Secondary Logo

Journal Logo


Imaging correlates of the blood–brain barrier disruption in HIV-associated neurocognitive disorder and therapeutic implications

Chaganti, Jogaa; Marripudi, Karthika; Staub, Lukas P.b; Rae, Caroline D.c; Gates, Thomas M.d; Moffat, Kirsten J.a; Brew, Bruce J.a,d

Author Information
doi: 10.1097/QAD.0000000000002300
  • Free



HIV type-1 (HIV) invades the central nervous system (CNS) early after primary infection leading eventually to HIV-associated neurocognitive disorder (HAND) in approximately 30–50% of infected individuals even in the context of sustained viral suppression with combined antiretroviral therapy (cART era) [1,2]. Neuroradiological, neuropathological and cerebrospinal fluid (CSF) studies have demonstrated that blood–brain barrier (BBB) compromise is a consistent feature in HAND correlating with both its presence and severity [3,4]. However, these data are largely derived from patients who had not been treated with cART and who had ongoing viraemia. Although it is known that cART can improve and sometimes reverse BBB compromise at least by CSF and standard MRI measures, the extent to which this occurs using more sensitive MRI techniques and the relationship to the presence and severity of HAND in the context of viral suppression is unknown [5].

The BBB disruption in HAND is challenging to assess, as the degree of compromise can be very mild and the conventional tools to measure the disruption have limited sensitivity. Dynamic contrast enhanced (DCE) MRI is a novel technique that has been shown to be a powerful and sensitive tool to measure the integrity of the BBB. K-trans, a metric derivative of DCE, is a marker of capillary permeability. It is being used widely in various diseases, such as brain tumour multiple sclerosis and Alzheimer's disease, to demonstrate subtle disruptions [6–11]. The metrics of BBB disruption can be obtained using different models. At a very low level of disruption of the BBB, the two-compartment Patlock model is considered to be the most robust method to generate reproducible K trans results [12,13].

We hypothesized that K-trans would detect compromise of the BBB in incident HAND patients even with viral suppression in both CSF and blood and would correlate with CSF neuroinflammatory biomarkers, neuropsychological measures and markers of neuroinflammation from magnetic resonance (MR) spectroscopy derived metabolites [14–18].

Materials and methods

Twenty HIV patients with HAND and an equal number of age and sex-matched controls (all white men age fifth decade) were enrolled. All patients had been on a stable cART regimen for at least the previous 12 months (All the patients received cART, with minor changes in the treatment regimens. The regimen included protease inhibitor duranavir or ritunavir, NRTI abacavir or lamuvudine, integrase inhibitor raltegravir) and all had HIV RNA levels less than 20 copies/ml in both blood and CSF. All had recent onset of HAND symptoms within the past 6 months. All were assessed by a neurologist and all had blood and CSF analyses to exclude other causes of cognitive impairment. All had albumin, beta-2 microglobulin and neopterin measured in both blood and CSF. All were diagnosed with HAND by the American Academy of Neurology criteria [19]; 18 patients were diagnosed utilizing a brief research battery [20,21]. In cases wherein results were inconclusive for HAND with the brief research battery, those were referred for full neuropsychological assessment to confirm the diagnosis. The remaining two patients were known to have HIV-associated dementia (HAD) based on clinical neurological review (BJB) and complete neuropsychological assessment. Out of these 20 patients with HAND, seven were identified with mild neurocognitive disorder (MND); nine patients with asymptomatic neurocognitive impairment (ANI) and four patients with HAD. Individuals with a prior history of drug abuse, significant head injury, psychiatric illness and hepatitis C virus (HCV) infection were excluded. None of the patients had any white matter abnormalities or microhaemorrhages on anatomical MRI protocols. Twenty HIV seronegative age-matched controls were included in the study. These controls were recruited from SVH radiology/medical imaging department. Their neurocognitive functioning was not assessed but assumed to be at least above average for age given their educational and occupational attainment. Local ethics approval was obtained from Hospital Human Research Ethics Committee and all the participants provided written informed consent prior to enrolment (Table 1).

