Objective: To evaluate computed tomography (CT) perfusion using first pass methods for lung nodule characterization.
Methods: Fifty-seven patients with 51 malignant and 6 benign nodules underwent first-pass, dynamic contrast-enhanced-CT (50 mL, 3-5 mL/s.). Kinetic analysis tools were CT Perfusion 3 (GEMS, Milwaukee, WI), a distributed parameter model approach, yielding blood volume (BV; mL/100 g), blood flow (BF; mL/min/100 g), mean transit time (1/s), and permeability surface area (mL/min/100 g), and an in-house Patlak-style analysis yielding fractional BV (mL/100 g) and an estimate of extraction (Kps, mL/100 g/min).
Results: CT Perfusion 3 parameters in malignant and benign nodules were: mean transit time 10.1 +/- 0.9 1/s versus 11.1 +/- 3.1 1/s (ns), permeability surface 23.3 +/- 9.1 mL/min/100 g versus 19.6 +/- 10.3 mL/min/100 g (ns), BF 111.3 +/- 8.7 mL/min/100 g versus 39.1+/- 5.7 mL/min/100 g (P < 0.001), BV 9.3+/- 0.7 mL/100 g versus 4.1 +/- 1.1 mL/100 g (P < 0.002); Patlak parameters were: Kps 13.3 +/- 1.2 mL/100 g/min versus 3.9 +/- 0.8 mL/100 g/min (P < 0.001), BV 8.4 +/- 0.8 mL/100 g versus 3.6 +/- 1.3 mL/100 g (P < 0.01). The two kinetic methods show good agreement for BV estimation (Bland-Altman plot). The limits of agreement (bias +/-2 standard deviation of bias) were 1.2 +/- 5.3 mL/100 g.
Conclusion: CT Perfusion using first pass modeling appears feasible for lung nodule characterization. Given the short acquisition duration used, weaknesses of the modeling methods are exposed. Nonetheless, microvascular characterization in terms of BF, BV, or Kps appears useful in distinguishing malignant from benign nodules.