Neuropediatrics 2000; 31(3): 159-161
DOI: 10.1055/s-2000-7489
Letter to the Editor

Georg Thieme Verlag Stuttgart · New York

Malignant Rhabdoid Tumor of the Brain: Quantitative 1H MR-Spectroscopy and Cytogenetics

T. A. G. Huisman1 , S. Brandner2 , F. Niggli 3 , D. R. Betts3 , E. Boltshauser4 , E. Martin1
  • 1 Department of Diagnostic Imaging and Radiology, University Children's Hospital Zurich, Switzerland
  • 2 Institute of Neuropathology, University Hospital Zurich, Switzerland
  • 3 Department of Oncology, University Children's Hospital Zurich, Switzerland
  • 4 Department of Neurology, University Children's Hospital Zurich, Switzerland
Weitere Informationen

Publikationsverlauf

Publikationsdatum:
31. Dezember 2000 (online)

Sir,

Malignant rhabdoid tumor (MRT) is a rare, highly malignant tumor in childhood which was originally described as a neoplasm originating in the kidney [[1]]. The tumor was originally believed to be a rhabdomyosarcomatoid subtype of Wilms' tumor. Studies of Haas et al [[4]] characterised MRT as an entity distinct from Wilms' tumor. As light microscopy findings were suggestive of myoblastic differentiation, the term rhabdoid was chosen. Immunohistochemical and ultrastructural findings however did not support a muscle cell origin. Since its recognition as an own tumor entity in 1978, primary MRT has been reported at almost any site of the body including liver, limbs, pelvis, heart, skin, tongue, uterus, prostate, urinary bladder, orbit and central nervous system (CNS) [[5], [6], [8], [9], [10]]. In addition, a possible association between MRT of the kidney with different CNS tumors was reported [[3]]. The histogenesis of MRT is still unclear. Multiple theories have been proposed. Observations of Bonnin et al [[3]] with a greater than expected association between renal MRTs and primitive neuroectodermal tumors of the CNS could support the theory that MRTs are neuroectodermal in origin. The diversity of locations of primary MRTs in the body favors however a mesenchymal origin [[9]]. Weeks et al [[12]] proposed that MRT represents a histogenetic and clinical entity with considerable morphological diversity, and that extrarenal MRT will emerge as a phenotypic concept encompassing a spectrum of histogenetic and clinical diversity. Cytogenetic studies, however, indicate that renal rhabdoid tumor and CNS rhabdoid tumor may be different [[6]].

Radiologically, cerebral MRT can be indistinguishable from primitive neuroectodermal tumor (PNET) which is the most common malignant central nervous system tumor occurring during the first decade of life [[6]].

As the biological features and outcome of MRT and PNET are dramatically different, a correct and early diagnosis is essential for adequate therapy. Magnetic resonance imaging (MRI) constitutes one of the primary, non-invasive tools in the work-up of these children. As 1H-magnetic resonance spectroscopy (MRS) of the tumor tissue can be performed in the same session, the additional information of MRS could possibly be helpful in a non-invasive differentiation between MRT and PNET.

We performed MRI and MRS in a 2-years and 4-months old boy with acute left hemiparesis, dysarthria and intermittent headache. MRI (Fig. [1 a] , b) showed a large partially cystic, partially solid tumor within the parieto-occipital white matter which proved to be an MRT by histology. Imaging characteristics did not allow a differentiation between PNET and MRT. Quantitative 1H-MRS (Point-resolved spectroscopy, PRESS, TR =  6000 msec, TE = 135 msec, Voxel 1.5 ml, calibration by external phantom with creatine 50 mmol/l, 2.0 Tesla Bruker MRI-Unit, Karlsruhe, Germany) of the solid tumor component showed an increased concentration (age-matched) of cholin (2.4 mmol/l). This could indicate an increased membrane turnover. In addition, an absent creatine- and N-acetyl-aspartate peak confirming the lacking neuronal differentiation of the tumor-cells, and a high concentration of intratumoral lipids which is usually seen in highly malignant CNS tumors like glioblastoma multiforme, support the highly malignant nature of the tumor. Future series of q1H-MRS should prove whether these results or so-called “fingerprint” are characteristic for MRT or whether in analogy to the imaging findings a wide variability exists in q1H - MRS of MRT.

Fig. 1 a Axial T1 spin-echo MRI, TR = 600 msec, TE = 20 msec. Partially cystic (hypointense) and solid tumor within the right hemisphere. The solid component appears isointense to the gray matter. A small vessel with flow-related enhancement can be seen in the center of the tumor (arrow).

A histological differentiation between MRT and PNET can also be difficult as MRTs are characterized by nests of rhabdoid cells next to fields that resemble classic PNETs [[6]]. To avoid a misdiagnosis, the pathologist should be provided with a large enough tumor specimen to give him the possibility to encounter the rhabdoid cells next to the PNET areas. Diagnosis can be facilitated with monoclonal antibody examinations using immunoperoxidase techniques. The three antibodies of greatest value are epithelial membrane antigen (EMA), vimentin and smooth muscle actin [[6]]. Positivity for this triad is not found in typical PNETs, in fact, EMA is essentially never positive in PNETs. Accordingly, our child showed a prominent immunoreactivity for EMA, vimentin and smooth muscle actin. In contrast to previous findings, a significant immunoreactivity for myoglobin was seen in our case. The exact meaning of this finding remains unclear.