Table 1:
Demographic, HIV and neurocognitive characteristics of study sample.

Neurocognitive evaluation

Neurocognitive testing was performed using a concise research test battery described previously that has been validated for use in the HIV-positive population [19,20]. Briefly, the test battery consisted of a short-computerized battery (CogState) supplemented with a small selection of standardized pencil-and-paper tests to cover five cognitive domains (Table 2). CogState is sensitive to HAND [22] and uses playing cards as stimuli to minimize the influences of culture and education on testing. Other tests were selected to provide greater coverage of other cognitive domains affected by HAND (psychomotor speed and motor coordination). In addition, premorbid intellectual functioning was estimated using National Adult Reading Test (NART). Psychological measures included depression, anxiety and stress scales (DASS- 21) and mini international psychiatric interview (MINI v5.0.). Functional decline for the purposes of HAND status classification was measured via a standard Independence of activities of daily living (IADL) questionnaire [23].

Table 2:
Tests included in brief neurocognitive research battery.

Neuroimaging measures

All imaging was performed with a 3.0-Tesla MRI (Ingenia Philips Healthcare, Best, The Netherlands) using a 24-channel head coil. Routine anatomical scanning using volumetric T1, T2, fluid-attenuation inversion recovery, diffusion weighted imaging and intracranial 3D time of flight magnetic resonance angiography was acquired. Volumetric T1 protocol was performed with following parameters 3DT-1 spoiled gradient recalled acquisition in steady state (SPGR): 128 sagittal slices, 1 mm isotropic, SPGR time to repeat/time to echo (TR/TE): Shortest, field of view: 240, Matrix: 256/256.

Dynamic contrast enhanced perfusion MRI

The DCE-MRI sequence was obtained using 3D T1-weighted spoiled gradient echo sequence in the axial plane covering the entire brain [TR and TE = shortest (Act TR/TE 15/3.0 ms, temporal resolution 5.8, flip angle = 15°, matrix = 184 x 141, number of slices = 23, slice thickness = 4 mm, number of signal averages = 1, temporal resolution = 5.8/dynamic, number of dynamics = 90,and scanning time = 9.06 min]. Contrast injection was commenced 6 s after the start of the dynamic MRI acquisitions, given in the form of a bolus injection of gadobutrol (Gadovist, Bayer, California, USA) at a concentration of 0.1 mmol/kg of body weight at 3 ml/s. Following the DCE-MRI scan, postcontrast-enhanced volumetric T1-weighted images were acquired as part of the routine clinical examination.

Magnetic resonance spectroscopy

We also obtained single-voxel magnetic resonance spectroscopy (MRS) in the basal ganglionic region and anterior frontal white matter at the ventricular level. MRS: Single-voxel 1H spectra were acquired using a short TE PRESS sequence (TE/TR 30/1800 ms, bandwidth 2000 Hz, 2048 data points). A voxel size of 30/15/10 mm (AP/RL/FH) was prescribed in two regions: anterior frontal centrum semiovale (white matter) at the level of the frontal horns of the lateral ventricles, and right and left) and basal ganglia (deep gray matter). Field homogeneity and water suppression were adjusted using automated algorithms from Philips. The spectra were obtained with TE/TR = 30/1800 ms, flip angle of 90°, bandwidth = 2000 Hertz, 128 averages.

Image processing

Dynamic contrast enhanced perfusion

DCE studies were processed with nordicICE [nordicICE (NICE) 4.0.4; NordicNeuroLab, Bergen, Norway], a propriety software that includes brain extraction, motion correction and image registration. NordicICE is a proprietary software and was employed to measure the DCE-derived metric K trans. The software has inbuilt features to correct the leakage correction and removal of negative slope values (values below zero), which were used to offset the blood plasma volume intercept. A two-compartment pharmacokinetic model was applied in the region of interests by using the Patlak graphical approach based on linear fitting of scatter plots [12,13], which was found to be the most appropriate model in a low-leakage regimen [12,13]. This Patlak graphical approach provided the BBB leakage rate and the local blood plasma volume. The slope of this fit is the BBB leakage rate (assuming a tissue density of 1 g/ml), and the intercept is the local blood plasma volume. All the patients were tested for fit of the model (chi-square goodness of model fit). Moreover, at very low-level BBB disruption, the flow is not a significant contributor to K trans (cf. to enhancing MS plaques and tumours) [24,25].

