Low-grade Astrocytoma

Recent advances in neurodiagnostic technology have caused a resurgence of interest in the question of what constitutes the optimal treatment of the cerebral astrocytoma. When isotope brain scans and cerebral angiography were the mainstays of brain tumor detection, the majority of patients were diagnosed late in the course of their disease and presented with large, high-grade lesions. The treatment of anaplastic gliomas and glioblastoma multiforme has consequently received most attention during the past several decades. With the advent of computed tomography (CT) and magnetic resonance imaging (MRI), however, we are now detecting smaller, less malignant lesions in younger patients. By the use of new techniques such as stereotactic biopsy, we have been able to obtain pathologic confirmation of tumor foci in many more instances than was possible previously. When the pathologic analysis indicates that a low-grade astrocytoma is present, the question arises as to what constitutes the best available therapy. By this, we mean that therapy which offers the longest and best quality of survival with the lowest risk of side effects. The latter issue is especially important, because these tumors affect patients who are relatively young and thus who can be expected to have a fairly long survival.

These tumors have been called astrocytomas in the threetier World Health Organization classification. They correspond to the grade I and grade II astrocytomas of the Kernohan classification, or what have been called "low-grade" astrocytomas in the past. Under the more recent classification schema proposed by Daumas-Duport et al., these tumors would be called grade 2 astrocytomas. This group includes lesions that previously were classified descriptively by such terms as fibrillary and protoplasmic.

Some types of astrocytomas would appear to have a well-defined prognosis that is unique unto themselves and therefore will not be included in this review. Specifically, the "gemistocytic" astrocytoma seems to have a high incidence of conversion into more malignant forms and thus a worse prognosis than the other low-grade astrocytomas; conversely, the "pilocytic " astrocytoma appears to have an excellent prognosis no matter how radical the surgical resection or the type of postoperative adjuvant therapy. In a similar manner, tumors such as the ganglioglioma and pleomorphic xanthoastrocytoma, which are sometimes discussed with the low-grade astrocytomas because of their relatively favourable prognosis, have unique features that make them best considered as unique clinical entities.

Supratentorial astrocytomas account for between 10 and 15 percent of all brain tumors and constitute between 25 and 35 percent of all gliomas. They occur predominantly in middle age, with the peak incidence falling between 35 and 45 years. They arise predominantly within the convexity of the brain, roughly in proportion to the relative mass of the different lobes. The frontal lobe is the most common location, followed by the temporal lobe. On gross examination, they are often firm, whitish tumors. When growth occurs deep within the brain, formation of small and large cysts will often occur.

The most common presenting symptom is epilepsy; other common ones are headache, mental changes, or focal neurological deficit. In the past, the neuroradiologic procedures used to diagnose the lesion included isotope brain scanning and cerebral angiography. The isotope brain scan might or might not demonstrate the lesion; the angiogram would usually demonstrate a mass lesion without evidence of abnormal vascularity. In recent years, the diagnostic procedures of choice have become CT and/or MR imaging. Because it is not uncommon to find low-grade astrocytomas that are detected only on MRI (after normal CT scans), most believe that MRI is the most sensitive test available today to diagnose these lesions. At surgery, the use of intraoperative ultrasonography has been reported to be extremely helpful in outlining the extent of the lesion.

Whether or not enhancement of the lesion on the CT scan is correlated with a poorer prognosis in patients with low-grade astrocytomas is controversial. Silverman and Marks reported that contrast enhancement had no prognostic value in patients with these tumors. But Piepmeier, in a larger series, concluded that those patients whose tumors enhanced on CT scanning had a poorer prognosis than those whose lesions did not enhance after the administration of an intravenous contrast agent. This poorer prognosis was evident even when adjustment was made for the age of the patient, which is the strongest of all prognosticators. This latter finding has recently been supported strongly by the retrospective study of McCormack et aI., who found that contrast enhancement on the CT scan was associated with almost seven times the risk of tumor recurrence relative to those tumors not showing such enhancement.

