Optimizing Pulmonary Embolism Computed Tomography in the Age of Individualized Medicine: A Prospective Clinical Study

The aim of the study was to simultaneously optimize contrast media (CM) injection and scan parameters for the individual patient during computed tomography pulmonary angiography (CTPA).

In this study (NCT02611115), 235 consecutive patients suspected of having pulmonary embolism were prospectively enrolled. Automated kV selection software on a third-generation multidetector computed tomography adapted tube voltage to the individual patient, based on scout scans. The contrast injection protocol was adapted to both patient body weight and kV-setting selection via a predefined formula, based on previous research. Injection data were collected from a contrast media and radiation dose monitoring software. Attenuation was measured in Hounsfield units (HU) in the pulmonary trunk (PT); attenuation values 200 HU or greater were considered diagnostic. Subjective image quality was assessed by using a 4-point Likert scale at the level of the PT, lobar, segmental, and subsegmental arteries. Results between groups were reported as mean ± SD.

Two hundred twenty-two patients (94%) were scanned at a kV setting below 100 kV: n = 108 for 70 kV, n = 82 for 80 kV, and n = 32 for 90 kV. Mean CM bolus volume (in milliliters) and total iodine load (in grams of iodine) for 70 to 90 kV were as follows: 24 ± 3 mL and 7 ± 1 g I, 29 ± 4 mL and 9 ± 2 g I, and 38 ± 4 mL and 11 ± 1 g I, respectively. Mean flow rates (in milliliters per second) and iodine delivery rates (in grams of iodine per second) were 3.0 ± 0.4 mL/s and 0.9 ± 0.1 g I/s (70 kV), 3.6 ± 0.4 mL/s and 1.0 ± 0.1 g I/s (80 kV), and 4.7 ± 0.5 mL/s and 1.3 ± 0.1 g I/s (90 kV). Mean radiation doses were 1.3 ± 0.3 mSv at 70 kV, 1.7 ± 0.4 mSv at 80 kV, and 2.2 ± 0.6 mSv at 90 kV. Mean vascular attenuation in the PT for each kV group was as follows: 397 ± 101 HU for 70 kV, 398 ± 96 HU for 80 kV, and 378 ± 100 HU for 90 kV, P = 0.59. Forty-six patients (21%) showed pulmonary embolism on the CTPA. One scan (90 kV) showed nondiagnostic segmental pulmonary arteries, and 5% of subsegmental arteries were of nondiagnostic image quality. All other segments were considered diagnostic-excellent subjective image quality.

Simultaneously optimizing both CM injections and kV settings to the individual patient in CTPA results in diagnostic attenuation with on average 24 to 38 mL of CM volume and a low radiation dose for most patients. This individualized protocol may help overcome attenuation-variation problems between patients and kV settings in CTPA.