Cerebral CT Perfusion in Acute Stroke: The Effect of Lowering the Tube Load and Sampling Rate on the Reproducibility of Parametric Maps

Abstract

The aim of this study was to define lower dose parameters (tube load and temporal sampling) for CT perfusion that still preserve the diagnostic efficiency of the derived parametric maps. Ninety stroke CT examinations from four clinical sites with 1 s temporal sampling and a range of tube loads (mAs) (100–180) were studied. Realistic CT noise was retrospectively added to simulate a CT perfusion protocol, with a maximum reduction of 40% tube load (mAs) combined with increased sampling intervals (up to 3 s). Perfusion maps from the original and simulated protocols were compared by: (a) similarity using a voxel-wise Pearson’s correlation coefficient r with in-house software; (b) volumetric analysis of the infarcted and hypoperfused volumes using commercial software. Pearson’s r values varied for the different perfusion metrics from 0.1 to 0.85. The mean slope of increase and cerebral blood volume present the highest r values, remaining consistently above 0.7 for all protocol versions with 2 s sampling interval. Reduction of the sampling rate from 2 s to 1 s had only modest impacts on a TMAX volume of 0.4 mL (IQR −1–3) (p = 0.04) and core volume of −1.1 mL (IQR −4–0) (p < 0.001), indicating dose savings of 50%, with no practical loss of diagnostic accuracy. The lowest possible dose protocol was 2 s temporal sampling and a tube load of 100 mAs.

Impact of temporal resolution on perfusion metrics, therapy decision, and radiation dose reduction in brain CT perfusion in patients with suspected stroke

“…Consequently, the definition of the lowest feasible sampling rate is an ongoing debate. There has been a prevailing notion that a temporal sampling below 2 s is required [ 10 , 12 , 13 ]. Some studies reported miscalculation of perfusion values and underestimation of ischemic volume when increasing the temporal sampling beyond this point [ 7 , 10 ].…”

“…Whereas previous studies on the concept of increased temporal sampling in CTP relied on rather small datasets [ 7 – 9 , 11 , 12 ], we examined a large dataset from a clinical routine which included a wide range of scan qualities. Here, we found that impaired data quality (for example due to motion, late contrast arrival, or low contrast to noise ratios) resulted in stronger artifacts when reducing the temporal resolution.…”

“…The radiation exposure is determined by various parameters, including tube current, tube voltage, slice coverage, temporal acquisition window, and temporal sampling interval. Consequently, multiple strategies have been proposed for dose reduction [ 4 ], in particular by increasing the temporal sampling intervals [ 7 – 12 ]. In retrospective modeling in small patient series, the principle feasibility of a lower temporal sampling of the perfusion bolus was shown: Wiesmann et al observed that the diagnostic value of the calculated perfusion maps at a temporal resolution of 3.0 s is equivalent to the original data set with 1.0 s [ 8 ].…”

Structured Reporting of Computed Tomography in the Staging of Neuroendocrine Neoplasms: A Delphi Consensus Proposal

“…In the protocol optimization stage, enhanced communication between different centres can lead to an overall quality improvement through optimization of radiation dose and contrast administration, higher patient safety and overall better diagnostic quality ( 35 , 36 ). Moreover, such enhanced communication can facilitate comparing the results obtained from different studies, thus reducing the variability due to different imaging protocols ( 37 , 38 ).…”

Quantification and Classification of Contrast Enhanced Ultrasound Breast Cancer Data: A Preliminary Study

“…Computationally, TMSP is the time required for the first derivative of Equation (1) to reach its maximum value. These equations have been used to describe the dispersion of a bolus as it passes through a series of compartments, mainly for perfusion modeling [ 33 , 34 , 35 , 36 ]. Further information and graphical representation of the quantitative markers for EMG can be found in [ 37 , 38 ].…”