Pediatric Brain Tumors: Magnetic Resonance Spectroscopic Imaging

Abstract

The incidence of cancer in the United States in children under 15 years of age has risen in recent years (Ries et al., 1991). This is largely due to the increased incidence of lymphoblastic leukemia and tumors of the brain and nervous system, as opposed to Wilms tumors, soft tissue and bone sarcomas, lymphomas and Hodgkin's disease or other malignancies of childhood. Between 1973 and 1988, the incidence of childhood nervous system tumors jumped by 30% (Bleyer, 1993). Every year, more than 1,500 children are diagnosed with brain tumors (Pollack, 1994). Because a child is more likely to develop cancer during the first 5 years of life, the etiology of these early cancers is likely different from those later in life and of different factors. While childhood tumors are more aggressive, their long-term control is often possible (Albright, 1993). Cancer Statistics Review reports an overall decrease in childhood cancer mortality, although brain and nervous system cancer deaths have decreased less than those due to other malignancies (Ries et al., 1991). These results indicate the need for useful in vivo biomarkers to allow the evaluation of treatment protocols for pediatric brain tumors. Given the inherent difficulties of sequential biopsies to monitor therapeutic response in children with brain tumors, non-invasive and nonirradiating imaging methods are needed to provide additional diagnostic indices or biomarkers beyond simple tumor volume measurements. Brain tumor treatment in most modern centers is managed through a tumor board, which typically rely in part on available proton Magnetic Resonance Spectroscopic Imaging (MRSI) results, especially for inoperable tumors that can be difficult to biopsy. Additionally, where progression or treatment response is questioned, serial in vivo MRSI is preferred over serial biopsy or PET/SPECT, which is irradiating, expensive and often unavailable. Non-invasive and non-irradiating in vivo MRSI can be performed as an adjunct to Magnetic Resonance Imaging (MRI), and is thus cost effective and the method of choice in children under 5 years, when radiation is a serious concern. Although MRSI does not obviate the utility of biopsy, it is suggested that it has the potential to replace serial biopsy and is an excellent alternative to biopsy in inoperable or unbiopsied tumors. The MRSI data when combined with anatomical or other type MR images provide unique information regarding brain tumor biochemistry in inoperable tumors and, might complement neuropathology, guide biopsies, and monitor therapy for operable brain tumors. The combination of such non-invasively acquired prognostic information and the high-resolution anatomical imaging provided by