Accounting for truncation artifacts in angiographic perfusion

Unlocking perfusion metrics from routine digital subtraction angiography (DSA) could transform how neurovascular disease is managed. Truncation artifact, defined as premature termination of image acquisition, is a key source of error in CT and MR perfusion imaging and requires prolonged imaging. However, length requirements for deriving perfusion metrics from DSA remain undefined. This study investigates the minimum image acquisition length required for interpreting mean transit time (MTT) from DSA. We analyzed 55 outpatient angiograms performed for surveillance an average of 687 days (SD 541 days) after aneurysmal rupture. Truncation artifact was simulated by progressively shortening the angiography runs after bolus arrival. A gamma function was used to assess if extrapolation of the truncated data could approximate full length acquisitions. Extrapolation of the truncated datasets with the gamma function produced highly reliable estimates of the MTT with at least 6 seconds of post-bolus data. The fraction of pixels fit by the gamma function was more sensitive to truncation and > 7 seconds of data was required to fit >90% of the pixels. These findings provide a framework for evaluating the adequacy of image acquisition time and lay the foundation for retrospective perfusion analysis using DSA.