Abstract
1703
Objectives: FLT PET has been used to non-invasively assess tumor proliferation in non-small cell lung cancer (NSCLC) with promising results. In addition, quantitative FLT PET imaging may provide a means of accurately assessing tumor response to cell cycle targeted therapies. However, current efforts at quantification are limited to metrics involving the entire tumor treated as a single volume of interest (VOI). We investigated the feasibility of generating parametric images mapping the voxel-by-voxel flux (KFLT) for FLT phosphorylation from dynamic FLT PET images, using a two-compartment model. Other parametric images such as that of K1 (FLT transport) were generated for a comprehensive understanding of the FLT kinetic model. Methods: Voxel-by-voxel kinetic analysis of FLT dynamic studies was performed using the VOXULUS kinetic modeling software (Philips Medical Systems) for 17 NSCLC patients, generating parametric images of the FLT model. VOXULUS is unique in that it uses closed-form analytical solutions to model time activity curves (TACs), permitting accelerated processing of parametric images. In addition, as cross-validation, VOXULUS modeling results generated from TACs derived from tumor VOIs were compared with the modeling results obtained using the PMOD software package (PMOD group, Switzerland). Results: Parametric images of the rate-constants characterizing the FLT model (KFLT, K1, k2, k3, k4 and Vb) were generated for all patients. KFLT images appeared visually similar to the 30-60 minute FLT distribution images, with K1 images providing information related to FLT transport. KFLT measures estimated from tumor VOIs with VOXULUS agreed very well with modeling results obtained using the PMOD package for multiple VOI definitions (Spearman’s rank rho > 0.9). Conclusions: Using voxel-by-voxel compartmental analysis, quantitative KFLT parametric maps can be used to visualize proliferative regions in NSCLC patients. This method correlates well with compartmental analysis performed on VOIs, but presents the advantage of demonstrating biological heterogeneity within a tumor. In turn, this will facilitate future efforts for improved tumor volume definitions in radiation therapy planning procedures.
Research Support (if any): This study is supported by NIH grants CA 107264 and CA 115559.
- Society of Nuclear Medicine, Inc.