Optimal Duration of Acquisition for Dynamic Perfusion CT Assessment of Blood-Brain Barrier Permeability Using the Patlak Model

Study Design

Imaging data obtained as part of standard clinical stroke care at our institution were retrospectively reviewed with the approval of the institutional review board. At our institution, patients with suspicion of acute stroke and no history of significant renal insufficiency or contrast allergy routinely undergo a stroke CT survey, including noncontrast CT (NCCT) of the brain, PCT at 2 cross-sectional positions, CT angiography (CTA) of the cervical and intracranial vessels, and postcontrast cerebral CT, obtained in this chronologic sequence.

We retrospectively identified a consecutive series of 23 patients admitted to the University of California San Francisco Medical Center from July 2007 to August 2008 who met the following inclusion criteria: 1) admission to the emergency department with signs and symptoms suggesting anterior circulation stroke within 12 hours after symptom onset; 2) documentation of acute ischemic anterior circulation stroke by both admission stroke protocol and clinical examination; 3) no evidence of intracerebral hemorrhage on the admission NCCT; and 4) no significant motion artifacts following application of the registration algorithm. Patients’ charts were reviewed for demographic and clinical data.

Imaging Protocol

PCT studies were obtained on 16-section (4 patients) and 64-section (19 patients) CT scanners. Each PCT study involved successive gantry rotations performed in cine mode during intravenous administration of iodinated contrast material. Images were acquired and reconstructed at a temporal sampling rate of 1 image per second for the first 37 seconds and 1 image every 2 seconds for the next 33 seconds. Additional gantry rotations were obtained at 90, 120, 150, 180, 210, and 240 seconds. Acquisition parameters were 80 kilovolt (peak) and 100 mAs. Two successive PCT series at 2 different levels were performed following the NCCT and before the CTA. At each PCT level, two 10-mm-thick sections (16-section CT scanners) or eight 5-mm-thick sections (64-section CT scanners) were assessed. The first PCT series was performed at the level of the third ventricle and the basal ganglia, and the second PCT series, above the lateral ventricles. For each PCT series, a 40-mL bolus of iohexol (300 mg/mL of iodine, Omnipaque; GE Healthcare, Piscataway, NJ) was administered into an antecubital vein by using a power injector at an injection rate of 5 mL per second for all patients. CT scanning was initiated 7 seconds after start of the injection of the contrast bolus. Data from both boluses were used because prior work has demonstrated that there is no significant parenchymal saturation effect from the first bolus and no underestimation of BBBP values from data from the second bolus.¹⁴

Image Postprocessing

PCT data were analyzed using commercially available PCT software (Brain Perfusion; Philips Healthcare, Cleveland, Ohio). This software relies on the central volume principle, which is the most accurate for low injection rates of iodinated contrast material.¹⁵ The software obtains mathematic descriptions of the time-attenuation curves for each voxel, by applying curve fitting by least mean squares, after correcting for motion and noise reduction through an anisotropic edge-preserving spatial filter. A closed-form (noniterative) deconvolution is then applied to calculate the mean transit time (MTT) map.¹⁶ The deconvolution operation requires a reference arterial input function (most often within the anterior cerebral artery), automatically selected by the PCT software within a region of interest drawn by the user. The cerebral blood volume (CBV) map is calculated from the area under the time-attenuation curves, with a correction factor for the leakage of contrast into the interstitial space calculated from the Patlak model applied to the delayed acquisition. The PCT infarct core and salvageable brain tissue are automatically calculated by the software by using CBV thresholds and MTT thresholds reported in the literature as the most accurate (PCT salvageable brain tissue: MTT > 145% of the contralateral side values plus CBV ≥ 2.0 mL × 100 g⁻¹; PCT infarct core: MTT > 145% of the contralateral side values plus CBV < 2.0 mL × 100 g⁻¹).¹⁷

BBBP measurements were extracted from PCT data by using a second prototype software developed by Philips Healthcare. This software is based on the Patlak model.^18,19 Applying the Patlak model to PCT involves performing linear regression by using data calculated from the PCT datasets. The slope of these regression lines was interpreted as a blood-to-brain transfer constant and used as an indicator of absolute BBBP values. Relative BBBP values, defined as absolute BBBP values divided by cerebral blood flow values in the corresponding voxels, were also recorded.

Using the prototype software, we calculated BBBP values from the gold standard delayed acquisition dataset, which incorporated all time points from 90 to 240 seconds. Additionally, we calculated BBBP values by using truncated datasets based on analysis windows from 90 to 210 seconds (excluding the data from the last gantry rotation), 90 to 180 seconds (excluding the data from the last 2 gantry rotations), 90 to 150 seconds (excluding the data from the last 3 gantry rotations), and 90 to 120 seconds (excluding the data from the last 4 gantry rotations). BBBP values were recorded in the infarct and penumbra, as automatically delineated by the software, and in mirrored regions of interest in the contralateral nonischemic tissue (also automatically delineated by the software).

