Abstract
Context.—Significant interobserver variability exists with respect to the diagnosis of oligodendroglial neoplasms, especially their distinction from astrocytoma and mixed oligoastrocytoma. Combined loss of the short arm of chromosome 1 and the long arm of chromosome 19 has been shown to be both relatively specific to oligodendroglioma and, when present, a marker of improved prognosis in patients with these tumors. In addition, 1p/19q loss has been shown to be a marker of “classic” oligodendroglial histology. These findings raise questions as to the role of 1p/19q testing in clinical practice, both as a prognostic marker and as a potential diagnostic marker among infiltrating glial neoplasms.
Objective.—This review discusses the issues raised above and tries to clarify the current status of 1p/19q evaluation in the diagnosis of oligodendroglioma.
Data Sources.—Sources for this review include recent literature as well as the experience of 3 practicing neuropathologists.
Conclusions.—1p/19q status is an important marker in oligodendroglioma. Loss of 1p/19q is associated with classic oligodendroglioma histology as well as improved prognosis. The combined 1p/19q marker will continue to be a clinically useful marker of prognosis and could potentially be incorporated into diagnostic criteria in the future.
HISTORICAL PERSPECTIVE
Oligodendrogliomas are diffusely infiltrating gliomas representing the second most common primary parenchymal brain tumor after astrocytomas/glioblastomas. They were first described by Bailey and Bucy,1 who postulated a histogenetic link between these tumors and oligodendrocytes. Subsequently, anecdotal studies in the late 1980s and early 1990s using the chemotherapy regimen of procarbazine, lomustine/CCNU, and vincristine (PCV) indicated that, compared with their astrocytic counterparts, anaplastic oligodendrogliomas are much more likely to respond favorably2,3; this was subsequently confirmed in a small clinical trial.4 Concurrently, molecular studies from multiple laboratories showed that the combined loss of the short arm of chromosome 1 (1p) and the long arm of chromosome 19 (19q) was a distinct genetic pattern encountered in a large subset of cases diagnosed as oligodendroglioma. A landmark study of 35 anaplastic oligodendroglioma patients published in 1998 provided the link between molecular genetics and clinical outcome by showing that those patients whose tumors exhibited 1p/19q loss experienced favorable therapeutic response and longer survival times than patients whose tumors lacked this change.5 Since that time, there has been great interest in the relationships among molecular genetics, histologic diagnosis, and clinical outcome in subsets of diffuse glioma. The powerful prognostic/predictive value of 1p/19q status has led to the development of clinical testing at a number of centers. Initially associated with response to a specific chemotherapy regimen, recent data suggest that oligodendroglioma is a tumor that is more responsive than astrocytoma to a variety of current therapeutic modalities in general. These recent findings raise the issue as to whether oligodendrogliomas are in general more responsive to treatment as compared to exhibiting an overall lower biologic potential when compared with other diffuse gliomas. This review summarizes the current literature on the topic and provides a perspective on the diagnosis of oligodendroglioma and the clinical relevance of molecular genetic testing of these tumors.
HISTOLOGIC FEATURES OF OLIGODENDROGLIOMAS
Oligodendroglioma (World Health Organization [WHO] grade 26) occurs in the white matter and cortex of the cerebral hemispheres and shows a monotonous pattern on low power with occasional “clonal” nodules of higher cellularity. The nuclei are round and regular, and clear perinuclear haloes are present in most paraffin-embedded specimens. Nevertheless, it is important to realize that this typical “fried egg” appearance is a formalin fixation artifact and is therefore not seen in frozen sections, smears, or rapidly fixed specimens. Therefore, the overall architecture (monomorphism or lack thereof) as well as nuclear cytology are the 2 most important features for diagnosing oligodendroglioma. Cells with identical nuclear features and small rounded bellies of eosinophilic cytoplasm are known as minigemistrocytes or microgemistocytes and are considered part of the morphologic spectrum of oligodendroglioma. The WHO defines anaplastic (WHO grade 3) as having a slight increase in nuclear variability, although the nuclei remain largely round to oval and retain an overall monotony without displaying the greater pleomorphism and atypia of astrocytic tumors. Often, the nuclei become larger and more vesicular with increased prominence of nucleoli. Accumulation of amphophilic cytoplasm may impart an epithelioid or plasmacytoid appearance. Anaplastic oligodendrogliomas usually show increased mitotic activity and/or endothelial hyperplasia (microvascular proliferation), sometimes retaining the overall chicken-wire pattern, or vessels may appear glomeruloid. In the WHO scheme, a specific mitotic cutoff for anaplasia was not provided7; however, it has been suggested that the presence of at least 6 mitoses per 10 high-power microscopic fields, even if focal, predicts a reduced survival time.8 Tumor necrosis may also be present.
