10-Year Trend of Abdominal Magnetic Resonance Imaging Compared With Abdominal Computed Tomography Scans in Inflammatory Bowel Disease

Abstract

Background

Patients with inflammatory bowel disease (IBD) frequently undergo multiple computed tomography (CT) examinations. With the widespread availability of magnetic resonance imaging (MRI), it is unclear whether the use of CTs in IBD has declined. We aimed to analyze the trends of CT and MRI use in a large cohort of IBD patients in a 10-year period.

Methods

We retrospectively analyzed adults ≥18 years of age using a de-identified database, IBM Explorys. Patients with ≥1 CT of the abdomen (± pelvis) or MRI of the abdomen (± pelvis) at least 30 days after the diagnosis of Crohn’s disease (CD) or ulcerative colitis (UC) were included. We examined the factors associated with patients undergoing multiple CTs (≥5 CTs of the abdomen) and performed a trend analysis from 2010 to 2019.

Results

Among 176 110 CD and 143 460 UC patients, those with ≥1 CT of the abdomen annually increased from 2010 to 2019 with mean annual percentage change of +3.6% for CD and +4.9% for UC. Similarly, annual percentage change for patients with ≥1 MRI (CD: +15.6%; UC: +22.8%) showed a rising trend. There was a 3.8% increase in CD patients receiving ≥5 CTs of the abdomen annually compared with a 2.4% increase among UC patients in the 10-year period. Age ≥50 years, men, African Americans, public insurance payors, body mass index ≥30kg/m2, and smoking history were associated with ≥5 CTs.

Conclusions

There is a considerable increase in the number of CT scans performed in IBD patients. Further studies can explore factors influencing the use of CT and MRI of the abdomen in IBD patients.

Introduction

Inflammatory bowel disease (IBD) is characterized by chronic inflammation of the gastrointestinal system. It has a relapsing remitting course in which medical imaging plays an important role in the diagnosis of the disease, detection of complications, and assessment of the treatment response. Computed tomography (CT) scans have played an important role in the diagnosis and management of IBD for several decades. However, it has been shown that 1 in 10 IBD patients is at risk of exposure to high diagnostic medical radiation,¹ and three-quarters of the radiation exposure comes from CT scans.² While the average radiation exposure from natural radiation is 1-3 mSv, the median effective dose of an abdominal CT scan is about 8-10 mSv, with a higher radiation dose for CT with contrast and multiphase CTs.^3,4 Radiation from CT scans can induce chromosome aberration including translocation, dicentric rings, insertion, and breaks in human chromosomes.⁵ Cumulative radiation exposure at a young age increases the risk of a variety of solid and hematological malignancies.^6,7 Because many IBD patients are diagnosed at a young age,⁸ they are exposed to a higher amount of radiation early in their life and are at a risk for cumulative radiation exposure that is associated with stochastic risks.⁹ Patients with IBD are more vulnerable to ionizing radiation than the general population due to inherent neoplastic risks from chronic inflammation and immunosuppressive therapy.¹⁰

Magnetic resonance imaging (MRI) allows for cross-sectional imaging in multiple planes in the absence of ionizing radiation¹¹ and has emerged as an important alternative imaging technique to CT scans. Several studies have reported comparable efficacy of CT and MR enterography (MRE) in assessing small bowel disease activity in Crohn’s disease (CD).¹² MRE also has the advantage of providing dynamic assessment of small bowel motility by which it can distinguish active inflammatory disease from chronic fibrostenotic disease.^13,14 Currently, MRE is recommended over CT enterography (CTE) in estimating response to medical treatment in asymptomatic CD¹⁵ as well as in evaluating CD in young patients (<35 years of age).^16,17 Similarly, pelvic MRI is the preferred modality in the assessment of perianal and perirectal complications of CD.¹⁸ In the light of these considerations, the use of CT scans in IBD is expected to decrease. Because data on the recent utilization of CT and MRI in IBD are limited, we aimed to analyze the trends of CT and MRI use in a large cohort of IBD patients over 10 years.

