Supplemental Imaging Methods and Statistics

Supplemental Imaging Methods and Statistics

Supplemental Imaging Methods and Statistics:

MRI perfusion was performed using dynamic T1 permeability technique to assess immediate biological activity followed by T2* perfusion technique. Using the Vitrea®workstation (Vital Images, Minnetonka, MN) and a perimeter technique, we performed volumetric tumor analyses on axial FLAIR sequences, and post-gadolinium axial T1 images with user-assisted semi-automated software; and then recorded tumor volumes on FLAIR, enhancement on T1 gadolinium images and cyst/necrosis on T1 gadolinium images.

Tumor Response Criteria

Complete Response (CR): Complete disappearance on MR of all tumor and mass effect, on a stable or decreasing dose of corticosteroids (or receiving only adrenal replacement doses), accompanied by a stable or improving neurologic examination, and maintained for at least 6 weeks. If CSF was positive it must become negative.

Partial Response (PR): Greater than or equal to 50% reduction in tumor size based on the area calculated using the maximal perpendicular cross-sectional measurements on MR, on a stable or decreasing dose of corticosteroids, accompanied by a stable or improving neurologic examination, and maintained for at least 6 weeks.

Stable Disease (SD): Neurologic exam is at least stable and maintenance corticosteroid dose not increased to maintain neurologic function, and MR/CT imaging meets neither the criteria for PR nor the criteria for Progressive Disease.

Progressive Disease (PD): Progressive neurologic abnormalities or worsening neurologic status not explained by causes unrelated to tumor progression (e.g., anticonvulsant or corticosteroid toxicity, electrolyte disturbances, sepsis, hyperglycemia, etc.), OR the appearance of a new lesion, OR a greater than 25% increase in the bi-dimensional measurement on MR over the smallest sum observed.

Note: The standard criterion for disease progression is 25% increase in tumor size. However, because AZD2171 is a cytostatic agent, it is possible that there may be a lag time between the initiation of therapy and antitumor effect. Removing a patient from treatment as soon as the tumor increases in size by 25% might have resulted in the study drug being terminated prematurely. It is possible that if these patients were maintained on the study drug that their disease might eventually regress. Thus, patients were allowed to remain on therapy until the tumor had increased at least 50% in size from baseline. However, patients who exhibited significant clinical symptoms from any degree of tumor enlargement were considered to have progressive disease and were taken off therapy.

Evaluation of permeability, perfusion, and diffusion parameters: Perfusion images were transferred to a UltraSPARC II™ workstation (Sun Microsystems, Santa Clara, California) and relative cerebral blood volume (rCBV) maps generated from the dynamic susceptibility-weighted perfusion MRI data. An MR perfusion region of interest was placed in the solid part of the target tumor lesion with the highest cerebral blood volume (CBV) and divided by a region of interest (ROI) from the frontal white matter and the ratio value recorded. Regarding diffusion images, a ROI (3mm in diameter) within the solid part of the tumor (determined from the T1, T2 FLAIR, T2 and post-gadolinium T1 sequences) from the apparent coefficient diffusion map was divided by the value from a ROI obtained in the frontal lobe white matter. Of note, the tumor ROI determined in diffusion analysis corresponded to the ROI generated in perfusion analysis. MR permeability imaging was performed with 3D axial T1 dynamic contrast enhanced sequences with kinetic modeling yield permeability (Kps) and CBV measurements. Regions of interest were placed within the areas of highest permeability on the Kps maps, and Kps values and fractional cerebral blood volume (fCBV) values were recorded. While ROI measurements may be difficult for small lesions 1-2mm in diameter, the reproducibility of these measurements for larger tumors is robust due to the same operator performing all ROI measurements.

FDG PET scans were acquired on a variety of scanners (Advance NXI [GE Healthcare], Discovery LS [GE Healthcare], Discovery STE [GE Healthcare], G-PET [Philips], HR1 [Siemens], and HiRez Bioscan [Siemens]). The consistency of the PET data was maintained by adherence to a standard acquisition protocol and quality assurance program which included daily blank scans and quarterly normalization, calibration, and preventive maintenance. Patients fasted for 4 hours before PET. The baseline brain PET scan was acquired in 3D mode for 10 minutes, 40–60 minutes after the intravenous administration of FDG (5.55 MBq/kg) (minimum dose, 18 MBq; maximum dose, 370 MBq). Attenuation correction was performed using a 3-minute segmented transmission scan with 68Ge/68Ga rods when the scanner was PET only or a CT-based approach when images were acquired on a PET/CT scanner. The acquired data were reconstructed using Fourier rebinning, followed by a 2D ordered subset expectation maximum reconstruction algorithm.

All PET and fused PET/MR images were evaluated by a pediatric neuroradiologist and nuclear medicine physicist, for intensity and uniformity of tracer uptake in the tumor. The intensity of FDG uptake was judged on a subjective scale (1-no uptake, 2- similar to white matter (WM), 3-more than WM but less than gray matter (GM), 4–similar to GM, 5-more than GM). Uniformity was defined as the percentage of the tumor (as delineated on the FLAIR MR) demonstrating FDG uptake and was subjectively graded on a 4-point scale (grade 0:<25%, grade 1: 25-50%, grade 2: 50-75%, and grade 3: >75%). In addition, an objective grading was also used. Mean and maximum pixel values within two dimensional ROIs drawn about the tumor (T) normalized by values form contra-lateral, normal GM and WM ROIs deemed Mean and Max T/WM and T/GM, respectively.

Statistics: The Wilcoxon signed rank test was used to look for changes between two time points for MRI FLAIR volume, enhancing volume, diffusion ratio and perfusion ratio. The change ratios for MRI variables were calculated at the same timepoints as the ones used for biological markers described above. A similar analysis was used to test the fold change between two time points for CBV 2D mean, Kps 2D mean, CBV 3D mean, and Kps 3D mean as well as PET parameters tumor vs. white matter (TM/WM) and tumor vs. gray matter (TM/GM) ratios.. The Wilcoxon rank sum test was used to compare TM/WM and TM/GM with Uniformity. Spearman rank correlation was used to investigate the association between biology markers and MRI parameters and between PET variables with MRI variables (e.g. Diffusion Ratio and Perfusion Ratio).

The Cox proportional hazard model was used to investigate the association of MRI and PET variables with progression free survival (PFS) in an exploratory fashion. An association with overall survival was not performed since the number of deaths reported while on study was too small to render such analysis viable. The p-values were not adjusted for the multiple testing.