Originally published as JCO Early Release 10.1200/JCO.2003.05.885 on August 11 2003

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Perry, J. R. Articles by Cairncross, J. G. Search for Related Content PubMed PubMed Citation Articles by Perry, J. R. Articles by Cairncross, J. G. Social Bookmarking                            What's this?

Glioma Therapies: How to Tell Which Work?

¹ Department of Medicine, Sunnybrook and Women’s Health Sciences Centre, University of Toronto and Toronto-Sunnybrook Regional Cancer Centre, Toronto, Ontario, Canada
² Department of Clinical Neurosciences, University of Calgary and Tom Baker Cancer Centre, Calgary, Alberta, Canada

THIRTEEN YEARS ago, in the Journal of Clinical Oncology, Macdonald et al¹ recommended imaging-based criteria for response assessment in clinical trials of new therapies for malignant glioma. These criteria were modeled after those used in general oncology, emphasizing the importance of controlling for the beneficial effects of corticosteroids on brain tumor images and symptoms, and defining response as a measurable change in the maximum cross-sectional tumor area. The minimum criterion for response was a 50% decrease in the size of the enhancing tumor on consecutive computed tomography (CT) or magnetic resonance imaging (MRI) scans performed at least 1 month apart. In addition, response to treatment was to be declared only if the corticosteroid dose was stable or reduced and the neurologic condition of the patient was stable or improved. The importance of a proper, corticosteroid-controlled, pretreatment baseline CT or MRI scan was also emphasized by Macdonald et al.¹ When these response criteria were applied rigorously to clinical trials of chemotherapy for anaplastic oligodendroglioma, investigators observed a direct correlation between the completeness of the radiographic response and the duration of tumor control.

The Macdonald criteria have served us well and have been an especially useful guide to response assessment in the context of phase II trials of cytotoxic chemotherapies for recurrent malignant glioma, the clinical setting for which the criteria were designed. They brought a degree of rigor to clinical trials in neuro-oncology, facilitating communication and comparison of results. However, over time, it has become apparent that the Macdonald criteria have certain shortcomings. For example, they cannot be applied to all clinical situations in which investigational therapies for brain tumor need to be tested. The contrast-enhancing component of the tumor that is measured by these criteria does not encompass all of the biologically and clinically relevant disease. Moreover, the Macdonald criteria are wholly reliant on CT and MRI, which are anatomic brain imaging methods that readily detect structural abnormalities, but lack tumor specificity. Brain tumor tissue cannot easily be distinguished from inflammatory changes or other brain pathologies.^2,3 Furthermore, there can be substantial interobserver variability in the interpretation of both CT and MRI.⁴ These limitations were relatively minor issues during the last 13 years, but today, three important trends are forcing neuro-oncologists to reexamine how response should be measured.

First, especially for patients with malignant glioma, there is a desire to be able to assess response to new therapies in a wider range of clinical situations; for example, when an active recurrent tumor cannot be easily measured, or shortly after initial surgery when brain tumor images are difficult to interpret and radiotherapy is concurrently administered. The Macdonald criteria are not useful in either of these common circumstances. To address this shortcoming, clinical investigators have introduced an alternative definition of response for patients with malignant glioma. In some studies, the proportion of patients whose tumors have not progressed 6 months after starting treatment has replaced a 50% decrease in maximum cross-sectional area as the primary response end point. Early experience with this definition is encouraging. Provided clinicians are careful to avoid the many well-known pitfalls in the interpretation of CT and MRI brain scans (eg, response caused by spontaneous resolution of postsurgical changes, progression caused by radiation effects on brain tissue, or control caused by gradual corticosteroid escalation) and are careful to interpret results cautiously, especially when the clinical trial does not include a randomized control group, tumor control at 6 months is probably a meaningful and reproducible measure of the biologic effect of a new therapy for malignant glioma.⁵

Second, some clinicians are now prescribing chemotherapy rather than radiotherapy when surgical management alone is no longer sufficient to control certain types of low-grade glioma, notably oligodendroglioma. Traditionally, radiotherapy has been the first postsurgical treatment for patients with symptomatic, enlarging low-grade gliomas. With initial chemotherapy, radiotherapy can be held in reserve. Because supratentorial low-grade gliomas are usually nonenhancing, ill-defined, and difficult to measure, response to chemotherapy can be difficult to judge using the Macdonald criteria. However, the most important consequence of this change in clinical practice—early chemotherapy for low-grade oligodendroglioma—lies elsewhere. In some cases, neuro-oncologists have documented unequivocal clinical improvement (eg, improved cognition and fewer seizures) that is unaccompanied by a significant reduction in the size of the brain tumor. Moreover, even minor reductions in tumor size may not become apparent until the treatment has been underway for 6 months or longer. These circumstances preclude an early assessment of response and emphasize that clinically important benefits of a brain tumor therapy can sometimes escape detection by either CT or MRI.