While morphologically classic grade 2 and grade 3 oligodendrogliomas are relatively straightforward diagnostically with little interobserver variability, the working neuropathologist also commonly encounters more ambiguous tumors with features intermediate between typical oligodendroglioma and typical astrocytoma. In such cases, some but not all features of classic oligodendroglioma may be present. Alternatively, classic features may be present in a subset of cells, but they are intermingled with tumor cells showing typical astrocytic features. A rare case may exhibit separate, discrete areas of classic oligodendroglioma and classic astrocytoma. There is a lack of consensus on the optimal diagnostic approach in these diagnostically challenging cases. The current WHO criteria recognize the entity mixed oligoastrocytoma (MOA), which is the term often applied to such lesions. Nevertheless, whether MOA truly represents a single, definable clinicopathologic entity or whether it simply reflects morphologic ambiguity in otherwise pure tumors remains unclear. Additional investigation into the molecular genetic signatures that characterize the low-grade to intermediate-grade astrocytomas, oligodendrogliomas, and morphologically ambiguous lesions will be required to resolve this controversial issue.
LOSS OF 1p AND 19q IN OLIGODENDROGLIOMAS
The molecular genetic signature of oligodendrogliomas at the DNA level is combined 1p and 19q losses, typically involving the entire chromosomal arm at both sites. While isolated 19q loss occurs in astrocytoma and MOA, combined 1p/19q loss (especially loss that involves the entire arms) is rare in gliomas other than oligodendroglioma. Combined 1p/19q loss has been reported in 50% to 80% of cases carrying a diagnosis of oligodendroglioma, with frequencies as high as 90% when using strict diagnostic criteria. The incidence of 1p/19q loss is considerably lower (20%–30%) in cases carrying an MOA diagnosis as well as diffuse gliomas with ambiguous morphologies in general.9–12 Chromosome 1p and 19q status has been assessed using a variety of molecular techniques, including loss of heterozygosity,13–15 comparative genomic hybridization,13,16–19 quantitative microsatellite analysis,20 and fluorescence in situ hybridization.21,22 Furthermore, while loss of heterozygosity studies have traditionally required comparisons of allele dosages in the tumor versus constitutional DNA, this has been modified recently with real time quantitative polymerase chain reaction for increased applicability to routine clinical practice,20 where constitutional DNA is not always readily available. Examples of fluorescence in situ hybridization in clinical samples are shown in Figure 1, A through F. These molecular diagnostic techniques have also been used in attempts to identify the presumed tumor suppressor genes that reside on 1p and 19q. Unfortunately, this search has been hampered by the fact that most oligodendrogliomas lose the entire arm of each chromosome, rendering them uninformative for positional cloning strategies attempting to define a minimal deletion region. Given the rarity of bona fide oligodendrogliomas that show partial 1p and/or 19q loss, astrocytic tumors (which show frequent 19q loss and occasional 1p loss) were included in the search because they more frequently show partial loss of these arms. However, it is not clear whether the minimal deletion regions on 1p and 19q defined by nonoligodendroglial tumors will actually lead to the identification of classic oligodendroglioma-associated tumor suppressor genes. Oligodendrogliomas tend to occur in adults and are rare in children. Interestingly, pediatric oligodendrogliomas rarely exhibit combined 1p/ 19q loss, and those 1p/19q-loss oligodendrogliomas that do occur in children do not seem to confer a favorable prognosis.23–25
