Intestinal ultrasound in inflammatory bowel disease: diagnostic accuracy and impact on clinical decision making in a New Zealand tertiary centre

Inflammatory bowel disease (IBD), comprising Crohn’s disease (CD) and ulcerative colitis (UC), is a chronic immune-mediated condition that carries lifelong morbidity and significant healthcare burden. Global prevalence continues to rise,^2–4 with some regions still reporting increasing incidence and a shift toward earlier age at diagnosis.^2–4 This growing prevalence wave is expected to further strain health systems^2–4 and increases the need for monitoring that is timely, non-invasive, reproducible and reliable.^5,6

Minimising exposure to ionising radiation is an important goal. Across the lifespan of a chronic disease, diagnostic imaging exposure can accumulate.^7–9 Computed tomography (CT) involves ionising radiation,⁹ and cohort studies in IBD, particularly CD, demonstrate non-trivial cumulative effective doses, with a measurable proportion of patients exceeding thresholds such as 50mSv.^7,8 These findings support adopting radiation-sparing strategies when clinically appropriate.^7–9

Modern IBD care has shifted towards treat-to-target approaches, where escalation and optimisation of therapy are guided by objective measures of inflammation rather than symptoms alone.^5,6 In practice, this requires a monitoring tool that can be embedded in routine clinic workflows and repeated frequently.

Intestinal ultrasound (IUS) is a radiation-free bedside modality providing real-time assessment of bowel wall thickness, vascularity, motility and complications.^10,11 International guidance increasingly positions IUS within diagnostic and monitoring pathways.^10–14 The American Gastroenterological Association (AGA) Clinical Practice Update summarises evidence supporting IUS integration into care,¹⁰ and the AGA also acknowledges pragmatic faecal calprotectin (FCP) thresholds, including 150mcg/g, used in clinical workflow.¹ The 2025 British Society of Gastroenterology guideline incorporates biomarkers and non-invasive tools within treat-to-target algorithms.¹² In 2025, ECCO–ESGAR–ESP–IBUS (European Crohn’s and Colitis Organisation; European Society of Gastrointestinal and Abdominal Radiology; European Society of Pathology; International Bowel Ultrasonography Group) consensus documents further detailed standardised IUS technique, reporting and quality assurance.^13,14 Prospective studies, including TRUST&UC, demonstrate responsiveness to change and clinically useful performance for disease activity assessment and follow-up.¹⁵ In selected contexts, IUS combined with biomarkers can help rule out active inflammation,^1,10 informing triage and resource use.

New Zealand has a high IBD burden in international comparisons, with linkage studies estimating prevalence around 390 per 100,000 and incidence around 28 per 100,000 person-years.^16,17 To our knowledge, IUS capacity remains concentrated in a small number of centres, and point-of-care use is not yet routine. Given the high incidence and young age distribution at diagnosis, radiation-free assessment and rapid point-of-care decision making are particularly relevant. We therefore audited a consecutive series of IUS examinations at Wellington Regional Hospital, evaluating correlation with follow-up assessments and quantifying clinical decision-making outcomes, with a focus on point-of-care use.

Methods

Design and settings

We conducted a retrospective audit of consecutive clinical IUS examinations performed at Wellington Regional Hospital between December 2021 and February 2025. Scans were undertaken as point-of-care assessments during outpatient clinics and on dedicated elective IUS lists. All examinations were performed by a single gastroenterologist.

Each examination was treated as a distinct event. Data were extracted from IUS reports and the local electronic clinical record (Single Clinical Portal). Variables collected included age, sex, diagnosis, encounter setting (inpatient or outpatient), scan scope (full or focussed), image quality and complications. Pregnancy status was coded as pregnant (yes), female of childbearing age (<50 years) but not pregnant (no), or female ≥50 years or male (not applicable). Anastomosis status was coded as visualised on IUS (yes), documented in clinical notes but not visualised (no) or no known anastomosis (not applicable).

Disease activity on IUS was classified as inactive (normal or quiescent appearances, or a fibrostenotic phenotype without sonographic features of active inflammation) or active (any degree of sonographic inflammation from mild to severe). Inflammation at an anastomosis was classified as active.

Follow-up investigations within 6 months after IUS included FCP, endoscopy and cross-sectional imaging.

Outcomes

Primary outcomes were correlation between IUS activity and follow-up investigations using two biomarker thresholds: 1) FCP <50mcg/g (local reference) and 2) FCP <150mcg/g (pragmatic threshold set in guidelines).¹ Where correlation was uncertain, cases were flagged and reviewed by the supervising gastroenterologist after review of clinical records. Some examinations were labelled “not commentable” and excluded from correlation analyses when meaningful comparison was not possible due to challenging post-operative anatomy, medication non-compliance, substantial therapy change or significant timing mismatch between IUS and follow-up testing.