CT scanning in a typical case reveals a nonenhancing lesion whose density is lower than that of the surrounding brain. A mass effect upon surrounding ventricular structures is common. When the MR images, the lesion typically presents as a low-intensity area on the T1-weighted images, whereas there is almost always an increase in signal intensity corresponding with an increase in relaxation time on T2-weighted images. The area of increased signal is usually homogeneous and well circumscribed, with no evidence of hemorrhage or necrosis. In many cases, it is difficult to differentiate on MR scans the tumor itself from surrounding areas of edema. Although the data are still not definitive, it does not appear that the use of an MRI contrast agent such as gadolinium will appreciably improve the ability to detect small lesions. Recent studies employing serial stereotactic biopsies of various regions of CT- and MR-defined abnormalities in patients with gliomas have indicated that there is infiltration of tumor cells into areas that were previously thought to represent only edematous white matter.

There also would appear to be a role for positron emission tomography (PET) scanning in the diagnosis and treatment of these patients. A low-grade astrocytoma will be hypometabolic and therefore "cold" on PET scanning. However, when dedifferentiation to a more malignant state occurs within a low-grade astrocytoma, this area will be hypermetabolic and consequently will appear as a "hot spot" on PET scanning. This information may be extremely valuable in determining a site for stereotactic biopsy and/or determining whether the patient should be treated with postoperative radiation therapy.

A great deal of effort has been expended to determine which factors might be of significance in determining the prognosis in a particular patient. All of the recent studies seem to agree that a young age at the time of diagnosis is by far the most important factor that correlates with a prolonged survival. Laws et al. found that other factors correlating with a prolonged survival were gross total surgical removal, lack of a major preoperative neurological deficit, long duration of symptoms prior to surgery, seizures as a presenting symptom, lack of a major postoperative neurological deficit, and having had surgery performed within recent decades. There has not been agreement, however, as to the positive prognostic impact of some of these latter factors; for instance, although Soffietti et al. found a longer survival in those with gross total removal, neither Weir and Grace nor Piepmeier found that the extent of surgical resection was significantly correlated with length of survival. However, most, but not all studies have indicated that patients who receive a biopsy only have a poorer prognosis than those who receive gross total resection. The issue of whether there is an advantage to attempting a gross total removal as compared to a subtotal resection has not been answered. In spite of this, most would agree that lobectomy has no place in the surgical treatment of these tumors and that a gross total resection should be attempted if it can be accomplished without a significant risk of producing a neurological deficit.

Scherer was one of the first to emphasize that great care must be taken to examine all areas of a lesion before deciding that anaplastic foci are not present. In his classic study, he made careful sections of the cerebral hemispheres of 18 patients with astrocytomas and found foci of anaplasia in 13. In a similar manner, Russell and Rubenstein examined 55 autopsy specimens of patients who had been diagnosed clinically as having an astrocytoma. In more than 50 percent of these cases, areas of anaplastic change were found. Looked at from another perspective, Russell and Rubenstein analyzed a series of 129 autopsied cases of glioblastoma multiforme and concluded that approximately 28 percent could be considered to have arising from a preexisting astrocytoma.

Muller and associates examined 72 patients whose pathologic diagnosis at the time of initial operation was astrocytoma. At the time of recurrence, 14 percent of the tumors were unchanged pathologically, but 55 percent were now classified as anaplastic astrocytoma, and 30 percent were now classified as glioblastoma multiforme. The time between the initial pathologic diagnosis and the second operation averaged 31 months. The authors concluded that, in approximately two-thirds of all astrocytomas (i.e., including both the astrocytomas and the anaplastic astrocytomas), one can expect an increase in malignancy (i.e., dedifferentiation). In 79 patients with recurrent tumor growth documented either at subsequent surgery or at autopsy, Laws et al. found that a change to astrocytoma grade III or grade IV had occurred approximately 50 percent of the time. Soffietti et al. reported that 79 percent of their astrocytoma patients who had recurrent tumor were found to have anaplastic areas at either reoperation or autopsy. Vertosick et al., in their series of 25 patients with welldifferentiated astrocytomas, found that none of the eight deaths was due to the progressive growth of the low-grade astrocytoma. Rather, seven of the eight deaths occurred in patients whose tumors had dedifferentiated into an anaplastic astrocytoma or a glioblastoma. McCormack et al. also found that dedifferentiation occurred in six of seven patients with recurrent tumor.

