PURPOSE

18F-labeled deoxy-fluorothymidine (FLT), a marker of cellular proliferation, has been used in PET tumor imaging. Here, the FLT kinetics of malignant brain tumors were investigated.

METHODS

Seven patients with high-grade tumors and two patients with metastases had 12 studies. After 1.5 MBq/kg 18F-FLT had been administered intravenously, dynamic PET studies were acquired for 75 min. Images were reconstructed with iterative algorithms, and corrections applied for attenuation and scatter. Parametric images were generated with factor analysis, and vascular input and tumor output functions were derived. Compartmental models were used to estimate the rate constants.

RESULTS

The standard three-compartment model appeared appropriate to describe 18F-FLT uptake. Corrections for blood volume, metabolites, and partial volume were necessary. Kinetic parameters were correlated with tumor pathology and clinical follow-up data. Two groups could be distinguished: lesions that were tumor predominant (TumP) and lesions that were treatment change predominant (TrcP). Both groups had a widely varying k1 (transport across the damaged BBB, range 0.02-0.2). Group TrcP had a relatively low k3 (phosphorylation rate, range 0.017-0.027), whereas k3 varied sevenfold in group TumP (range 0.015-0.11); the k3 differences were significant (p < 0.01). The fraction of transported FLT that is phosphorylated [k3/(k2+k3)] was able to separate the two groups (p < 0.001).

CONCLUSION

A three-compartment model with blood volume, metabolite, and partial volume corrections could adequately describe 18F-FLT kinetics in malignant brain tumors. Patients could be distinguished as having: (1) tumor-predominant or (2) treatment change-predominant lesions, with significantly different phosphorylation rates.