Evaluation of Pediatric CNS Malignancies with ^99mTc-Methoxyisobutylisonitrile SPECT

MATERIALS AND METHODS

RESULTS

A total of 29 pairs of MRI and ^99mTc-MIBI scans from 20 children with CNS malignancy were included in the study. Age, sex, diagnosis, and tumor location are summarized in Table 1. The mean age of the children was 9.9 ± 4.4 y, and the age range was 5–17 y. This population included 12 boys and 8 girls.

A wide variety of pediatric brain neoplasms took up ^99mTc-MIBI. As summarized in Table 2, these included brain stem gliomas, pilocytic astrocytoma, glioblastoma multiforme, and choroid plexus carcinoma. Tumors with no demonstrable uptake of ^99mTc-MIBI when tumor was evident on MRI included optic glioma, medulloblastoma, desmoplastic neuroepithelial tumor, craniopharyngioma, spinal anaplastic astrocytoma, low-grade astrocytoma, and ependymoma. Therefore, of the 29 sestamibi studies completed, 13 were true-positive, 13 were false-negative, and 3 were true-negative when positive MRI findings were used as definitive evidence of tumor. A false-negative result indicated either that no significant amount of radiotracer was taken up by the tumor or that the radiotracer had subsequently been expelled from the tumor at the time of imaging. In no instance was a positive ^99mTc-MIBI result accompanied by a negative MRI result.

An example of a positive sestamibi scan is shown in Figure 1, which depicts both the gadolinium-enhanced T1- and T2-weighted MR images and the ^99mTc-MIBI SPECT images for patient 9, a 13-y-old girl with a glioblastoma multiforme tumor of the right frontotemporal region. In this patient, a close morphologic correlation was observed between the positive ^99mTc-MIBI image and both MR images. MR and SPECT images of a brain stem glioma can be seen in Figure 2.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Glioblastoma multiforme of right frontotemporal region, revealed on T1-weighted MRI with gadolinium (A) and T2-weighted MRI (B), is also well defined on sestamibi scan (C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Serial pairings of sagittal MRI and SPECT images for 15-y-old boy with exophytic brain stem glioma. Third column is black-and-white version of original SPECT images. Baseline images (A) are compared with scans at 8 (B) and 12 (C) mo, both of which show increased signal on ^99mTc-MIBI SPECT but no change on MRI. Final panel (D) demonstrates increased tumor on MRI at 16 mo, accompanied by decreased signal from SPECT.

Astrocytomas constituted the largest proportion of tumors in this group of children (23 pairs of scans from 15 children). Uptake of sestamibi appeared to depend in part on the histologic grade of the tumor. Of the 9 patients with intracranial astrocytomas for which histologic grading was available, 4 failed to demonstrate any positive sestamibi scans. In all of these cases, tumors were of grade I or II. Conversely, both patients with grade IV astrocytomas had positive sestamibi scans. Furthermore, of the remaining 7 nonastrocytoma tumors, the only positive sestamibi scans were in a patient with a high-grade (III) choroid plexus carcinoma.

Brain stem gliomas were the most consistently positive type of tumor in our limited sample, with 5 of 8 cases demonstrating uptake of the sestamibi tracer. All 3 of the brain stem gliomas that had negative sestamibi scans were located in the midbrain.

Normal cranial structures found to take up ^99mTc-MIBI in this group of patients included the choroid plexus and the pituitary gland. In none of the 20 tumors imaged did this uptake by normal tissues interfere with the interpretation of the scan. Only 1 case of spinal cord tumor was included, and it failed to show up on SPECT despite its high grade.

Multiple pairs of MR and ^99mTc-MIBI images were available for 6 children. A summary of the respective changes in both modes of imaging, the time course of these changes, and the progression or nonprogression of their illness is included in Table 3. In 2 of these 6 patients, sestamibi images remained negative at all times. In the remaining 4 patients, sestamibi activity was found to correlate well with changes found on MRI over the same period. Furthermore, in patient 3, increased tumor was noted on 2 consecutive sestamibi scans when MRI done at the same time revealed no interval change. On the final pairing for this patient, this trend was reversed: Increased tumor was noted on MRI, whereas the sestamibi scan showed decreased tumor. This progression is illustrated in Figure 2, which shows pairings of MR and SPECT images from time zero through 8, 12, and 16 mo, respectively. Throughout these changes, clinical evidence of disease progression was not seen until a few days before this child died. Therefore, the sestamibi scan was the first to show tumor progression, followed by MRI and finally by clinical signs.

DISCUSSION

Our results demonstrate the potential of ^99mTc-MIBI to image many pediatric brain tumors. A positive SPECT scan using this tracer appears to depend in part on tumor histology, including both the type and the histologic grade of the tumor. Although MRI appears to be superior in many cases, the lack of consistent correlation that we observed between MRI and SPECT suggests that unique information may be available from each mode of imaging.