Piepmeier found that malignant transformation was seen in only 13 percent of patients at the time of second operation or autopsy. This number is almost certainly low, however, because the patient population reviewed in his study had a median follow-up of only 5 years and such malignant dedifferentiation will undoubtedly occur in more patients as the follow-up time increases.

Although this is still a subject of some debate and no definite answer can be given, it is probably fair to suggest that the presence of anaplastic areas at the time of a second resection or biopsy in a patient with a previously diagnosed low-grade astrocytoma is not necessarily due to an initial sampling error. Rather, in up to one-half of the cases, dedifferentiation of the low-grade astrocytoma to a more malignant form occurs.

Recent basic research may offer insight as to the moleculargenetic basis of the process of dedifferentiation. In the case of colonic cancer, the progression from the more benign polyp to the overt cancer is reflected in an increasing number of genetic abnormalities within the genome of the tumor cell. Several studies would seem to indicate that a similar process may be occurring in the progression (or dedifferentiation) of an astrocytoma through the stage of anaplastic astrocytoma and ultimately to glioblastoma multiforme. More specifically, Fults et al. found mutations in the p53 tumor suppressor gene in 0 percent of low-grade astrocytomas, 36 percent of anaplastic astrocytomas, and 28 percent of glioblastomas. while abnormalities on chromosome 10 were found in 0 percent of low-grade astrocytomas, 23 percent of anaplastic astrocytomas. and 61 percent of glioblastomas. Sidransky et al. found that in patients who were found to have undergone dedifferentiation at a second operation, the percentage of cells showing a mutation at the p53 gene had increased markedly, perhaps because of a selective growth advantage of these mutated cells.

A reasonable hypothesis of the events underlying dedifferentiation, then, would be that a glial cell experiences a small number of "genetic hits" to change it into a low-grade astrocytoma. If these same cells then experience further alterations in their genetic make-up (e.g., by mutation in a tumor suppressor gene such as p53), they will then undergo dedifferentiation to a more malignant phenotype.

The clinical implications of this work would seem to be that one should remove as many of these minimally genetically altered cells as possible, because over time they may accumulate further genetic abnormalities that cause them to become more malignant. This could be accomplished by surgical removal, or possibly by radiation, if it could destroy small foci of anaplastic cells, as has been suggested in a pathologic study.

It is a general rule of surgical oncology that surgery should be carried out as early in the course of malignancy as possible. However, it has never been proved that earlier treatment of a low-grade astrocytoma produces an increase in life-span as measured from the time of diagnosis. Furthermore, because more and more patients are having their tumors detected while they are neurologically intact and because operative intervention in some locations will carry a significant risk of postoperative morbidity, some have made the case that surgery should be delayed in lesions that do not show a change in appearance on sequential radiologic studies. One retrospective study compared the outcome of a group of 26 patients who had radiologic evidence strongly suggestive of a low-grade astrocytoma and who were initially not operated upon to a similar group of 20 patients who were subjected to immediate surgical intervention. These authors found no significant difference in the incidence of dedifferentiation or survival time between the two groups and thus could not demonstrate that deferring surgery worsened the outcome for patients with low-grade astrocytomas.

Nevertheless, it is still impossible to conclusively determine pathology by radiographic procedures, as is proven by the fact that one study found that enhancement was absent in more than 30 percent of patients with highly anaplastic astrocytoma and even in 4 percent of patients with glioblastoma multiforme. Because the adjuvant treatment recommended in the latter two instances would likely be different than in low-grade astrocytoma, in most cases the desire to forego surgery is outweighed by the necessity of obtaining a precise pathologic diagnosis.