Methods

Database

This is a retrospective cross-sectional observational study performed using a de-identified database, IBM Explorys. Explorys extracts electronic health record (EHR) data from participating intuitions using health data gateway server behind their firewall. These data are then de-identified to meet the standards of the Health Insurance Portability and Accountability Act. The EHR data are then categorized to facilitate searching and indexing.^19,20 International Classification of Diseases diagnoses are mapped into SNOMED CT (Systematized Nomenclature of Medicine–Clinical Terms) hierarchy, and procedures are indexed using Current Procedural Terminology.¹⁹ This Health Insurance Portability and Accountability Act–enabled platform unifies de-identified clinical data across 26 healthcare networks comprising 360 hospitals and over 64 million unique patients from academic and community hospitals.¹⁹ Explorys rounds to the nearest 10 for statistical de-identification. Because Explorys aggregates data without any patient identifiable information, it does not require approval from the individual institutional review board.

Study Population

Adults ≥18 years of age who had a hospital encounter (inpatient, outpatient, surgery, procedure, emergency or urgent care visit) in 2010-2019 were reviewed in the study. In this population, SNOMED CT diagnosis codes for ulcerative colitis (UC) and CD were employed to identify the study population. Among those with UC or CD, those who underwent ≥1 CT of the abdomen (with or without pelvis, with or without contrast) or MRI of the abdomen (with or without pelvis, with or without contrast) at least 30 days after UC or CD diagnosis were analyzed on a yearly basis. Demographic characteristics such as 10-year age range, sex, ethnicity, race, insurance, obesity, smoking, and IBD complications such as fistula of intestine, perforation of intestine, small bowel obstruction, abscess of intestine, and perianal abscess were included. Individuals were considered obese if they had at least 1 body mass index (BMI) in the obese category (BMI ≥30kg/m²).

Statistical Analysis

The primary aim of the study was to analyze the yearly trends of IBD individuals with ≥1 CT of the abdomen and ≥1 MRI of the abdomen in the 10-year period. The secondary aim was to determine the percentage of IBD individuals who underwent ≥5 abdominal CTs during the study period. The prevalence of CD and UC in the study population was calculated by dividing the individuals with CD or UC who had at least 1 of the reviewed encounters by the total number of individuals who had an encounter during the study period. A demographic comparison was performed with individuals who had ≥5 CT scans (multiple CT scans) and those who had <5 CT scans for CD and UC using univariate chi-square test. The yearly trend from 2010 to 2019 was performed using Cochrane-Armitage test for the following (1) individuals with CD and UC who underwent ≥1 CT and ≥1 MRI; (2) individuals with 1-4 CT, ≥5 CT, and ≥10 CT scans in a year; and (3) individuals who developed IBD-related complications. Statistical analyses were performed using Microsoft Excel 2016 (Microsoft Corporation, Redmond, WA) and Addinsoft xlstat 2019.1 (Addinsoft, Paris, France). A P value <.05 was considered statistically significant.

Results

Among 39 717 520 adults, 176 110 (0.44%) with CD and 143 460 (0.36%) with UC had at least one of the studied encounters in the last 10 years. About 9.4% of the individuals with IBD had ≥5 abdominal CTs in the last 10 years, with CD significantly higher than UC (10.6% vs 8.2%; P<.001). There was a significant increase in the percentage of individuals with ≥1 CT of the abdomen and MRI of the abdomen from 2010 to 2019.

Demographics

The baseline demographics between the CD and UC study population are shown in Table 1. Demographic details of individuals with ≥5 CT scans were compared with those with <5 CTs of the abdomen in Table 2. Age >50 years, men, African Americans, public insurance holders (Medicare and Medicaid), smoking, and BMI ≥30kg/m² were significantly associated with higher rates of abdominal CTs in CD and UC.

