Journal of Neuro-Oncology 59: 63–69, 2002. © 2002 Kluwer Academic Publishers. Printed in the Netherlands.

Clinical Study

Good clinical course in infants with desmoplastic cerebral neuroepithelial tumor treated by surgery alone Kazuhiko Sugiyama1 , Kazunori Arita1 , Takeshi Shima2 , Mitsuo Nakaoka3 , Takashi Matsuoka3 , Eiji Taniguchi1 , Tatsunori Okamura1 , Humiyuki Yamasaki1 , Yoshinori Kajiwara1 and Kaoru Kurisu1 1 Department of Neurosurgery, Hiroshima University Medical Hospital, Hiroshima; 2 Department of Neurosurgery, Chugoku Rosai Hospital, Kure; 3 Department of Neurosurgery, Matsue Red Cross Hospital, Matsue, Japan

Key words: desmoplastic astrocytoma, desmoplastic cerebral neuroepithelial tumor, desmoplastic ganglioglioma, infant, pre-operative diagnosis, surgery Summary We investigated why surgery alone provides for a benign clinical course in patients with desmoplastic infantile ganglioglioma and astrocytoma (DIG/A). The clinical course of 4, less than six-month-old girls, surgically treated at our institutions, was evaluated retrospectively. All presented with the clinical symptom of increasing head circumference. CT and MRI scans revealed a solid tumor attached to the dura that was surrounded by large, multiple cysts, in fronto-temporo-parietal lobe. Gross total removal succeeded in all 4 cases because the solid components of the tumor were very firm in contrast to the soft adjacent brain tissue. Microscopically, the surgical specimens consisted almost entirely of dense fibrous connective tissue containing generally elongate cells with inconspicuous cytoplasm. Most of these cells were immunopositive for GFAP. There was no evidence of tumor cells in the cyst wall. In 3 cases, some small neurons were positive for neurofilament immunostain. A high proportion of undifferentiated small cells in a less demoplastic area far from the dura were immunopositive for MIB-1. All of the 4 patients have been free of recurrence for more than five years. In patients with DIG/A, there are 5 reasons for a good clinical course. [1] At surgery, the tumor margin is clearly discernible because of the difference between the solid tumor and the soft adjacent brain tissue. [2] The tumor is located in the superficial cerebral hemisphere. [3] Large, multiple cysts surround the tumor. [4] The growth point appears to be adjacent to the cysts. [5] The cyst walls are free of invading tumor cells. Introduction In the last quarter of the 20th century, knowledge about desmoplastic neuroepithelial tumors has improved remarkably [1,2]. These tumors include pleomorphic xanthoastrocytoma (PXA) [3], gliofibroma [5], desmoplastic infantile ganglioglioma (DIG) [5], and desmoplastic infantile astrocytoma (DIA) [6] (desmoplastic cerebral astrocytoma of infancy [7, 8], and superficial cerebral astrocytoma attached to the dura [9]). These desmoplastic neuroepithelial tumors are distinct pathological entities, respectively, and could be subsumed into one clinical category: most of them develop in the cerebral hemisphere and occur in less than 30-yearold patients, especially infants and young children, and most have a good clinical course after surgery alone [10,11]. Although the postoperative status of patients with DIG and DIA (DIG/A) is usually highly satis-

factory, the reasons for the generally good treatmentoutcomes have not been previously discussed. In this paper, we present the clinicopathological aspects of DIG/A and discuss the reasons for its generally benign clinical course.

Materials and methods Between 1991 and 1995, 4 female infants were treated at our hospitals for pediatric brain tumors. The diagnosis was DIG/A, established by routine histological and immunohistochemical techniques based on welldefined pathological criteria [6]. From their medical records, we ascertained age at presentation, presenting symptoms and their duration from initial onset, radiological features, surgical findings, and histological characteristics. The 4 patients have been followed up

64 for a median of 88 months. All underwent imaging studies with either CT or MRI scans at 6- to 10-month intervals.

Results The overall median age of the patients was 2.5 months (range 2–5 months). The most common symptom was

A

C

increasing head circumference (4 cases), followed by hemiparesis (2 cases), failure to thrive (2 cases), and irritability (2 cases). In case 2, there was ultrasonographic evidence, obtained at 35 weeks of gestation, of a fetus with a large brain. CT and MRI demonstrated an irregularly shaped solid tumor attached to the dura. Many large cysts in the frontal or fronto-parieto-temporal lobe surrounded the tumor. The solid components exhibited strong and homogeneous contrast enhancement

B

D

Figure 1. These tumors presented on CT or MRI as hypodense or hypointense cystic mass with enhancement of the solid nodule and the periphery after administration of contrast material or gadolinium. (A) Case 1, Axial plain CT. (B) Case 2, Axial enhanced CT. (C) Case 3, Coronal T1-weighted post-gadolinium image. (D) Case 4, Axial T1-weighted post-gadolinium image.