Fluorescence in situ hybridization to assess 1p/19q status in gliomas. A probe for 1p32 (green) is paired with 1q42 (red) (A, C, and E), and 19q13.4 (red) is paired with 19p13 (green) (B, D, and F). Examples are as follows: A, 1p deleted; B, 19q deleted; C, 1p intact; D, 19q intact; E and F, 1p and 19q deleted with near-tetraploid cells (note relative loss of green [E] and red [F] signals, respectively) (fluorescent probe–stained cells, original magnification ×1000)
Fluorescence in situ hybridization to assess 1p/19q status in gliomas. A probe for 1p32 (green) is paired with 1q42 (red) (A, C, and E), and 19q13.4 (red) is paired with 19p13 (green) (B, D, and F). Examples are as follows: A, 1p deleted; B, 19q deleted; C, 1p intact; D, 19q intact; E and F, 1p and 19q deleted with near-tetraploid cells (note relative loss of green [E] and red [F] signals, respectively) (fluorescent probe–stained cells, original magnification ×1000)
Loss of 1p/19q is also associated with other recurring changes in a nonrandom fashion. Mutations in TP53, common in astrocytomas, are inversely associated with 1p/ 19q loss.24,26–28 Loss of chromosome 10q, especially common in glioblastoma, is also inversely associated with 1p/ 19q loss,26,29–31 as is loss of 9p and amplification of epidermal growth factor receptor.27,28,32 In fact, epidermal growth factor receptor amplification and 10q deletions are sufficiently rare in oligodendroglial tumors that they should prompt the consideration of alternate diagnoses, such as small cell glioblastoma, an astrocytoma variant with features that significantly overlap with those of anaplastic oligodendroglioma.29 Detailed comments regarding markers that distinguish astrocytoma from oligodendroglioma are beyond the scope of this article, and the reader is referred to a recent review on the topic.33 Recently, it has also been suggested that 1p/19q loss is associated with lower expression of MGMT.8,34–36 A recent study comparing MGMT promoter methylation status with 1p/19q status in oligodendrogliomas found a significant positive correlation of promoter hypermethylation with 1p/19q loss,36 which is of interest given the recent finding that MGMT promoter methylation is associated with chemosensitivity in glioblastoma.35
1p/19q AS A DIAGNOSTIC MARKER
The fact that combined 1p/19q loss among gliomas is fairly specific for oligodendroglial tumors indicates its utility as a diagnostic marker. In that regard, it can serve as an adjunct to histology and may serve as a point of reference for glioma diagnosis. Neuroepithelial tumors that may mimic oligodendroglioma on a morphologic basis are shown in the Table. With respect to the differential diagnosis of oligodendroglioma, in daily practice the most important distinction is that of diffuse astrocytoma. Features that distinguish these entities are listed in the Table. Features of classic oligodendroglioma (Figure 2, A through H) and typical astrocytoma (Figure 3, A through D) are shown as examples. As discussed above, a fraction of grade 2 to 3 diffuse gliomas show features intermediate along the continuum between oligodendroglioma and astrocytoma (Figure 3, E through H). In our experience it is likely that the tumors shown in Figure 3, E through H would receive nonconcordant diagnoses when viewed by a group of neuropathologists.
