists who were not blinded to IUS, PCDAI, or CRP results.

2.3. Intestinal ultrasound

All IUS examinations were performed by a gastroenterologist [M.D.] who was previously trained and certified according to the International Bowel Ultrasound Group [IBUS] criteria, who is also a certified expert trainer within IBUS. M.D. was blinded to the clinical, laboratory, and endoscopy data available at the time each IUS was performed for each patient. Children did not undergo any fasting or bowel preparation prior to the IUS examination. The IUS examinations were performed with a Samsung RS85 Prestige located in the IBD clinic exam room using a convex probe [3–10 MHz] for initial global assessment and a linear probe [2–14 MHz] for detailed measurements. The IUS technique included a brief survey of the pelvis followed by a complete greyscale and colour Doppler evaluation of the colon beginning with the sigmoid colon in the left lower quadrant of the abdomen until the TI was identified in the right lower quadrant of the abdomen superior to the iliac vessels and iliopsoas muscle. Standard assessments were reported for the TI: [1] BWT [mm], measured and recorded as the average of four measurements, two in the longitudinal plane, and two in the cross-sectional plane; [2] bowel wall hyperaemia as measured by the presence or absence of colour Doppler signal, with a velocity rate scale of ±5.2 cm/s, and graded according to the semi-quantitative modified Limberg score [MLS: 0-III]; [3] presence or absence of mesenteric inflammatory fat wrapping; [4] loss of preservation or preservation of bowel wall layer stratification; and [5] presence or absence of reactive lymph nodes in the mesentery. The International Bowel Ultrasound Segmental Activity Score [IBUS-SAS], a comprehensive score from 0 to 100 (IBUS-SAS [0–100] = 4*BWT + 15*inflammatory fat presence + 7*MLS + 4*bowel wall stratification) developed by international experts,¹⁸ was calculated for the terminal ileum at each study timepoint.

2.4. Ileocolonoscopy

Ileocolonoscopy was performed at T2T within 12 months of biologic therapy initiation independently at the discretion of and by the treating gastroenterologist and within 21 days of the T2T IUS examination. The proceduralist performing the T2T ileocolonoscopy and calculating the SES-CD was blinded to results of the T2T IUS examination. The SES-CD¹⁹ for the TI was calculated according to findings at the time of colonoscopy by the proceduralist.

2.5. Outcomes

The primary outcome of the study was accuracy of optimal IUS cut-points to predict T2T endoscopic remission in the TI [SES-CD ≤ 2] for [1] percentage change in BWT from baseline to week 8, [2] absolute change in BWT from baseline to week 8, and [3] absolute BWT at week 8. Secondary outcomes included the accuracy of optimal percentage and absolute change in IBUS-SAS, PCDAI and CRP from baseline to week 8 and absolute IBUS-SAS, PCDAI and CRP at week 8 to predict T2T endoscopic remission in the TI, and association between changes in IUS parameters, IBUS-SAS PCDAI, and CRP between each study follow-up timepoint and T2T endoscopic remission in the TI.

2.6. Statistical considerations

Descriptive statistics summarized the data as frequencies and percentages for categorical variables and median and interquartile range [IQR] for continuous variables. Differences between children who achieved T2T endoscopic remission vs children who did not at baseline were tested by Fisher’s exact test for dichotomous variables or Mann–Whitney U-test for continuous non-parametric variables. Optimal cut points for area under the receiver operating curve [AUROC] were used to determine the accuracy, sensitivity, specificity, positive predictive value [PPV], and negative predictive value [NPV] to predict endoscopic remission based on the Youden index. Wilcoxon rank tests or Mcnemar tests were used to compare paired samples where appropriate. A result was considered statistically significant at p < 0.05. Statistical analyses were conducted utilizing R studio v.2022.12.0 + 353.

2.7. Ethical statement

Informed consent was obtained from all patients and this study was approved by the institutional review board at the Icahn School of Medicine at Mount Sinai [IRB #17-01304].

3. Results

From September 2020 to September 2021, 44 children with active ileal CD initiating biologic therapy were enrolled. All children completed a T2T ileocolonoscopy at a median timepoint of 8 [7–11] months after treatment initiation. Endoscopic remission was achieved in 29 [66%] children. At baseline, children who achieved endoscopic remission had a shorter disease duration (2 [1–5.8] months vs 25.5 [3.3–40.2] months, p = 0.04), lower PCDAI (20 [20–30] vs 35 [30–50], p = 0.017), lower TI BWT (4.4 [4.0–5.2] mm vs 5.9 [5.3–6.3] mm, p = 0.004), and lower percentage of inflammatory fat presence (13/29 [45%] vs 13/15 [87%], p = 0.01) compared with the 15 children who did not achieve endoscopic remission. Overall, 12/14 [86%] children treated with IFX, 7/9 [78%] with ADA, and 10/21 [48%] with UST achieved endoscopic remission [Table 1].

