Intracranial hemangiopericytomas: correlation of topoisomerase IIα expression with biologic behavior

1. Introduction 

Hemangiopericytomas of the central nervous system are aggressive meningeal tumors with a high tendency for local recurrence and extraneural metastasis despite radical excision and radiotherapy. The 5-year survival rates vary from 67% to 93% [5], [7]. Although associated with prolonged survival, no statistically significant curative benefits have been noted for gross total resection, external brain radiotherapy, stereotactic radiotherapy, and currently used chemotherapeutic regimens [5].

Topoisomerase IIα is a molecular and immunohistochemical marker of cell proliferation [30]. It is also the target of various chemotherapeutic agents such as teniposide, etoposide, 4′-(9-acridinylamino)methanesulfon-m-anisidide, doxorubicin, ellipticine, and mitoxantrone. In various cell lines established from human cancers, the cellular resistance or sensitivity to these agents depends on the nuclear level of topo IIα. Determining the nuclear levels of topo IIα in tissue specimens of human cancers has therefore been speculated to predict efficacy of topo IIα–targeted chemotherapeutic agents [8], [16], [26]. We report a retrospective analysis of intracranial hemangiopericytomas managed over the last 2 decades. The immunoexpression of topo IIα was studied in these tumors to determine if there was any correlation of its expression to biologic behavior.

2. Materials and methods 

This was a retrospective analysis of 27 intracranial hemangiopericytomas, diagnosed between January 1990 and June 2003. Patients with a diagnosis of intracranial hemangiopericytoma in whom the paraffin blocks or slides of the tumor were unavailable were excluded from the study.

3. Results 

The male/female ratio was 15:12. The age at presentation ranged from 11 to 52 years with a mean of 31.33 years. The duration of symptoms ranged from 11 to 36 months (mean, 9.18 months). On radiological imaging, all 27 cases were contrast-enhancing dural-based tumors. Details of the site and size of tumor and extent of surgical excision are given in Table 1. Six of the tumors were convexity-based masses, another 7 were parasagittal, 5 arose in the middle cranial fossa, and the remaining tumors were seen in the tentorium, cerebellopontine angle, sellar region, orbit, and the basifrontal region. Nineteen of 27 cases were larger than 5 cm. Most cases had a radical excision (18/27). Of the 2 tumors that had a subtotal excision, one was in the middle cranial fossa, and the other arose in the tentorium. Of the 3 sellar tumors, 2 had a biopsy, and one had a partial excision. In the case of 2 parasagittal tumors, owing to increased vascularity and blood loss, surgery was limited to a biopsy in one case and a partial excision in the other. Eighteen of 27 cases received postoperative radiation therapy, of which 11 had undergone a radical tumor excision.

Table 1. The distribution of hemangiopericytomas by site, type of excision, and size

	Type of excision	Size of tumor (cm)			Convexity	Parasagittal	Middle cranial fossa	Tentorium	Cerebellopontine angle	Sellar region	Basifrontal	Orbit
		<3	3-5	>5
	Biopsy		2	1		1				2
	Partial		1	3		1	2			1
	Subtotal			2			1	1
	Radical	1	4	13	6	5	2	1	2		1	1

Microscopically, the tumors were cellular and composed of cells with scant ill-defined cytoplasm and oval nuclei. The tumors were richly vascular with predominantly thin-walled vascular channels, some having a staghorn appearance (Fig. 1A). Reticulin was seen investing blood vessel walls and individual cells, especially around blood vessels (Fig. 1B). The mitotic index was variable. All cases had a similar immunohistochemical profile with the tumors being positive for vimentin (Fig. 1C) and negative for epithelial membrane antigen. The endothelial cells were positive for CD34 (Fig. 1D), and the tumor cells were focally positive for CD34 in 4 cases.

View Large Image Download to PowerPoint

Fig. 1. Hemangiopericytoma. A: Illustrating high cellularity and staghorn configuration of blood vessels (hematoxylin-eosin stain, magnification ×90). B: Dense perivascular and pericellular reticulin (Gordon and Sweet reticulin stain, magnification ×400). C: Cytoplasmic immunopositivity with vimentin (streptavidin-biotin-peroxidase, magnification ×400). D: Immunopositivity of blood vessels for CD34 (streptavidin-biotin-peroxidase, magnification ×90). E: Topoisomerase IIα index of less than 5% (streptavidin-biotin-peroxidase, magnification ×400). F: Topoisomerase IIα index of 5% or greater (streptavidin-biotin-peroxidase, magnification ×400).

