MDCT vs. MRI for incidental pancreatic cysts: measurement variability and impact on clinical management

To investigate incidental pancreatic cysts (IPCs) size discrepancy in a cohort of patients receiving both computed tomography (CT) and magnetic resonance imaging (MRI) and its impact on clinical management based on the 2010 American College of Radiology (ACR) guidelines.

This was a HIPAA-compliant, retrospective, IRB-approved study. Informed consent was waived. Patients with known IPCs and at least one case-pair, consisting of an abdominal CT and MRI examination within 180 days between 05/1999 and 12/2011, were included. Maximum diameter of cysts was measured in both the CT and MR examinations. A subset of 30 patients was measured by three radiologists independently to assess inter-observer variability. Absolute difference in diameter measurements between CT and MRI was calculated. Influence of cyst size, cyst location, and patient characteristics such as weight, height, and body mass index (BMI) on variability of size measurements were evaluated. Clinical impact in terms of current ACR guidelines was assessed.

Overall, 267 case-pairs of cysts in 113 patients were included in this study. 59/267 cysts were visualized on MRI but not on CT (22.1%, 95% CI 17.1%-27.1%; 32 patients, 64.6 ± 11.7 years, BMI 28.5 ± 4.8 kg/m²), with a median MRI cyst size of 7.8 mm, IQR 6.0-9.0 mm, range 2-17.8 mm. 208 case-pairs in 113 patients with a mean BMI of 26.9 ± 5.1 kg/m² (range 16.9-39.5 kg/m²) and mean cysts size of 13.4 ± 8.1 mm (range 3-49 mm) were seen in both CT and MRI. The mean absolute size difference for IPCs measured on MRI and CT was 2.1 ± 1.8 mm (median 1.5 mm, IQR 0.9-2.9 mm, range 0-9 mm). Absolute size difference between CT and MRI measurements increased with size of the cyst (r = 0.31, p < 0.001), whereas location of the cyst did not influence the absolute difference between CT and MRI measurements (p = 0.44). Patient weight and BMI had a negative correlation with the difference in cyst size between CT and MRI (weight r = -0.17, p = 0.023; BMI r = -0.17, p = 0.027), with cyst measurements being larger on MRI in thin patients and on CT in obese patients. Inter-reader variability was excellent (ICC = 0.99). In 12/208 (5.7%, 95% CI 2.7%-9.1%), variability between CT and MRI would have changed ACR-based follow-up recommendation.

There was a median difference of 1.5 mm between measurements of IPCs on CT vs. MRI. If imaging modality was switched during follow-up, variability of measurements may lead to inappropriate change of follow-up regimen in 6% of all cases. A single follow-up CT for incidental IPCs does not seem sufficient due to a high number (22%) of missed IPCs on CT.