Table 1.

Baseline characteristics.

Clinical characteristics	T2T endoscopic remission [N = 29]	No T2T endoscopic remission [N = 15]	p-Value
Female, N [%]	9 [31]	8 [53]	0.197
Age, years, median [IQR]	14 [13–16]	13 [11–17]	0.7011
Disease duration, months, median [IQR]	2 [1.5–5.8]	25.5 [3.3–40.2]	0.040
CD characteristics
Lower gastrointestinal involvement, N [%]
L1: Ileal	29 [100]	15 [100]	–
Upper gastrointestinal involvement, N [%]	7 [24]	3 [20]	1.00
Perianal involvement, N [%]	1 [3]	1[7]	1.00
Behaviour phenotype, N [%]
Non-stricturing, non-penetrating	29 [100]	15 [100]	1.00
Previous biologic therapy, N [%]
0	22 [76]	7 [47]	0.092
1	5 [17]	7 [47]	0.071
2	2 [7]	1 [6]	1.00
Biologic therapy initiated, N [%]
Infliximab	12 [41]	2 [13]	0.089
Adalimumab	7 [24]	2 [13]	0.695
Ustekinumab	10 [35]	11 [74]	0.025
Concomitant steroids, N [%]	8 [28]	6 [40]	0.500
Baseline PCDAI, median [IQR]	20 [20–30]	35 [30–50]	0.017
Baseline CRP, mg/L, median [IQR]	8.5 [1–22.1]	23.1 [8.7–39.4]	0.057
Baseline TI IUS parameters
Bowel wall thickness, mm, median [IQR]	4.4 [4.0–5.2]	5.9 [5.3–6.3]	0.004
Hyperaemia on colour Doppler signal, N [%]	28 [97]	15 [100]	1.00
Modified Limberg score [0–III], N [%]
0	1 [3]	0 [0]	1.00
I	9 [31]	2 [13]	0.281
II	8 [28]	3 [20]	0.722
III	11 [38]	10 [67]	0.112
Loss of bowel wall stratification, N [%]	22 [76]	12 [80]	1.00
Presence of inflammatory fat, N [%]	13 [45]	13 [87]	0.010
Presence of lymphadenopathy, N [%]	6 [21]	6 [40]	0.284

PCDAI, Pediatric Crohn’s Disease Activity Index; CRP, C-reactive protein; IUS, intestinal ultrasound; IQR, interquartile range.

3.1. Association of TI BWT and T2T endoscopic remission

Between baseline and each follow-up timepoint, children who achieved endoscopic remission had a significantly lower absolute TI BWT (baseline: 4.4 [4.0–5.3] mm vs 5.9 [5.2–6.3] mm, p = 0.004; week 8: 2.9 [2.3–3.2] mm vs 6.0 [5.3–7.0] mm, p < 0.0001; 6 months: 2.0 [1.5–2.4] mm vs 5.2 [3.9–5.7] mm, p < 0.0001; T2T: 1.8 [1.2–2.2] mm vs 5.2 [5.0–6.1] mm, p < 0.0001) compared with children who did not achieve endoscopic remission [Figure 1A]. Children who achieved endoscopic remission had a greater percentage change in TI BWT vs children who did not achieve endoscopic remission (week 8: −38% [-−30 to −48] vs 0% [–3 to 12] change, p < 0.0001; 6 months: −53% [−46 to −66] vs 7% [−3 to 11] change, p < 0.0001; T2T: −61% [−53 to −70] vs −19% [−9 to −40] change, p < 0.001; Figure 1B). Children who achieved endoscopic remission also had a greater absolute change in BWT from baseline to week 8 (−1.8 [−1.2 to −2.3] mm vs 0.0 [−0.2 to 0.7] mm, p < 0.0001), and from baseline to 6 months (−2.4 [−2.1 to −2.8] mm, vs 0.0 [−1.0 to 0.1] mm, p < 0.0001) but not from baseline to T2T compared with children who did not achieve endoscopic remission. There was a significant percentage change in BWT from week 8 to 6 months (−22% [−10 to −36] vs −8% [−3 to −18], p = 0.02), but no significant percentage change for other timepoints in children who achieved endoscopic remission vs children who did not [Table 2].

Table 2.

Association change in terminal ileum bowel wall thickness, PCDAI, and CRP in children who achieved endoscopic remission vs children who do not achieve endoscopic remission.

