Comparison of Contrast Agents in Liver MR for the Detection of Hepatic Metastases

Evaluation of Liver MR With an Abbreviated Gadobenate Dimeglumine Hepatobiliary Phase Protocol in Comparison to Liver MR With Gadoxetate Disodium for the Detection of Hepatic Metastases

If an abbreviated HBP protocol liver MR with gadobenate dimeglumine is shown clinically comparable to standard of care liver MR with gadoxetate disodium for detecting hepatic metastasis from colorectal cancer, its use will save time, cost, and patients' effort.

The goal is to: Estimate and compare the diagnostic performance, including sensitivity, specificity, positive/negative predictive value, and area under the receiver operating characteristics (AUROC), of abbreviated protocol liver magnetic resonance (MR) with hepatobiliary phase (HBP) using gadobenate dimeglumine for detecting liver metastases, with 1) abbreviated protocol liver MR with HBP using gadoxetate disodium, 2) standard of care complete protocol liver MR using gadoxetate disodium, and 3) complete protocol liver MR using gadobenate dimeglumine. Estimate and compare quantitative measures of HBP images (liver enhancement ratio, lesion contrast to noise and signal to noise ratios [CNR and SNR]) for both gadobenate dimeglumine and gadoxetate disodium. Qualitatively assess the preference, or lack thereof, of radiologists regarding the images generated by abbreviated protocol liver MR with HBP using gadobenate dimeglumine versus abbreviated protocol liver MR with HBP using gadoxetate disodium.

This is a prospective study. The subjects will be randomized into two groups. Both groups will undergo two complete protocol liver MRs for known or suspected CRC metastasis, one exam with gadoxetate disodium and the other exam with gadobenate dimeglumine, within an interval of 3-10 days, but in opposite order, determined randomly. From the complete protocol liver MR with both contrast agents, an image set of abbreviated HBP protocol liver MR for each agent will be obtained (Fig.1).Two week-intervals will be placed between the evaluations of different image sets of a patient, and the images will be presented in a different random order to each evaluating radiologist. The schedule of random order of protocol liver MRs and presentation of images to radiologists will be determined by the project biostatistician.

We have attempted to mitigate any potential conflict of interest through randomization of the order of exams and blinding the image evaluators to contrast agents.

The subjects will be randomized into two groups. Both groups will undergo two complete protocol liver MRs for known or suspected CRC metastasis, one exam with gadoxetate disodium and the other exam with gadobenate dimeglumine, within an interval of 3-10 days, but in opposite order, determined randomly.

The subjects will be randomized into two groups. Both groups will undergo two complete protocol liver MRs for known or suspected CRC metastasis, one exam with gadoxetate disodium and the other exam with gadobenate dimeglumine, within an interval of 3-10 days, but in opposite order, determined randomly.

Gadoxetate disodium is now mainly used for the purpose of HBP liver MR imaging to save MR scanner time and total examination time.

The most commonly used MR contrast agent in abdominal imaging is gadobenate dimeglumine, which has mainly the characteristics of an extracellular agent used for most indications of MR examinations.

The primary performance diagnostic will be sensitivity, similar to the recently published retrospective study by Canellas et al. (2019). In addition to other diagnostic performance metrics of interest (e.g., specificity, AUROC), lesions will be analyzed descriptively in terms of number of metastases detected and lesion size.

Mixed effects regression models will again be used to compare the three outcomes between methods, accounting for correlated data. The specific link function of the regression models will depend on the distributional characteristics of each outcome (e.g., logit link for dichotomous outcomes; linear regression for continuously measured outcomes).

The quality of amHBP and aeHBP images will be assessed with ordinal response mixed effect models that include right/left image as a covariate. We will assess if there was any reader specific and/or general bias to prefer an image on the left or the right of a screen, regardless of the amHBP or aeHBP, and consider this when modeling the probability of preference of amHBP over aeHBP. We will estimate the relative probabilities of no-preference, amHBP preference, or aeHBP preference.