65

B

A

Figure 2. (A) Intraoperative view after dural incision showing solid component of the tumor (center) and adjacent brain parenchyma (right) (case 3). (B) Gross appearance of the excised tumor: multilobular tumor surface and white to gray cut surface (case 3).

A

B

C

D

Figure 3. Dense desmoplastic and collagen-rich tissue, often exhibiting a spindle, short fusiform cell or storiform pattern, made up the bulk of the lesions. (A) Case 1, H&E, ×50. (B) Case 2, H&E, ×100. (C) Case 3, H&E, ×100. (D) Case 4, H&E, ×100.

66

A

B

C

D

Figure 4. (Case 3) (A) Marked desmoplastic reaction (silver impregnation for reticulin, ×25). (B) The astrocytic nature of the tumor was revealed with GFAP immunostain (×100). (C) Ganglion cells in desmoplastic area were identified by their immunoreactivity for neurofilament (×100). (D) The spindle cells in desmoplastic area had a low positivity rate for MIB-1 (1.5%, MIB-1 immunostain).

(Figure 1). No calcification was noted on plain CT (Figure 1A). All patients underwent total removal of the tumor. After craniotomy, ultrasonographic imaging enabled us to aspirate xanthochromic fluid and to open the dura mater safely. The superficial solid portion of the tumor was commonly extracerebral and involved the cerebral cortex. Separation from the brain parenchyma was easy because of the solid nature of the tumor compared to the soft adjacent tissue. There appeared to be no involvement of major vessels. A biopsy of cyst walls was obtained in all cases (Figure 2A). All tumors were large, measuring up to 7 cm in diameter. Macroscopically, they were multilobular or irregular in shape and white to gray in color. Their texture was extremely firm due to stromal reaction (Figure 2B). Routine histopathologic staining revealed that fibroblast-like spindle cells with elongated nuclei and

cellular processes were arranged in bundles with abundant interwoven collagen fibers (Figure 3). The predominant feature was desmoplasia in combination with neuroepithelial and fibroblastic elements (Figure 4A). Immunoperoxidase stains showed diffuse, strong positivity for GFAP in many of the spindle cells (Figure 4B). In 3 cases, neoplastic neurons were identified by their immunoreactivity for neurofilaments (Figure 4C). There were focal less-desmoplastic areas with undifferentiated small cells (UDSCs) present in all cases (Figure 5A,B). Most of these cells were also positive for GFAP, and negative for neurofilament (Figure 5C). The UDSCs had a higher positivity rate for MIB-1 than the predominating spindle cells in the desmoplastic area. (Figures 4D, 5D). This difference between two types of the tumor cells was significant by the Mann–Whitny U test (p = 0.02). There was no histological evidence of tumor cell invasion into the cyst wall.

67

A

B

C

D

Figure 5. (Case 3) (A) Focal less desmoplastic area consisting of undifferentiated small cells (UDSCs, right) (H&E, ×25). (B) High power view of less desmoplastic area: diffuse proliferation of the UDSCs (H&E, ×50). (C) The UDSCs were immuno-positive for GFAP (GFAP immunostain, ×50). (D) The UDSCs had a high positivity rate for MIB-1 (4.2%, MIB-1 immunostain). Table 1. Clinicopathological summary of four patients Case #

Age (mo.)/sex

Location

Tx

Astrocytic diff.

Neuronal diff.

MIB-1

Follow-up

Status

1

4/F

F

GTR

(+++)

(+)

2/F

F-T-P

GTR

(++)

(+)

3

5/F

F-T-P

GTR

(+++)

(+)

4

3/F

F-T-P

GTR

(++)

(−)

alive (111 mo.) alive (98 mo.) alive (78 mo.) alive (51 mo.)

N.D.

2

2.1%: PSPCs 4.9%: UDRCs 2.7%: PSPCs 5.2%: UDRCs 1.5%: PSPCs 4.2%: UDRCs 1.5%: PSPCs 3.1%: UDRCs

Mild M.R. Hemiparesis N.D. N.D.

mo.: month; Tx: treatment; GTR: gross total removal; diff.: differentiation; PSPCs: prominent spindlecells; UDRCs: undifferentiated round cells; N.D.: normal development; M.R.: mental retardation; F: frontal; F-T-P: fronto-temporo-parietal.