	Endoscopic remission	No endoscopic remission	p-Value
BWT, median [IQR]
Baseline BWT	4.4 [4.0–5.3]	5.9 [5.2–6.3]	0.0042
Week 8 BWT	2.9 [2.3–3.2]	6.0 [5.3–7.0]	<0.0001
6 Months BWT	2.0 [1.5–2.4]	5.2 [3.9–5.7]	<0.0001
T2T BWT	1.8 [1.2–2.2]	5.2 [5.0–6.1]	<0.0001
Absolute BWT change from baseline to week 8 [mm]	−1.8 [−2.3 to −1.2]	0.0 [−0.2 to 0.7]	<0.0001
Absolute BWT change from baseline to 6 months [mm]	−2.4 [−2.8 to −2.1]	0.0 [−0.1 to 1.0]	<0.0001
Absolute BWT change from baseline to T2T [mm]	−2.6 [−3.3 to −2.1]	−1.3 [−5.6 to −0.9]	0.075
Absolute BWT change from week 8 to 6 months [mm]	−0.6 [−1.1 to −0.3]	−0.5 [−0.8 to −0.2]	0.560
Absolute BWT change from week 8 to T2T [mm]	−0.8 [−1.4 to −0.5]	−1.2 [−6.1 to −0.6]	0.234
Absolute BWT change from 6 months to T2T [mm]	−0.2 [−0.4 to 0]	−0.3 [−3.4 to 0.0]	0.454
% BWT change from baseline to week 8	−38 [−48 to −30]	0 [−3 to 12]	<0.0001
% BWT change from baseline to 6 months	−53 [−66 to −46]	−7 [−11 to 3]	<0.0001
% BWT change from baseline to T2T	−61 [−70 to −53]	−10 [−40 to −9]	0.0003
% BWT change from week 8 to 6 months	−22 [−36 to −10]	−13 [−27 to −5]	0.007
% BWT change from week 8 to T2T	−33 [−48 to −19]	−24 [−40 to −12]	0.025
% BWT change from 6 months to T2T	−10 [−21 to 0]	−6 [−20 to 0]	0.24
IBUS-SAS, median [IQR]
Baseline IBUS-SAS	44 [32–80]	83 [74–84]	0.006
Week 8 IBUS-SAS	12 [10–27]	79 [72–86]	<0.0001
6 Months IBUS-SAS	8 [6–10]	76 [73–82]	<0.0001
T2T IBUS-SAS	8 [5–9]	73 [69–73]	<0.0001
Absolute IBUS-SAS change from baseline to week 8	−22 [−49 to −15]	0 [−5 to 4]	<0.0001
Absolute IBUS-SAS change from baseline to 6 months	−37 [−60 to −25]	−4 [−7 to −1]	<0.0001
Absolute IBUS-SAS change from baseline to T2T	−39 [−70 to −25]	−11 [−11 to −2]	<0.0001
Absolute IBUS-SAS change from week 8 to 6 months	−4 [−18 to −1]	−1 [−2 to 3]	0.009
Absolute IBUS-SAS change from week 8 to T2T	−6 [−23 to −2]	−1 [−8 to 4]	0.046
Absolute IBUS-SAS change from 6 months to T2T	−1 [−2 to 0]	0 [−2 to 0]	0.80
% IBUS-SAS change from baseline to week 8	−60 [−82 to −37]	0 [−4 to 6]	<0.0001
% IBUS-SAS change from baseline to 6 months	−82 [−88 to −77]	−4 [−10 to 1]	<0.0001
% IBUS-SAS change from baseline to T2T	−87 [−91 to −79]	−13 [−13 to −2]	<0.0001
% IBUS-SAS change from week 8 to 6 months	−37 [−63 to −13]	−1 [−3 to 4]	<0.0001
% IBUS-SAS change from week 8 to T2T	−48 [−84 to −22]	−2 [−10 to 9]	0.0001
% IBUS-SAS change from 6 months to T2T	0 [−21 to 0]	0 [−27 to 0]	0.09
PCDAI, median [IQR]
Baseline PCDAI	20 [20–30]	35 [30–50]	0.018
Week 8 PCDAI	10 [0–10]	20 [13–33]	0.005
6 Months PCDAI	0 [0–10]	15 [10–29]	0.002
T2T PCDAI	0 [0–5]	10 [0–20]	0.010
Absolute PCDAI change from baseline to week 8	−13 [−25 to −8]	−10 [−15 to −9]	0.431
Absolute PCDAI change from baseline to 6 months	−20 [−30 to −10]	−10 [−29 to −4]	0.417
Absolute PCDAI change from baseline to T2T	−20 [−30 to −15]	−23 [−30 to −15]	0.714
Absolute PCDAI change from week 8 to 6 months	0 [−5 to 0]	0 [−9 to 0]	0.360
Absolute PCDAI change from week 8 to T2T	−5 [−10 to 0]	−1 [−14 to 0]	0.973
Absolute PCDAI change from 6 months to T2T	0 [−5 to 0]	0 [0 to 8]	0.173
% PCDAI change from baseline to week 8	−50 [−100 to −43]	−33 [−50 to −20]	0.045
% PCDAI change from baseline to 6 months	−100 [−100 to −50]	−48 [−78 to −10]	0.007
% PCDAI change from baseline to T2T	−100 [−100 to −83]	−60 [−100 to −45]	0.036
% PCDAI change from week 8 to 6 months	0 [−100 to 0]	0 [−30 to 0]	0.064
% PCDAI change from week 8 to T2T	−67 [−100 to 0]	−8 [−43 to 0]	0.217
% PCDAI change from 6 months to T2T	0 [−33 to 0]	0 [0 to 38]	0.109
CRP [mg/L], median [IQR]
Baseline CRP	8.5 [1.0–22.1]	23.1 [8.7–39.4]	0.058
Week 8 CRP	1.0 [0.3–4.0]	2.8 [1.2–16.7]	0.059
6 Months CRP	0.9 [0.2–3.9]	8.8 [1.4–31.4]	0.007
T2T CRP	1.3 [0.3–4.0]	4.6 [0.2–11.8]	0.403
Absolute CRP change from baseline to week 8	−1.8 [−21.2 to 0.2]	−13.7 [−23.2 to −1.4]	0.334
Absolute CRP change from baseline to 6 months	−4.8 [−21.2 to 0]	−5.3 [−21.1 to 16.1]	0.632
Absolute CRP change from baseline to T2T	−2.8 [−20.6 to 0]	−2.4 [−15.9 to 1.2]	0.518
Absolute CRP change from week 8 to 6 months	0 [−1.6 to 0.1]	2.3 [−0.1 to 12.3]	0.004
Absolute CRP change from week 8 to T2T	0 [−1.3 to 0.2]	−0.3 [−2.0 to 1.8]	0.897
Absolute CRP change from 6 months to T2T	0 [−0.3 to 0.7]	−0.9 [−7.8 to −0.5]	0.036
% CRP change from baseline to week 8	−67 [−94 to 13]	−66 [−94 to −18]	0.892
% CRP change from baseline to 6 months	− 80 [−96 to 0]	−73 [− 90 to 118]	0.271
% CRP change from baseline to T2T	−69 [−95 to 0]	−15 [−95 to 66]	0.567
% CRP change from week 8 to 6 months	0 [−55 to 22]	35 [−64 to 118]	0.008
% CRP change from week 8 to T2T	0 [−41 to 18]	−4 [−75 to 173]	0.941
% CRP change from 6 months to T2T	0 [−33 to 59]	−38 [−80 to −23]	0.092

