Key words: recurrent glioblastoma multiforme, bevacizumab, irinotecan, monitoring therapeutic response, magnetic resonance imaging

Transcription

1 The Neuroradiology Journal 21: , 2008 www. centauro. it Peritumoral Apparent Diffusion Coefficient as a Metric of Response in Patients with Recurrent Glioblastoma Multiforme Treated with Bevacizumab and Irinotecan J.B. ANDRE, S. LU, K. SPEARMAN, S.N. RAVAL* Department of Radiology, Monmouth Medical Center; Long Branch, NJ (JA, SL, KS) * David S. Zocchi Brain Tumor Center, Monmouth Medical Center; Long Branch, NJ (SR), USA Key words: recurrent glioblastoma multiforme, bevacizumab, irinotecan, monitoring therapeutic response, magnetic resonance imaging SUMMARY Bevacizumab and irinotecan have shown promising results in patients with recurrent glioblastoma multiforme (GBM), which traditionally carries a poor prognosis after first-line therapies have been exhausted. Retrospectively documenting the short-term effects of this chemotherapeutic regimen on recurrent GBM, as evidenced by comparative magnetic resonance images obtained two weeks prior to, and one-month following initiation of treatment, we hypothesize that peritumoral apparent diffusion coefficient (ADC) values will decrease on post-treatment scans. Brain MR data were collected from August 2005 to December 2006, in which post-contrast T1-weighted images demonstrated measurable enhancement or GBM tumor mass. Pre- and post-treatment MR images for ten consecutive patients were collected, each having failed temozolomide and radiation therapy. Pre- and post-treatment recurrent GBM bulk tumor and peritumoral T2 signal abnormality were measured in three dimensions. Diffusion of peritumoral T2 signal abnormality was evaluated on pre- and post-treatment ADC. All patients witnessed a significant decrease in tumor bulk ranging from 15.3% to 96.7% with a mean reduction of 48.2%, having received an average of two cycles of chemotherapy. FLAIR images demonstrated a mean volumetric reduction in peritumoral T2 signal abnormality of 44.3%. ADC measurements demonstrated an average reduction in peritumoral ADC of 20.6%, which was statistically significant (p-value <.005). Recurrent GBM tumor bulk demonstrated a 48.2% mean reduction, with corresponding decrease in peritumoral ADC values of 20.6%, suggesting that ADC may represent a valuable metric in the evaluation of the chemotherapeutic response of recurrent GBM, when treated with bevacizumab and irinotecan. Introduction Primary glioblastoma multiforme (GBM), a rapidly progressive, WHO grade IV astrocytoma carries a poor prognosis after first-line therapies have been exhausted, even in the setting of gross total resection. The median survival of patients with GBM is 14 weeks with conservative therapy, 20 weeks with surgical resection alone, 36 weeks with surgery and radiation, and weeks with the addition of chemotherapy 123. Adjuvant external beam radiation therapy (RT) with concurrent temo- zolomide chemotherapy is the current standard of care, even following complete tumor resection. However, the optimum management for patients with recurrent GBM specifically is unclear and there is no widely accepted standard therapy. Therapeutic options include conventional chemotherapy, re-irradiation, re-operation, and experimental treatments. Bevacizumab and irinotecan have show promising results in patients with recurrent GBM 4. GBMs are highly vascular and demonstrate widespread expression of proangiogenic vascular endothelial growth factor (VEGF), 350

2 J.B. Andre Peritumoral Apparent Diffusion Coefficient as a Metric of Response in Patients with Recurrent Glioblastoma Multiforme Treated... which plays an important role in tumor angiogenesis and maintenance of existing tumor vessels 5. Increased VEGF expression in human patient astrocytoma has been associated with a poor prognosis 6. Bevacizumab is a humanized monoclonal IgG1 antibody that binds to and inhibits the biological activity of human VEGF- A, thereby interfering with the growth of new blood vessels, potentially helping to block further growth of tumor, and potentially assisting with synergistic chemotherapeutic effectiveness. Irinotecan is a topoisomerase-1 inhibitor mainly used in the treatment of colorectal cancer, particularly when combined with other chemotherapeutic agents. The first prospective phase II trial of Bevacizumab and Irinotecan in recurrent malignant astrocytomas has shown radiographic response in 63% of patients (20 of the included 32 patients) receiving this chemotherapeutic regimen every two weeks for a