Cytogenetic analysis of the tumor specimen revealed a balanced translocation between the long arms of chromosome 12 and 22, t (12;22) (q24;q11), (Fig. [2]). Cytogenetic and molecular analyses have shown that deletion of region 11.2 of the long arm of chromosome 22 (22 q11.2) is a recurrent genetic characteristic of MRT, indicating that this locus may encode a tumor suppressor gene [[11]]. Abnormalities of chromosome 22 have been seen in a variety of CNS tumors, suggesting that the putative mechanism of tumorogenesis is inactivation of a still to be defined tumor suppressor gene located in the distal long arm of chromosome 22 [[2], [7]]. Abnormalities of the q-arm terminal region of chromosome 12, as seen in our patient, are however rare in brain tumors. Only one case was recently reported with an identical reciprocal translocation t(12;22) (q24.3; q11.2 - 12) in MRT of the brain. The identification of this subtle translocation further sublocalizes the chromosomal breakpoint and could be of potential diagnostic value in cerebral MRTs [[8]].

Fig. 2 Cytogenetic analysis of tumor tissue revealed a balanced translocation between the long arms of chromosome 12 and 22.

In conclusion: MRI of MRT is non-specific, PNETs can present with similar imaging findings. Quantitative 1H - MRS could possibly narrow the differential diagnosis. Our findings of q1H - MRS support the lacking neuronal differentiation of the tumor cells. Larger series of q1H - MRS in MRT have to show if a characteristic 1H-spectrum (fingerprint) exists for MRT. Neuropathological examination is still mandatory to define the histology of the tumor. This case is the first to be reported with a significant positive immunohistochemistry for myoglobin. Future reports and investigations could possibly explain the meaning of this finding. Our patient is the second to be reported with a balanced translocation between chromosomes 12 and 22.

References

  • 1 Beckwith J B, Palmer N F. Histopathology and prognosis of Wilms' tumor. Results from the First National Wilms' Tumor Study.  Cancer. 1978;  41 1937-1948
  • 2 Biegel J A, Parmiter A H, Sutton L N, Rorke L B, Emanuel B S. Abnormalities of chromosome 22 in pediatric meningeomas.  Genes, Chromosomes and Cancer. 1994;  9 81-87
  • 3 Bonnin J M, Rubinstein L J, Palmer N L, Beckwith J B. The association of embryonal tumors originating in the kidney and in the brain.  Cancer. 1984;  54 2137-2142
  • 4 Haas J E, Palmer N F, Weinberg A G, Beckwith J B. Ultrastructure of malignant rhabdoid tumor of the kidney.  Hum Pathol. 1981;  12 646-657
  • 5 Hanna S L, Langston J W, Parham D M, Douglass E C. Primary malignant rhabdoid tumor of the brain: Clinical imaging and pathologic findings.  AJNR. 1993;  14 107-115
  • 6 Rorke L B, Packer R J, Biegel J A. Central nervous system atypical teratoid/rhabdoid tumors of infancy and childhood: definition of an entity.  J Neurosurg. 1996;  85 56-65
  • 7 Rubio A. 4-year old girl with ring chromosome 22 and brain tumor.  Brain Pathology. 1997;  7 1027-1028
  • 8 Sawyer J R, Goosen L S, Swanson C M, Tomita T, de Leon G A. A new reciprocal translocation (12; 22) (q24. 3; q11. 2 - 12) in a malignant rhabdoid tumor of the brain.  Cancer Genet Cytogent. 1998;  101 62-67
  • 9 Sotelo-Avila C, Gonzalez-Crussi F, deMello D, Vogler C, Gooch W M, Gale G. et al . Renal and extrarenal rhabdoid tumors in children: a clinicopathologic study of 14 patients.  Semin Diagn Pathol. 1986;  3 151-163
  • 10 Tortori-Donati P, Fondelli M P, Rossi A, Colosimo C, Garre M L. et al . Atypical teratoid rhabdoid tumors of the central nervous system in infancy.  Int J Neuroradiol. 1997;  3 327-338
  • 11 Versteege I, Sevenet N, Lange J, Rousseau-Merck M F, Ambros P, Handgretinger R. et al . Truncating mutations of hSNF5/INI1 in aggressive paediatric cancer.  Nature. 1998;  394 203-206
  • 12 Weeks D A, Beckwith J B, Mierau G W. Rhabdoid tumor. An entity or a phenotype?.  Arch Path Lab Med. 1984;  113 113-114

M. T. A. G. Huisman

Neuroradiology Section and MGH-NMR Center Massachusetts General Hospital and Harvard Medical School Mailcode CNY 149-2301

Boston, MA 02119

USA