Kensuke Tateishi

Nobutaka Kawahara

Tissue hypoxia is a typical feature of highly malignant tumors, and glioblastoma, the most malignant tumor in the central nervous system, is not an exception. Indeed, glioblastomas are characterized by necrotic regions that distinguish them from less malignant gliomas. In addition, treatment-resistant cancer stem cell–like cells are supposed to reside and survive within hypoxic regions. Because of recent advances in tracer research, PET is now being used to depict tissue hypoxia in vivo.

Therefore, if the hypoxic regions in gliomas can be detected by PET, it would provide crucial clinical information for treatment. Although several studies using nitroimidazole analogs, such as FMISO and FRP-170, have recently been applied to assess tissue hypoxia in gliomas, we focused on using Cu-diacetyl-bis (N4-methylthiosemicarbazone) (Cu-ATSM), which has different retention mechanisms in hypoxic cells with a potentially higher signal-to-noise ratio. Using this tracer, we could demonstrate hypoxic regions within gliomas, which were confirmed by the expression of hypoxic inducible factor-1α (HIF-1α), a biomarker for hypoxia. We also showed that hypoxic regions depicted by Cu-ATSM PET were highly specific in glioblastoma (WHO grade IV), and could not be obtained using ¹¹C-methionine (MET) or ¹⁸F-fluorodeoxyglucose (FDG) PET.

Our findings have several implications that may influence our clinical decision-making. First, our findings would help in efficient and early

diagnosis. In some glioblastomas located in the eloquent regions, tissue biopsy is performed to confirm diagnosis. In such procedures, targeting the hypoxic region for biopsy would greatly increase the reliability of pathologic diagnosis and grading, particularly in secondary glioblastoma. Second, our findings would help in preferential surgical resection of these hypoxic regions when total resection of enhanced tumor is not possible because of potential neurologic deficits. This preferential resection would decrease the remaining tumor cells that are resistant to postoperative chemotherapy and radiation therapy.

The current findings provide an important insight on the tumor biology of glioblastoma. We generally assume that glioblastomas would be heterogeneous tumors with different metabolic, proliferative, or genetic components; however, little is known about whether hypoxic regions depicted by this imaging are different from the other regions in these aspects. To answer these questions, we are currently analyzing whether the hypoxic region is metabolically distinct using MET and FDG-PET. It would also be required to examine whether these regions really harbor highly proliferative or tumor-initiating cells that are resistant to chemoradiation in the clinical setting. We are conducting several studies in these directions, and hope that hypoxic imaging studies, including Cu-ATSM PET, will contribute to elucidation of the mechanism of treatment resistance in glioblastoma.

Kensuke Tateishi, MD
Nobutaka Kawahara, MD, PhD

Read this article at AJNR.org . . .