Type of Surgery

Most, but certainly not all, of the retrospective studies that have been carried out have indicated that patients who underwent a "gross total" removal of their lesion experienced a longer survival than those who did not. We must be quite careful in evaluating such data, however, because it is quite likely that the patients in these two groups were not comparable -i.e., in those patients whose tumors were widely infiltrating into vital areas, surgical judgment was at variance with an attempt at "gross total" removal. However, given the general oncologic principle that one should try to obtain the maximum reduction of tumor burden possible and the known propensity of residual cells to undergo malignant dedifferentiation, it would appear prudent to attempt a "gross total" removal in those lesions where this can be done without producing .a postoperative neurological deficit.

If a standard craniotomy is carried out, in many instances the ultrasonic aspirator will be helpful. As in the case of surgery for other intrinsic brain tumors, the tumor should be entered as close to the center as possible and then progressively removed toward the periphery. In many instances, it is quite difficult to be certain as to the interface between frank tumor tissue and normal surrounding brain. As indicated previously, studies by Kelly and others have shown that in low-grade astrocytomas there may be no such clear interface. Consequently, the surgeon should err on the conservative side when potentially important areas of brain are nearby. Other technical adjuncts that may be helpful include the use of an ultrasound device for tumor localization and cortical electrophysiologic monitoring to outline contiguous eloquent areas of the brain.

Because in many instances it is quite difficult for the surgeon to determine the precise location of a low-grade tumor at surgery, these tumors in many cases are ideal for a "stereotactic craniotomy." In this instance, the quite clear delineation of the tumor as seen on the CT or MRI scan can be used to allow one to stereotactically pass a catheter into the center of the lesion. At the time of craniotomy, the surgeon then follows this catheter to the tumor, and thus is certain that he or she is removing the abnormal area that has been seen on the radiologic study. Here come the importance of real-time intraoperative MRI guided surgery.

In the present era, many patients have MRI or CT scans that show quite small lesions with no evidence of mass effect. In such an instance, CT- or MRI-guided stereotactic biopsy is an excellent means of obtaining a definitive tissue diagnosis without subjecting the patient to the risks and inconvenience of a standard craniotomy. In many cases such patients may be discharged from the hospital the following day. Within the last several years, stereotactic biopsy has proven itself to be an extremely accurate, low-risk technique that is therefore ideal for the definitive diagnosis of many low-grade astrocytomas.

The Role of Postoperative Adjuvant Radiation Therapy

Perhaps the most controversial area in the treatment of low-grade astrocytomas is the question of whether postoperative radiation therapy should be used as an adjunctive form of therapy. The answer to this question should be relatively easy to come by. Ideally, such an answer would be forthcoming from what is probably our most powerful tool for scientifically answering clinical questions such as this one: the randomized, controlled, prospective clinical trial. In this case, one would have to carry out a multigroup, long-term (perhaps as long as 10-year) study in which two large groups of patients (containing individuals who are balanced with respect to important variables such as age, tumor location, and histologic classification) were treated identically in every respect (i.e., extent of operation, use of steroids, etc.) ex dept that one group received an exactly specified course of radiation therapy and the other group did not. Whether there was a statistically significant difference in the length and/or quality of survival between these two groups could then be determined. Such a study has never been completed, although at the present time such cooperative studies are being planned in the United States by the Brain Tumor Cooperative Group and are presently being carried out in Europe. Unfortunately, the results of these studies will not be available for many years to come.

Because no single neurosurgeon's experience is adequate to answer properly how patients with a low-grade astrocytoma should be optimally treated postoperatively, and the results of present cooperative trials will not be available for many years, we are faced with the question of how to presently manage this group of patients. The imperfect present-day solution would seem to be a review of the major studies in this area to see whether they can furnish any guidance.

