DOI: 10.1002/alz.081815 ISSN: 1552-5260

Characterizing the effects of chemotherapy on the brain of patients with head and neck cancer using FDG PET

Steven Bishay, Trent Schwartz, David Samuel Smith, Lok Hin Lee, Eben Rosenthal, Timothy J. Hohman, Angela L. Jefferson, Mary Ellen Koran
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



Chemotherapy has known adverse effects on the brain. From small (N<30) studies primarily in breast cancer, we know that post‐chemotherapy brain changes can be detected with 18F ‐fluorodeoxyglucose positron emission tomography (FDG PET), but this has been relatively unexplored in other cancers, and has not been explored in patients with head and neck cancer (HNCa) (Chen, 2018). Patients with HNCa have 2‐3 times the risk of dementia after therapy compared to controls and may be particularly vulnerable given underlying lower levels of education and heavy tobacco and alcohol use (Zer, 2018). We will determine which areas of the brain are most vulnerable to the chemotherapy in HNCa with FDG PET.


Patients with HNCa were identified from Vanderbilt’s Cancer Registry, limited to patients with FDG PET in the last year (2021). Clinical data was mined from the electronic health record. Images were processed with well‐established SPM12 methods (Friston, 1995). A voxel‐wise ANCOVA was performed on FDG PET images before and after chemotherapy, including covariates of age, sex, ethnicity, and time between chemotherapy and follow‐up imaging using conservative family wise error of <0.05 and voxel extent threshold of 10.


68 patients with HNCa met inclusion criteria and all patients received Carboplatin‐based chemotherapy (Table 1). Compared to baseline, 17 clusters met conservative significance thresholds with decreases in glucose metabolism following treatment (Figure 1). These included large regions (46397 1.5mm3 voxels) in the frontal and temporal lobes (pFWE = <0.001). Specifically, there was significantly decreased glucose metabolism in the dorsolateral prefrontal cortex (BA 9), frontoparietal cortex (BA 10), orbital frontal cortex (BA 11), medial and superior temporal gyri (BA 21, 22), dorsal (BA 24), and the subgenual (BA 25), pregenual (BA 32), and rostral (BA 33) anterior cingulate cortex post‐chemotherapy.


To our knowledge, this is the first study to explore the effects of chemotherapy in HNCa, a population particularly vulnerable to these effects. We found that regions in the forebrain and temporal lobes are most vulnerable. Studies like these are important to establish and confirm spatial distributions of these effects which will be crucial for development of diagnostics and monitoring tools for these effects.

More from our Archive