3.2. Prognostic variables 

Despite a mitotic index of less than 5 per 10 hpf, 47% of cases recurred (Fig. 2A). In comparison, only 20% of cases with a TPI of less than 5% recurred (Fig. 2B). On comparing the 2 methods of assessing proliferation, mitotic index vis-a-vis TPI using κ statistics, the agreement between the 2 methods was poor (0.42) (Fig. 3). Adjuvant radiation therapy reduced the incidence of recurrence (Fig. 4). However, irrespective of the extent of excision, 44% of cases that underwent radiation therapy recurred.

View Large Image Download to PowerPoint

Fig. 2. A: Incidence of recurrence in relation to mitotic index. B: Incidence of recurrence in relation to TPI.

View Large Image Download to PowerPoint

Fig. 3. Relation of mitotic index to TPI.

View Large Image Download to PowerPoint

Fig. 4. Relation of radiotherapy to incidence of recurrence.

Using the Cox proportional hazard model, it was found that the incidence of tumor progression or recurrence was 2.9 (0.64, 13.27) times greater in patients with a TPI 5% or greater as compared with a TPI of less than 5%. The incidence of tumor progression or recurrence was 1.63 (0.50, 4.54) if the mitotic index was greater than 5 per 10 hpf, 0.96 (0.29, 3.18) for partial or subtotal excision and biopsy, and 0.96 (0.34, 2.69) for those with no radiation therapy. The overall 5-year recurrence-free survival was 44.3% (Fig. 5). Although radical excision, mitotic index of less than 5 per 10 hpf, TPI of less than 5%, radiation therapy, patient age and sex, and tumor location did not provide a statistically significant effect on risk of recurrence, a TPI of less than 5%, radical excision, and radiation therapy were variables that had a higher recurrence-free survival (Fig. 6).

View Large Image Download to PowerPoint

Fig. 5. Kaplan-Meier plot: overall recurrence-free rates.

View Large Image Download to PowerPoint

Fig. 6. Recurrence-free rates at 5 years for select variables (survival analysis). RT indicates radiation therapy.

4. Discussion 

Meningeal hemangiopericytomas account for less than 1% of all intracranial tumors. They tend to occur at a younger age than meningiomas and more often in men than in women. The mean age at diagnosis was 43 years in one series [7] and 44.9 years in another series [1]. In the present study, the mean age was slightly higher at 49 years. The histogenesis of hemangiopericytomas is uncertain, but its microscopic and ultrastructural features suggest an origin from meningeal capillary pericytes [22]. They are most often seen as single tumors attached to the cranial or spinal dura. The symptoms of meningeal hemangiopericytomas are indistinguishable from those caused by meningiomas. Computed tomography and magnetic resonance imaging show an intensely contrast-enhancing, well-defined, dural-based tumor with surrounding edema [7].

The conventional treatment of these tumors includes complete microsurgical removal, followed by local irradiation to the tumor bed [1], [6], [7], [11], [21]. Stereotactic radiotherapy has been used for treating tumor residues, recurrences, and local metastases [4], [5], [6]. Adjuvant external brain radiation therapy, especially at doses greater than 50 Gy, and stereotactic radiosurgery have resulted in dramatic tumor shrinkage, tumor obliteration, and lengthened disease-free interval in several series [5], [29]. Nevertheless, there still remain cases that develop recurrences and new metastasis, and in our series, irrespective of the extent of excision, more than 40% of cases that underwent radiation recurred. Ecker et al [5] also found no statistically significant survival benefit associated with gross total resection. Galanis et al [6] and Ecker et al [5] have concluded that chemotherapy is ineffective in treatment of recurrent hemangiopericytomas with a few exceptions. However, in the series reported by Ecker et al [5], one patient had a shrinkage of a pulmonary metastasis with doxorubicin, and another patient remained disease-free at 5-year follow-up after initial adjuvant chemotherapy with agents including doxorubicin and etoposide. It is plausible that the previously mentioned cases that had a good response to chemotherapy had a high topo IIα content and hence responded to the chemotherapy directed against topo IIα.