Statistical Analysis

We calculated the correlation between permeability values from each of the truncated analysis windows and permeability values from the gold standard 90- to 240-second delayed acquisition.

We measured the quality of the linear fit to quantify how well the assumptions of the Patlak model were met by data extracted from the datasets of 90–240, 90–210, 90–180, 90–150, and 90–120 seconds. To do this, we used the root-mean-square error (√MSE), which is a measure of variability of data points around a fitted straight line: A value close to zero indicates a smaller spread of data points around the line, corresponding to a better fit.

Absolute and relative BBBP values and corresponding √MSE values extracted from different durations of delayed PCT analysis windows in ischemic and nonischemic regions of interest were compared by using generalized estimating equation models with robust variance estimation, with fixed effects for patients, type of CT scanner, and type of regions of interest. Because the distribution of the parameters was not normal but skewed, we report estimated mean values (obtained by log transformation of the data) rather than simple means. For all values, 95% confidence intervals (CI) were also calculated.

Radiation Dose and Motion Artifacts

We calculated the effective radiation dose associated with the gold standard 90- to 240-second acquisition and the percentage decrease in radiation associated with truncated analysis windows (datasets of 90–210, 90–180, 90–150, and 90–120 seconds).

The number of cases with motion artifacts interfering with the processing was recorded for the gold standard 90- to 240-second acquisition and for the truncated analysis windows (datasets of 90–210, 90–180, 90–150, and 90–120 seconds).

Patients and Imaging Studies

Twenty-three patients who matched our inclusion criteria were retrospectively identified. Patient characteristics are summarized in Table 1. The median time from symptom onset to PCT was 2.25 hours (range, 1–11.75 hours). The Patlak analyses were performed in a total number of 230 ischemic regions of interest and 230 nonischemic regions of interest.

BBBP Measurements for Different Durations of the Delayed PCT Acquisition

BBBP values for truncated analysis windows of 90–210 seconds and 90–180 seconds were highly correlated with the gold standard 90- to 240-second delayed acquisition (Table 2). Shorter analysis windows (90–150 seconds and 90–120 seconds) were less correlated, especially in the infarct core (Table 2).

Absolute (Table 3) and relative (Table 4) BBBP values were overlapping for the gold standard 90- to 240-second delayed acquisition and for the 90- to 210-second analysis window. For shorter analysis windows (90–180 seconds and, especially, 90–150 and 90–120 seconds), absolute and relative BBBP values were overestimated in the infarct, penumbra, and nonischemic tissue (Fig 1). In terms of mean absolute BBBP values, the differences between the gold standard 90- to 240-second delayed acquisition and the 90- to 210-second analysis window were 12% and 15% in the infarct and penumbra, respectively. The differences in terms of mean absolute BBBP values in the infarct and penumbra jumped sharply to 36% and 42%, 76%, and 89%, and 146% and 166% for the 90- to 180-, 90- to 150-, and 90- to 120-second analysis windows, respectively. Very similar differences were observed for the relative BBBP results.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

Absolute and relative permeability maps for the gold standard 90- to 240-second acquisition and for the truncated acquisitions (90–210, 90–180, 90–150, and 90–120 seconds). Absolute and relative BBBP values are overlapping for the gold standard 90- to 240-second delayed acquisition and for the 90- to 210-second truncated analysis window. For shorter analysis windows (90–180 seconds and, especially, 90–150 and 90–120 seconds), absolute and relative BBBP values are overestimated. The map depicting PCT-determined infarct and penumbra at the same level is also presented for reference.

Quality of the Linear Regression according to the Patlak Model

√MSE measurements (Table 5) were low and overlapping for the gold standard 90- to 240-second delayed acquisition and for the 90- to 210-second analysis window. For shorter analysis windows (including 90–180 seconds and, especially, 90–150 and 90–120 seconds), √MSE measurements were higher, indicating a poorer linear fitting.

Radiation Dose and Motion Artifacts

The effective radiation dose associated with the gold standard 240-second acquisition was calculated (3.2 mSv). For the 90- to 210-, 90- to 180-, 90- to 150-, and 90- to 120-second analysis windows, the decrease in radiation dose was 1.7%, 3.3%, 5%, and 6.7%, respectively.

Motion artifacts significant enough to interfere with permeability processing even after application of the registration algorithm were seen in 3 patients. In each of these cases, suitable permeability processing was not possible with either the full 90- to 240-second acquisition or any of the truncated analysis windows due to motion throughout the entirety of the delayed acquisition.