What is immediately apparent in carrying out such a review, however, is that the reports previously published have almost universally not satisfied even the minimal criteria that could be set forth for a study that could properly answer this question. More specifically, the previous studies have been retrospective analyses in which the irradiated and nonirradiated groups of patients have not been similar in important characteristics (e.g., age or Karnofsky rating). The pathologic classification of the lesions has been different (e.g., varying numbers of grade I and grade II tumors). The location and size of the tumors have been different, and the extent of operation has not been uniform (e.g., biopsy versus complete resection). Finally, the parameters of the treatment being tested (i.e., radiation therapy) have not been standardized with respect to total dose, duration of therapy, field size, etc. With these objections in mind, one of the earliest reports was that of Levy and Elvidge, who reviewed 176 cases that were treated at the Montreal Neurological Institute between 1940 and 1949. These authors found what has been confirmed subsequently by many other authors: that the "gemistocytic" type of astrocytoma has a poorer prognosis than that of other variants and that patients with cerebellar astrocytomas did better than those with cerebral lesions, even in the face of incomplete removal. Several years later, Bouchard and Peirce reviewed all patients seen at this same institution over a much longer period and compared the survival of 81 low-grade astrocytoma patients who had received radiation therapy with a group of 71 patients who had not. They found that, although the 3-year survival rate was virtually identical (i.e., 62 versus 59 percent), the 5-year survival statistics showed an increased longevity in those who had received radiation therapy (i.e., 49 versus 38 percent). From these were given radiation therapy. He found that, irrespective of whether biopsy or resection was the surgical procedure used, the addition of radiation therapy caused an increase in survival (biopsy, 10 versus 2 months; resection, 32 versus 23 months). In addition, this study was the first indicating that patients whose tumor was resected rather than just biopsied did better no matter what other, therapy was used.

Uihlein et al. published the first of three major studies utilizing the clinical material of the Mayo Clinic. They reviewed 83 patients with astrocytoma treated between 1955 and 1959. Thirty-three of their patients underwent operation alone, and 50 were treated with operation followed by radiation therapy. They found that 65 percent of those treated with operation alone were alive at 5 years and only 54 percent of those treated by operation and radiation therapy were alive at 5 years. If anything, this indicated a decreased survival after the addition of radiation therapy. However, when they separated the irradiated cases into those that had received 3500 rad (35 Gy) or more and those that had received a lower dosage, the 5-year survival rates were 63 and 42 percent. From this analysis, they concluded that there is a "suggestion" that irradiation may be helpful in the treatment of the low-grade astrocytoma.

In 1982, Bloom reviewed the experience at the Royal Marsden Hospital in treating brain tumors with radiation therapy. His treatment group consisted of 120 patients with grade I or grade II lesions. Although survival data are given only for those treated with operation and radiation therapy (grade I, 5-year survival of 33 percent; 10-year survival, 16 percent; grade II, 5-year survival, 21 percent; 10-year survival, 6 percent), he concluded that "delay of recurrence and greater survival can be expected following postoperative radiotherapy than after surgery alone."

In 1984, Laws et al. again used the patient population at the Mayo Clinic to review 461 astrocytoma patients treated between 1915 and 1975. These cases were selected from a much larger group of patients and represented only those with supratentorial tumors who survived at least 30 days postoperatively and for whom follow-up data were available. Multiple prognostic factors were analyzed for possible correlation with an increase in survival. The authors found that the age of the patient was the most important variable and surpassed all others in its positive correlation with long-term survival. In addition, they interpreted the data as supporting radical operation and a beneficial effect of radiation therapy only in those patients with poor prognostic factors (e.g., older age).

In 1985, Garcia and coworkers reported a retrospective study of 86 adults treated at Washington University between 1950 and 1979. Although the number of patients with well-differentiated astrocytomas was small, they found that those with a juvenile pilocytic type of astrocytoma did well regardless of treatment and did not require radiation therapy, a conclusion that has been confirmed in other studies.

Piepmeier reviewed the records of 60 patients with low-grade astrocytomas seen at the Yale-New Haven Hospital between 1975 and 1985. In this retrospective review, there was no significant difference found in survival between those patients who received radiation therapy in addition to surgery and those who did not. What is important in this study is that all patients who were irradiated received between 50 and 60 Gy delivered over 5 to 6 weeks to fields that were constructed by using CT scanning to include the tumor plus a wide margin of surrounding brain. One caveat expressed by the author, however, was that because the patient population reviewed in this paper was treated over the last decade, the mean follow-up time was slightly less than 5 years and thus this may have been insufficient time to allow a potential effect of radiation to become evident. However, it should also be noted that most previous studies which did indicate a beneficial effect of radiation therapy did so mainly at 5 years, with such beneficial effect decreasing at 10 years and longer.