six week cycle 7. In this trial and other phase II trials of chemotherapies for recurrent malignant astrocytoma, the Macdonald criteria have been historically used to assess response in a clinical setting, and take into account steroid use and clinical findings 8. These criteria utilize the largest cross-sectional area of post-contrast tumor enhancement, in which foci of enhancement represent a breakdown in the blood-brain barrier. Although established over 15 years ago, the Macdonald criteria substantially reduced the frequency of false-positive responses in chemotherapeutic trials. However, it is now recognized that false-negative responses may also occur in the setting of successful chemotherapy. In part, this is because abnormal enhancement is nonspecific and often cannot differentiate tumor progression from therapy-related changes. Because of the limitations of conventional post-contrast T1-weighted images, additional MR sequences have been employed such as fluid-attenuated inversion recovery (FLAIR) and diffusion weighted imaging. Vasogenic edema describes increased fluid content in extracellular spaces secondary to disturbances in vascular permeability, implying underlying reactive, reversible, pathologic processes. In the special case of astrocytomas, it is accepted that peritumoral edema is best referred to as infiltrative edema, because it represents both vasogenic edema as well as infiltrating tumor cells behind the blood-brain barrier usually invading along white matter tracts. Peritumoral FLAIR signal intensity abnormality has been shown by Tang et Al to have a positive predictive value for malignant astrocytoma of 84%, when differentiating from metastases 9. While FLAIR imaging provides valuable information regarding the presence and volumetric extent of peritumoral edema, it possesses no units, and therefore does not proffer a quantitative measurement for changes in peritumoral T2 signal. Thus, it is difficult to quantifiably measure therapy-related changes in peritumoral T2 signal abnormality immediately adjacent to, or at some specified distance from a given tumor s epicenter. MR diffusion imaging has been performed to study water mobility in normal brain tissue, cerebral infarction, multiple sclerosis, astrocytomas, brain abscesses and in differentiating arachnoid from epidermoid cysts. Multiple studies have been performed to assess the utility of diffusion weighted imaging (DWI) in evaluating and monitoring cancer therapy. True hyperintensity on diffusionweighted images (with corresponding apparent diffusion coefficient hypointensity) is thought to reflect restricted diffusion of water molecules in cells compromised by cellular edema. The apparent diffusion coefficient (ADC) is calculated from DWI and is a measurable value of the conglomerate water molecules mobility in tissue. In contrast to DWI, ADC represents diffusion effects only, in which T2 effects have been removed, such that abnormalities appear as low signal. Several studies have demonstrated that diffusion-weighted imaging alone fails to differentiate tumor types 10. ADC has been shown to have a limited role specifically in differentiating high grade astrocytomas from surrounding peritumoral edema which is thought to be secondary to the signal intensity heterogeneity caused by tumoral necrosis and/ or susceptibility artifact 11. Perfusion MR imaging, specifically relative cerebral blood volume (rcbv) measurement, has shown promising results regarding initial tumor characterization and glioma grade, accomplished by radiographic evaluation of vascular morphology and degree of angiogenesis, which correlate well with histopathologic grade 12,13. While this method of tumor analysis shows high correlation with initial glioma grading, it has been suggested that surgery and post-operative changes may affect tumor vascularity and cellularity, possibly compromising evaluation of the post-resection bed 12. An established role for this advanced MR imaging modality has not been defined in the setting of post-resection suspected clinical recurrence of high-grade as- 351