The K-trans images were interrogated by placing multiple regions of interest (ROI) in the following areas of the brain [basal ganglia (caudate and lentiform nucleus), frontal white matter, thalami, splenium of corpus callosum, occipital white matter and posterior limb of internal capsule] by two radiologists, one with 25 years of experience and one with 1 year of experience (JC, KM) (Fig. 1). The volumes of the regions of interest were 0.7 μl and whenever the area was smaller due to volume loss, the ROI was adjusted to reduce the effects of CSF (Fig. 1). After extraction of negative values, the remaining cumulative sum of the bins was defined as the BBB leakage volume fraction (Figs. 2 and 3 a,b). K-trans values were obtained from identical regions of the brain from the opposite hemispheres and the average values were taken to compare with the normal controls (e.g. average values of right and left caudate and lentiform nuclei and bilateral deep frontal white matter) (Table 3).

Fig. 1:
Areas of interrogation K-trans maps.
Fig. 2:
Dynamic contrast enhancement curves for patient with HAD: in the basal ganglia (a) and frontal white matter (b) demonstrating the temporal time course of the signal intensity changes represented with green line and yellow line showing the ‘fit of the model’.
Fig. 3:
(a,b) Distribution of K-trans values [(y-axis) and the number of voxels (x-axis)] in the frontal white matter and caudate nucleus in patient with ANI.
Table 3:
Distribution of K-trans in the patients/controls and the literature values in the basal ganglia and frontal white matter.


All spectra were taken for analysis with LCModel (ver: LCModel: 6.3–11) using water scaling option. Raw data were selected when the metabolites could be accepted for further statistical analysis only if the standard deviation provided by LCModel was within 20%. LCModel fits a linear combination of the full spectral pattern for each metabolite included in the basis set. The absolute values of N-acetyl aspartate (NAA), creatine (Cr), glutamate/glutamine (Glx), glutamate (Glu), glutamine (Gln), choline (Ch) and myo-inositol (MI) were calculated by LCModel relative to free water (55 mol/l) and are reported in millimoles uncorrected for T1 and T2 relaxation (Fig. 4).

Fig. 4:
Bar graph showing distribution of K-trans in the basal ganglia and frontal white matter in patients and controls (patients -in orange and controls in blue).

Statistical analyses of K-trans in the basal ganglia and frontal white matter

Given the small size of the sample and nonnormality of data in descriptive analyses, nonparametric analyses were used for group comparisons.

For the primary endpoint, a two-sample Wilcoxon rank-sum test was used to show the differences in K-trans measurements between the patients and controls in the predefined regions of the brain.

In the second analysis, we assessed whether K-trans was associated with neurocognitive scores. Pearson's correlation coefficient (rho) was used to assess the correlation between K-trans and neurocognitive scores in patients. Patients were then divided into two groups by a split based on their neurocognitive scores. ANI group and MND and HAD group. Kruskal–Wallis test was used as a global test to assess K-trans differences between the two groups, and Wilcoxon rank-sum tests were carried out for pairwise comparisons using Bonferroni correction.

Last, a sensitivity analysis was done to assess whether K-trans was associated with four metabolites (NAA, Glu+Gln, myo-inositol, choline) and four laboratory markers (nadir CD4+ cell count, CSF albumin, CSF neopterin, CSF/serum ratio). We first compared the differences in metabolites between patients and controls using Wilcoxon rank-sum test. To explore multivariable relationships, generalized linear models (GLMs) were then built with K-trans as the response variable and group (patients vs. controls) and the metabolite as explanatory variables.

