Visualization of subtle contrast-related intensity changes using temporal correlation.

Contrast-enhanced magnetic resonance imaging (MRI) is a promising method for investigating the breakdown of the blood-retinal barrier (BRB). However, subtle intensity changes due to low concentrations of contrast agent can be difficult to detect without observer bias. In this study, we developed a temporal correlation method for detecting these subtle signal intensity changes. The method was evaluated in eyes with chemically induced retinal lesions of known size. A time series of MRI data were collected following i.v. administration of different doses of gadolinium-diethylaminetriaminepentaacetic acid (0.05, 0.1, 0.5 mmol/kg). These time course images were analyzed by temporal correlation to a reference enhancement curve. The reference curve was generated based on a validated theoretical enhancement curve. The temporal correlation method detected signal intensity changes in cases where the changes were too subtle to be visible on a postinjection image or a subtraction image (obtained by subtracting the precontrast image from the final image in the time course set). In addition, assessment of leakage was performed by viewing each image in the set with an eight gray-level palette. Areas of leakage identified in this manner corresponded to those identified by temporal correlation, a finding which supports the validity of the temporal correlation method. These results suggest that temporal correlation may be a time-efficient way to screen large numbers of image data sets using an objective, user-independent criterion.