In 1989 Shaw et al. once again reviewed the patients at the Mayo Clinic and reported on 167 patients, of whom 139 (83 percent) received surgery plus radiation therapy with a mean tumor dose of 50 Gy. The 5-year survival rate for those receiving high dose (>53 Gy) radiation therapy was 68 percent, whereas the survival rate was 47 percent for those who received low-dose irradiation (<53 x Gy) and 32 percent for those who had surgery but were not irradiated. The comparable 10-year survival rates were 39 percent, 21 percent, and 11 percent, respectively. In contrast to these data for the grade I and grade II astrocytomas indicating a beneficial effect of radiation therapy, they found that postoperative irradiation was not associated with improved survival in the patients with pilocytic astrocytomas.

Hirsch et al. in 1989 reported on 22 paediatric patients who were operated on for grade I or grade II astrocytomas. None of these patients was initially given radiation therapy. Because only three recurrences (8 percent) were seen in the entire group of 42 patients (which included 8 patients with oligodendroglioma and 12 patients with oligoastrocytoma), the authors concluded that postoperative radiation therapy should not be given to paediatric patients with low-grade cerebral gliomas.

In 1990, North et al. reported on a series of 77 patients from the Johns Hopkins School of Medicine who were treated with a uniform radiation therapy dose of 50 to 55 Gy over a period of 51/2 to 6 weeks. Most importantly, in this study quality of life was determined at 1 to 2 years postoperatively and at last follow-up at 2 to 12 years after surgery. They noted that mental retardation was observed in 50 percent of the children who had received radiation therapy. Overall, however, 80 percent of short-term survivors and 67 percent of long-term survivors were intellectually and physically intact and without major neurological deficit.

Also in 1990, Whitton and Bloom reviewed 88 adults with cerebral low-grade gliomas who were treated with postoperative radiotherapy at the Royal Marsden Hospital between 1960 and 1985. Treatments were given five times a week to a total dose of 50 to 55 Gy. They were able to confirm that age was a very important prognostic factor, but indicated that it was still unclear whether or not postoperative radiation therapy was effective.

In 1991, Vertosick et al. analyzed treatment results in 25 patients with well-differentiated cerebral astrocytomas. The median survival for their entire group of patients is 8.2 years, which is the longest that has thus far been reported. They attributed this long-term survival to earlier diagnosis in the CT /MRI scan era rather than to the specific efficacy of any modern form of adjuvant therapy. Approximately 70 percent of their patients received postoperative radiation therapy, whereas 30 percent did not. In this series, the use of radiation therapy did not have a significant impact upon the time to tumor dedifferentiation or the time to death, although they cautioned that the number of patients in each group was small.

In 1992, McCormack et al. carried out a retrospective review of 53 patients with supratentorial astrocytomas. Because fully 98 percent of their patients received postoperative radiation therapy, it could not be determined whether such patients lived longer than those who did not receive such adjuvant therapy.

There have been several recent reports on the use of alternative forms of radiation therapy in the treatment of low-grade astrocytomas. Three authors have reported on the use of interstitial radiation therapy with implanted radioactive iodine seeds. Frank et al. detailed a series of 45 patients, and concluded that its use should be limited to patients less than 40 years of age whose tumors are not in the optic chiasm, hypothalamus, or lower brain stem. Voges et al. reported on its use in 13 children and indicated that tumor shrinkage was seen on CT scans in all children by 6 months postimplantation. In 1991, Mundinger et al. reported on the use of interstitial radiation in 89 patients harbouring nonresectable low-grade brain stem astrocytomas. Because these tumors differ from cerebral tumors in many respects, one cannot extrapolate from these data as to the possible effectiveness of this technique in the treatment of cerebra/low-grade tumors. The paper does, however, indicate that interstitial radiation therapy when carried out with I²⁵¹ via an implanted catheter is a safe and feasible technique.