3 Peritumoral Apparent Diffusion Coefficient as a Metric of Response in Patients with Recurrent Glioblastoma Multiforme Treated... J.B. Andre trocytoma. Moreover, there is little consensus as to what maximal rcbv value might constitute a threshold for definitive tumor recurrence in the post-resection setting. Some authors suggest there are clear differences in the diffusion characteristics of the vasogenic edema surrounding intraparenchymal lesions when compared with those of normal white matter 14. The compromised ability of the ADC map to reliably differentiate central tumoral necrosis from immediate peritumoral edema naturally leads to an examination of the peritumoral region on ADC map as a more reliable way to quantify peritumoral edema and chemotherapeutic response. Along these lines, changes in peritumoral diffusion parameters measured shortly after initial treatment may prove to be a more quantifiable and reliable method to predict treatment response, perhaps allowing for an earlier detection of a successful regimen, as suggested by Jordan, et Al 15. In this study, we sought to retrospectively document and evaluate the short term effects that bevacizumab and irinotecan exert on recurrent GBM, as evidenced by pre- and posttreatment post-contrast T1-weighted, FLAIR, and apparent diffusion coefficient map images in order to better define pragmatic and reliable methods for measuring treatment response in this cohort of patients. Materials and Methods Patient Selection Brain MR data were retrospectively collected from August 2005 to December 2006 of patients with suspected recurrence of GBM. Only post-contrast spin-echo T1-weighted images demonstrating measurable enhancement and/or GBM tumor mass met initial inclusion criteria, which also required that both pre- and post-treatment images be available for direct measurement by a neuroradiologist at our institution. Ten consecutive patients, each with recurrent high-grade astrocytoma, confirmed by biopsy or tumor resection at the time of suspected clinical recurrence, met inclusion criteria for the study. GBM Chemotherapy Regimen All patients received external bean radiation at the time of their initial diagnosis. After surgical resection of the original tumor, each subsequently failed at least one course of temozolomide chemotherapy. Institutional review board approval was obtained. Under the direct guidance and supervision of a neurooncologist (SR) each patient was treated with Bevacizumab 5 mg/kg IV infusion over 60 minutes and Irinotecan 125 mg/kg IV infusion over 120 minutes every two weeks until disease progression or development of unacceptable toxicity. Dose modifications were made for hematologic and gastrointestinal toxicity, whose details are reported elsewhere 16. A staging MR was performed after two cycles of chemotherapy. Imaging Parameters Using a General Electric Sigma 1.5 T scanner, sequential pre- and post-gadolinium-enhanced spin-echo T1-weighted images were acquired separately in three planes with an imaging parameter protocol that included a slice thickness of 5 mm, an interslice gap of 1.5 mm, TR=450, TE=20, FOV of 24x24 cm, matrix 256x256, and NEX=1. Sequential spin-echo T2-weighted fluid attenuated inversion recovery (FLAIR) images were acquired in two planes using a slice thickness of 5 mm and an interslice gap of 1.5 mm, TR=9,000, TE=120, FOV of 24x24 cm, matrix 256x256, and NEX=1. Multiplanar reformat FLAIR images were generated at a separate workstation. ADC map images were acquired (before administration of contrast material) using b values of 0 and 1000 mm 2 /s, applied in the three directions, with a bandwidth of 125 KHz, FOV of 36x36 cm, matrix 256x256, TR=10,000, TE=121, Slew rate of 77 mt/s, and NEX=1. Post-processing of ADC map images was performed off-line on the Silicon Graphics Octane workstation. Data Acquisition Pre-treatment MR imaging was performed two weeks prior to the initiation of bevacizumab and irinotecan therapy (Mean: 13 days, Median: 10 days). Post-treatment scans were performed at approximately one month following initial treatment (Mean: 30 days, Median: 28 days). Under the direct supervision of a fellowship-trained neuroradiologist (SL), one investigator (JA) measured pre- and posttreatment tumor volume on T1-weighted postcontrast images, and peritumoral T2 signal abnormality on FLAIR images, in anteroposterior (AP), transverse, and cranio-caudal (CC) dimensions respectively. Volumetric tumor data were calculated based on an ellipsoid tumor configuration: 352