Secondary outcomes were whether IUS altered clinical management and time to clinical decision making.

Statistics

Categorical variables are reported as counts and percentages. For primary correlation outcomes, Wilson 95% confidence intervals are reported.¹⁸

Ethics

This project was deemed out of scope by the Health and Disability Ethics Committees. The local audit committee confirmed that the project met criteria for audit and quality improvement and did not require local ethics committee review. All data were de-identified prior to analysis.

Results

Study cohort and characteristics

Across the study period there were 373 examinations in 253 patients. Baseline examination characteristics are shown in Table 1.

View Table 1–5.

IUS findings

Disease activity on IUS is outlined in Table 2. Recorded disease complications were present in 59 examinations (15.8%), most commonly stricture (n=49). Ten patients were pregnant at time of IUS, 130 females of childbearing age were not pregnant and there were 233 not applicable cases. Known anastomoses were visualised in 108 patients, were known but not visualised in 11 and not present in 254.

Follow-up investigations

Follow-up investigations within 6 months were available for 160/373 scans (42.9%). Some examinations had more than one follow-up investigation within 6 months. The distribution of these is outlined in Table 3.

FCP was the most common investigation. Most investigations occurred 8–30 days after the index IUS (64/160), with another 48/160 occurring within the first 7 days.

Primary outcome

Excluding the five “not commentable” examinations, correlation was 82.6% (128/155) using FCP <50mcg/g, and 91.0% (141/155) using FCP <150mcg/g. Including all follow-up examinations produced similar results (N=160, correlation with FCP <50mcg/g=80%, FCP <150mcg/g 88.1%).

Secondary outcomes

Across the full cohort, IUS altered clinical management in 134 of 373 examinations (35.9%). Among examinations that led to a change, 57 occurred in point-of-care sessions in-clinic (45% of total point-of-care scans) and 77 occurred in elective sessions on a dedicated list (31% of total elective scans). Treatment-altering decisions were made on the same day in 70 of 134 decision-altering cases (52.2%). The remaining decisions were distributed as outlined in the Table 5 below.

Discussion

In this real-world tertiary setting, IUS correlated well with follow-up assessments commonly used in routine care, and frequently informed management decisions. This supports its value as a point-of-care tool in a system with limited endoscopy resource and cross-sectional imaging capacity.^19,20

These findings align within a rapidly maturing evidence base supporting IUS for IBD assessment and monitoring.^10,13,14,21 Prospective studies such as TRUST&UC have shown that changes in IUS parameters, including bowel wall thickness, track endoscopic outcomes over time.¹⁵ Reviews and meta-analyses report that, in appropriately trained hands, IUS performs comparably to endoscopy for activity assessment to magnetic resonance enterography (MRE) or CT enterography (CTE) for evaluating disease activity.^21,22 International guidance has increasingly incorporated IUS within diagnostic and monitoring pathways,^10–14 and supports pragmatic biomarker-based strategies (e.g., FCP thresholds) in selected contexts to reduce reliance on invasive or radiation-based testing when the likelihood of active inflammation is low.¹

Interpretation of correlation in this audit must consider the limitations of the reference standards available in routine care. Methodologically, an ideal reference standard for IUS would be cross-sectional imaging such as CTE or MRE because these modalities, like IUS, evaluate transmural inflammation and complications.^13,23 ECCO–ESGAR guidance emphasises the central role of cross-sectional imaging for small bowel CD and complications, with modality selection balancing diagnostic yield with resource implications.²³ In the United Kingdom multicentre METRIC diagnostic accuracy study, both MRE and IUS demonstrated high sensitivity for detecting small bowel CD, although MRE showed higher overall accuracy.²² In this cohort, cross-sectional imaging following IUS occurred infrequently (9.4%), reflecting both real-world constraints and service intent of reducing radiology demand. The most common comparators were FCP and endoscopy, which primarily reflect mucosal inflammation and are therefore imperfect surrogates for the transmural signal IUS is designed to detect.^13,23,24 Against this background, correlation estimates in the 80–90% range are reassuring.