Examples of classic oligodendroglioma. Oligodendroglial neoplasms showing nuclear roundness/uniformity, perinuclear haloes, and architectural features such as chicken-wire vasculature and perineuronal satellitosis are shown in A through H (hematoxylin-eosin–stained sections, original magnification ×400)
Examples of classic oligodendroglioma. Oligodendroglial neoplasms showing nuclear roundness/uniformity, perinuclear haloes, and architectural features such as chicken-wire vasculature and perineuronal satellitosis are shown in A through H (hematoxylin-eosin–stained sections, original magnification ×400)
Examples of astrocytic and morphologically ambiguous infiltrating glial neoplasms. Typical astrocytic tumors (A through D) show nuclear pleomorphism and fibrillary and/or gemistocytic background. E through H show morphologically ambiguous gliomas. Some oligodendroglial features (eg, nuclear roundness or perinuclear haloes) are present, but the tumors lack monomorphism and in our experience are difficult to classify (hematoxylin-eosin–stained sections, original magnification ×400)
Examples of astrocytic and morphologically ambiguous infiltrating glial neoplasms. Typical astrocytic tumors (A through D) show nuclear pleomorphism and fibrillary and/or gemistocytic background. E through H show morphologically ambiguous gliomas. Some oligodendroglial features (eg, nuclear roundness or perinuclear haloes) are present, but the tumors lack monomorphism and in our experience are difficult to classify (hematoxylin-eosin–stained sections, original magnification ×400)
The minimal diagnostic criteria among neuropathologists for acceptance as oligodendroglioma among grade 2 to 3 diffuse gliomas ranges from “relaxed” to “strict.” In a relaxed approach, a tumor that exhibits 1 or 2 features of oligodendroglioma (eg, perinuclear haloes and a subset of cells with rounded nuclei) while at the same time exhibiting other, more astrocytic features (eg, nuclear pleomorphism) is an oligodendroglioma. At the other end of the spectrum is a strict approach, which limits the diagnosis of oligodendroglioma only to those cases that exhibit most or all of the classic features. The unfortunate result of varying approaches is that for many diffuse grade 2 to 3 gliomas, subtle oligodendroglial features exist in the eye of the beholder, and one person's oligodendroglioma may be another person's astrocytoma. Adding to this concern is the variable use of the term MOA, either to reflect the morphologic ambiguity or to convey prognostic optimism to the patient and treating physician and perhaps leave open the door to chemotherapy, to which oligodendrogliomas were initially considered to be particularly responsive. Such considerations have led to the suggestion that the lack of reproducibility in the classification of many grade 2 to 3 diffuse gliomas presents an opportunity where genetics may be used as a key diagnostic feature, because genetic signatures are in general more objective. In this approach such morphologically ambiguous diffuse gliomas would be classified not according to the subjectivity/individual style of the pathologist but rather according to 1p/19q status. Nonetheless, even this approach is partially flawed because not all classic oligodendrogliomas harbor this genetic signature and a smaller subset of diffuse gliomas lacking these deletions nevertheless follows a favorable clinical course. As such, additional biomarkers are clearly needed (discussed further below).
Small cell variants of anaplastic astrocytoma and glioblastoma are an important differential diagnostic consideration because these tumors share several morphologic features with oligodendroglioma, including monotony, cytologic blandness, and calcifications. The nuclear features are perhaps the most important morphologic consideration in distinguishing these entities, which is clinically important given the glioblastoma-like aggressive behavior of small cell astrocytoma. Here again, a potential role for ancillary genetic studies has been highlighted by recent findings that small cell astrocytomas do not exhibit 1p/ 19q loss but nearly always have 10q loss and frequently show epidermal growth factor receptor amplification.36
Additional diagnostic considerations include dysembryoplastic neuroepithelial tumor (DNET), neurocytoma, and clear cell ependymoma. Distinguishing oligodendroglioma from DNET may also be difficult and is similarly critical, given that DNET is a benign (WHO grade 1) lesion requiring no additional treatment beyond surgery, whereas even the grade 2 oligodendrogliomas progress and are eventually fatal in most cases. Although DNET is generally limited to the cortex and displays characteristic patterned mucin-rich nodules and “floating neurons,” the oligodendroglial element is cytologically identical to oligodendroglioma and the distinction between these 2 entities can be difficult to impossible in incomplete or fragmented samples. In that regard, it has been shown that DNET does not exhibit 1p/19q loss,33,37,38 indicating the usefulness of a positive 1p/19q test in this situation. Since many patients with DNET are children, and childhood oligodendroglioma also does not frequently show combined 1p/ 19q loss, it should be emphasized that a negative test (ie, a finding that the tumor is 1p/19q intact) in a pediatric patient is not helpful.