The primary analysis will use pathology as the gold standard, if available, but will revert to long term imaging in the absence of pathology. To address the potential impact of this limitation, we will also conduct an exploratory analysis to compare the sensitivity and specificity of the imaging methods by each gold standard.

We will also conduct an exploratory analysis to test the validity of long term imaging as a gold standard against available pathology reports.

In addition to analyses for the primary aims, additional analyses will be conducted to examine patient time associated with each imaging method; time metrics can be compared between imaging methods in absolute terms, but also proportionally to sensitivity.

In addition to analyses for the primary aims, additional analyses will be conducted to examine cost associated with each imaging method; cost metrics can be compared between imaging methods in absolute terms, but also proportionally to sensitivity.

Inclusion Criteria: Colorectal cancer patients Age 18-80 years No prior treatment including surgery Prior imaging with suspected liver metastasis Exclusion Criteria: Age < 18 years or > 80 years eGFR < 30 ml/min/1.73 m2 Previous reaction to gadolinium contrast agents History of claustrophobia or movement disorders likely to impact image quality Non-MR safe implants or metallic foreign bodies

Brismar TB, Dahlstrom N, Edsborg N, Persson A, Smedby O, Albiin N. Liver vessel enhancement by Gd-BOPTA and Gd-EOB-DTPA: a comparison in healthy volunteers. Acta Radiol. 2009 Sep;50(7):709-15. doi: 10.1080/02841850903055603.

Baek SE, Park MS, Hong HS, Choi JY, Chung YE, Lim JS, Kim MJ, Kim KW. Characterisation of small hypoattenuating hepatic lesions in multi-detector CT (MDCT) in patients with underlying extrahepatic malignancy: added value of contrast-enhanced MR images. Eur Radiol. 2010 Dec;20(12):2853-61. doi: 10.1007/s00330-010-1872-x. Epub 2010 Jul 9.

Schulz A, Viktil E, Godt JC, Johansen CK, Dormagen JB, Holtedahl JE, Labori KJ, Bach-Gansmo T, Klow NE. Diagnostic performance of CT, MRI and PET/CT in patients with suspected colorectal liver metastases: the superiority of MRI. Acta Radiol. 2016 Sep;57(9):1040-8. doi: 10.1177/0284185115617349. Epub 2015 Nov 29.

Zhang L, Yu X, Huo L, Lu L, Pan X, Jia N, Fan X, Morana G, Grazioli L, Schneider G. Detection of liver metastases on gadobenate dimeglumine-enhanced MRI: systematic review, meta-analysis, and similarities with gadoxetate-enhanced MRI. Eur Radiol. 2019 Oct;29(10):5205-5216. doi: 10.1007/s00330-019-06110-1. Epub 2019 Mar 26.

Canellas R, Patel MJ, Agarwal S, Sahani DV. Lesion detection performance of an abbreviated gadoxetic acid-enhanced MRI protocol for colorectal liver metastasis surveillance. Eur Radiol. 2019 Nov;29(11):5852-5860. doi: 10.1007/s00330-019-06113-y. Epub 2019 Mar 19.

Hardie AD, Naik M, Hecht EM, Chandarana H, Mannelli L, Babb JS, Taouli B. Diagnosis of liver metastases: value of diffusion-weighted MRI compared with gadolinium-enhanced MRI. Eur Radiol. 2010 Jun;20(6):1431-41. doi: 10.1007/s00330-009-1695-9. Epub 2010 Feb 11.

Kenis C, Deckers F, De Foer B, Van Mieghem F, Van Laere S, Pouillon M. Diagnosis of liver metastases: can diffusion-weighted imaging (DWI) be used as a stand alone sequence? Eur J Radiol. 2012 May;81(5):1016-23. doi: 10.1016/j.ejrad.2011.02.019. Epub 2011 Mar 4.

Zech CJ, Herrmann KA, Reiser MF, Schoenberg SO. MR imaging in patients with suspected liver metastases: value of liver-specific contrast agent Gd-EOB-DTPA. Magn Reson Med Sci. 2007;6(1):43-52. doi: 10.2463/mrms.6.43.