Just watched a terrific session by Dominik Fleischmann, MD, on controlling arterial enhancement in subsecond scans -- definitely a growing issue as MDCT technology continues to improve. How do you inject for fast acquisitions, like the single-heartbeat scans made possible by 320-slice CT?
Fleischmann presented three possibilities: injecting for the duration of the scan, doing a moderate delayed injection like in 64-slice protocols, and individually tailoring injections to each patient. Though the first two methods can be made to work, obviously the ideal method would be #3: individualized injection protocols. But how can this be done when cardiac output and central blood volume are unknown?
The answer, of course, is the word of the day: iterative reconstruction. "What you could do is look at the patient's time attenuation response to a test bolus and extract the patient characteristic function," Fleischmann said. "Then you could use this information to iteratively find an individualized injection profile."
He would know: he's tried it in a few patients and found that ultimately, you can predict and tailor the injection protocol individually for a given patient. Peak enhancement can be predicted within 75 HU, and time course can be determined within plus or minus two seconds.
"Maybe in the future there will be an easy way to individually tailor the injections," Fleischmann speculated. Sounds like we're getting close!