Neurocytoma is usually distinguishable from oligodendroglioma based on the location (usually intraventricular in neurocytomas). However, extraventricular neurocytoma is a rare consideration that may be difficult to distinguish from oligodendroglioma. In contrast to the latter, neurocytomas are typically demarcated, show rosettes and/or neuropil islands, and display diffuse synaptophysin immunoreactivity. Adding to the complexity, however, is the recent finding that a subset of oligodendrogliomas exhibits neurocytic differentiation, suggesting that these 2 entities may be more closely related than previously appreciated.39 In fact, some degree of neuronal differentiation may be a more general phenomenon than once believed, especially in 1p/19q-loss oligodendrogliomas.40 In contrast, it has been shown that central neurocytomas (as well as clear cell ependymomas) do not harbor 1p/19q loss.41
Together, these data demonstrate the utility of 1p/19q status as a diagnostic marker in selected cases. Additionally, it can be considered as a point of reference with which to measure diagnostic criteria. Since a major goal of classification of grade 2 to 3 gliomas is to stratify patients into clinically distinct groups, it would be prudent to ensure that the majority of cases diagnosed as oligodendroglioma in daily practice exhibit combined 1p/19q loss because this signature is associated with both classic oligodendroglial features and improved outcome.
The use of 1p/19q may also be helpful as a marker in tumors at recurrence. A recent study compared 1p/19q status in longitudinal samples from 31 patients. In general, combined 1p/19q was a stable genetic change, concordant in the initial versus recurrent samples.42 Moreover, it has been shown to represent a prognostically useful marker even at the time of recurrence; a recent study of recurrent, mostly anaplastic oligodendrogliomas found median survival times after recurrence of 3.9 versus 1.0 years, with and without deletions respectively (P < .001).39,43 This can be particularly clinically useful because of the tendency of oligodendroglioma to lose oligodendroglial features (ie, become more “astrocytic”) after therapy. In a situation where the primary tumor is not available for review, 1p/ 19q status on the recurrent specimen could provide supportive diagnostic data.
1p/19q AS A PROGNOSTIC/PREDICTIVE MARKER
Initial results suggested that 1p/19q status is a predictive (in addition to prognostic) marker of response to PCV chemotherapy.44 Subsequent experience indicates that the therapeutic benefit of 1p/19q loss is not restricted to patients receiving PCV chemotherapy. For example, recent data suggest that patients with such tumors also respond favorably to the less toxic agent temozolomide.5,44 However, the most definitive study includes results from a phase III clinical trial from the Radiation Therapy Oncology Group (RTOG 94-02), which compared radiation therapy alone versus chemotherapy plus radiation therapy. Of the 291 enrolled patients, 1p/19q status was determined for 206. In both study arms, patients whose tumors exhibited 1p/19q loss survived longer than patients with 1p/ 19q-intact tumors,45 suggesting that the therapeutic sensitivity of 1p/19q-loss oligodendroglioma is not restricted to chemotherapy but also extends to radiosensitivity. Other studies have shown that patients with 1p/19q-loss oligodendrogliomas have improved progression-free survival and their tumors have more indolent behavior even before the initiation of treatment.46,47 Therefore, it seems that patients with the “genetically favorable” form of oligodendroglioma are likely to enjoy a more prolonged survival, almost regardless of the therapeutic approach that is selected. As suggested earlier, it is not clear whether oligodendrogliomas with 1p/19q loss represent a tumor type that is more responsive to cytotoxic therapies or whether these tumors are more biologically indolent. Nevertheless, the results of the Radiation Therapy Oncology Group study clearly show that 1p/19q loss is an independent prognostic marker in anaplastic oligodendroglioma. It has also been suggested that patients with 1p/19q-deleted oligodendroglioma are more likely to present with seizures, compared to patients with 1p/19q-intact tumors, who are more likely to present with a focal neurologic deficit.47
1p/19q AND IMAGING FEATURES
The status of 1p/19q is also statistically associated with location in the cerebrum. One study compared 1p/19q status with tumor location (temporal vs nontemporal) in 203 astrocytomas, oligoastrocytomas, and oligodendrogliomas. Tumors located in the temporal lobe were significantly less likely to harbor 1p/19q loss than were those in nontemporal locations.48 Imaging parameters also vary based on 1p/19q status. Loss of 1p/19q has been associated with an indistinct border of the mass lesion with brain parenchyma, as well as mixed intensity signals on both T1- and T2-weighted imaging.49 A molecular imaging study has shown that 1p/19q-loss oligodendroglioma shows increased [18F]fluorodeoxyglucose uptake, as well as evidence of increased metabolism compared to 1p/19q-intact tumors.
