Machine learning prediction of neurocognitive impairment among people with HIV using clinical and multimodal magnetic resonance imaging data

Effects of clinical, comorbid, and social determinants of health on brain ageing in people with and without HIV: a retrospective case-control study

BACKGROUND

Neuroimaging reveals structural brain changes linked with HIV infection and related neurocognitive disorders; however, group-level comparisons between people with HIV and people without HIV do not account for within-group heterogeneity. The aim of this study was to quantify the effects of comorbidities such as cardiovascular disease and adverse social determinants of health on brain ageing in people with HIV and people without HIV.

METHODS

In this retrospective case-control study, people with HIV from Washington University in St Louis, MO, USA, and people without HIV identified through community organisations or the Research Participant Registry were clinically characterised and underwent 3-Tesla T1-weighted MRI between Dec 3, 2008, and Oct 4, 2022. Exclusion criteria were established by a combination of self-reports and medical records. DeepBrainNet, a publicly available machine learning algorithm, was applied to estimate brain-predicted age from MRI for people with HIV and people without HIV. The brain-age gap, defined as the difference between brain-predicted age and true chronological age, was modelled as a function of clinical, comorbid, and social factors by use of linear regression. Variables were first examined singly for associations with brain-age gap, then combined into multivariate models with best-subsets variable selection.

FINDINGS

In people with HIV (mean age 44·8 years [SD 15·5]; 78% [296 of 379] male; 69% [260] Black; 78% [295] undetectable viral load), brain-age gap was associated with Framingham cardiovascular risk score (p=0·0034), detectable viral load (>50 copies per mL; p=0·0023), and hepatitis C co-infection (p=0·0065). After variable selection, the final model for people with HIV retained Framingham score, hepatitis C, and added unemployment (p=0·0015). Educational achievement assayed by reading proficiency was linked with reduced brain-age gap (p=0·016) for people without HIV but not for people with HIV, indicating a potential resilience factor. When people with HIV and people without HIV were modelled jointly, selection resulted in a model containing cardiovascular risk (p=0·0039), hepatitis C (p=0·037), Area Deprivation Index (p=0·033), and unemployment (p=0·00010). Male sex (p=0·078) and alcohol use history (p=0·090) were also included in the model but were not individually significant.

INTERPRETATION

Our findings indicate that comorbid and social determinants of health are associated with brain ageing in people with HIV, alongside traditional HIV metrics such as viral load and CD4 cell count, suggesting the need for a broadened clinical perspective on healthy ageing with HIV, with additional focus on comorbidities, lifestyle changes, and social factors.

FUNDING

National Institute of Mental Health, National Institute of Nursing Research, and National Institute of Drug Abuse.

Machine Learning Approaches to Understand Cognitive Phenotypes in People With HIV

Cognitive disorders are prevalent in people with HIV (PWH) despite antiretroviral therapy. Given the heterogeneity of cognitive disorders in PWH in the current era and evidence that these disorders have different etiologies and risk factors, scientific rationale is growing for using data-driven models to identify biologically defined subtypes (biotypes) of these disorders. Here, we discuss the state of science using machine learning to understand cognitive phenotypes in PWH and their associated comorbidities, biological mechanisms, and risk factors. We also discuss methods, example applications, challenges, and what will be required from the field to successfully incorporate machine learning in research on cognitive disorders in PWH. These topics were discussed at the National Institute of Mental Health meeting on "Biotypes of CNS Complications in People Living with HIV" held in October 2021. These ongoing research initiatives seek to explain the heterogeneity of cognitive phenotypes in PWH and their associated biological mechanisms to facilitate clinical management and tailored interventions.

The Evolution of Assessing Central Nervous System Complications in Human Immunodeficiency Virus: Where Do We Go From Here?

In this fifth decade of the human immunodeficiency virus (HIV) epidemic, central nervous system (CNS) complications including cognitive impairment and mental health remain a burden for people with HIV (PWH) on antiretroviral therapy. Despite the persistence of these complications, which often co-occur, the underlying pathophysiology remains elusive and consequently treatments remain limited. To continue to grow our understanding of the underlying mechanisms of CNS complications among PWH, there is a need to reexamine our current approaches, which are now more than 2 decades old. At the 2021 National Institutes of Health-sponsored meeting on Biotypes of CNS Complications in PWH, the Neurobehavioral Working Group addressed the following: (1) challenges inherent to determining CNS complications; (2) heterogeneity in CNS complications; and (3) problems and solutions for examining integrated biotypes. The review below provides a summary of the main points presented and discussed by the Neurobehavioral Working Group at the meeting.

