Closed for Enrollment

• Detection of the Occurrence of Infiltration of Gadolinium Injection in Brain MR Scans Using Artificial Intelligence Rochester, Minn.

  The aim of this study is to develop machine-learning algorithms and infrastructure through the analysis of MR image data for training and testing, for the detection of intravenous infiltration of MR contrast agent. This aim will be achieved through the analysis of both non-infiltrated and infiltrated brain MR images.  This analysis will involve segmenting and analyzing the nasal mucosa region following contrast agent administration. The developed machine learning model will be trained to be able to differentiate between non-infiltrated (contrast-enhanced) and infiltrated MR images and thus determine if infiltration has occurred.  The training and testing data will include both prospective and retrospective MR data.  This IRB application outlines the part of the study which will involve the collection prospective data for training the machine learning model.

  Machine learning can be used to make a binary decision regarding the occurrence of infiltration of the MR contrast agent, gadolinium, through analysis of MR Brain scan images.