Quantitative assessment of lung cancer perfusion using MDCT: does measurement reproducibility improve with greater tumor volume coverage?

Abstract

To date, quantitative assessment of tumor vascularity using perfusion CT has been limited to a single tumor level, with the potential for measurement error in heterogeneous tumors. We aimed to determine if greater z-axis tumor coverage improves the reproducibility of perfusion CT measurements in lung cancer. Paired perfusion studies were performed on 10 patients who had histologically confirmed advanced non-small cell lung cancer. Using 16-MDCT, multiple sequential volumetric acquisitions encompassing the entire tumor were acquired after infusion of i.v. contrast material. Using Patlak analysis, median values of tumor permeability (mL/100 mL/min) and blood volume (mL/100 mL) were measured for 10-mm z-axis coverage, and for 40-mm z-axis coverage in each of the paired perfusion studies. Measurement reproducibility was evaluated using Bland-Altman statistics. Mean difference (95% limits of agreement) for tumor permeability was 1.4 (-4.0 to 6.8) for 10-mm coverage and 0.8 (-3.6 to 5.2) for 40-mm coverage. Mean difference (95% limits of agreement) for blood volume was 1.9 (-5.1 to 8.9) for 10-mm coverage and 1.4 (-3.7 to 6.6) for 40-mm coverage. The coefficient of variation for permeability was 18.7% for 10-mm coverage, improving to 11.9% for 40-mm coverage. The coefficient of variation for blood volume was 41.7% for 10-mm coverage, improving to 32.6% for 40-mm coverage. Our results show that an improvement in tumor perfusion measurement reproducibility may be achieved with greater z-axis coverage.