1p/19q AND HISTOLOGIC FEATURES
The status of 1p/19q is significantly associated with histologic features in oligodendroglioma. One study found 1p/19q loss in 19 of 22 cases of “classic” oligodendroglioma, while only 6 of 22 nonclassic tumors had this pattern.49 Studies employing experienced neuropathologists find a positive association between a consensus diagnosis of classic oligodendroglioma and the presence of 1p/19q loss.50,51 However, it is important to recognize that the correlation of histology and genetics, while statistically significant, is not one-to-one. The example of pediatric oligodendrogliomas was mentioned earlier. In addition, 10% to 20% of classic oligodendrogliomas are intact for 1p/ 19q. Similarly, nonclassic tumors occasionally exhibit 1p/ 19q loss. A recent study, which included a total of 131 anaplastic oligodendroglioma tumor samples, compared consensus review from 5 neuropathologists with 1p/19q status and patient survival. Tumors were dichotomized by histology into classic and nonclassic categories based on the majority opinion. Figure 4, A, shows a positive relationship of the opinion of 5 neuropathologists as to the presence of classic oligodendroglial features, compared to the likelihood of 1p/19q loss. When cases were condensed into 3 groups, based on the number of pathologists who considered the case as classic, a significant survival difference was seen between the cases scored as classic by 4 of 5 or 5 of 5 pathologists, as compared with the cases considered classic by 3 or fewer neuropathologists (Figure 4, B). From a patient survival standpoint, it appears that only the most classic oligodendroglial neoplasms exhibit favorable survival, which may tend to argue for the use of strict criteria in daily practice if the goal is to stratify patients into clinically distinct groups. Although 1p/19q status is a predictor of survival in anaplastic oligodendroglioma as a whole, it did not predict improved survival in the subset of tumors with nonclassic histology.52 Although this finding awaits independent confirmation, it has potential implications for whether case selection (on the basis of histologic features) should be used in clinical 1p/19q testing. On the other hand, these deletions are considerably less common in the nonclassic examples and therefore, larger cohorts may be needed to achieve statistical significance. For example, in a recent study of more than 900 high-grade glioma patients, one of us (A.P., unpublished data, 2006) found that 1p/19q codeletion by fluorescence in situ hybridization was an independent prognostic variable on multivariate analysis, even in the group of mixed oligoastrocytomas. On the other hand, A.P. has encountered these codeletions in <1% of classic astrocytomas, suggesting that the yield may be too small for cost efficiency in routine clinical testing of such cases. A retrospective study of 90 morphologically ambiguous infiltrative gliomas revealed an incidence of 1p/19q loss in only 9%, again suggesting that on a practical basis, clinical 1p/19q testing should be limited to cases exhibiting oligodendroglial histologic features.
