TY - JOUR
T1 - MR spectroscopy and diffusion imaging in people with human immunodeficiency virus
T2 - Relationships to clinical and immunologic findings
AU - Morgello, Susan
AU - Buyukturkoglu, Korhan
AU - Murray, Jacinta
AU - Veenstra, Mike
AU - Berman, Joan W.
AU - Byrd, Desiree
AU - Inglese, Matilde
N1 - Publisher Copyright:
© 2021 American Society of Neuroimaging.
PY - 2022/1/1
Y1 - 2022/1/1
N2 - Background and purpose: People with human immunodeficiency virus (HIV; PWH) present a complex array of immunologic and medical disorders that impact brain structure and metabolism, complicating the interpretation of neuroimaging. This pilot study of well-characterized multi-morbid PWH examined how medical and immunologic factors predicted brain characteristics on proton MR spectroscopy (1H-MRS) and diffusion-weighted imaging (DWI). Methods: Eighteen individuals on combination antiretroviral therapy (cART), with mean age of 56 years, underwent medical history review, neuroimaging, and on the day of imaging, blood draw for assay of 20 plasma cytokines and flow cytometric characterization of peripheral blood mononuclear cell subsets. Predictors of n-acetyl aspartate, choline, myoinositol, glutamate/glutamine, fractional anisotropy and mean diffusivity were identified through bivariate correlation; those significant at p <.1000 were advanced to multivariate analysis, with models created for each neuroimaging outcome. Results: Monocyte subsets and diverse cytokines accounted for 16 of 25 (64%) variables predicting 1H-MRS spectra in frontal gray and white matter and basal ganglia; monocyte subsets did not predict any DWI characteristic. In contrast, age, presence of hypertension, and duration of HIV infection accounted for 13 of 25 (52%) variables predicting diffusion characteristics in the corpus callosum, thalamic radiations, and basal ganglia but only 3 of 25 (12%) predictors of 1H-MRS features. Conclusions: 1H-MRS neurometabolites were most often predicted by immunologic factors sensitive to temporal variation, whereas DWI metrics were more often related to longer-term disease state. In multi-morbid cART-era populations, selection and interpretation of neuroimaging modalities should account for complex temporal and pathogenetic influences of immunologic abnormality, disease state, and aging.
AB - Background and purpose: People with human immunodeficiency virus (HIV; PWH) present a complex array of immunologic and medical disorders that impact brain structure and metabolism, complicating the interpretation of neuroimaging. This pilot study of well-characterized multi-morbid PWH examined how medical and immunologic factors predicted brain characteristics on proton MR spectroscopy (1H-MRS) and diffusion-weighted imaging (DWI). Methods: Eighteen individuals on combination antiretroviral therapy (cART), with mean age of 56 years, underwent medical history review, neuroimaging, and on the day of imaging, blood draw for assay of 20 plasma cytokines and flow cytometric characterization of peripheral blood mononuclear cell subsets. Predictors of n-acetyl aspartate, choline, myoinositol, glutamate/glutamine, fractional anisotropy and mean diffusivity were identified through bivariate correlation; those significant at p <.1000 were advanced to multivariate analysis, with models created for each neuroimaging outcome. Results: Monocyte subsets and diverse cytokines accounted for 16 of 25 (64%) variables predicting 1H-MRS spectra in frontal gray and white matter and basal ganglia; monocyte subsets did not predict any DWI characteristic. In contrast, age, presence of hypertension, and duration of HIV infection accounted for 13 of 25 (52%) variables predicting diffusion characteristics in the corpus callosum, thalamic radiations, and basal ganglia but only 3 of 25 (12%) predictors of 1H-MRS features. Conclusions: 1H-MRS neurometabolites were most often predicted by immunologic factors sensitive to temporal variation, whereas DWI metrics were more often related to longer-term disease state. In multi-morbid cART-era populations, selection and interpretation of neuroimaging modalities should account for complex temporal and pathogenetic influences of immunologic abnormality, disease state, and aging.
UR - http://www.scopus.com/inward/record.url?scp=85114864632&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85114864632&partnerID=8YFLogxK
U2 - 10.1111/jon.12931
DO - 10.1111/jon.12931
M3 - Article
C2 - 34520593
AN - SCOPUS:85114864632
SN - 1051-2284
VL - 32
SP - 158
EP - 170
JO - Journal of Neuroimaging
JF - Journal of Neuroimaging
IS - 1
ER -