RSNA 2018

Abstract Archives of the RSNA, 2018


SSM19-01

The Gadolinium Deposition Debate Revisited Using a Fully Automated Big Data Approach

Wednesday, Nov. 28 3:00PM - 3:10PM Room: N227B



Participants
Julian Gehweiler, Basel, Switzerland (Presenter) Nothing to Disclose
Thomas J. Re, MD, Princeton, NJ (Abstract Co-Author) Consultant, Siemens AG
Michael Amann, Basel, Switzerland (Abstract Co-Author) Nothing to Disclose
Sabine Schadelin, MSc, Basel, Switzerland (Abstract Co-Author) Nothing to Disclose
Ruben Cabezon, Basel, Switzerland (Abstract Co-Author) Nothing to Disclose
Bram Stieltjes, MD,PhD, Basel, Switzerland (Abstract Co-Author) Nothing to Disclose

For information about this presentation, contact:

julian.gehweiler@usb.ch

PURPOSE

Numerous independent studies reported increasing T1w signal intensity (SI) of the dentate nucleus (DN) and the globus pallidus (GP) after repetitive administrations of linear Gadolinium-based contrast agents (LGBCA). Also, a recent study indicates that a high number of macrocyclic GBCA (MGBCA) administrations may also lead to an increase of SI. However, most previous studies were performed at small scale and predominantly by using manual ROI-based methods. In this study, we present results from a large-scale study that uses a completely automated high performance computing pipeline.

METHOD AND MATERIALS

In-house developed PACS/RIS mining software was used to locate and retrieve all native T1W MPRAGE datasets performed on patients that received at least 2 exams with GBCA. Datasets were histogram-normalized and segmented utilizing FSL and a cerebellum atlas. Calculations were performed overnight on a high performance computing cluster. SI increase was analyzed in a mixed effects model.

RESULTS

3021 consecutive patients receiving 11,922 MRI brain scans were included. Repeated administration of GBCA was associated with increased overall SI for both LGBCA (Estimate 0.23, 95% CI 0.07-0.39, p=.006) and MGBCA (Estimate 0.06, 95% CI 0.02-0.09, p<.001). Stratified among brain nuclei, significant SI increase was found only after LGBCA application in the DN (Estimate 0.6, 95% CI 0.077-1.12, p=.027) and the GP (Estimate 0.38, 95% CI 0.03-0.72, p=.033) but not for MGBCA. On the other hand, significant SI increase was found only after MGBCA in the thalamus, putamen, amygdala, caudate, hippocampus, and accumbens (Estimates up to 0.29, p<.001) but not for LGBCA.

CONCLUSION

Our findings confirm previous reports on increased SI after LGBCA application in the DN and GP, but not after MGBCA. However, we found increase in SI for other brain nuclei only for MGBCA. Based on our findings, Gadolinium accumulation in the brain may be more widespread than assumed and could occur with the use of both L- and MGBCA.

CLINICAL RELEVANCE/APPLICATION

Gadolinium accumulation in the brain may be more widespread than assumed and could occur with the use of both L- and MGBCA.