All statistical analyses were performed using SAS statistical software (version 9.4; SAS Institute Inc., Cary, North Carolina, USA).


Comparison of K-trans between patients and controls

Table 3 summarizes the distribution of K-trans measurements in the two groups. There was a statistically significant difference between the two groups both in basal ganglia (Z = 4.88, P < 0.0001) as well as in the frontal white matter (Z = 4.73, P < 0.0001), with the control groups showing significantly lower levels of K-trans than the patient group (Table 3, Fig. 5).

Fig. 5:
(a) The bar diagrams with error bars showing significantly different distribution of the K-trans with P value of less than 0.0001. (b) The scatter plot of the neurocognitive scores correlation with K trans demonstrating a week negative correlation (r values -0.056 and -0.129).

K-trans and neurocognitive scores in the basal ganglia

In the basal ganglia, the global test (Kruskal–Wallis) indicated that there are significant differences between the groups in K-trans (chi-square with 2 DF = 24.07, P < 0.0001). Pairwise tests showed that, after Bonferroni correction, the controls had significantly lower K-trans values than both patient subgroups (controls vs. ‘ANI’ P < 0.0003; controls vs. ‘MND and HAD’ P = 0.0009), but K-trans in the two patient subgroups did not differ significantly (P = 1.0). There was also no clear statistically significant linear correlation between K-trans and neurocognitive scores (Pearson's rho = −0.06, P = 0.8133) (Figs. 6 and 7 a,b and 7, Table 4).

Fig. 6:
Bar plots with error maps of the distribution of K-trans in patients and controls with impaired representing ANI and significantly impaired reflecting group of MND and HAD.Distribution of K-trans in controls and two patient subgroups.
Fig. 7:
MR spectroscopic localization in the frontal white matter and basal ganglia and spectra-derived (LC model) demonstrating increased choline relative to normal controls).
Table 4:
Distribution of K-trans in controls and two patient subgroups.

K-trans and neurocognitive scores in the frontal white matter

Again, there were significant differences identified in between the controls and patients on the global test (Kruskal-–) (chi-square with 2 DF = 22.48, P < 0.0001). Pairwise tests showed that, after Bonferroni correction, the controls had significantly lower K-trans values than the MND and HAD group patient subgroup (P = 0.0006) and ‘ANI ’ patients (P < 0.0003), but K-trans did not differ between the two patient subgroups (P = 1.0) with a Pearson's correlation coefficient rho = −0.13, P = 0.5891 (Figs. 6a,b and 7, Table 4).

Sensitivity analysis: association between K-trans and metabolites/laboratory markers

Laboratory markers of the patient group showed similar results for basal ganglia and frontal white matter measurements. Nadir CD4+ cell count and CSF albumin were not associated with changes in K-trans, while CSF neopterin and CSF/serum albumin ratio were higher CSF neopterin values correlated with higher K-trans (BG P = 0.0193; FWM P = 0.0014) as did a higher CSF/serum albumin ratio (basal ganglia P = 0.023; frontal white matter P = 0.0032).

Spectroscopy-derived metabolites demonstrated no significant change in NAA, Glu+ Glx as well as myo-inositol compared with controls. However, choline was higher in HAND patients both in the basal ganglia (Z = 2.9895, P = 0.0028) and frontal white matter (Z = 5.0728, P < 0.0001) when compared with the controls.

K-trans has been positively correlated with metabolite choline values (higher choline is associated with higher K-trans). Although there was trend in association, the multivariate general linear model did not demonstrate any causal association between the two (Figs. 4 and 8).

Fig. 8:
Bar plot with error maps and scatter plots showing the choline distribution in normal controls and patients.