BWT, bowel wall thickness; IBUS-SAS, International Bowel Ultrasound Segmental Activity Score; PCDAI, Pediatric Crohn’s Disease Activity Index; CRP, C-reactive protein; IUS, intestinal ultrasound; IQR, interquartile range.

Figure 1.

Terminal ileum bowel wall thickness stratified by endoscopic remission. [A] Lower bowel wall thickness was associated with children who achieved endoscopic remission vs children who did not achieve endoscopic remission at each study timepoint. Baseline: 4.4 [4.0–5.2] mm vs 5.9 [5.3–6.3] mm, p = 0.004; Week 8: 2.9 [2.3–3.2] mm vs 6.0 [5.3–7.0] mm, p < 0.0001; 6 months: 2.0 [1.5–2.4 mm] vs 5.2 [3.9–5.7] mm, p < 0.0001; and T2T: 1.8 [1.2–2.2] mm vs 5.2 [5.0–6.1] mm, p < 0.0001. [B] Percentage decrease in bowel wall thickness from baseline was associated with endoscopic remission at each follow-up timepoint. Week 8: 37.5% [29.5–48] decrease vs 0.0% [−11.5 to 2.7] decrease, p < 0.0001; 6 months: 53% [46–66] decrease vs 7% [−3 to 11] decrease, p < 0.0001; T2T: 60.5% [53–70] vs 18.5% [9–40], p < 0.001].

