Study of intrapatient variability and reproducibility of quantitative tumor perfusion parameters evaluated with dynamic contrast-enhanced ultrasonography

Objectives Dynamic contrast-enhanced (DCE) ultrasonography (US) is a functional imaging technique enabling quantitative assessment of solid tumor perfusion in metastatic patients treated with antiangiogenic therapies. The objective of this prospective single-center study was to evaluate in real-life conditions (in routine clinical practice) the intrapatient variability and reproducibility of DCE-US parameters. Materials and Methods Each patient provided written informed consent and had 2 DCE-US examinations (preprandial and postprandial) at baseline, day 15, and 1 month after treatment initiation. Perfusion curves were recorded after Sonovue injections to determine 7 perfusion parameters. Dynamic contrast-enhanced US examinations were analyzed in pairs: Preprandial and postprandial. Log transformed values were used to determine the variability of the pairs (within-subject coefficient of variation) and their reproducibility (Spearman correlation coefficient). Results We included 60 patients (23 colon cancers, 36 kidney cancers, and 1 breast cancer) treated with axitinib (26 patients), sunitinib (27 patients), and other antiangiogenic treatments (7 patients). The 60 patients included 38 men (63%) and 22 women (37%) with a median age of 62 (range, 25-82 years). Thirty patients had hepatic and 30 had extrahepatic target lesions. Data were analyzed for 128 pairs of DCE-US: 45 (baseline), 45 (day 15), and 38 (1 month). Preprandial and postprandial values were not significantly different. For area under the curve and area under the washout, the correlation coefficient between preprandial and postprandial values was 0.89; the associated within-subject coefficients of variation were 61% and 64%, respectively. However, the range of individual variations (postprandial value/preprandial value) was less than 2 logs for a range of parameter values of about 4 logs. Variability was independent of the metastatic site. Conclusions This study showed that area under the curve and area under the washout are the 2 most reproducible DCE-US parameters.