^99mTc-MIBI is established as a marker of CNS malignancy (2,11,14,16,24–27), but further data are necessary to clarify the role in specific tumor types and the clinical value when these patients are followed over time. In our series, 9 of 20 lesions were positive, including both low-grade and high-grade astrocytic tumors. This finding is consistent with adult data showing positivity for multiple tumor types, including astrocytoma (11,20) and meningioma (11,13,22). However, in our study group, 11 of 20 tumors evident on MRI were not evident on SPECT. Those diagnoses included craniopharyngioma, ependymoma, desmoplastic neuroepithelial tumor, medulloblastoma, and some low-grade gliomas. This finding suggests a selectivity in tracer uptake or extrusion, which might be due to several factors. Regional blood flow, blood-brain barrier (BBB) permeability, tumor cell membrane adenosine triphosphatase pump, and mitochondrial content may influence uptake of radiotracers (5). Tracer needs to reach the tumor cells first through the bloodstream, and a correlation between cerebral blood flow and ^99mTc-MIBI reactivity has been shown previously (27). The tracer must then leave the vascular lumen and cross the BBB. The inability of ^99mTc-MIBI to enhance normal structures within the brain aside from the choroid plexus supports the importance BBB integrity.

In addition to being sealed by tight junctions, BBB endothelial cells also express the multidrug resistance 1 gene, the product of which is an adenosine triphosphatase membrane pump extruding a variety of toxins from the cells. ^99mTc-MIBI is one of these substrates (16). Inhibition of multidrug resistance has been shown to delay excretion of ^99mTc-MIBI (28). Without inhibition, the pump prevents the tracer from reaching the interstitial space. Choroid plexus tissue, the vasculature of which expresses neither tight junctions nor multidrug resistance 1, accumulates ^99mTc-MIBI, supporting this concept. Consequently, tumors arising from this tissue are ^99mTc-MIBI positive as well (14).

The next barrier that the tracer must pass to enter tumor cells is the cell membrane. Passage through the cell membrane may be influenced by several factors, including multidrug resistance, mentioned above. In addition, both regional blood flow distribution and mitochondrial oxidation capacity correlated well with ^99mTc-MIBI uptake in cardiac myocytes (29,30). Intracellular parameters may also determine ^99mTc-MIBI accumulation, which occurs in both mitochondria and cytosol, as well as in association with specific cellular proteins (31). In tumor cell lines, most ^99mTc-MIBI has been demonstrated to associate with the mitochondria (32). Therefore, metabolically active and highly malignant tumors may be more strongly positive than are lower-grade tumors. Our finding of no positivity in craniopharyngioma and some low-grade tumors, combined with well-established positivity in high-grade glioma, fits this picture. However, other investigators have also demonstrated negativity in ependymoma and medulloblastoma (26), tumors that are both well vascularized and metabolically active, suggesting that the answer lies beyond tumor grade alone.

Similarly, the variable results we observed in low-grade gliomas, only half of which were positive for tracer uptake, merit further investigation. These tumors are frequently located in sites where complete surgical resection is difficult, such as the optic chiasm, basal ganglia, or brain stem. The biologic behavior of brain tumors remains difficult to predict by histology alone, and the success of nonsurgical treatments is limited. Therefore, clarifying the nature of the variability of SPECT findings in pediatric low-grade glioma has the potential to add a clinically valuable piece of information to allow a more accurate prediction of clinical behavior over time.

The effect of previous surgery, radiation, or chemotherapy on ^99mTc-MIBI uptake is poorly established in the literature. In pediatric brain tumors, the progression of tumor at the edge of a previously resected site may be preferentially imaged by ^99mTc-MIBI as opposed to ²⁰¹Tl (2). Isolated case reports suggest that the ^99mTc-MIBI signal is decreased in tumor cells after radiation treatment (27,33). Decreased SPECT activity after therapy despite no change in tumor size by CT or MRI has been shown (27), suggesting that such imaging may be a more sensitive measure of response to treatment. Our results from 6 patients for whom consecutive pairs of images were available are insufficient to significantly support an advantage of sestamibi scans. However, in 1 patient with high-grade glioma, progression on ^99mTc-MIBI appeared to precede any evidence of progression on MRI or clinically. Therefore, similarities in the information available from SPECT may exist between adult and pediatric patients with brain tumors.

CONCLUSION

Our results demonstrate the potential uses of ^99mTc-MIBI SPECT in the imaging of astrocytic pediatric brain malignancy. We suggest that larger studies be undertaken to better assess the clinical utility of this tool, particularly in low-grade gliomas of childhood.