Evaluating breast lesions with ultrafast DCE-MRI: The impact of temporal resolution on pharmacokinetics

This study aimed to evaluate the impact of temporal resolution on pharmacokinetic parameters and the performance in diagnosing breast cancer.

This prospective study included 89 consecutive patients with 97 lesions (76 malignant and 21 benign). Ultrafast dynamic contrast-enhanced (UF-DCE) imaging was performed on a 3-T MRI scanner using CAIPIRINHA-Dixon-TWIST-VIBE (CDTV) with a native temporal resolution of 4.5 s per phase. The CDTV data were downsampled at various temporal resolutions ranging from 9.0 to 58.5 s, creating eight additional datasets. Pharmacokinetic parameters, including K^trans, k_ep, and v_e, were calculated using Tofts model and population-based arterial input function (AIF). Paired-sample t-test with Bonferroni correction and ROC curves were used for statistical analysis.

As temporal resolution decreased from 4.5 s, K^trans and k_ep values increased, while v_e values decreased. The AUCs of K^trans, k_ep, and v_e at 4.5-s temporal resolution were 0.648, 0.778, and 0.685, respectively. Temporal resolutions shorter than 18 s showed less than a 10 % deviation in K^trans and k_ep of quantitative parameters compared to the 4.5-s reference. No significant difference was noted among AUC pairs for any of the parameters (corrected p > 0.00625).

Our results revealed that the temporal resolution significantly impacted pharmacokinetic parameters based on the population-based AIF from UF-DCE MRI. However, the diagnostic ability to characterize breast cancer might not be substantially affected. A temporal resolution shorter than 18 s is recommended for accurate calculation of pharmacokinetic parameters.