Third, the next generation of drug therapies for brain tumor are unlikely to be cytotoxic chemotherapies. Targeted molecules with specific and novel mechanisms of action are envisioned. It is unclear at this time whether targeted agents will reduce the size of the anatomic brain tumor image: Will compounds that block growth signal pathways shrink brain tumors? Experience with imatinib mesylate (Gleevec; Novartis, East Hanover, NJ) for gastrointestinal stromal tumors suggests that some targeted compounds will have beneficial effects that lead to smaller tumors. In all likelihood, however, CT and MRI, by themselves, will be insufficient methods to detect all important brain tumor responses.

The Macdonald criteria took a big step in the right direction by substantially reducing the frequency of false-positive responses in phase II studies of chemotherapy for malignant glioma. The criteria were restrictive, however—a shortcoming that has been addressed by expanding the concept of response to include other measurements of tumor control. Reliance on CT and MRI became a significant issue when it was recognized that false-negative responses might be occurring in the context of successful chemotherapy for low-grade oligodendroglioma. Fortunately, solutions to this additional concern also are beginning to emerge. New MRI sequences⁶ and functional brain imaging methods promise to augment response assessment in neuro-oncology. Indeed, in this issue of the Journal of Clinical Oncology, Vos et al⁷ report that functional imaging of malignant gliomas using thallium-201 single-photon emission computed tomography (²⁰¹Tl-SPECT) is a better early predictor of tumor control and patient survival than either CT or MRI.

In a series of patients with recurrent malignant gliomas treated with paclitaxel or procarbazine, lomustine, and vincristine, Vos et al⁷ found that maximal tumor intensity on the baseline and posttreatment ²⁰¹Tl-SPECT scans was a better predictor of both the percentage of patients who were alive and progression free at 6 months and overall survival time than was CT or MRI. Lessening of the maximal tumor intensity after two cycles of chemotherapy (response) was associated with significantly better tumor control and longer patient survival. Such was not the case for CT or MRI, for which tumor regression did not correlate with either of these measures of improved patient outcome. Increasing tumor intensity on ²⁰¹Tl-SPECT (progression) was found to be a powerful predictor of poor patient outcome. This association was also true for CT or MRI.

Vos et al⁷ have reaffirmed than poor outcome is easy to predict in neuro-oncology. Ominous prognostic factors include glioblastoma pathology, nonseizure symptoms, older age, poor performance status at diagnosis, extensive residual tumor after surgery, corticosteroid dependence after radiotherapy, and evidence of early progression on follow-up CT or MRI. To this long list, Vos et al⁷ have added ²⁰¹Tl-SPECT activity. In support of ²⁰¹Tl-SPECT scanning, they emphasize the importance of being able to discontinue ineffective therapies at the earliest possible moment, especially when such treatments also have toxic effects. Although this assertion is true, what neuro-oncologists really need for their patients are new effective therapies for malignant glioma, not additional refinements of an already robust collection of predictors of poor outcome. This cynical view is tempered, however, by a key finding in the study by Vos et al⁷: unlike CT or MRI, favorable changes in the ²⁰¹Tl-SPECT scan, a metabolic brain image, were associated with significantly better patient outcomes. This finding alone should prompt additional evaluation of ²⁰¹Tl-SPECT and other functional methods of brain imaging, including positron emission tomography and magnetic resonance spectroscopy, in the assessment of response to new therapies for brain tumor. In this regard, it might be especially instructive to analyze functional imaging changes in anaplastic oligodendrogliomas that harbor allelic loss of chromosomal arms 1p and 19q. Unlike glioblastomas, these malignant gliomas demonstrate response by traditional CT or MRI criteria. Experience with responding patients may facilitate the interpretation of posttreatment ²⁰¹Tl-SPECT changes in patients with refractory tumors, such as glioblastoma.