Oligodendroglial histologic features compared with 1p/19q loss and overall survival. A, Relationship of histology review score and probability of 1p/19q loss. Anaplastic oligodendroglial tumors were reviewed by 5 neuropathologists who determined whether each case exhibited classic oligodendroglial histology. Each case was given a score (from 0 to 5) based on the number of pathologists who scored the case as classic. The review scores are plotted as a function of the proportion of cases showing loss of 1p/19q. The curve based upon the estimated proportion (solid line) as determined by logistic regression, along with 95% confidence intervals (dashed lines), is shown. B, Survival according to histology. The 6 possible review scores (ranging from 0 to 5) were condensed into 3 groups, and cases were plotted according to overall survival. There is a statistically significant difference (P < .01 log-rank test) according to the number of neuropathologists who considered the case classic. This clinical correlation would tend to support the use of strict criteria for the diagnosis of oligodendroglioma
Oligodendroglial histologic features compared with 1p/19q loss and overall survival. A, Relationship of histology review score and probability of 1p/19q loss. Anaplastic oligodendroglial tumors were reviewed by 5 neuropathologists who determined whether each case exhibited classic oligodendroglial histology. Each case was given a score (from 0 to 5) based on the number of pathologists who scored the case as classic. The review scores are plotted as a function of the proportion of cases showing loss of 1p/19q. The curve based upon the estimated proportion (solid line) as determined by logistic regression, along with 95% confidence intervals (dashed lines), is shown. B, Survival according to histology. The 6 possible review scores (ranging from 0 to 5) were condensed into 3 groups, and cases were plotted according to overall survival. There is a statistically significant difference (P < .01 log-rank test) according to the number of neuropathologists who considered the case classic. This clinical correlation would tend to support the use of strict criteria for the diagnosis of oligodendroglioma
1p/19q TESTING IN CLINICAL PRACTICE
A number of centers, including those of the authors, offer the evaluation of 1p/19q status as a clinical test. The methods range from fluorescence in situ hybridization to allelic imbalance to quantitative polymerase chain reaction studies. Each method is relatively reliable, and each has its own distinct advantages and disadvantages. Most available tests rely on paraffin sections or on DNA extracted from paraffin sections, enabling them to be performed on gliomas encountered in daily practice. Currently, no firm guidelines exist with respect to the selection of cases for 1p/19q testing, but it is likely that such guidelines will be developed in the future as more experience is gained. In any event, evaluation of 1p/19q status in the context of characteristic clinical imaging and histologic features seems prudent, given the clinical variability of patients with oligodendroglioma. The identification of additional biomarkers, which account for the even greater variability of other diffuse glioma subtypes, awaits further investigation.
UNANSWERED QUESTIONS
Several important but as yet unresolved issues remain with respect to 1p/19q loss in oligodendroglioma:
First, the presumed tumor suppressor genes whose losses promote oligodendroglial pathogenesis have yet to be identified. Finding the genes critical to oligodendroglioma pathogenesis may lead to therapeutic targets to better treat this lethal disease. Related to this is the issue of whether the loss of genetic material on chromosomes 1p/19q somehow predisposes to therapeutic sensitivity, as opposed to merely representing a genetic marker for a tumor that happens to be more sensitive for other reasons.
Second, what is an MOA? It is currently unclear whether this entity actually exists on a genetic level as distinct from either astrocytoma or oligodendroglioma. It is apparent from studies using microdissected tissue that both components of the MOA appear to be part of the same neoplastic process (ie, show similar genetic alterations).53 Perhaps more likely the term MOA is a reflection of the limitations of morphology alone to accurately classify the subset of nonclassic diffuse gliomas and highlights the need for molecular diagnostics as an adjunct to diagnosis. Future consensus conferences should take a critical look as to the diagnostic criteria for MOA and whether the inclusion of this entity in WHO criteria makes sense from either a practical or conceptual point of view.