3.2. Association of IBUS-SAS and T2T endoscopic remission

Between baseline and each follow-up timepoint, children who achieved endoscopic remission had a significantly lower absolute IBUS-SAS (baseline: 44 [32–80] vs 83 [74–84], p = 0.006; week 8: 12 [10–27] vs 79 [72–86], p < 0.0001; 6 months: 8 [6–10] vs 76 [73–82], p < 0.0001; T2T: 8 [5–9] vs 73 [69–73], p < 0.0001) compared with children who did not achieve endoscopic remission. Children who achieved endoscopic remission had a greater percentage change in IBUS-SAS vs children who did not achieve endoscopic remission (week 8: −60% [−37 to −82] vs 0% [−6 to 4], p < 0.0001; 6 months: −82% [−77 to 88] vs −4% [−1 to −10], p < 0.0001; T2T: −87% [−79 to −91] vs −13% [−2 to −13], p < 0.0001). Children who achieved endoscopic remission also had a greater absolute change in IBUS-SAS between each study time point except for 6 months and T2T [Table 2].

3.3. Association of PCDAI and CRP and T2T endoscopic remission

Children who achieved T2T endoscopic remission had a lower absolute PCDAI at each timepoint vs children who did not (baseline: 20 [20–30] vs 35 [30–50], p = 0.018; week 8: 10 [0–10] vs 20 [13–33], p = 0.005; 6 months: 0 [0–10] vs 15 [10–29], p = 0.002; T2T: 0 [0–5] vs 10 [0–20], p = 0.01). Absolute CRP was only significantly lower at 6 months in children who achieved T2T endoscopic remission vs children who did not (0.9 mg/L [0.2–3.9] vs 8.8 mg/L [1.4–31.4], p = 0.007). There was a greater percentage decrease, but not absolute decrease, in PCDAI from baseline to each follow-up timepoint in children who achieved T2T endoscopic remission vs children who did not and there was no significant difference in either percentage or absolute change in CRP in children who achieved T2T endoscopic remission vs children who did not from baseline to each follow-up timepoint [Table 2].

3.4. Accuracy of week 8 TI BWT, PCDAI, and CRP to predict T2T endoscopic remission

A ≥18% decrease in TI BWT from baseline to week 8 was accurate to predict T2T endoscopic remission with an AUROC of 0.99 [95% CI 0.98–1.00], 100% sensitivity, 93% specificity, 97% PPV, and 100% NPV [Figure 2]. An absolute decrease in TI BWT ≥ 1.0 mm from baseline to week 8 was accurate to predict T2T endoscopic remission with an AUROC of 0.99 [95% CI 0.98–1.0], 100% sensitivity, 93% specificity, 97% PPV, and 100% NPV. An absolute decrease in TI BWT ≥ 1.0 mm and a decrease in bowel wall hyperaemia by ≥1 point based on the MLS from baseline to week 8 was accurate to predict endoscopic remission with an AUROC of 0.90 [95% CI 0.81–0.99], 86% sensitivity, 93% specificity, 96% PPV, and 78% NPV.

Figure 2.

Percentage change in bowel wall thickness from baseline to predict endoscopic remission. Baseline to week 8: 18.2% decrease in BWT predicted endoscopic remission with an AUROC of 0.995 [95% CI 0.984–1.000], 100% sensitivity, 93.3% specificity, 96.7% PPV, and 100% NPV. Baseline to 6 months: 33.3% decrease in BWT predicted endoscopic remission with an AUROC of 0.988 [95% CI 0.962–1.000], 100% sensitivity, 93% specificity, 97% PPV, and 100% NPV. Baseline to T2T: 46.5% decrease in BWT predicted endoscopic remission with an AUROC of 0.893 [95% CI 0.698–1.000], 96% sensitivity, 90% specificity, 96% PPV, and 90% NPV.

An absolute TI BWT cut-off at week 8 of ≤5.2 mm was accurate to predict T2T endoscopic remission with an AUROC of 0.96 [95% CI 0.90–1.0], 100% sensitivity, 80% specificity, 91% PPV, and 100% NPV.

A ≥26% decrease in IBUS-SAS (AUROC of 0.94 [95% CI 0.86–1.0], 93% sensitivity, 93% specificity, 96% PPV, and 88% NPV), a ≥46% decrease in PCDAI (AUROC of 0.67 [95% CI 0.49–0.84], 72% sensitivity, 64% specificity, 81% PPV, and53% NPV) and a ≥84% decrease in CRP (AUROC of 0.49 [95% CI 0.31–0.67] 45% sensitivity, 67% specificity, 72% PPV, and 39% NPV) from baseline to week 8 predicted T2T endoscopic remission. An absolute IBUS-SAS ≤58 (AUROC 0.95 [95% CI 0.87–1.0], 97% sensitivity, 87% specificity, 93% PPV, and 93% NPV), absolute PCDAI ≤ 15 (AUROC 0.76 [95% CI 0.58–0.94], 83% sensitivity, 67% specificity, 83% PPV, and 67% NPV) and an absolute CRP ≤ 14 mg/L (AUROC 0.68 [95% CI 0.56–0.86] 100% sensitivity, 33% specificity, 74% PPV, and 100% NPV) at week 8 predicted T2T endoscopic remission [Table 3].