Because FCP was a major comparator, we pre-specified two thresholds that reflect distinct clinical uses. Our local laboratory reference of FCP <50mcg/g was used as the conservative definition. A higher threshold of around 150mcg/g is commonly used in clinical algorithms to help rule out active inflammation in selected contexts.^1,24 For example, AGA guidance in UC supports FCP <150mcg/g (together with normal inflammatory markers) as a strategy to rule out active inflammation in patients in symptomatic remission and avoid routine endoscopy, while higher values should prompt further evaluation.¹ ECCO–ESGAR diagnostic guidance emphasises interpreting biomarkers in clinical context²³ and practical calprotectin reviews note that thresholds around 150mcg/g provide reasonable diagnostic accuracy in many settings, with optimal cutoffs varying by phenotype and context.^24,25  

FCP thresholds are not absolute, particularly for isolated small bowel CD, where clinically meaningful activity can occur despite low FCP values²⁶ and in patients who are low or non-producers of calprotectin.²⁴ Consensus guidelines highlight that while FCP <150mcg/g has good negative predictive value for significant colonic inflammation,¹ clinically meaningful small bowel activity can occur at lower values,²⁶ and serial measurements interpreted alongside symptoms and objective assessment remain essential.^1,24 This likely contributed to some discordant cases where FCP suggested remission while IUS demonstrated convincing features of active transmural inflammation.

Patterns of discordance were consistent with recognised limitations of comparator tests and anatomical blind spots of transabdominal ultrasound.¹¹ Several non-correlating cases involved isolated small bowel disease, where FCP may be low despite active inflammation on IUS and, in some cases, supportive endoscopic or cross-sectional findings.²⁶ This aligns with published work showing reduced calprotectin performance in isolated ileal disease and that clinically meaningful small bowel activity can occur despite low FCP values.²⁶ In these situations, calprotectin is a poor surrogate for transmural inflammation and ultrasound may more accurately represent inflammatory burden.^13,23,24 Conversely, rectal predominant disease was more likely to be discordant, consistent with broader experience that identifies the rectum as a relative blind spot for transabdominal ultrasound.¹¹

A small number of examinations were deemed as “not commentable” because technical factors or clinical timelines meant meaningful comparison with a single follow-up modality was not possible; these are best addressed transparently in the methods rather than forced into correlation categories. Taken together, these observations suggest discordance often reflects limitations of comparators or mismatched targets (mucosal versus transmural disease) rather than inaccuracy of ultrasound.^13,23 This is an important contextual point when interpreting IUS performance within real-world treat-to-target workflows.^5,6

Secondary outcomes support relevance at a service level. IUS led to direct treatment-altering decisions in approximately one in three examinations overall. This is consistent with real-world reports of point-of-care IUS influencing management and reducing reliance on magnetic resonance imaging and endoscopy.^19,20 In our audit, many decisions were implemented on the same day. This supports the potential of point-of-care scanning to shorten the interval between assessment and action when governance permits decisions to be made in clinic.^19,20 Where decisions were implemented later, this likely reflected workflow realities rather than uncertainty, particularly when scans were performed by a gastroenterologist who was not the primary long-term clinician. As a result, our “time to decision” reflects time to implementation rather than the moment when a recommendation became apparent. Future audit cycles could distinguish “time to recommendation” from “time to implementation” to better quantify the benefit of point-of-care scanning.²⁰

Our definition of “decision altering” focussed on therapy change or further investigation. Decisions to continue current therapy following a reassuring scan, which are often clinically valuable and may be the reason for referral, were not captured. This likely under-estimates the clinical utility of IUS, particularly in resource-constrained settings, where avoiding unnecessary escalation or invasive testing is a meaningful outcome.

This audit has several strengths. It includes consecutive examinations in routine practice across both elective and point-of-care settings, and use of pragmatic correlation definitions that reflect how biomarkers are interpreted in practice. We reported “not commentable” cases transparently rather than forcing classification. We also included outcomes meaningful to service delivery, including decision-altering impact and time to decision making, complementing prospective and trial-based evidence in recent IUS and imaging consensus statements.^11,13,14

However, several limitations should be acknowledged. This was a single-centre, operator-dependent audit, and generalisability to other centres and operators requires caution. Verification bias is possible because correlation analyses were restricted to examinations with follow-up investigations within 6 months of IUS; some patients with recent pre-ultrasound investigations were not included. Another limitation is reliance on FCP and endoscopy as primary comparators rather than cross-sectional imaging^13,23,24 (used sparingly in this cohort in keeping with service goals and real-world constraints). Documentation-based timing meant we could not reliably distinguish “time to recommendation” from “time to implementation”.

Despite these limitations, within a system seeking to minimise radiation exposure,^7–9 manage constrained radiology and endoscopy capacity and support timely decisions for a young and complex IBD population,^16,17 these findings provide pragmatic evidence supporting wider implementation of point-of-care IUS in New Zealand.^10,11,13,14 Emerging work on transmural healing and treat-to-target strategies that integrate IUS, biomarkers and cross-sectional imaging suggests such programmes may improve monitoring and support treat-to-target care.^5,6,13,14 Ongoing evaluation at this centre, including standardisation of reporting and expanded operator capacity, will be important as IUS becomes more established in local practice.