Abstract
The correlation between treatment plans based upon 18F-boronophenylalanine positron emission tomography scans and microdosimetry for boron neutron capture therapy
Trent L. Nichols1 and George W. Kabalka2
1Research Staff, Measurement Science and Systems Engineering Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 37831-6075, U.S.A.
2Robert H, Cole Professor of Neuroscience, the University of Tennessee Medical Center, Knoxville, Tennessee, 37920, U.S.A. and the Department of Chemistry, the University of Tennessee, Knoxville, Tennessee, 37996, U.S.A.
Introduction: As with all radiation therapies, an accurate treatment plan is crucial to delivering the optimal tumor dose while minimizing damage to normal surrounding tissue. The complex radiation dosimetry of boron neutron capture therapy (BNCT) requires accurate dosimetry for efficacy. One method to do treatment planning is to load the patient with boronophenylalanine (BPA) prior to debulking surgery to obtain tissue specimens. Those specimens can then be analyzed for boron content. This technique has two major problems. One is that no information is obtained about the boron concentration away from the tumor bed. The second problem is that the resection process disrupts the local vasculature, consequently treatment plans are based on prior rather than the current biochemical milieu. Labelling BPA with positron emitting fluorine-18 allows positron emission tomography (PET) scans to be performed just prior to the treatment with the actual anatomy.
Materials and Methods: Eight patients (7 with glioblastoma and 1 with metastatic melanoma) were evaluated with 18F-BPA PET. The patients with glioblastomas were studied pre- and post-resection as well as post BNCT and the melanoma patient was studied pre- and post-BNCT. The BPA was fluorinated and solubilized with fructose using previously reported protocols. The 18F-BPA fructose (~10 mCi ) was then administered as an intravenous bolus to each patient .
The patients were scanned within 2 hours using an ECAT EXACT 921 whole-body PET system that produced 47 image slices over a 16.2 cm axial field of view. The spatial resolution was 6.5 mm in the xy plane and 7 mm along the z-axis. Dynamic emission images were acquired using 12 10 s frames, five 60 s frames, and six 5 min frames, followed by 10 min frames up to 2 h post injection for as long as the patient would lie on the gantry. All images were reconstructed with conventional software using the measured attenuation correction with filtered back projection, a zoom factor of 2, and a 0.35 pixel Hann filter. Treatment plans were performed using only the PET images where the boron distribution is used to linearly scale the treatment plan with simulation environment for radiotherapy applications (SERA). Both sets of isodose curves were compared both analytically and visually.
A microdosimetric model was written in FORTRAN for a regular body centered cubic geometry of ellipsoidal cells. The effect of separation, boron concentration, and eccentricity were studied for boron concentrations that are consistent with known BPA concentrations. The distribution of the BPA was also studied for differing concentrations in the nucleus, cytoplasm, and interstitium.
Results and discussion: The isodose contours generated using 18F-BPA PET scans are significantly different when compared to curves obtained using traditional treatment plans. The 18F-BPA contours are more irregular and much smaller. Thus the central portion of the tumor receivex a lethal dose whereas the tumor margins receive a much smaller dose that will not be lethal to all malignant cells in that area. The clinical implication is that the tumor will recur in the periphery of the tumor bed which is what is seen clinically. The significance of these findings is amplified when coupled with microdosimetric calculations using ellipsoidal cells. The microdosimetric calculations demonstrate that, when malignant cells are more than a few microns apart or uptake boron avidly, the dose decreases rapidly. With a disrupted normal vascular supply due to resection of the tumor , combined with the likelihood that some cells will be in a resting phase at the time of BPA infusion, some of the malignant cells in the periphery will not receive a significant dose. The clinical implication is that infiltrative tumor such as glioblastoma will require fractionation of the beam and it is likely that BNCT with BPA will not provide long term cures. The implications of this work needs to be reconciled with the relatively good results from much of the BNCT research from Japan. Thus, 18F-BPA PET scans are requisite to adequate treatment planning but more research needs to be done to further our understanding of the radiation dose distribution in BNCT with BPA.