Impact of Clinical Outcomes and Socioeconomic Status on CT versus MR Utilization in Thoracic Aortic Surveillance Imaging

Background: Computed tomography (CT) and magnetic resonance (MR) angiography are crucial for diagnosing and surveilling thoracic aortic diseases (TAD). Clinical guidelines recommend baseline and follow-up imaging to monitor aortic growth. Limited data exist on the utilization of CT versus MR in patients with TAD [1]. Additionally, the impact of socioeconomic disparities on access to CT and MR remains poorly understood [2]. Our study aims to provide a descriptive analysis of an institutional database of over 50,000 patients undergoing aorta evaluation, better characterize MR and CT utilization, and explore the impact of income and demographics on imaging utilization.

Methods: A retrospective query of the electronic health record was conducted at a tertiary academic medical system, selecting all aortic angiography imaging studies acquired from 2011 to 2023. Studies were classified by modality (CT or MR), scan date, and patient demographics (age, sex, and ZIP code). Patient outcomes were tracked, and median household income was estimated by ZIP code–based 2023 U.S. Census data. Major aortic events included thoracic aortic aneurysms, aortic dissections, and aortic surgical interventions.

For baseline and follow-up analyses, the scans were first classified as baseline and follow-up, and then CT and MR scan counts were totaled for each stratum. For outcome analyses, each patient’s first diagnosis date for the specific major aortic event was defined as their index event. Scans for each patient were then categorized as pre- or post-event by modality, with scan counts recorded for each category.

Results: A total of 51,490 unique patients (73,614 scans, 56.1% male, mean age 59.3 ± 16.6 years) were identified. CT comprised the majority of baseline scans (85.6%), while MR accounted for 14.4% (Table 1). Younger subjects (18-31 years) who underwent follow-ups had four times the odds of MR than CT (odds ratio (OR) 4.45) than older subjects (74-89 years).

Across patients with any major aortic events (n=11,744), the mean CT scan count from pre-event to post-event increased (+169%), while MR increased more (+318%).

Furthermore, analysis by income quartile demonstrated patients in Q4 were 62.3% more likely to undergo MR than those in Q1 (relative risk (RR) 1.623, 95% CI 1.560–1.689, p < 0.001). Logistic regression confirmed that higher income was an independent predictor, providing 8% higher odds of MR utilization per $10,000 income increase (p < 0.001) (Figure 1).

Conclusion: In our cohort, CT dominated baseline imaging, particularly for initial scanning in TAD. However, MR utilization increased for follow-up scans, particularly among younger subjects, likely reflecting an awareness of CT radiation risk in younger patients who require multiple follow-ups. Follow-up imaging increased after TAD diagnosis (aneurysm, dissection), and income trends suggest that patients with higher incomes are more likely to be imaged with MR. Overall, findings suggest that MR may be underutilized in TAD.

Age and Sex-Stratified CT and MR Utilization at Baseline and Follow-Up Imaging Encounters

Stratum	Total imaging studies in stratum	Baseline CT total scan count1 (% of baseline scans)	Baseline MR total scan count (% of baseline scans)	Follow-up CT total scan count (% of follow-up scans)	Follow-up MR total scan count (% of follow-up scans)
All	73614	44098 (85.6)	7392 (14.4)	14338 (64.8)	7786 (35.2)
18-31 years	6164	4322 (83.2)	874 (16.8)	450 (46.5)	518 (53.5)
32-45 years	9808	6006 (81.8)	1340 (18.2)	1201 (48.8)	1261 (51.2)
46-59 years	18133	10948 (84.2)	2051 (15.8)	3010 (58.6)	2124 (41.4)
60-73 years	24737	13715 (85.7)	2294 (14.3)	5838 (66.9)	2890 (33.1)
74-89 years	14772	9107 (91.6)	833 (8.4)	3839 (79.4)	993 (20.6)
Male	43880	24388 (84.4)	4496 (15.6)	9398 (62.7)	5598 (37.3)
Female	29713	19697 (87.2)	2895 (12.8)	4934 (69.3)	2187 (30.7)
Unspecified	21	13 (92.9)	1 (7.1)	6 (85.7)	1 (14.3)

¹ Baseline percentages are calculated from the count of unique subjects (n = 51490); follow-up percentages are calculated from the count of subjects with follow-up scans (n = 22124), i.e., a subset of all subjects. The first instance of imaging for a subject was considered baseline, and any subsequent instance was follow-up.
Differences in CT and MR Utilization Before and After Major Aortic Events

Stratum	Total subjects in outcome group	Pre-event CT mean scan count (sd)	Post-event CT mean scan count (sd)	Pre-event MR mean scan count1 (sd)	Post-event MR mean scan count (sd)	CT utilization change (%)	MR utilization change (%)
All events	11744	0.29 (0.55)	0.78 (1.19)	0.11 (0.44)	0.46 (1.21)	169	318
Aneurysm2	9082	0.22 (0.53)	0.89 (1.28)	0.10 (0.46)	0.58 (1.34)	305	480
Dissection without surgery	414	0.17 (0.45)	0.78 (1.16)	0.03 (0.24)	0.11 (0.42)	359	267
Dissection with surgery3	169	0.08 (0.33)	1.40 (1.34)	0.01 (0.02)	0.21 (0.83)	1650	2000
Surgery	2248	0.61 (0.56)	0.35 (0.58)	0.15 (0.41)	0.07 (0.31)	-42.6	-53.3

¹ The difference in pre-event MR vs post-event MR mean scan counts were statistically significant (p < 0.001 by t-test) for all strata. (sd = standard deviation)

² A series of ICD codes was defined for each aortic event for individual extraction. 

³ Subjects who had a dissection outcome and a surgical intervention event. The dissection event was considered here as the threshold between pre-event and post-event, rather than the surgery event.



(Left) CT and MR mean scan count across income-stratified quartiles. The number of total scans (n) for each modality and quartile is displayed above the bars. Q1 represents the lowest income quartile, while Q4 represents the highest income quartile. (Right) A forest plot of the adjusted odds ratios (95% CI) given by the logistic regression model.