4 www. centauro. it The Neuroradiology Journal 21: , 2008 Figure 1: Tumor Volume on Post-contrast T1WI Tumor Volume (cm 3 ) Pre-Treatment Post-Treatment Patient Figure 1 Enhancing tumor volume as measured from gadolinium-enhanced spin-echo T1-weighted images, obtained both preand post-treatment with bevacizumab and irinotecan. Patients are labeled 1 through 10, respectively. 4/3 π (AP radius Transverse radius CC radius) In patients with tumor recurrence immediately adjacent to a resection cavity, the volume of the resection cavity was subtracted from the volume of the enhancing parenchymal lesion, to ensure a more accurate representation of the measured enhancing tissue volume. On ADC map images, one author (JA) applied four circular regions of interest (ROIs) to the peritumoral T2 signal intensity abnormality and four such ROIs within the corresponding contralateral normal-appearing white matter (NAWM), under the guidance of a board-certified neuroradiologist (SL). The author placed fixed 4.0-mmdiameter ROIs anteriorly, medially, posteriorly, and laterally circumferentially adjacent to the enhancing portion of the tumor to obtain the most random reproducible and consistent ADC sampling among all patients. For cases with more cortically based tumors, ROIs were placed within white matter only, such that inapplicable ROIs were distributed evenly in the surrounding white matter. Careful attention was made to avoid the placement of ROIs within gray matter, in order to prevent erroneous sampling of inherent gray matter ADC characteristics. Statistics Statistical analysis was performed using GraphPad Software to generate descriptive statistics, confidence intervals of the mean, and Wilcoxon Matched-Pairs Signed-Ranks Test. Results After receiving an average of two cycles of bevacizumab and irinotecan, all patients showed a significant decrease in tumor bulk volume ranging from 15.3% to 96.8%, with a mean reduction of 48.2% (SD=25.8%, SEM=8.2%, 95% CI= % decrease). The Wilcoxon Matched-Pairs Signed-Ranks test yielded a two-tailed p-value <= These data are illustrated in figure 1. Individual patient volumetric post-contrast T1-weighted image data are summarized in table 1, and are referred to as Volume on Post-contrast T1WI. These measurements are taken from pre- and posttreatment MR images in patients 1 10, labeled Pre Tx and Post Tx, respectively, and are measured in cubic centimeters (cm 3 ). Individual patient-associated percent decrease of tumor bulk after treatment with two-cycles of bevacizumab and irinotecan is labeled % decrease. Examples of the radiographic representation of the post therapeutic response on MR images are provided in figure 2. Multiplanar reformatted FLAIR images demonstrated a mean reduction in peritumoral T2 signal abnormality of 44.3%, ranging from 21.4 to 83.0% (SD=18.9%, SEM=6.0%, 95% CI= %). The calculated Wilcoxon Matched-Pairs Signed-Ranks test yielded a corresponding twotailed p-value <= One example of posttreatment volumetric reduction in peritumoral T2 signal abnormality, as evidenced on FLAIR 353

5 Peritumoral Apparent Diffusion Coefficient as a Metric of Response in Patients with Recurrent Glioblastoma Multiforme Treated... J.B. Andre Table 1 Synopsis of MR image data including pre- and post-treatment T1-weighted, FLAIR, and ADC map image data. Patient Volume on Postcontrast T1WI Pre Tx (cm 3 ) Post Tx (cm 3 ) % Decrease Volume on FLAIR Pre Tx (cm 3 ) Post Tx (cm 3 ) % Decrease Mean ADC Pre Tx (mm 2 /s) Post Tx (mm 2 /s) % Decrease 01 04,820 01,210 74,9% 087, ,810 83,0% 1,701 1,212 28,7% ,020 22,900 18,3% 203, ,010 53,7% 1,312 1,249 04,8% ,200 06,490 42,1% 071, ,570 44,4% 1,580 1,270 19,6% ,880 96,470 15,3% 557, ,100 34,7% 1,457 1,252 14,1% ,540 09,710 63,4% 177, ,840 23,5% 1,832 1,447 21,0% ,700 02,250 39,2% 028, ,320 32,6% 1,364 1,300 04,7% ,100 11,250 25,5% 214, ,640 21,4% 1,694 0,977 42,3% ,740 03,340 56,8% 306, ,910 43,9% 1,388 1,079 22,3% ,130 4,19E-03 96,8% 133, ,200 40,5% 1,569 1,121 28,6% ,150 07,120 49,7% 132, ,820 65,4% 1,172 0,933 20,4% Average 022,528 16,074 48,2% 191, ,322 44,3% 1,507 1,184 20,6% Median 012,675 06,805 45,9% 155, ,605 42,2% 1,513 1,231 20,7% imaging, can be seen in figure 2. FLAIR image data are summarized in Table 1 in which volumetric peritumoral T2 signal abnormality is depicted based on measurements taken from multiplanar reformat FLAIR image data (labeled Volume on FLAIR and measured in cm3), with associated percent decrease in posttreatment peritumoral T2 signal abnormality. A graphical summary of FLAIR image data is provided in figure 3. Measurement of peritumoral T2 hyperintensity on ADC map image data is summarized in figure 4, revealing a mean reduction in ADC of 20.6% (SD=11.3%, SEM=3.6%, 95% CI= % decrease). Comparing the pre- and post-treatment peritumoral T2 signal hyperintensity on ADC map, the Wilcoxon Matched- Pairs Signed-Ranks test yielded a two-tailed p- value< Individual patient-specific ADC map image data are shown in table 1, in which Mean ADC represents the average measurement of pre- and post-treatment peritumoral T2 signal abnormality as measured