A review on multimodal machine learning in medical diagnostics

Nowadays, the increasing number of medical diagnostic data and clinical data provide more complementary references for doctors to make diagnosis to patients. For example, with medical data, such as electrocardiography (ECG), machine learning algorithms can be used to identify and diagnose heart disease to reduce the workload of doctors. However, ECG data is always exposed to various kinds of noise and interference in reality, and medical diagnostics only based on one-dimensional ECG data is not trustable enough. By extracting new features from other types of medical data, we can implement enhanced recognition methods, called multimodal learning. Multimodal learning helps models to process data from a range of different sources, eliminate the requirement for training each single learning modality, and improve the robustness of models with the diversity of data. Growing number of articles in recent years have been devoted to investigating how to extract data from different sources and build accurate multimodal machine learning models, or deep learning models for medical diagnostics. This paper reviews and summarizes several recent papers that dealing with multimodal machine learning in disease detection, and identify topics for future research.

Effectiveness of facial palsy protocol among patient with mucormycosis following COVID-19: A case study

A 41-year-old male with a history of diabetes mellitus presented with right facial palsy post COVID-19 Associated Mucormycosis. A 4-week physiotherapeutic intervention; ice therapy, Mime therapy, Facial Soft Tissue Manipulation, and Facial Proprioceptive Neuromuscular Stimulation, showed improvement in the symptoms of patient and scores of House- Brackman Facial Grading Scale.

Multimodal magnetic resonance neuroimaging measures characteristic of early cART-treated pediatric HIV: A feature selection approach

Children with perinatally acquired HIV (CPHIV) have poor cognitive outcomes despite early combination antiretroviral therapy (cART). While CPHIV-related brain alterations can be investigated separately using proton magnetic resonance spectroscopy (1 H-MRS), structural magnetic resonance imaging (sMRI), diffusion tensor imaging (DTI), and functional MRI (fMRI), a set of multimodal MRI measures characteristic of children on cART has not been previously identified. We used the embedded feature selection of a logistic elastic-net (EN) regularization to select neuroimaging measures that distinguish CPHIV from controls and measured their classification performance via the area under the receiver operating characteristic curve (AUC) using repeated cross validation. We also wished to establish whether combining MRI modalities improved the models. In single modality analysis, sMRI volumes performed best followed by DTI, whereas individual EN models on spectroscopic, gyrification, and cortical thickness measures showed no class discrimination capability. Adding DTI and 1 H-MRS in basal measures to sMRI volumes produced the highest classification performancevalidation accuracy = 85 % AUC = 0.80 . The best multimodal MRI set consisted of 22 DTI and sMRI volume features, which included reduced volumes of the bilateral globus pallidus and amygdala, as well as increased mean diffusivity (MD) and radial diffusivity (RD) in the right corticospinal tract in cART-treated CPHIV. Consistent with previous studies of CPHIV, select subcortical volumes obtained from sMRI provide reasonable discrimination between CPHIV and controls. This may give insight into neuroimaging measures that are relevant in understanding the effects of HIV on the brain, thereby providing a starting point for evaluating their link with cognitive performance in CPHIV.

Putative role of mitochondria in SARS-CoV-2 mediated brain dysfunctions: a prospect

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the cause of the COVID-19 pandemic. Though the virus primarily damages the respiratory and cardiovascular systems after binding to the host angiotensin-converting enzyme 2 (ACE2) receptors, it has the potential to affect all major organ systems, including the human nervous system. There are multiple clinical reports of anosmia, dizziness, headache, nausea, ageusia, encephalitis, demyelination, neuropathy, memory loss, and neurological complications in SARS-CoV-2 infected individuals. Though the molecular mechanism of these brain dysfunctions during SARS-CoV-2 infection is elusive, the mitochondria seem to be an integral part of this pathogenesis. Emerging research findings suggest that the dysfunctional mitochondria and associated altered bioenergetics in the infected host cells lead to altered energy metabolism in the brain of Covid-19 patients. The interactome between viral proteins and mitochondrial proteins during Covid-19 pathogenesis also provides evidence for the involvement of mitochondria in SARS-CoV-2-induced brain dysfunctions. The present review discusses the possible role of mitochondria in disturbing the SARS-CoV-2 mediated brain functions, with the potential to use this information to prevent and treat these impairments.