Our study showed that there are significant differences in the K-trans between the normal controls and HAND patients in both the basal ganglia and frontal white matter indicating that there is increased capillary permeability in these regions. This was in contradistinction to other brain regions, which were not different from the controls. K-trans was strongly correlated with the traditional BBB marker, the CSF/serum albumin ratio confirming its value as a marker of BBB impairment. Importantly, both K-trans and CSF/serum albumin ratios were abnormal across the patient group despite viral suppression in both CSF and blood (Fig. 2). Thus, BBB impairment is frequent despite virally efficacious cART. CSF neopterin strongly correlated with high K trans but not with serum neopterin or nadir CD4+ cell count, indicating that the ongoing BBB disruption at least is partly mediated by the brain parenchymal rather than systemic inflammatory response or historical immune compromise. Two other results did not achieve statistical significance, likely because of small sample size, but are worth mentioning for their potential pathogenetic implications. The highest mean K-trans was associated with lowest mean neurocognitive scores indicating that the increased permeability is associated with more neurocognitive impairment and the areas of abnormal K-trans (basal ganglia and frontal white matter) also demonstrated increased choline, a phospholipid associated with membrane inflammation and linked to HAND [15,16]. In addition, the K-trans values of both those with mild and significantly compromised NC are far higher than normal values that are derived from the normal appearing grey and white matter [KTrans (10–4/min)] Grey matter: [0.17 ± 0.81 (0.000017), normal white matter: 0.20 (0.04–0.84), Grey matter: 0.08 (0.05–0.89)] [8].

Imaging studies to assess the BBB disruption both in the era of pre and post c-ART era are very limited. To the best of our knowledge, there have been no studies performed using dynamic contrast enhancement to study the BBB in HAND. However, there were two studies that used conventional contrast enhanced MRI to measure BBB disruption [3,26]. Importantly, neither of these studies focussed on patients who were virally suppressed. In the pre c-ART era, Berger et al.[3] studied HAND using contrast-enhanced MRI and demonstrated the persistence of the contrast in the basal ganglia relative to normal white matter in delayed studies. The superior sagittal sinus contrast enhancement was used as a function of time for normalization. This study concluded that the delayed enhancement was strongly suggestive of BBB disruption and the enhancement was proportional to the severity of the dementia including the milder forms of HAND, though ANI patients were not included in the study [3].

The second study by Avison et al.[26] examined the relationship between BBB breakdown using contrast-enhanced MRI with plasma viral load and neurological status. The study had a mixed cohort of patients who were c-ART naive and those who were on c-ART with HAND. In addition, MRS-derived metabolites were analysed. The fractional enhancement of the volume of tissue in immediate and delayed postcontrast studies was measured and showed proportional increase in fractional enhancement with more severe BBB disruption and inverse correlation with cognitive scores. MRS-derived markers of inflammation (choline/creatine ratios) and astrogliosis (myoinositol/creatine ratios) were reduced in patients on c-ART as compared to those who were not and were correlated with BBB compromise. In both these studies, the increased enhancement was limited to the basal ganglia and frontal white matter.

There are several aspects of biological and clinical significance to these results that are hypothesis generating. First, the demonstration of BBB impairment in the frontal white matter and basal ganglia and nowhere else in the context of viral suppression suggests that HIV is already present in these areas rather than HIV being present in the systemic circulation with subsequent spread to involve the brain more diffusely. This would be in accord with the two previously discussed studies of BBB impairment in HIV [27]. However, the presence of BBB impairment even in the mildest form of HAND is unique and further supports HIV already being in the brain in this population. Clinically, therefore, treatment of HIV as early as possible is advisable.

Second, the absence of a relationship between BBB impairment and serum neopterin, a sensitive marker of inflammation in the systemic circulation, argues against systemic inflammation as the principle driver for BBB impairment; rather, it would appear that HIV in the brain is the driver. This would imply that the microbial translocation hypothesis for continued systemic inflammation in the context of virally suppressive cART has less relevance to HAND-related CNS inflammation.

Third, K-trans defined BBB impairment may mean that the use of cART with good CNS penetration may not be required in all patients. The ability of a drug to cross the BBB is dependent upon the following properties: lipid solubility, particle size, charge and polar surface area (PSA), the latter perhaps being dominant [28–30].