on ADC map image data (labeled Pre Tx and Post Tx, respectively) measured in millimeters 2 / second (mm 2 /s). This data is depicted with comparative percent decrease in mean ADC values (labeled as % decrease ). Figure 5 provides a pre-treatment radiographic example of a large left frontoparietal rim-enhancing recurrent GBM and the typical associated peritumoral T2 signal abnormality as evidenced on ADC map. Two additional patient examples of preand post-treatment ADC map image data are provided in figure 6, both of which demonstrate significant decrease in peritumoral T2 signal abnormality on post-treatment images (B and D, respectively), relative to pre-treament images (A and C, respectively). Discussion Our study sought to evaluate the chemotherapeutic response of GBMs as represented on pre- and post-treatment MR data, as it might pertain to clinical efficacy of therapeutic regimen. The initial reported results from several phase II trials demonstrate good correlation between imaging findings such as those reported by our group, and observed clinical improvement, with sometimes associated increased progression-free survival. The radiographic findings presented above, in patients treated with an average of two cycles of bevacizumab and irinotecan, are comparable to those reported in the literature in concurrent phase II clinical trials. Of the ten 354

6 www. centauro. it The Neuroradiology Journal 21: , 2008 A B C D Figure 2 A-D Baseline (A and B) and post-treatment (C and D) MRs of patient treated with bevacizumab and irinotecan. T1- weighted post-contrast images (A and C) demonstrate an enhancing recurrent GBM in the right temporal lobe. FLAIR images (B and D) demonstrate the extent of peritumoral T2 abnormality. Note the significant reduction in both the contrast-enhancing tumor mass as well the peritumoral T2 signal abnormality on post-treatment images (C and D respectively). patients included in our study, all witnessed a significant decrease in post-contrast T1- weighted tumor mass (48.2% reduction), with significant associated decrease in peritumoral T2 signal abnormality as evidenced on both FLAIR images (44.3% reduction) and on ADC map (20.6% reduction). These changes were all statistically significant (p<.005). The marked reduction in enhancing tumor bulk as evidenced on post-treatment T1- weighted images may be secondary to myriad effects. These may include, but are not limited to: regression of tumor, increased intra- and/ or peritumoral necrosis, collapse of the tumor cavity, regression of tumor cells within the tumor bed, tumor cell apoptosis, and possibly de- 355

7 Peritumoral Apparent Diffusion Coefficient as a Metric of Response in Patients with Recurrent Glioblastoma Multiforme Treated... J.B. Andre Figure 3: Peritumoral T2 Signal Abnormality on FLAIR imaging Peritumoral T2 signal abnormality volume (cm 3 ) Pre-Treatment Post-Treatment Patient Figure 3 Volumetric peritumoral T2 signal abnormality, as measured from multiplanar reformat FLAIR images in patients labeled 1-10, respectively. 2 Figure 4: ADC of Immediate Peritumoral Region and Contralateral NAWM 1,8 1,6 1,4 ADC (x 10-3 mm 2 /s) 1,2 1 0,8 0,6 Peritumor pre-tx NAWM pre-tx Peritumor post-tx NAWM post-tx 0,4 0, Patient Figure 4 Average change in ADC of the immediate peritumoral T2 signal abnormality compared with contralateral normal appearing white matter (NAWM), as demonstrated on ADC map in pre- and post-treatment MR images (labeled pre-tx and post-tx, respectively). ADC is measured in millimeters 2 /second (mm 2 /s). 356

8 www. centauro. it The Neuroradiology Journal 21: , 2008 A B Figure 5 A,B Radiographic correlation of a large left frontoparietal rim-enhancing recurrent GBM as seen on pre-treatment postcontrast T1WI (A). Associated ADC map image data (B) demonstrates the GBM (arrow) and large area of peritumoral T2 signal abnormality surrounding the tumor, affecting nearly the entire visualized left hemisphere. creased vascularity secondary to the presumed anti-angiogenic effects of bevacizumab. The reader is reminded that peritumoral edema, as depicted on MR imaging in astrocytomas, has been reported to contain infiltrating neoplastic cells 17. It has been suggested that bevacizumab may target VEGF directly to decrease vascular permeability and cerebral edema as well as enhancing the effects of irinotecan to reduce tumor bulk. The correlation of decreased post-treatment peritumoral T2 signal abnormality supports this theory, although this proffered relationship is neither specific nor exclusive. Additionally, the observed