Quantitative studies using positron emission tomography (PET) for the diagnosis and therapy planning of oncological patients.

UNLABELLED: Positron emission tomography (PET) has found wide-spread use in oncology due to the relatively high accuracy in the staging, differential diagnosis and therapy monitoring. Most PET studies are performed as a whole body scan. In selected cases a semiquantitative analysis is performed, which is based on the calculation of standardized uptake values (SUV). The present studies were undertaken in order to evaluate the impact of dynamic PET studies in malignant diseases with respect to tumor diagnosis and therapy management. Dynamic data acquisition is superior to static images because is provides information about the tracer distribution with respect of time and space, in a region of interest. The impact of different compartmental and non-compartmental approaches for the diagnostics and therapy planning was also studied. The radiopharmaceuticals used for patient studies were: O-15-water, C-11-ethanol, F-18-fluorodeoxyglucose (FDG), F-18-fluorouracil (F-18-FU), and 6-F-18-fluoro-L-DOPA. A new evaluation strategy of dynamic PET studies based on an integrated evaluation including both compartment and non-compartment models as well as the use of SUV is presented. Furthermore, the parametric imaging including Fourier-analysis and regressions analysis was used.

RESULTS: PET-studies with labeled cytostatic agents provide informations about the transport and elimination of a cytostatic agent and help to predict the therapeutic outcome. The retention of the radiolabeled cytostatic agent F-18-FU in liver metastases of colorectal cancer was low after systemic application. Lesions with retention values >3.0 SUV and with <2.0 SUV correlated with negative and positive growth rates, respectively. A high F-18-FU retention (>2.96 SUV) was associated with longer survival times (>21 months). In contrast, patients with lower F-18-FU retention values (<1.2 SUV) survived no longer than one year. A higher diagnostic accuracy was obtained by using an integrated evaluation including both compartment and non-compartment models. (18)F-FDG studies for the diagnosis of soft tissue sarcomas showed a sensitivity and specificity of 91% and 88% for the primary tumors and 88% and 92% for the recurrences, respectively. Using a combination of SUV and transport rates, it was possible to further classify malignant soft tissue tumors with regard to tumor grading percentages of 84% of the G III, 37.5% of the G II, 80% of the G I tumors, as well as 50% of the lipomas and 14.3% of scar tissue were correctly classified using the integrated evaluation. In patients with bone tumors, integrated evaluation was also superior to SUV or visual evaluation leading to a sensitivity of 76% (for SUV: 54%), a specificity of 97% (for SUV: 91%) and an accuracy of 88% (for SUV: 75%). The diagnostic efficacy of SUV and of the fractal dimension of the time activity data of FDG was evaluated in 159 patients with 200 lesions of different tumors with respect to differential diagnosis and the prognosis of therapeutic outcome. Discriminant analysis revealed a diagnostic accuracy of 76.65% for all patients, 67.7% for the untreated group of patients and 83.44% for the pretreated patients. The advantage of parametric imaging is the visualization of one isolated parameter of the tracer s kinetic, like the phosphorylation in case of (18)F-FDG. Furthermore, the delineation of a tumor is better due to the absence of background activity. The presented data also demonstrate that parametric imaging based on Fourier transformation may be useful for the evaluation of the pharmacokinetics and effectiveness of regional therapeutic procedures. In conclusion, a semiquantitative analysis of PET data sets based on SUV is in general helpful and should be performed under standardized conditions, concerning the time after tracer application, the blood glucose level in case of (18)F-FDG, partial volume correction and the choice of reconstruction parameters. The combination of two SUV s, an early and a late one is a simple and usefull approach for the evaluation of a dynamic series in a clinical environment. PET studies with labeled cytostatic agents provide information about the transport and elimination of a cytostatic agent and help to predict the therapeutic outcome. Non-compartment models require further evaluation.