Multi-label, multi-domain learning identifies compounding effects of HIV and cognitive impairment

Older individuals infected by Human Immunodeficiency Virus (HIV) are at risk for developing HIV-Associated Neurocognitive Disorder (HAND), i.e., from reduced cognitive functioning similar to HIV-negative individuals with Mild Cognitive Impairment (MCI) or to Alzheimer's Disease (AD) if more severely affected. Incompletely understood is how brain structure can serve to differentiate cognitive impairment (CI) in the HIV-positive (i.e., HAND) from the HIV-negative cohort (i.e., MCI and AD). To that end, we designed a multi-label classifier that labels the structural magnetic resonance images (MRI) of individuals by their HIV and CI status via two binary variables. Proper training of such an approach traditionally requires well-curated datasets containing large number of samples for each of the corresponding four cohorts (healthy controls, CI HIV-negative adults a.k.a. CI-only, HIV-positive patients without CI a.k.a. HIV-only, and HAND). Because of the rarity of such datasets, we proposed to improve training of the multi-label classifier via a multi-domain learning scheme that also incorporates domain-specific classifiers on auxiliary single-label datasets specific to either binary label. Specifically, we complement the training dataset of MRIs of the four cohorts (Control: 156, CI-only: 335, HIV-only: 37, HAND: 145) acquired by the Memory and Aging Center at the University of California - San Francisco with a CI-specific dataset only containing MRIs of HIV-negative subjects (Controls: 229, CI-only: 397) from the Alzheimer's Disease Neuroimaging Initiative and an HIV-specific dataset (Controls: 75, HIV-only: 75) provided by SRI International. Based on cross-validation on the UCSF dataset, the multi-domain and multi-label learning strategy leads to superior classification accuracy compared with one-domain or multi-class learning approaches, specifically for the undersampled HIV-only cohort. The 'prediction logits' of CI computed by the multi-label formulation also successfully stratify motor performance among the HIV-positive subjects (including HAND). Finally, brain patterns driving the subject-level predictions across all four cohorts characterize the independent and compounding effects of HIV and CI in the HAND cohort.

Decoding pathogenesis factors involved in the progression of ATLL or HAM/TSP after infection by HTLV-1 through a systems virology study

Background

Human T-cell Leukemia Virus type-1 (HTLV-1) is a retrovirus that causes two diseases including Adult T-cell Leukemia/Lymphoma (ATLL cancer) and HTLV-1 Associated Myelopathy/Tropical Spastic Paraparesis (HAM/TSP, a neurodegenerative disease) after a long latency period as an asymptomatic carrier (AC). There are no obvious explanations about how each of the mentioned diseases develops in the AC carriers. Finding the discriminative molecular factors and pathways may clarify the destiny of the infection.

Methods

To shed light on the involved molecular players and activated pathways in each state, differentially co-expressed modules (DiffCoEx) algorithm was employed to identify the highly correlated genes which were co-expressed differently between normal and ACs, ACs and ATLL, as well as ACs and HAM/TSP samples. Through differential pathway analysis, the dysregulated pathways and the specific disease-genes-pathways were figured out. Moreover, the common genes between the member of DiffCoEx and differentially expressed genes were found and the specific genes in ATLL and HAM/TSP were introduced as possible biomarkers.

Results

The dysregulated genes in the ATLL were mostly enriched in immune and cancer-related pathways while the ones in the HAM/TSP were enriched in immune, inflammation, and neurological pathways. The differential pathway analysis clarified the differences between the gene players in the common activated pathways. Eventually, the final analysis revealed the involvement of specific dysregulated genes including KIRREL2, RAB36, and KANK1 in HAM/TSP as well as LTB4R2, HCN4, FZD9, GRIK5, CREB3L4, TACR2, FRMD1, LHB, FGF3, TEAD3, GRIN2D, GNRH2, PRLH, GPR156, and CRHR2 in ATLL.

Conclusion

The identified potential prognostic biomarkers and therapeutic targets are proposed as the most important platers in developing ATLL or HAM/TSP. Moreover, the proposed signaling network clarifies the differences between the functional players in the activated pathways in ACs, ATLL, and HAM/TSP.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12985-021-01643-8.