The definition of recurrence and the duration of follow-up vary considerably between series, the time-to-recurrence after apparent complete removal also varies remarkably, but the mean recurrence-free intervals of 47 months [7] and 78 months [11] suggest that meningeal hemangiopericytomas tend to grow more rapidly than meningiomas [11]. In the present study, the mean time to recurrence was 49 months. The reported incidence of recurrence varies from 50% to 80% [7], [11], [19], [28]. Tumor recurrence was seen in 55% of cases in our series, which compares well with other reports. The 5-year recurrence-free survival in our series was 44.3%. This is in contrast to the 5-year recurrence-free survival of 89% in the series reported by Ecker et al [5]. The same group had, in an earlier study, reported a 5-year survival of 67%. Although the reasons for improved patient survival in the later series were not readily apparent, refined surgical techniques with low intraoperative mortality, radiosurgery, dedicated neurointensive care management, improved general care of patients with cancer, a higher percentage of patients receiving more than 50 Gy of adjuvant radiation therapy, and the use of radiosurgery were thought to be factors that may have contributed to improving survival rates.

Hemangiopericytomas frequently metastasize extraneurally. Metastasis has been described in 64% to 69% [7], [31] at 15 years most frequently, in descending order of frequency, to the bones, lung, and liver. The median interval to metastatic spread was 84 to 99 months in different studies. The one case in our series with metastatic tumor in the lung presented as early as 17 months after the first surgery. In a series of 28 patients who had survived primary removal, the probability of tumor related death was 61% at 15 years [31]. The mean survival time after diagnosis of metastasis was 2 years [7].

Studying predictors of biologic behavior in hemangiopericytomas, Guthrie et al [7] found that although histologically low-grade hemangiopericytomas were associated with a prolonged survival, no statistically significant survival benefit was noted with low proliferation indices. However, a later study from the same institution, done on a larger series of hemangiopericytomas, found no significant survival benefit for low-grade histological characteristics, although early tumor recurrence was seen in high-grade tumors [5]. In other studies, lower than average values for Ki-67 index [24], [31] and S-phase fractions [31] were only marginally associated with a better prognosis in terms of increased disease-free survival and lower rates of metastasis, as well as longer overall survival.

Topoisomerase IIα and IIβ isoforms are class II enzymes that play a crucial role in DNA replication. They function in the segregation of newly replicated chromosomal pairs, in chromosomal condensation and in altering DNA superhelicity. Expression of topo IIα increases in the late S phase and decreases rapidly at the end of the M phase [30]. It is a reliable proliferation marker [3], [12], [13], [14], [15], [27] and is a potential predictive factor for prognosis [2], [10], [17], [20], [23], [25]. Topoisomerase IIα content in tumor cells is also considered to be a critical determinant for anticancer drug cytotoxic activity or drug sensitivity/resistance of various human tumors [16], [18], [23], [26]. In view of the dual role that topo IIα plays as a cell proliferation marker and as a possible indicator of chemosensitivity, we investigated its expression in hemangiopericytomas in relation to conventional clinicopathologic parameters and patient outcome. In our study, the TPI was found to have clear prognostic utility with the incidence of tumor progression or recurrence being 2.9 times greater if the TPI was 5% or greater as compared with a TPI of less than 5%. The recurrence-free survival was also found to be longer for those with a TPI of less than 5%. Because intracranial hemangiopericytomas are relatively rare, most studies have been retrospective, covering a span of a decade or more. It is likely that statistically significant survival benefits have not been elicited in prior studies owing to the choice of proliferation markers. For example, the prognostic reliability of MIB-1 proliferation index has been found to be inconsistent because of its degradation during prolonged storage, making it unreliable for archival cases [9], [30], [32]. Topoisomerase IIα is known to have several advantages over other proliferation markers currently in use, notably providing a better estimate of the number of cycling cells and exhibiting a more distinct, less variable nuclear-staining pattern [12], [27]. The antigen labeled by topo IIα is extremely robust, resisting degradation by fixation and enzymatic tissue desegregation. Although limited, the observations on using topo IIα as a marker indicate that, in older sections and stored precut sections, topo IIα shows much better nuclear staining than MIB-1 [9], [30], [32].

We conclude that a high proportion of intracranial hemangiopericytomas recur despite radical excision and adjuvant radiation therapy. The TPI of 5% or greater could be used as a predictor of recurrence. Our study suggests that there may be a role for adjuvant chemotherapy using anti–topo IIα agents in cases that express high levels of topo IIα.