Patient-Specific Timing for Pulmonary CT Angiography: Human Validation of a Physiologically Derived Post-Trigger Delay Model.

Objective

To evaluate the ability of a physiologically derived post-trigger delay (PTD) model to predict pulmonary arterial peak enhancement timing using human dynamic CT perfusion data and to compare physiologically derived PTD estimates with commonly used fixed-delay protocols.

Methods

In this retrospective study, 39 patients who underwent dynamic pulmonary CT perfusion imaging were evaluated. Pulmonary arterial enhancement curves were generated from the main pulmonary artery (MPA) using attenuation measurements obtained from 19 dynamic CT volume scans and fitted using a gamma-variate function. Reference peak enhancement timing was determined from the fitted enhancement curves. A physiologically derived PTD was calculated using a previously established relationship incorporating contrast injection duration and a dispersion component derived from first-pass contrast transport dynamics. Fixed PTD protocols of 4, 5, 6, and 7 s were evaluated as reference benchmarks. Enhancement and contrast-to-noise ratio (CNR) measurements were obtained in the MPA and five lobar pulmonary arteries. Protocol performance was assessed relative to the reference standard using nonparametric statistical testing, regression analysis, and Bland-Altman analysis, with enhancement and CNR measurements summarized as median [interquartile range] and paired differences reported with 95% confidence intervals.

Results

The mean physiologically derived PTD was 5.06 s, compared with a mean reference PTD of 5.18 s. In the MPA, the median absolute enhancement difference relative to the reference was 1.02 HU (95% CI, 0.42-4.66) for the physiologically derived PTD, compared with 50.76 HU (35.53-66.88), 1.05 HU (0.48-8.35), 25.60 HU (17.95-40.34), and 114.66 HU (78.45-138.55) for the 4-, 5-, 6-, and 7-s fixed-delay protocols, respectively. The corresponding median absolute CNR differences were 0.02 (0.01-0.10), 0.89 (0.67-1.23), 0.02 (0.01-0.16), 0.51 (0.32-0.78), and 1.96 (1.46-2.67). The physiologically derived PTD demonstrated significantly smaller differences from the reference than the 4-, 6-, and 7-s protocols (Holm-adjusted p < 0.05), with no significant difference compared with the 5-s protocol. Regression and Bland-Altman analyses demonstrated strong agreement with the reference standard.

Conclusions

A physiologically derived PTD demonstrated strong agreement with retrospectively determined pulmonary arterial peak enhancement timing in this human cohort. While performance was similar to a fixed 5-s delay under the relatively standardized injection conditions used in this study, the findings support the applicability of a previously established physiologic PTD relationship to human pulmonary arterial enhancement and provide validation of the model using human dynamic CT data.