There has also been one report on the use of stereotactic radiosurgery in the treatment of low-grade astrocytomas. Pozza et al. treated 14 patients with nonoperable low-grade astrocytomas with unconventionally fractionated stereotactic radiosurgery. A total of 16 to 50 Gy was administered in either one or two fractions 8 days apart. They indicated that 12 of 14 patients demonstrated a partial or complete response as demonstrated by CT scanning.

Finally, there was even been a recent case report on the use of re-irradiation in a patient with a low-grade astrocytoma who had been irradiated 8.5 years previously. In this single instance, there was no evidence of clinical or radiologic brain injury at the time of 3-year follow-up.

As indicated in the literature review above, the majority of the major English-language studies have found that radiation therapy is beneficial when added to surgery in the treatment of cerebral astrocytoma. One must, however, be extremely cautious in interpreting the retrospective data from these reports. As indicated previously, it is mandatory to take into account the various prognostic factors that may be present to differing degrees in the two groups of patients that are being compared. Age, functional status of the patient, extent of surgical removal, and pathologic grade (i.e., grade I or grade II) are at least some of the important variables that must be known. In almost none of the studies reviewed is this information readily available. In addition, all of suffer from being retrospective analyses in which the two groups are not strictly comparable with respect to various selection factors or even the treatment given. Consequently, any conclusions reached must be considered only tentative until the proper studies are carried out.

Future advances in technology may allow a subgroup of patients with low-grade astrocytomas to be selected who would most benefit from receiving postoperative radiation therapy. Currently, procedures have been developed that can measure the proliferative potential of low-grade astrocytomas:" using immunohistochemical techniques such as in vivo or in vitro:" labelling with bromodeoxyuridine (BUdR) or labelling with the monoclonal antibody Ki-67. A more simple technique may involve the measurement of nucleolar organizer regions. Preliminary data seem to reveal a correlation between an increase in proliferative potential and a poor prognosis. Furthermore, a study of 12 patients with low-grade astrocytomas who underwent PET scanning with ISF fluorodeoxyglucose (FDG) indicated that malignant change may be associated with a focal area of hypermetabolism that develops within an area that in general is hypometabolic. If this is confirmed in other studies, then perhaps only those patients whose tumors have a labelling index above a certain level or who have a hypermetabolic area on PET scanning should receive radiation therapy.

The issue of whether radiation therapy should be used in patients with low-grade astrocytomas is not one that can be taken lightly. In patients with anaplastic astrocytoma or glioblastoma multiforme, it is quite probable that the relatively short survival time prevents the long-term deleterious effects of radiation therapy from becoming evident. This would not be the case in the group of patients with grade I or grade II astrocytoma, who may have a 5-year survival rate of approximately 65 percent and a 10-year survival rate of perhaps 40 percent.

There have been many studies of complications produced by cerebral radiation therapy. One such study reported on patients in whom malignant gliomas developed after radiation therapy that had been previously administered for other conditions. At least seven such cases have been documented; patients who experienced this complication tended to be young, as is the case in most patients with low-grade astrocytomas who are given radiation therapy. A review from the Mount Sinai Hospital in New York City found seven cases of radiation-induced meningiomas. The overwhelming majority of these patients had received low-dose radiation therapy (8 Gy) to the scalp for tinea capitis. The second largest group, however, were patients who received high-dose radiation for primary brain tumors.

Although the reported incidence of radiation necrosis varies widely, white matter changes are being seen more and more frequently on MRI scans of patients who have previously undergone radiation therapy. A recent study indicates the presence of radiation necrosis in 9 percent of a series of 76 patients treated with whole-brain radiation for various intrinsic brain tumors. In this regard, it is of interest that a review of 371 irradiated brain tumor patients by Marks and Wong found the incidence of radiation necrosis to be 1.5 percent at 55 Gy and 4 percent at 60 Gy, with a substantial increase for higher doses. Because it is generally accepted that the risk of untoward sequelae from radiation therapy is greater after whole-brain radiation therapy than after more localized treatment, it would seem most prudent to carry out only localized radiation if one decides to use this adjuvant form of therapy.