All patients are alive without evidence of recurrence for a median of 88 months. One patient (case 2) is mentally retarded and manifests right hemiparesis. The status of the other 3 patients is satisfactory and they exhibit normal physical and mental development. These results were summarized in Table 1.

Discussion None of the 4 patients we studied experienced recurrence during the 5-year period since the total removal of their tumors. The surgical procedure was very simple: the solid elements of the tumors were totally

68 removed under a surgical microscope after aspiration of the cyst fluid and opening of the dura. In dissecting the tumor from the brain parenchyma, its firmness was the most valuable clue to determining the cleavage plane between the tumor and the surrounding brain. DIG/A have an intense reticulin network [1,2,5,6]. Our retrospective study suggests that the difference in firmness coincides with the tumor margin. Median survival time in supratentorial diffuse astrocytoma (WHO grade II) is less than 10 years [12], although the tumor cells in diffuse astrocytoma have a lower positivity rate for MIB1 than the typical spindle ones in DID/A [11,13]. If these tumors had the characteristic of microscopic focal infiltration at the interface between the desmoplastic area and the brain interface, gross total tumor resection based on the difference in tissue firmness would not provide for a benign clinical course. PXA has a higher rate of recurrence after total removal than DIG/A [3,6,10,11,14] because PXA has a less intense reticulin network and the tumor margin is relatively unclear [3,10,14]. Therefore, curative surgery is less frequent in patients with PXA [3,10]. Tumor location and the presence of multiple large cysts surrounding the tumor also play a role in the good outcomes obtained in patients with DIG and A. Since DIG/A frequently develop in the frontal and frontotemporo-parietal lobe [5,6,8,11], the tumor can be visualized just after dural incision. Before this, the larger cysts should be punctured and their fluid aspirated to decrease intracranial pressure. It is important to control intracranial pressure and inspect the entire tumor mass in the early stage of surgery [1]. Removal of the tumor is more difficult if it is located in the median surface of the cortex attached to the falx [15], or if there is involvement of major vessels and vital structures, e.g. Sylvian fissure [1,16], corpus callosum [15], basal ganglia [1,15,17], thalamus [1,17], and midbrain [1,15,17]. Growth of DIG/A occurs in the area where cells with high proliferative potential are amassed. The histogenesis of UDSCs is controversial [8,11], and the focus on these cells historically led to the failure to understand that DIG/A are astrocytic tumors [18]. Our morphologic evaluation included multiple immunostaining and indicated that UDSCs had a higher positive MIB-1 labeling index in the less-desmoplastic region than did the typical elongated spindle cells in the desmoplastic region. Our hypothesis that growth occurs in the area around the cysts rather than the dura mater is supported by the observation that upon silver impregnation for reticulin, the areas closer to the dura exhibit an increasingly intense desmoplastic reaction.

Lastly, in DIG/A, tumor cells do not appear to invade the cyst wall. There have been no CT and MRI studies documenting enhancement of the cyst wall in DIG and A [18,19,20]. In our study population, biopsy of the cyst wall demonstrated no tumor cells in the gliotic cerebral cortex. Although the mechanisms underlying the formation of the large multiple cysts in DIG and A remains to be elucidated [5,11], it is clear that the cyst wall is free of tumor cell invasion. The most important therapeutic modality in young patients with DIG/A is a correct pre-operative diagnosis [1]. Biopsy based on a mistaken pre-operative diagnosis of malignant tumor must be avoided [1,15,21] because it may adversely affect the young patients’ clinical course and compromise the opportunity for total tumor removal [15,21]. Physicians must be familiarized with the clinical characteristics of DIG/A because the therapeutic gold standard is total removal under the pre-operative diagnosis of desmoplastic neuroepithelial tumor. Acknowledgement This study was supported in part by a Research Grant for Cancer [18–15] from the Japanese Ministry of Health, Labour and Welfare. References 1. Lellouch-Tubuiana MC, Salazar C, Cinalli G, Renier D, Sainte-Rose C, Pierre-Kahn A, Zerah M: The management of desmoplastic neuroepithelial tumors in childhood. Child’s Nerv Syst 16: 8–14, 2000 2. Rushing EJ, Rorke LB, Sutton L: Problems in the Nosology of desmoplastinc tumors of childhood. Pediatr Neurosurg 19: 57–62, 1993 3. Kepes JJ, Rubinstein LJ, Eng LF: Pleomorphic xanthoastrocytoma: a distinctive meningocerebral glioma of young subjects with relatively favorable prognosis. Cancer 44: 1839–1852, 1979 4. Payson RA: Gliofibroma: A distinctive entity of a subtype of desmoplastic astrocytoma. Human Pathol 27: 610–617, 1996 5. VandenBerg SR, May EE, Rubinstein LJ, Herman MM, Perentes E, Vinores SA, Collins VP, Park TS: Desmoplastic supratentorial neuroepithelial tumors of infancy with divergent differentiation potential (“desmoplastic infantile ganglioglioma”). J Neurosurg 66: 58–71, 1987 6. Taratuto AL, VandenBerg SR, Rorke LB: Desmoplastic infantile ganglioglioma and astrocytoma. In: Kleihues P, Cavenee WK (eds) Pathology & Genetics. Tumours of the Nervous System. IARC Press, Lyon, 2000, pp 99–102 7. Luis DN, von Deimling A, Dickerson GR, Dooling EC, Seizinger BR: Desmoplastic cerebral astrocytomas of