decrease in peritumoral T2 signal supports the concept that the latter likely represents peritumoral infiltrative edema, which is not surprisingly reduced by an anti-angiogenic chemotherapy such as bevacizumab. This is perhaps best demonstrated by the reduction in diffusion as evidenced by decreased mean ADC on posttreatment images. The observation that pre-treatment diffusion of peritumoral T2 signal abnormality is significantly elevated relative to the measured contralateral NAWM suggests increased peritumoral water diffusivity. This is corroborated by the peritumoral T2 signal abnormality present on FLAIR images. The peritumoral T2 signal abnormality, as evidenced by both FLAIR images and ADC map, likely represents brain parenchymal reaction to the presence of a neoplastic and/or inflammatory process. The observation that this peritumoral T2 signal abnormality is reduced in size on FLAIR images, and reduced in extent on ADC map suggests that there is reduction in the effect that this neoplastic process exerts upon the surrounding brain parenchyma. It is unclear at present if this observation represents a decrease in the brain s reaction to the presence of tumor, potentially a by-product of decreased cytokine availability leaked into the brain parenchyma, which is assisted by the antiangiogenic effects of bevacizumab. Alternatively, this phenomenon may be evidence of decreased capillary leakage, manifesting as decreased peritumoral edema, secondary to inhibitory effects of bevacizumab on VEGF expression. Finally, the observed decrease in peritumoral T2 signal abnormality may simply represent a reduction in the brain parenchyma s response to the presence of a shrinking 357

9 Peritumoral Apparent Diffusion Coefficient as a Metric of Response in Patients with Recurrent Glioblastoma Multiforme Treated... J.B. Andre necrotic mass, which by sheer virtue of decreased size alone incites a reduced inflammatory response. While these observations can only be proven histologically, it is interesting to note that ADC values within the contralateral NAWM on both pre- and post-treatment images were not significantly different (difference in the means = ), suggesting that any observed differences in peritumoral ADC map values are representative of significant alteration in the peritumoral process alone, whether that be a reduction in edema, infiltrative tumor, some additional as yet undefined process, or some combination of these factors. It should be noted that additional physiologic parameters may affect ADC measurements including regional hydrostatic and osmotic gradients, regional blood flow and endoluminal surface area. While there is relatively good correlation between the volumetric reduction in tumor bulk and that of peritumoral T2 signal abnormality on FLAIR images (48.2% and 44.3% reduction, respectively), for the 10 patients, the reduction in peritumoral ADC is less remarkable, measuring 20.6%. While this is statistically significant, comparison between ADC map values and either tumor bulk or peritumoral T2 FLAIR signal abnormality on an individual basis did not reveal any significant correlation (See table 1). This can best be explained by the observation that both tumor bulk and volumetric size of peritumoral T2 signal abnormality on FLAIR images were measured from the outermost detectable periphery of the lesion or abnormality, while ADC map ROIs were always placed immediately adjacent to the enhancing rim of any given lesion, and thus in the proverbial center of the presumed inflammatory process. In nearly all cases, the volumetric reduction of peritumoral T2 signal abnormality was not extensive enough so as to affect the immediate peritumoral area, and hence measured ADC values were not likely to have been placed over white matter that had become free of infiltrating tumor cells and/or edema. This observation may help to explain the apparent incongruent relationship between tumor bulk reduction and that of peritumoral T2 signal abnormality, as measured on ADC map. Thus, the immediate peritumoral T2 signal abnormality as measured on ADC map may represent an additional, important variable in the radiographic monitoring of therapeutic effectiveness of bevacizumab and irinotecan. In measuring the effectiveness of bevacizumab and irinotecan in treating recurrent GBM, some of the post-treatment images demonstrated such robust and significant reduction in tumor bulk that orthogonal measurements were at times difficult to measure. This was most apparent when measuring the tumor bulk reduction in patient 9 (see