Table 3.

Accuracy of week 8 changes in TI BWT, PCDAI, and CRP to predict endoscopic remission.

Week 8 measure	Cut-point	AUROC [95% CI]	Sensitivity [%]	Specificity [%]	PPV [%]	NPV [%]
% Decrease in BWT	≥18	0.99 [0.98–1.0]	100	93	97	100
Absolute decrease in BWT [mm]	≥1.0	0.99 [0.98–1.0]	100	93	97	100
Absolute BWT [mm]	≤5.2	0.96 [0.90–1.0]	100	80	91	100
% Decrease in IBUS-SAS	≥26	0.94 [0.86–1.0]	93	93	96	88
Absolute decrease in IBUS-SAS	≥12.6	0.93 [0.86–1.0]	86	87	93	76
Absolute IBUS-SAS	≤58.2	0.95 [0.87–1.0]	97	87	93	93
% Decrease in PCDAI	≥46	0.67 [0.49–0.84]	72	64	81	53
Absolute decrease in PCDAI	≥20	0.57 [0.40–0.75]	48	80	82	44
Absolute PCDAI	≤15	0.76 [0.58–0.94]	83	67	83	67
% Decrease in CRP	≥84	0.49 [0.31–0.67]	45	67	72	39
Absolute decrease in CRP [mg/L]	≥8.6	0.41 [0.22–0.60]	100	13	69	100
Absolute CRP [mg/L]	≤13.8	0.68 [0.56–0.86]	100	33	74	100

BWT, bowel wall thickness; IBUS-SAS, International Bowel Ultrasound Segmental Activity Score; PCDAI, Pediatric Crohn’s Disease Activity Index; CRP, C-reactive protein; AUROC, area under the receiver operating curve; NPV, negative predictive value; PPV, positive predictive value.

3.5. Comparison of week 8 TI BWT cut-offs and current norms to predict T2T endoscopic remission

An optimal cut-off of a ≥18% decrease in TI BWT from baseline to week 8 (AUROC 0.99 [95% CI 0.98–1.00]) was similar in accuracy to a ≥20% decrease in TI BWT from baseline to week 8 (AUROC of 0.95 [95% CI 0.88–1.0]) and the defined ≥25% decrease in BWT for IUS response (AUROC 0.95 [95% CI 0.89–1.0]) to predict T2T endoscopic remission. A ≥1 mm decrease in TI BWT from baseline to week 8 was more accurate to predict T2T endoscopic remission than the defined ≥2 mm decrease in BWT for IUS response (AUROC 0.99 [95% CI 0.98–1.0] vs 0.69 [95% CI 0.53–0.82], p < 0.0001). A ≤5.2 mm TI BWT absolute cut-off was more accurate to predict T2T endoscopic remission than using a normal BWT cut-off of ≤3.0 mm (AUROC 0.96 [95% CI 0.90–1.0] vs 0.79 [0.69–0.90], p = 0.04) [Table 4].

Table 4.

Accuracy of week 8 TI BWT cut-points and current BWT, PCDAI, and CRP norms to predict T2T endoscopic remission.

Week 8 cut-point	AUROC [95% CI]	Sensitivity [%]	Specificity [%]	PPV [%]	NPV [%]
≥18% Decrease in BWT	0.99 [0.98–1.0]	100	93	97	100
≥20% Decrease in BWT	0.95 [0.88–1.0]	97	93	97	93
≥25% Decrease in BWT	0.95 [0.89–1.0]	90	100	100	83
≥1.0 mm Absolute decrease in BWT	0.99 [0.98–1.0]	100	93	97	100
≥2.0 mm Absolute decrease in BWT	0.69 [0.53–0.82]	38	100	100	44
Absolute BWT ≤5.2 mm	0.96 [0.90–1.0]	100	80	91	100
Absolute BWT ≤3.0 mm	0.79 [0.69–0.90]	66	93	95	58
PCDAI < 10	0.75 [0.61–0.89]	76	73	85	61
CRP ≤ 5 mg/L	0.61 [0.47–0.76]	73	55	83	40

BWT, bowel wall thickness; PCDAI, Pediatric Crohn’s Disease Activity Index; CRP, C-reactive protein; AUROC, area under the receiver operating curve; NPV, negative predictive value; PPV, positive predictive value.