This reference to the lessons that might be learned from studying oligodendrogliomas leads us to emphasize one final point. As discussed recently in the Journal of Clinical Oncology by Betensky et al,⁸ an additional shift in the traditional paradigm of response-based drug assessment will be required. The fact that histologically indistinguishable oligodendrogliomas can have distinct genetic signatures that correlate with different natural histories and responses to treatment means that a new therapy for this type of cancer might be curative in 10% of patients, whereas it might be totally ineffective for 90% of patients. When this new reality is compounded by the unavoidable inaccuracies in clinical response measurement, the risk of overlooking an effective therapy for a small subset of patients is high. We believe, therefore, like others,^9,10 that response criteria for brain tumor, however defined, must be applied in concert with molecular characterization of tumor tissue. Without this link, improvements in patient care will continue to evolve slowly, despite rapid progress in the understanding of cancer.

AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The authors indicated no potential conflicts of interest.

REFERENCES

1. Macdonald DR, Cascino TL, Schold SC Jr, et al: Response criteria for phase II studies of supratentorial malignant glioma. J Clin Oncol 8:1277–1280, 1990[Abstract]

2. Cairncross JG, Pexman JHW, Rathbone MP, et al: Postoperative contrast enhancement in patients with brain tumor. Ann Neurol 17:570–572, 1985[CrossRef][Medline]

3. Forsyth PAJ, Petrov E, Mahallati H, et al: Prospective study of postoperative magnetic resonance imaging in patients with malignant gliomas. J Clin Oncol 15:2076–2081, 1997[Abstract/Free Full Text]

4. Vos MJ, Uitdehaag BMJ, Barkhof F, et al: Interobserver variability in the radiological assessment of response to chemotherapy in glioma. Neurology 60:826–830, 2003[Abstract/Free Full Text]

5. Perry JR, DeAngelis LM, Schold SC Jr et al: Challenges in the design and conduct of phase III brain tumor therapy trials. Neurology 49:912–917, 1997[Free Full Text]

6. Mardor Y, Pfeffer R, Spiegelmann R, et al: Early detection of response to radiation therapy in patients with brain malignancies using conventional and high b-value diffusion-weighted magnetic resonance imaging. J Clin Oncol 21:1094–1100, 2003[Abstract/Free Full Text]

7. Vos MJ, Hoekstra OS, Barkhof F, et al: Thallium-201 single-photon emission computed tomography as early predictor of outcome in recurrent glioma. J Clin Oncol 21:3559–3565, 2003[Abstract/Free Full Text]

8. Betensky RA, Louis DN, Cairncross JG: Influence of unrecognized molecular heterogeneity on randomized clinical trials. J Clin Oncol 20:2495–2499, 2002[Abstract/Free Full Text]

9. Eisenhauer EA: Phase I and II trials of novel anti-cancer agents:Endpoints, efficacy, and existentialism. Ann Oncol 8:1047–1052, 1998

10. Lang FF, Gilbert MR, Puduvalli VK, et al: Toward better early-phase brain tumor clinical trials: A reappraisal of current methods and proposals for future strategies. Neuro-oncol 4:268–277, 2002[Abstract]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				A. Hormigo, B. Gu, S. Karimi, E. Riedel, K. S. Panageas, M. A. Edgar, M. K. Tanwar, J. S. Rao, M. Fleisher, L. M. DeAngelis, et al. YKL-40 and Matrix Metalloproteinase-9 as Potential Serum Biomarkers for Patients with High-Grade Gliomas. Clin. Cancer Res., October 1, 2006; 12(19): 5698 - 5704. [Abstract] [Full Text] [PDF]

				C. Walker, B. Haylock, D. Husband, K. A. Joyce, D. Fildes, M. D. Jenkinson, T. Smith, J. Broome, D. G. du Plessis, and P. C. Warnke Clinical use of genotype to predict chemosensitivity in oligodendroglial tumors Neurology, June 13, 2006; 66(11): 1661 - 1667. [Abstract] [Full Text] [PDF]

				C. Walker, D. G. du Plessis, D. Fildes, B. Haylock, D. Husband, M. D. Jenkinson, K. A. Joyce, J. Broome, K. Kopitski, J. Prosser, et al. Correlation of Molecular Genetics with Molecular and Morphological Imaging in Gliomas with an Oligodendroglial Component Clin. Cancer Res., November 1, 2004; 10(21): 7182 - 7191. [Abstract] [Full Text] [PDF]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Perry, J. R. Articles by Cairncross, J. G. Search for Related Content PubMed PubMed Citation Articles by Perry, J. R. Articles by Cairncross, J. G. Social Bookmarking                            What's this?