table 1), who witnessed a 96.8% reduction in tumor volume. A few limitations are inherent in the MR image acquisition process that are worth mentioning. First is the difficulty in positioning a patient within the MR gantry in such a way that both pre-treatment and post-treatment MR scans acquire images of a given lesion in an identical manner on two separate occasions. This is compounded by the observation that many intraaxial lesions assume an ellipsoid configuration that rarely matches the orthogonal axes imposed upon these lesions by the MR scanner. Furthermore, there is additional image variability created by patient motion between given sequence acquisitions, within the same MR scan, such that pre- and post-contrast images may not align exactly with FLAIR images, for example. This generates additional error when comparing lesion characteristics on associated sequences within the same MR scan data. To compensate for this potential introduction of error, volumetric measurements were obtained, thereby correcting for changes in patient positioning and head angle. An additional limitation to our study is the relatively subjective placement of ROIs in evaluating ADC map data. To control for differences in ADC values both within and between patients, we sampled the contralateral white matter that corresponded most closely to the peritumoral ROIs. Significant effort was made to precisely and accurately sample the contralateral white matter most representative of that adjacent to or surrounding the tumor bed, while avoiding adjacent gray matter, where applicable. Interestingly, while image resolution and voxel size might somewhat compromise the peritumoral measurement of orthogonal values, Oouchi et Al recently demonstrated that ADC values are not significantly affected by voxel size or shape 18. Given the significant reduction in reported post-therapy tumor volume and peritumoral T2 signal abnormalities, the small sample size of the study seems acceptable. 358

10 www. centauro. it The Neuroradiology Journal 21: , 2008 A B C D Figure 6 A-D Baseline (A and C) and post-treatment (B and D) ADC map image data of two patients treated with bevacizumab and irinotecan. Pre-treatment (A) and post-treatment (B) ADC map images for select patient 1 demonstrates GBM recurrence in the left temporal lobe, in which increased peritumoral ADC values are demonstrated on pre-treatment image (A), which has shown moderate decrease on post-treatment image (B), as evidenced by decrease in peritumoral signal. Pre-treatment (C) and post-treatment (D) ADC map images for select patient 2 demonstrates GBM recurrence in the right temporal lobe, in which increased peritumoral ADC signal abnormality is noted on pre-treatment image (C), which is markedly decreased on post-treament image (D), evidenced by decrease in peritumoral signal abnormality. This small sample represents our initial experience, and additional patients will be selected in the near future to expand the patient population. Withstanding the above limitations, the use of ADC as a metric of treatment response remains strikingly robust when correlated with the use of bevacizumab and irinotecan in the effective treatment of recurrent GBM tumor, and radiographic correlation. 359

11 Peritumoral Apparent Diffusion Coefficient as a Metric of Response in Patients with Recurrent Glioblastoma Multiforme Treated... J.B. Andre Conclusion As demonstrated by post-contrast T1-weighted, FLAIR, and apparent diffusion coefficient images, bevacizumab and irinotecan demonstrate significant anti-tumor activity against recurrent glioblastoma multiforme. Given the observed 48.2% mean reduction in recurrent GBM tumor bulk, 44.3% volumetric reduction in peritumoral T2 signal abnormality, and 20.6% reduction in peritumoral mean ADC in patients treated with an average of two cycles of this chemotherapeutic regimen, our data suggest that ADC may be used as a metric in monitoring the therapeutic effectiveness of this synergistic chemotherapeutic regimen and that changes in peritumoral ADC may correlate well with the chemotherapeutic alteration in the peritumoral process alone, whether that be a measurement of edema, infiltrative tumor, some additional as yet undefined process, or some combination of these factors. It is hoped that perhaps the future addition of perfusion MR imaging to the standard imaging regimen in this cohort of patients will help to answer these questions, particularly when compared to the initial results seen with the use of ADC map image data. Notwithstanding, further long term prospective