K-trans data are derived from the passage of gadobutrol across the BBB. It is a macrocyclic water soluble compound with a PSA of 194 A^2, which is much higher than all of the current c-ART medications, as these all have PSAs less than 115 A^2.

This study has several limitations. Its statistical power is restricted by the small sample. Neurocognitive scores and laboratory markers were only collected in the patient group but not in the controls. However, for this exploratory study, sufficient data were available to assess K-trans in the two groups. The DCE methods have inherent limitations regardless of the models used: they are influenced by cardiac output, vascular patency and haematocrit. All our cohorts were tested for patency using intracranial and extracranial MR angiography and ROIs were taken from similar sites in the opposite hemisphere and averaged values were taken as final value to overcome this pitfall.


Thus, BBB impairment is key to HAND pathogenesis even in the context of effective HIV suppression. Its modulation by cART provides a model for testable hypotheses for the treatment of early and established HAND.


Conflicts of interest

There are no conflicts of interest.


1. Heaton RK, Clifford DB, Franklin DR Jr, Woods SP, Ake C, Vaida F, et al. HIV-associated neurocognitive disorders persist in the era of potent antiretroviral therapy: CHARTER study. Neurology 2010; 75:2087–2096.
2. Heaton RK, Franklin DR, Ellis RJ, McCutchan JA, Letendre SL, Leblanc S, et al. HIV-associated neurocognitive disorders before and during the era of combination antiretroviral therapy: differences in rates, nature, and predictors. J Neurovirol 2011; 17:3–16.
3. Berger JR, Nath A, Greenberg RN, Andersen AH, Greene RA, Bognar A, Avison MJ. Cerebrovascular changes in the basal ganglia with HIV dementia. Neurology 2000; 54:921–926.
4. Berger JR, Avison M. The blood brain barrier in HIV infection. Front Biosci 2004; 9:2680–2685.
5. Anesten B, Yilmaz A, Hagberg L, Zetterberg H, Nilsson S, Brew BJ, Gisslén M. Blood–brain barrier integrity, intrathecal immunoactivation, and neuronal injury in HIV. Neurol Neuroimmunol Neuroinflamm 2016; 3:e300.
6. Ivey NS, MacLean AG, Lackner AA. AIDS and the blood-brain barrier. J Neurovirol 2009; 15:111–122.
7. Armitage PA, Farrall AJ, Carpenter TK, Doubal FN, Wardlaw JM. Use of dynamic contrast-enhanced MRI to measure subtle blood–brain barrier abnormalities. Magn Reson Imaging 2011; 29:305–314.
8. Heye AK, Culling RD, Valdés Hernández Mdel C, Thrippleton MJ, Wardlaw JM. Assessment of blood–brain barrier disruption using dynamic contrast-enhanced MRI. A systematic review. NeuroImage Clin 2014; 6:262–274.
9. Taheri S, Gasparovic C, Huisa BN, Adair JC, Edmonds E, Prestopnik J, et al. Blood-brain barrier permeability abnormalities in vascular cognitive impairment. Stroke 2011; 42:2158–2163.
10. Minagar A, Alexander JS. Blood-brain barrier disruption in multiple sclerosis. Mult Scler 2003; 9:540–549.
11. Farrall AJ, Wardlaw JM. Blood-brain barrier: ageing and microvascular disease -- systematic review and meta-analysis. Neurobiol Aging 2009; 30:337–352.
12. Larsson HB, Tofts PS. Measurement of blood-brain barrier permeability using dynamic Gd-DTPA scanning: a comparison of methods. Magn Reson Med 1992; 24:174–176.
13. Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab 1983; 3:1–7.
14. Carroll A, Brew B. HIV-associated neurocognitive disorders: recent advances in pathogenesis, biomarkers, and treatment. F1000Res 2017; 6:312.