3.7. Additional IUS parameters and T2T endoscopic remission

Children who achieved endoscopic remission were more likely to have normalization of all other IUS parameters measured at week 8 vs children who did not achieve endoscopic remission: presence of bowel wall hyperaemia ([11/29 [38%] vs 14/15 [93%], p = 0.004), loss of preservation of bowel wall stratification (8/29 [28%] vs 13/15 [87%], p < 0.0001), presence of inflammatory fat (5/29 [17%] vs 13/15 [87%], p < 0.0001), and presence of lymphadenopathy (2/29 [7%] vs 5/15 [33%], p = 0.023). At 6 months and T2T, children who achieved endoscopic remission were also more likely to have normalization of all other IUS parameters [Figure 3A and B].

Figure 3.

Normalization of IUS parameters in children who achieve endoscopic remission vs children who do not achieve endoscopic remission. Children who achieved endoscopic remission were more likely to have normalization of all other IUS parameters at each subsequent follow-up timepoint except for the presence of lymphadenopathy at 6 months and T2T.

3.8. Treatment intensification

Overall, 30/44 [68%] children underwent treatment intensification at or after week 8, nine of which were after week 8. There were no differences in the number of children who underwent treatment intensification who achieved endoscopic remission vs children who underwent treatment intensification who did not achieve endoscopic remission (18/29 [62%] vs 12/15 [80%], p = 0.31). There was no difference in the number of children who underwent treatment intensification who achieved a ≥18% decrease in BWT at week 8 compared with children who did not achieve a ≥18% decrease in BWT at week 8 (16/30 [53%] vs 11/14 [79%], p = 0.18).

Of the 30 children who underwent treatment intensification, the drivers were CRP only in six [20%], IUS, CRP, and PCDAI together in five [17%], and IUS and CRP together in five [17%]. Proactive therapeutic drug monitoring alone was utilized to drive treatment intensification in four [13%] children. IUS was utilized alone to drive treatment intensification in one [3%] child [Supplementary Table 1].

4. Discussion

In this prospective study of children with ileal CD undergoing treatment with biologic therapy, we demonstrated that early reduction in BWT on IUS was highly predictive of T2T endoscopic remission. A ≥18% decrease in BWT at week 8 at the end of biologic therapy induction predicted T2T endoscopic remission with the highest accuracy, with an AUROC of 0.99, 100% sensitivity, 93% specificity, 97% PPV, and 100% NPV. These results are in line with the prospective study performed by De Voogd et al.¹⁴ In a smaller cohort of 31 adults with CD [both ileal and colonic] undergoing induction with anti-TNF-α therapy and T2T endoscopy 12–32 weeks after treatment initiation, they found that an 18% decrease in BWT 4–8 weeks after treatment initiation was accurate to detect T2T endoscopic response with an AUROC of 0.77, 82% sensitivity, 71% specificity, 64% PPV, and 86% NPV. However, our study finds that an 18% decrease in ileal BWT from baseline to week 8 may also be accurate to predict T2T endoscopic remission, not just endoscopic response, in children with ileal CD.

Prior to the publication by De Voogd et al., international expert consensus guidelines defined IUS response as a ≥25% decrease in BWT at week 14 after treatment initiation. However, aiming for a 25% decrease in BWT and waiting until week 14 may not be necessary in children with CD, as both our study and the study by De Voogd et al. found earlier response to therapy at week 8 to be predictive of endoscopic outcomes. Additionally, targeting an earlier 20% decrease in BWT at week 8 in clinical practice may be able to guide treatment optimization towards a goal of T2T endoscopic remission. With these single-centre studies, consideration of prospective multicentre studies in both adults and children with ileal and colonic CD are clearly needed to better understand how early changes on IUS can be utilized to predict endoscopic and long-term outcomes.

Improved understanding of the natural history of transmural and endoscopic response to biologic therapies is needed in CD. There are many factors, including disease duration, severity, and phenotype, that probably have a large influence on the rate of both mucosal and transmural healing, and a potential lag that may exist between the two. For example, in the IUS sub-study from STARDUST of adults with CD undergoing treatment with UST, IUS response in the terminal ileum was found in only 6/26 [23%] patients, but there was a discrepant 73% endoscopic response in these patients. Overall, they found that terminal ileal BWT decreased by 11.9% at week 8 in the cohort with ileal CD, suggesting a potential lag in the timing of transmural healing and mucosal healing.^13,20 Additionally, the role of BWT at baseline requires further exploration. Our findings indicate that children with ileal CD with a lower BWT at baseline are more likely to achieve endoscopic remission after treatment with biologic therapy, regardless of early IUS responsiveness. Future multicentre and translational studies are needed on the role of BWT, and individual bowel wall layer thicknesses, to better classify CD, and predict response to therapy, long-term outcomes and how different immunophenotypes drive inflammation.

