To our knowledge, this is one of the widest cohorts of CD patients where the presence of SI was investigated by means of MRE. The prevalence of active SI defined according to the ASAS MRI study group criteria amounted to 10.5%. Few studies evaluated the phenomenon by means of MRI or MRE, with an overall prevalence of SI from 10.9 to 39.0% in CD patients [19,20,21,22,23,24,25,26,27,28]. In these studies, SI was associated with demographic features such as female sex, greater age, and body mass index, and clinical features including a previous ileocecal resection, a longer disease duration, stricturing disease behavior, and an ileocolic disease localization. Among the aforementioned studies, some included established SpA patients, and SI was associated with back pain, history of dactylitis and higher Bath Ankylosing Spondylitis Functional Index scores [19,20,21,22,23,24,25,26,27,28]. Unfortunately, these studies lack homogeneity due to different inclusion and exclusion criteria or unique definitions for determining the presence of SI. Furthermore, SI has never been associated with the presence of the MRE-assessed intestinal lesions. The only two studies which reported the prevalence of SI on MRE in a whole sample of CD patients clearly specifying the SI definition, the frequencies of the phenomenon were of 17.0 and 20.0% [23, 28]. Said frequencies are significantly higher than our findings, however, these studies based the definition of SI either on the presence of active SI according to the ASAS MRI criteria or on the presence of structural lesions alone, such as erosions, potentially overestimating the prevalence of SI. In fact, although erosions are considered structural lesions specific of axSpA, a recent systematic review of the literature has found that almost 10% of patients without axSpA might present erosions detected by MRI [29]. No associations between SI and clinical or demographic features have been found in our study. However, by comparing SI positive and negative groups based on the presence of inflammatory and structural intestinal lesions, an association of SI with the presence of asymmetric mural hyperenhancement was demonstrated for the first time. This finding was confirmed following the normalization of the result by age, sex, and disease duration with an odds ratio of 8.61. This finding could serve as a clue for SpA, prompting the gastroenterologist to test for “SpA red flags,” raise suspicion of clinical SI, and refer the patient to a rheumatologist; in fact, early diagnosis is important not only for improving patient outcomes, but also for enabling personalized medical management.

The significance of bowel wall hyperenhancement has been extensively evaluated in computed tomography enterography, where it has been associated with an increased risk of disease relapse, need for surgery, neutrophilic inflammation, and endoscopic disease activity [30,31,32,33]. Although fewer studies have investigated the significance of mural hyperenhancement on MRE, the available evidence similarly links it to higher endoscopic activity and disease recurrence [34, 35]. However, data specifically addressing asymmetric mural hyperenhancement remain limited. In existing literature, this pattern has primarily been established in the differential diagnosis of CD versus other inflammatory, neoplastic, or infectious conditions, where it is considered a relatively specific marker of CD [18, 36].

Although direct evidence on clinical implication of this specific pattern is lacking, its known specificity for CD and its association with markers of more active disease suggest that it may reflect a more aggressive inflammatory phenotype, potentially contributing to extraintestinal manifestations such as sacroiliitis. No significant associations were observed with other intestinal imaging markers on MRE.

This study has several strengths. Firstly, this is the second widest cohort described in literature which evaluated SI by means of MRE in CD patients and the first study evaluating the association of this phenomenon with inflammatory and structural intestinal lesions. Moreover, this study has been conducted in one of the national referral centers for IBDs.

The present study also has some limitations: The first concerns the retrospective design of the study which did not make it possible to collect the data regarding the rheumatic symptoms of the patients enrolled. Moreover, most patients of the sample had never been evaluated by a rheumatologist according to our healthcare software; therefore, the assumption was that said patients had never experienced any articular pain worthy of further investigation.

This assumption inevitably led us to also include patients with an occult inflammatory axial involvement, as demonstrated by the presence of patients with articular damage such as ankylosis (three patients with bilateral ankylosis, 1.5% of the whole sample).

Secondly, the most relevant risk factors of SpA, such as HLA locus B status or personal and/or family history of PsO, were not available for all patients due to the fact that these were not routinely assessed in CD patients by the gastroenterologists.

Moreover, the monocentric nature of the study may not be representative of the global CD population, despite the fact that patients had been enrolled in a third-level center for IBDs.

In its current state, SI does not seem to represent a predictor of SpA in CD patients, in contrast to previous EIMs and arthralgia, which are associated with an increased risk of developing CD-related arthritis [37]. It should be noted that despite the absence of this association, in the cohort described by Giovannini and colleagues, only six cases of peripheral SpA had developed during the follow-up; therefore, the association—or absence of association—with axSpA with SI could not be stated.

Lastly, considering that BME has multiple differential diagnoses, the etiology of the suspected inflammatory lesion should be differentiated according to the localization proposed by some authors [38].

As previously stated by other authors, wider cohorts are needed, by developing complex study protocols including all the potential variables which characterize both CD and SpA phenotypes and the most updated protocols for SIJs and spine evaluation [39, 40].