Optimizing photon-counting CT pulmonary angiography for robustness, combining mixed bolus, fixed scan delay and low...

This study's purpose was to create a robust, optimized CTPA protocol by exploring different combinations of keV reconstructions and contrast infusion parameters.

PCCT CTPAs between July 2023-January 2024 were retrospectively included, spanning three contrast injection protocols during a period of protocol optimization: cohort 1: 50/0mL contrast/saline mix using bolus tracking; cohort 2: 50/50 mL mix and cohort 3: 35/35 mL mix, both with fixed scan delay. Patients were matched by water-equivalent diameter between cohorts. Images were reconstructed at 40, 55 and 67 keV. Hounsfield unit (HU) values were measured in the main pulmonary artery and the segmental arteries. Three blinded radiologists rated subjective image quality.

150 patients (49 men, 101 women) were evaluated (mean age: 60.6 ± 16.6 years), 50 per cohort. Mean HU values were significantly higher in cohort 1 > 2 > 3 (p < 0.001) and were significantly higher at 40 > 55 > 67 keV (p < 0.001). At 40 keV, three (1.5 %) cohort 1 contrast injections were < 350 HU, with none < 350 HU in cohorts 2 or 3. Readers rated contrast best at 40 keV for vascular assessment (cohort 2: 4.7 ± 0.5, vs 4.5 ± 0.7 at 55 keV, vs 3.6 ± 1.0 at 67 keV, p < 0.001).

Our study suggests that the most robust PCCT contrast injection protocol for CTPA is a mixed injection of 50 or 35 mL of contrast with the same volume of saline, using a fixed scan delay. 40 keV vascular reconstructions are favored by radiologists and minimize non-diagnostic contrast injections.