Previously, a radiologist-performed IUS study by Dillman et al.,¹⁵ of 28 children with newly diagnosed ileal CD undergoing treatment with IFX and monitored with IUS at 2 weeks, 1 month, 3 months, and 6 months after treatment initiation, but without comparison to endoscopy, found that ileal BWT decreased by 23% from a median BWT of 5.6 to 4.3 mm by 1 month. Our study found that a 37% decrease in TI BWT from baseline to week 8 was associated with endoscopic remission, and while the study by Dillman et al. was not compared to T2T endoscopic outcomes, the decrease in BWT over time is similar to our study.

In our study, absolute TI BWT response and remission cut-points at week 8 were superior to traditional cut-points for PCDAI and CRP to predict T2T endoscopic remission. These results suggest that directly monitoring change in BWT could be used in an IUS-dominant tight control strategy to guide therapeutic optimization with the potential to drive endoscopic remission rates higher in children with ileal CD than monitoring with traditional biomarkers, but future head-to-head comparison studies for outcomes based on an IUS vs traditional biomarker monitoring strategy are needed.

The main limitation of our study was that treatment intensification was not protocolized and treatment intensification decisions were at the discretion of the treating gastroenterologist who was not blinded to the results of the IUS examination. By not protocolizing treatment decisions based on defined measures of IUS, clinical, and CRP response, we were unable to compare the effect of an early IUS target-based therapeutic intensification strategy on T2T endoscopic outcomes to more traditional targets of PCDAI and CRP. However, in our study there was no significant difference between children who underwent treatment intensification and children who did not who achieved either endoscopic remission or a >18% decrease in BWT at week 8. Furthermore, IUS response was only the sole driver of treatment intensification in one child. Future protocolized treatment intensification studies comparing IUS response to changes in clinical symptoms and serum and stool biomarker responses are needed and are currently in development in both children and adults with CD. Other limitations were the small sample size from a single centre with a lack of an external, multicentre validation cohort, and lack of central reading for endoscopy outcomes; although this was the largest prospective study to date, larger, prospective, multicentre studies in both children and adults are needed.

Overall, early reduction in TI BWT on IUS at the end of biologic therapy induction, of at least 18%, accurately predicts T2T endoscopic remission and aiming for a BWT reduction of greater than 25% may not be necessary. Absolute TI BWT response and remission cut-offs at week 8 are superior to traditional cut-offs to predict T2T endoscopic remission. Future studies on the accuracy of IUS target-based treatment optimization vs traditional biomarker and clinical targets to predict T2T endoscopic outcomes are needed.

Supplementary Data

Supplementary data are available online at ECCO-JCC online.

Conference

Digestive Disease Week 2023.

Funding

This work was supported by an internal institutional grant from the Icahn School of Medicine at Mount Sinai, the RFA Paediatrics Scholar Grant.

Conflict of Interest

Michael T. Dolinger: Consulting fees from Neurologica Corp., a subsidiary of Samsung Electronics Co., Ltd. Illya Aronskyy: nothing to disclose. Amelia Kellar: Nothing to disclose. Elizabeth A. Spencer: Consulting fees from Prometheus Biosciences. Nanci Pittman: Consulting fees from Prometheus Biosciences. Marla C. Dubinsky: Consulting fees from Abbvie, Allergan, Amgen, Arena Pharmaceuticals, AstraZeneca, Boehringer Ingelheim, Celgene, Ferring, Genentech, Gilead, Hoffmann-La Roche, Janssen, Pfizer, Prometheus Biosciences, Takeda, Target PharmaSolutions. Research funding from Abbvie, Janssen, Pfizer, Prometheus Biosciences, Takeda.

Acknowledgments

This work received generous support from Henry and Elaine Kaufman in addition to funding from an internal institutional grant.

Author Contributions

Michael T. Dolinger: Conceptualized, planned, and conducted the study, collected and interpreted the data, and drafted the manuscript. Illya Aronskyy: Collected, interpreted and analysed the data, and reviewed and edited the manuscript. Amelia Kellar: Interpreted the data, reviewed and edited the manuscript. Elizabeth Spencer: Interpreted the data, reviewed and edited the manuscript. Nanci Pittman: Interpreted the data, reviewed and edited the manuscript. Marla C. Dubinsky: Conceptualized and planned the study, interpreted the data, reviewed and edited the manuscript. Each author has approved the final draft of the manuscript submitted.

Data Availability

The data underlying this article will be shared on reasonable request to the corresponding author.