radiographic evaluation is necessitated regarding the parenchymal effects that bevacizumab and irinotecan exert on recurrent GBMs. Acknowledgements The clinical information presented in this manuscript was presented in greater detail at the 2007 annual meeting of the American Society of Clinical Oncology and may be found under the following citation: Journal of Clinical Oncology, 2007 ASCO Annual Meeting Proceedings Part I. Vol 25, No. 18S (June 20 Supplement), 2007: A portion of the presented radiographic material, including a subset of T1-weighted image data, with omission of all FLAIR and ADC map image data, was printed in the proceedings, but not presented at the 2007 annual meeting of the American Society of Clinical Oncology, and may be found under the following citation: Journal of Clinical Oncology, 2007 ASCO Annual Meeting Proceedings Part I. Vol 25, No. 18S (June 20 Supplement), 2007: A poster presentation of the FLAIR and T1- weighted image data, with specific omission of all ADC map data radiographic material is to be presented at the 2007 annual meeting of the Society of Neuro-oncology, and may be found at the society s website under the heading, Accepted Abstracts Ra-08. Radiographic Evaluation of Recurrent Glioblastoma Multiforme (GBM) In Patients Treated with Bevacizumab and Irinotecan. 360

12 www. centauro. it The Neuroradiology Journal 21: , 2008 References 1 Fine HA, Dear KB, Loweffler JS et Al: Meta-analysis of radiation therapy with and without adjuvant chemotherapy for malignant gliomas in adults. Cancer 71: 2585, Salcman M, Scholtz H, Kaplan RS et Al: Long-term survival in patients with malignant astocytoma. Neurosurgery 34: 213, Huncharek M, Muscat J: Treatment of recurrent high grade astrocytoma; results of a systematic review of 1415 patients. Anticancer Res 18: 1303, Stark-Vance V: Bevacizumab and CPT-11 in the treatment of relapsed malignant glioma. Neuro-Oncol 7: 369, Lamaszus K. Ulbricht U, Matschke J et Al: Levels of soluble vascular endothelial growth factor (VEGF) receptor 1 in astrocytic tumors and its relation to malignancy, vascularity and VEGF-A. Clin Cancer Res 9: , Godard S, Getz G, Delorenzi M et Al: Classification of human astrocytic gliomas on the basis of gene expression: a correlated group of genes with angiogenic activity emerges as a strong predictor of subtypes. Cancer Res 63: , Vredenburgh JJ, Desjardins A, Herndon JE et Al: Phase II Trial of Bevacizumab and Irinotecan in Recurrent Malignant Glioma. Clinical Cancer Research 13; , Macdonald DR, Cascino TL, Schold AC et Al: Response criteria for phase II studies of supratentorial malignant glioma. J Clin Oncol 8: , Tang YM, Ngai S, Stuckey S: 2003 The Solitary Enhancing Cerebral Lesion: Can FLAIR Aid the Differentiation between GLIOMA and Metastasis? Am J Neuroradiol 27: , Kono K, Inoue Y, Nakayama K et Al: The role of diffusion-weighted imaging in patients with brain tumors. Am J Neuroraiol 22: , Castillo M, Smith J., Kwock L et Al: Apparent Diffusion Coefficients in the Evaluation of High-grade Cerebral Gliomas. Am J Neuroradiol 22: 60-64, Law M, Yang S, Wang H et Al: Glioma Grading: Sensitivity, Specificity, and Predictive Values of Perfusion MR Imaging and Proton MR Spectroscopic Imaging Compared with Conventional MR Imaging. Am J Neuroradiol 24: , Law M, Young R, Babb J et Al: Comparing Perfusion Metrics Obtained from a Single Compartment Versus Pharmacokinetic Modeling Methods Using Dynamic Susceptibility Contrast-Enhanced Perfusion MR Imaging with Glioma Grade. Am J Neuroradiol 27: , Lu S, Ahn D, Johnson G et Al: Peritumoral Diffusion Tensor Imaging of High Grade Gliomas and Metastatic Brain Tumors. Am J Neuroradiol 24: , Jordan B, Runquist M, Raghunand N et Al: Dynamic Contrast-Enhanced and Diffusion MRI Show Rapid and Dramatic Changes in Tumor Microenvironment in Response to Inhibition of HIF-1a Using PX-478. Neoplasia 7: , Raval, SN, Hwang, SS, Dorsett, L: Bevacizumab and irinotecan in patients (pts) with recurrent glioblastoma multiforme (GBM). J of Clin Oncol, 2007 ASCO Annual Meeting Proceedings Part I 25: 2078, Kelly PJ, Daumas-Duport C, Kispert DB et Al: Imaging-based stereotactic biopsies in untreated intracranial glial neoplasms. J Neurosurg 66: , Oouchi H, Yamada K, Sakai K et Al: Diffusion Anisotropy Measurement of Brain White Matter Is Affected by Voxel Size: Underestimation Occurs in Areas with Crossing Fibers. Am J Neuroradiol 28: , Jalal B. Andre, MD Department of Radiology Monmouth Medical Center 300 Second Avenue, Long Branch NJ (SR), USA Tel.: (732) Fax: (732) jalala@gmail.com 361

13