UCLA

Positron emission tomography (PET) is a unique and powerful imaging technique that is used to visualize and quantify various biological processes in living subjects in health and disease. PET imaging can also provide biological information for the assessment of therapies. In this dissertation, we will cover three projects that utilize the quantitative capability of PET for studying two neurological disorders: Alzheimer's disease and brain tumors.

One of the goals in PET imaging is to produce an image volume that accurately describes the true distribution of the injected radioactivity. The correction factor that has the most significant impact on the quantitative aspects of a PET image is attenuation correction. Without it, the reconstructed images will give a distorted view of the true activity distribution. Head movement during a PET scan (especially a dynamic scan) can also lead to a loss in the information content contained in a PET image. This is especially true when scanning patients with dementia or movement disorders. The transmission scan, which is typically acquired at the start of a PET study, corrects for photon attenuation in each of the serial emission scans that are acquired afterwards.

In the first project of this dissertation, we developed a retrospective image-based movement correction (MC) method and evaluated its implementation on dynamic ¹⁸F-FDDNP PET images of cognitively intact control subjects and patients with Alzheimer's disease (AD), each with varying degrees of head movement. ¹⁸F-FDDNP is a PET probe that binds to beta-amyloid plaques and neurofibrillary tangles, the neuropathological hallmarks of AD. The MC method corrected for transmission-emission misalignments as well as for emission-emission misalignments that might have been present in the dynamic PET scan. The image quality, tracer kinetics, and diagnostic accuracy of the ¹⁸F-FDDNP PET images were significantly improved after applying the MC method.

In the second project of this dissertation, we investigated whether changes in ¹⁸F-FLT kinetic parameters, taken early after the start of therapy, could predict overall survival and progression-free survival in patients with recurrent malignant glioma undergoing treatment with bevacizumab (an angiogenesis inhibitor) and irinotecan (a chemotherapeutic agent). ¹⁸F-FLT is a radiotracer used in PET to measure cellular proliferation. We found that when a group of optimal kinetic parameter changes are incorporated into a linear discriminant function, one could accurately classify patients into their known survival groups. This method is advantageous because by reliably identifying short- and long-term survivors early during therapy, clinicians can discontinue ineffective treatment strategies and switch to more advanced treatment regimens that could improve patient outcome.

Our third project expanded on what we did in our second project in that we acquired longitudinal ¹⁸F-FDOPA PET scans in addition to ¹⁸F-FLT PET scans. We also tried to predict overall survival as a continuous outcome variable using multiple linear regression (as opposed to a dichotomous categorical variable with discriminant analysis from before). In brain tumors, ¹⁸F-FDOPA is used to image amino acid transport. We found that in patients with recurrent malignant glioma, kinetic information from ¹⁸F-FLT alone was more predictive than using information from ¹⁸F-FDOPA alone. Using both probes combined provided comparable results to using ¹⁸F-FLT alone, suggesting that a single tracer may provide sufficient information for predicting OS with reasonable accuracy.

The studies reported in this dissertation have demonstrated in three examples that the utility of kinetic quantification can significantly improve the value of PET for imaging biological functions in brain tissues in vivo. The developed methodologies in these examples are also expected to be useful in other quantitative brain PET imaging studies – for other applications or using other tracers.