Trend Analysis

The percentage of individuals who had ≥1 CT of the abdomen significantly increased from 2010 to 2019 (CD: 10.2% vs 13.9%; UC: 6.6% vs 10.0%), and the mean annual percentage change was +3.6% for CD and +4.9% for UC. Similar increase was noted for individuals who had ≥1 MRI (CD: 1.5% vs 4.1%; UC: 0.7% vs 2.3), with a higher mean annual percentage change (CD: +15.6%, UC: +22.8% (Figure 1). We found a significant increase in the percentage of individuals with ≥5 CT scans (CD: 1.3% vs 5.1%; UC: 1% vs 3.4%) and ≥10 CT scans (CD: 0.3% vs 1.5%; UC: 0.2% vs 0.9%) from 2010 to 2019 (Table 3), with P<.001 for the previous comparisons. However, the trend was not significant for those with 1-4 CTs. Among those with ≥5 CTs per year, there was no significant difference in the percentage of individuals diagnosed with complications between 2010 and 2019 (CD: 12.1% vs 12.0%; P=.959; UC: 11.1% vs 8.4%; P=.237) (Table 4).

Discussion

Using a large population-level database, we showed that the percentage of IBD patients receiving abdominal CTs significantly increased from 2010 to 2019 despite the considerable increase in the availability of MRI of the abdomen. A total of 9.4% of adults with IBD had ≥5 abdominal CTs in the 10-year study period, and there was at least a 3-fold increase in this percentage between 2010 and 2019.

Radiation in CD vs UC

CD often presents earlier than UC.⁸ CD, being a transmural disease, has higher complication rates in the form of strictures, fistulae, and abscesses compared with UC.²¹ CD has been identified as an independent risk factor of increased radiation, and CD patients are exposed to 1.3-4 times higher diagnostic medical radiation than UC patients, in line with our study results.^2,22-24 A recent population-based study by Nguyen et al²⁵ showed that 15.6% of CD patients were exposed to a cumulative radiation >50 mSv compared with 6.2% of UC patients over a 5-year period. In contrary to most studies, Israeli et al²⁶ did not find significant difference in the cumulative effective dose between UC and CD over an 80-month period. However, this study also represents data from a relatively high-risk group of IBD patients who presented to the emergency department (ED) with gastrointestinal complaints only.

Risk Factors of Undergoing Multiple CT Scans

Age ≥57 years was shown as an independent risk factor of higher radiation.²³ In our study, older age (≥50 years) was a significant risk factor for undergoing multiple CTs in CD (odds ratio, 1.26) and UC (odds ratio, 4.36), the association being stronger in UC. The difference could be due to UC patients presenting later than CD⁸ or to providers being more likely to order CT scans in older individuals. Men with CD have higher risk of penetrating disease, stricturing disease, perianal manifestation,²⁷ and major abdominal surgeries compared with women.²⁸ This was reflected as a positive trend for receiving higher number of CTs in men with CD similar to the literature.^23,29 Additionally, our study showed an elevated risk of undergoing multiple CTs for men with UC. It is likely that concern for pregnancy affected provider decisions in ordering CT scans for women of the reproductive age group.

While CD and men are established risk factors of high diagnostic medical radiation in IBD, we found that smoking and obesity are additional risk factors of undergoing multiple CTs that have not been extensively described in prior studies. Current smoking status is associated with 1.7 times higher risk of abdominal surgeries compared with never smokers.²⁸ Obese patients have more clinic visits, hospitalizations, and flares than normal-weight individuals.³⁰ These factors likely influenced more CT imaging in this population.

Trends

In our study, 7.6% of IBD patients (55% CD, 45% UC) had ≥5 abdominal CTs in 5 years (2015-2019). While the cause of this increasing trend remains unclear, individuals with ≥5 CTs in a year did not have an increasing trend in complication rates from 2010 to 2019. This shows that the increase in severity or complication rate cannot fully account for this rising trend. However, it is also possible that they were performed to assess disease activity and treatment response in individuals with severe disease.