Third, and related to the previous issue, is the interobserver variation in standards used for the diagnosis of oligodendroglioma, ranging from relaxed criteria to strict criteria. Issues 2 and 3 bring up the question of whether, in theory, a better classification scheme for grade 2 to 3 diffuse gliomas should be based heavily on 1p/19q status. The conundrum facing the neuro-oncology community is that while our grading scheme is currently based on morphology alone overall, 1p/19q status may prove to be a better predictor of outcome than morphologic interpretation among grade 3 gliomas. If a molecular criterion were adapted in the future, most of these diffuse gliomas could be neatly and reproducibly stratified into 1p/19q-loss vs 1p/19q-intact groups, with the clear advantages of reduced subjectivity and interobserver variability. In support of this scheme is the close relationship of combined 1p/19q loss with both classic oligodendroglioma histology and improved outcome. However, this may not represent the best solution for a number of reasons already discussed, as well as practical limitations of molecular testing in pathology laboratories. At minimum, however, a careful reassessment of WHO criteria, perhaps with more explicit standards as to the strictness of criteria for distinguishing oligodendroglioma from astrocytoma, is certainly warranted to enhance diagnostic reproducibility among pathologists. A diagnostic framework that supposes that “an infiltrating glioma is astrocytic until proven otherwise” may be the most prudent approach because it is currently supported by the existing data and also would have the secondary benefit of reducing interobserver variability in the classification of morphologically ambiguous grade 2 to 3 gliomas.
Fourth, since some patients with 1p/19q-loss tumors respond poorly to therapy (and some patients with 1p/ 19q-intact tumors respond well), additional biomarkers are needed to further enhance predictive accuracy. Studies incorporating gene expression microarray analysis32 represent a promising approach toward achieving this goal, and additional studies will likely help in this regard.
Fifth, although 1p/19q status provides the opportunity for patient selection with respect to therapeutic regimen, it is important to realize that no dramatic improvements in outcome have been seen since the initial finding that oligodendroglioma is more sensitive than astrocytoma to therapy. Finally, the robust relationship of 1p/19q status and outcome among oligodendrogliomas raises notable questions regarding standards of care in clinical practice; although molecular testing is becoming increasingly common, the majority of pathology laboratories still neither offer it locally nor send their cases out to larger reference laboratories.
CONCLUSIONS
The status of chromosomes 1p and 19q represents a robust marker of outcome in gliomas, especially in patients with anaplastic (WHO grade 3) oligodendrogliomas. It appears to be an overall prognostic marker, since 1p/19q loss is associated with improved outcome regardless of the specific therapeutic regimen. Testing for 1p/19q status in the clinical setting appears to be most useful in 2 situations. The most common is that of a tumor that appears as classic oligodendroglioma, where 1p/19q status is used as a prognostic marker and a potential guide to patient management. Second is the diagnostic utility for cases where a histologic mimic of oligodendroglioma or a morphologically ambiguous tumor is considered. Although the molecular targets of 1p/19q losses are unknown, as are the mechanistic relationships of 1p/19q loss with therapeutic sensitivity, 1p/19q status indeed serves as a marker that appears to be of clinical utility. Additional refinement of diagnostic criteria for oligodendroglioma, with reference to 1p/19q status, will improve the current lack of uniformity of standards for the diagnosis of this entity. Third, whether anaplastic diffuse gliomas should be defined primarily on the basis of genetics is a topic of active discussion and will likely be addressed more fully in the future. Finally, further work will surely identify additional clinically relevant genetic alterations that will allow for further refinement of the prognostic/predictive power of 1p/19q status in oligodendroglioma. In addition, 1p/19q status in oligodendroglioma will continue to serve as a useful paradigm for the use of molecular signatures to supplement clinicopathologic data in the diagnosis and management of human gliomas.
References
The authors have no relevant financial interest in the products or companies described in this article.
Author notes
Reprints: Kenneth Aldape, MD, M. D. Anderson Cancer Center, Department of Pathology and Brain Tumor Center, 1515 Holcombe Blvd, Houston, TX 60153 ([email protected])