The Role of Postoperative Adjuvant Chemotherapy

Over the years, there have been several anecdotal reports on the use of various chemotherapeutic agents in small numbers of patients with low-grade astrocytomas. Invariably, one has been unable to draw conclusions with respect to efficacy from such case reports. There has been one analysis that compared 75 patients who were treated with radiation plus intra-arterial BCNU as well as vincristine and procarbazine, to 57 patients treated with radiation alone. This study seemed to show a longer survival in those treated with this aggressive chemotherapy regimen.

There has, however, been a prospective randomized study that came to the opposite conclusion. This study. which was conducted by the Southwest Oncology Group, demonstrated that the addition of CCNU to radiation therapy did not result in an increase in survival. At the present time, therefore, it seems that there is no proven beneficial effect of chemotherapy in the treatment of patients with low-grade astrocytomas.

Treatment at Recurrence

Failures of the previously described treatment modalities are almost always due to local recurrence. This can be the result either of the continued growth of the low-grade neoplasm (which can result in the death of the patient if this tumor is located in the deeper part of the brain) or to dedifferentiation of a low-grade tumor into a malignant glioma.

The treatment of such a recurrence depends on establishing the tumor grade. This implies that repeat biopsy will be necessary in most cases. If the tumor remains low grade, then the patient may be followed by periodic CT/MRI scans and/or PET scans. Observation may also be warranted if the patient's clinical status is stable. If such a tumor is enlarging and causing a significant mass effect or CSF obstruction, then repeat resection alone should be considered. On the other hand, if the neuroradiologic studies, clinical course, and/or biopsy indicate that malignant transformation has occurred, a more aggressive course consisting of repeat surgical resection, interstitial radiation therapy, and/or chemotherapy may be considered. Because the time period between the initial radiation that may have been given and the recurrence may be quite long, re-irradiation may even be considered. A good result after re-irradiation has recently been described in a patient whose tumor recurred 8.5 years after the initial treatment.

Outcome

A review of the several series would indicate a 5-year survival rate of approximately 40 to 50 percent and a 10-year survival rate of approximately 20 to 30 percent. However, the recent series indicate a current median survival for the entire group of patients of approximately 7 years, with a 5-year survival of approximately 65 percent and a 10-year survival of approximately 40 percent.

Conclusions

Because the prospective randomized studies that are presently being carried out have not been completed, the optimal treatment of the patient with a low-grade astrocytoma remains controversial. Until more definitive data become available, certain tentative conclusions may be drawn:

1. An attempt should be made to obtain pathologic confirmation of the nature of a supratentorial lesion that is seen on CT or MRI scan and has at least some of the features of an intrinsic brain tumor.

2. Consistent with sound neurosurgical judgment as to postoperative sequelae, an attempt should be made to carry out gross total removal of a hemispheric astrocytoma or to remove as much tumor as possible.

3. In the case of such a gross total surgical removal (and even in its absence in the case of the cerebral pilocytic astrocytoma), radiation therapy can be withheld and the patient carefully followed with periodic CT and/or MRI scans. If the lesion does not show definite evidence of recurrence then radiation therapy should be withheld. If the cerebral low-grade astrocytoma is present in a paediatric patient (even if the resection has not been complete) then radiation should be withheld and the patient carefully followed with CT, MRI, and possibly PET scans.

4. It is likely that monoclonal antibodies and PET scanning will allow us to select a subpopulation of patients who would most likely benefit from postoperative radiation therapy.

5. At the present time, however, in cases where total removal cannot be accomplished, postoperative radiation therapy may be warranted.

6. Such radiation therapy should be given in a conventional fractionation schedule to a dose not exceeding 55 Gy. This radiation therapy should be given to a limited volume as determined by CT and/or MRI studies rather than to the whole brain. As future studies become available, it is quite possible that interstitial radiation therapy will have a role to play in the treatment of this tumor.

Using present-day techniques, an optimal treatment regimen for the patient who is diagnosed as having a low-grade astrocytoma will lead to a median survival of approximately 7.5-years with a 5-year survival of approximately 65 percent and a 10-year survival of approximately 40 percent. A more precise estimate of survival time can be made if the particular prognostic variables of the individual patient are known.