CT is available in >95% of institutions across the United States, with 94% offering 24/7 accessibility to the ED.³¹ Patients with IBD get their CTs either from the ED or from an inpatient setting more than a quarter of the time.²⁹ Aside from conventional CTs, CTE is increasingly popular in the management of IBD patients, and its use increased by 840% from 2003 to 2007.²⁹ While CT appears to be the most accessible imaging technique in the ED, there is a possibility that it is overutilized.³² There are numerous factors such as ambiguity in clinical presentation, uncertainty in decision making, provider type, practices adopted in training, and medical liability risk that have been reported to influence ordering CT studies.^33,34 Thus, the efficacy of the study, encounter location, test availability, provider type, personal preferences, and medical liability risk could collectively play a role in the increasing rates of CT scans and should to be studied in the future to get a thorough picture of the possible etiologies. Further, newer techniques in imaging such as iterative reconstruction algorithms can lead to noise reduction and aid in radiation reduction compared with conventional filtered back projection in abdominal imaging.³⁵

MRI has gained popularity as a radiation sparing modality in the evaluation of CD. In a single-center study, a 9-fold increase in MRI was reported among CD patients between 2006 and 2015.³⁶ Despite the increasing use of MRI in the past decade, likely as a result of better availability and insurance precertification, the rising trend of CT scans stays unmatched probably due to its shorter exam time,²⁹ greater accessibility, round-the-clock availability,³¹ ease in individuals with morbid obesity, and insurance limitations.

Limitation

This study comes with several limitations. The database is based on population-level data centered around diagnoses codes; hence, patient-level details such as, indication, location of imaging, interval between imaging, ordering provider, age at diagnosis of IBD, severity of disease, complication rates, treatment response, and IBD-related surgeries that require individual review of EHR were unavailable. We acknowledge this as a significant limitation, as frequency of imaging could be related to the severity of illness and treatment response rates. It is well known that the effective dose of an abdominal CT is not an absolute value. Patient size, implementation of dose-reduction technologies, and multiphase vs routine CT account for differences in the exposed radiation from abdominal imaging.³ However, the specific type of scan or the exposure from an individual imaging was not available. Further, we could not derive the trends of CTE and MRE specifically due to limitations of coding in the database. However, the code of CT of the abdomen (with or without pelvis) used in the study includes CTE, and similarly, MR of the abdomen (with or without pelvis) also includes MRE.

Despite these limitations, the study describes abdominopelvic imaging trends on a large cohort of IBD patients in a real-world setting. The prevalence of CD and UC in the entire database with 71 829 080 unique adults (not limited to the study population) was 319 per 100 000 and 265 per 100 000, respectively, at the time of the study, which is consistent with the literature.³⁷ Prior studies based on Explorys have also reported prevalence of IBD and celiac disease, which were comparable to national estimates.^38,39

Future Directions

While CT has proved critical in the timely management of IBD patients, especially in the ED and inpatient settings,²⁶ it can increase the cumulative radiation exposure in a chronic medical condition. Development of risk-stratification scoring systems could lead to judicious use of CT scans in most clinical settings. Additionally, application of societal guidelines into practice^15-17 and EHR-related strategies to calculate the cumulative exposure can help providers to analyze the risk-benefit ratio and search for alternative tests such as MRI and colonoscopy with ileoscopy and capsule endoscopy before ordering imaging with ionizing radiation whenever available.

Conclusions

CT scans have been the cornerstone of medical imaging since 1970s and are an invaluable tool in the clinical decision making of IBD patients. The yearly trend of abdominal CTs continues to increase despite the growing popularity and availability of MRI.^29,31 Even though we were unable to adjust for disease severity and complications, the proportion of IBD patients who underwent multiple CT exams (≥5 abdominal CTs) increased substantially in the past decade. Further studies are needed to explore the interventions to improve appropriate imaging utilization in patients with IBD.

Funding

No funding sources declared.

Conflicts of Interest

None declared.

References

Author notes