15. Andersson LM, Hagbwerg L, Fuchs D, Svennerholm B, Gisslen M. Increased blood brain-barrier permeability in neuroasymptomatic HIV-1-infected individuals-correlation with cerebrospinal fluid HIV-1RNA and neopterin levels. J Neurovirol 2001; 7:542–547.
16. Price RW, Peterson J, Fuchs D, Angel TE, Zetterberg H, Hagberg L, et al. Approach to cerebrospinal fluid (CSF) biomarker discovery and evaluation in HIV infection. J Neuroimmune Pharmacol 2013; 8:1147–1158.
17. Krut J, Mellberg T, Price RW, Hagberg L, Fuchs D, Rosengren L, et al. Biomarker evidence of axonal injury in neuroasymptomatic HIV-1 patients. PLoS One 2014; 9:e88591.
18. Edén A, Marcotte TD, Heaton RK, Nilsson S, Zetterberg H, Fuchs D, et al. Increased Intrathecal immune activation in virally suppressed HIV-1 infected patients with neurocognitive impairment. PLoS One 2016; 11:e0157160.
19. Antinori A, Arendt G, Becker JT, Brew BJ, Byrd DA, Cherner M, Wojna VE. Updated research nosology for HIV-associated neurocognitive disorders. Neurology 2007; 69:1789–1799.
20. Heaton RK, Miller SW, Taylor MJ, Grant IM. Revised comprehensive norms for an expanded Halstead-Reitan Battery: demographically adjusted neuropsychological norms for African American and Caucasian adults. Lutz, Florida: PAR; 2004.
21. Gates TM, Cysique LA, Siefried KJ, Chaganti J, Moffat KJ, Brew BJ. Maraviroc-intensified combined antiretroviral therapy improves cognition in virally suppressed HIV-associated neurocognitive disorder. AIDS 2016; 30:591–60013.
22. Cysique LA, Maruff P, Darby D, Brew BJ. The assessment of cognitive function in advanced HIV-1 infection and AIDS dementia complex using a new computerized cognitive test battery. Arch Clin Neuropsychol 2006; 21:185–194.
23. Lane T, Moore D, Batchelor J, Brew B, Cysique L. Facial emotional processing in HIV infection: relation to neurocognitive and neuropsychiatric status. Neuropsychology 2012; 26:713–722.
24. Barnes SR, Ng TS, Montagne A, Law M, Zlokovic BV, Jacobs RE. Optimal acquisition and modeling parameters for accurate assessment of low Ktrans bloodbrain barrier permeability using dynamic contrast-enhanced MRI. Magn Reson Med 2016; 75:1967–1977.
25. Cramer SP, Larsson HB. Accurate determination of blood-brain barrier permeability using dynamic contrast-enhanced T1-weighted MRI: a simulation and in vivo study on healthy subjects and multiple sclerosis patients. J Cereb Blood Flow Metab 2014; 34:1655–1665.
26. Avison MJ, Nath A, Greene-Avison R, Schmitt FA, Greenberg RN, Berger JR. Neuroimaging correlates of HIV-associated BBB compromise. J Neuroimmunol 2004; 157:140–214.
27. Wiley CA, Soontornniyomkij V, Radhakrishnan L, Masliah E, Mellors J, Hermann SA, et al. Distribution of brain HIV load in AIDS. Brain Pathol 1998; 8:277–284.
28. Geldenhuys WJ, Mohammad AS, Adkins CE, Lockman PR. Molecular determinants of blood–brain barrier permeation. Therap Deliv 2015; 6:961–971.
29. Yilmaz A, Price RW, Gisslén M. Antiretroviral drug treatment of CNS HIV-1 infection. J Antimicrob Chemother 2012; 67:299–311.
30. Cheng KT. Gadobutrol. Molecular Imaging and Contrast Agent Database (MICAD) [Internet]. Bethesda, MD: National Center for Biotechnology Information (US); 2004-2013. 15 February 2006. [Updated 18 October 2007]. NBK23589 [Accessed July 2019].

blood–brain barrier; dynamic contrast enhanced MRI; HIV; HIV-associated neurocognitive disorder; K-Trans

Copyright © 2019 Wolters Kluwer Health, Inc.