69

8.

9.

10.

11.

12.

13.

14.

15.

infancy: a histopathologic, immunohistochemical, ultrastructural, and molecular genetic study. Human Pathol 23: 1402–1409, 1992 de Chadarevian JP, Pattisapu JV, Faerber EN: Desmoplastic cerebral astrocytoma of infancy. Light microscopy, immunohistochemistry, and ultrastructure. Cancer 66: 173–179, 1990 Taratuto AL, Monges J, Lylyk P, Leiguarda R: Superficial cerebral astrocytoma attached to dura. Report of six cases in infants. Cancer 54, 2505–2512, 1984 Giannini C, Sheithauer BW, Burger PC, Brat DJ, Wollan PC, Lach B, O’Neill BP: Pleomorphic xanthoastrocytoma: do we really know about it? Cancer 85: 2033–2045, 1999 VandenBerg SR: Desmoplastic infantile ganglioglioma and desmoplastic cerebral astrocytoma of infancy. Brain Pathol 3: 275–281, 1993 Philippon JH, Clemenceau SH, Fauchon FH, Foncin JF: Supratentorial low-grade astrocytomas in adults. Neurosurgery 32: 554–559, 1993 Coons SW, Johnson PC: Regional heterogeneity in the proliferative activity of human gliomas as measured by the Ki-67 labeling index. J Neuropathol Exp Neurol 52: 609–618, 1993 Kros JM, Vecht CJ, Stefanke SZ: The pleomorphic xanthoastrocytoma and its differential diagnosis: A study of five cases. Human Pathol 22: 1128–1135, 1991 Duffner PK, Burger PC, Cohen ME, Sanford RA, Krischer JP, Elterman R, Aronin PA, Pullen J, Horowitz ME, Parent J, Martin P, Kun LE: Desmoplastic infantile

16.

17.

18.

19.

20.

21.

ganglioglioma: an approach to therapy. Neurosurgery 34: 583–589, 1994 Jay V, Edwards V, Rutka J, Mosskin M, Hwang P, Resch L: Unique desmoplastic cerebral tumor in a patient with complex partial seizures. Pediatr Dev Pathol 1: 234–242, 1998 Fan X, Larson TC, Jennings M, Tulipan NB, Toms SA, Johnson MD: December 2000: 6 month old boy with 2 weeks histology of progressive lethargy. Brain Pathology 11: 265–266, 2001 Horten BC, Rubinstein LJ: Primary cerebral neuroblastoma. A clinicopathological study of 35 cases. Brain 99: 735–756, 1976 Martin DS, Awwad EE, Pittman T: Desmoplastic infantile ganglioglioma: CT and MR features. AJNR 12: 1195–1197, 1991 Tenreiro-Picon OR, Kamath SV, Knorr JR, Ragland RL, Smith TW, Lau KY: Desmoplastic infantile ganglioglioma: CT and MRI features. Pediatr Radiol 25: 540–543, 1995 Duffner PK, Krischer JP, Horowitz ME, Cohen ME, Burger PC, Friedman HS, Kun LE, the Pediatric Oncology Group: Second malignancies in young children with primary brain tumors following treatment with prolonged postoperative chemotherapy and delayed irradiation: a Pediatric Oncology Group study. Ann Neurol 44: 313–316, 1998

Address for offprints: Kazuhiko Sugiyama, Department of Neurosurgery, Hiroshima University Medical Hospital,1-2-3Kasumi,Minamiku, Hiroshima 734-8551, Japan; Tel.: 81(Japan)-82-257-5227; Fax: 81(Japan)-82-257-5229; E-mail: [email protected]