Longitudinal Imaging Analyses Research Articles

Disease activity and therapeutic response to pegcetacoplan for geographic atrophy identified by deep learning-based analysis of OCT

PurposeTo quantify morphological changes of the photoreceptors (PR) and retinal pigment epithelium (RPE) layers under pegcetacoplan therapy in geographic atrophy (GA) using deep learning-based analysis of optical coherence tomography (OCT) images. DesignPost-hoc longitudinal image analysis SubjectsPatients with GA due to age-related macular degeneration from two prospective randomized phase III clinical trials (OAKS and DERBY) MethodsDeep learning-based segmentation of RPE loss and PR degeneration, defined as loss of the ellipsoid zone (EZ) layer on OCT, over 24 months on SD-OCT images Main outcome measuresChange in the mean area of RPE loss and EZ loss over time in the pooled sham arms and the monthly (PM)/every other month (PEOM) treatment arms Results897 eyes of 897 patients were included. There was a therapeutic reduction of RPE loss growth by 22%/20% in OAKS and 27%/21% in DERBY for PM/PEOM compared to sham, respectively, at 24 months. The reduction on the EZ level was significantly higher with 53%/46% in OAKS and 47%/46% in DERBY for PM/PEOM compared to sham at 24 months. The baseline EZ-RPE difference had an impact on disease activity and therapeutic response. The therapeutic benefit for RPE loss growth increased with larger EZ-RPE difference quartiles from 21.9%, 23.1%, 23.9% to 33.6% for PM vs. sham (all p<0.01) and from 13.6% (p=0.11), 23.8%, 23.8% to 20.0% for PEOM vs. sham (p<0.01) in quartiles 1,2,3 and 4, respectively, at 24 months. Regarding EZ layer maintenance, the therapeutic reduction of loss increased from 14.8% (p=0.09), 33.3%, 46.6% to 77.8% (p<0.0001) between PM and sham and from 15.9% (p=0.08), 33.8%, 52.0% to 64.9% (p<0.0001) between PEOM and sham for quartiles 1-4 at 24 months. ConclusionOCT-based AI analysis objectively identifies and quantifies PR and RPE degeneration in GA. Reductions in further PR degeneration consistent with EZ loss on OCT are even higher than the effect on RPE loss in phase 3 trials of pegcetacoplan treatment. The EZ-RPE difference has a strong impact on disease progression and therapeutic response. Identification of patients with higher EZ-RPE loss difference may become an important criterion for the management of GA secondary to AMD.

Cortical volumetric changes after cochlear implantation in postlingually deaf adults: correlation with speech perception abilities

This study aims to analyse the volumetric changes in brain MRI after cochlear implantation (CI), focusing on the speech perception in postlingually deaf adults. We conducted a prospective cohort study with 16 patients who had bilateral hearing loss and received unilateral CI. Based on the surgical side, patients were categorized into left and right CI groups. Volumetric T1-weighted brain MRI were obtained before and one year after the surgery. To overcome the artifact caused by the internal device in post-CI scan, image reconstruction method was newly devised and applied using the contralateral hemisphere of the pre-CI MRI data, to run FreeSurfer. We conducted within-subject template estimation for unbiased longitudinal image analysis, based on the linear mixed effect models. When analyzing the contralateral cerebral hemisphere before and after CI, a substantial increase in superior frontal gyrus and superior temporal gyrus (STG) volumes was observed in the left CI group. A positive correlation was observed in the STG and post-CI word recognition score in both groups. As far as we know, this is the first study attempting longitudinal brain volumetry based on post-CI MRI scans. We demonstrate that better auditory performance after CI is associated with structural restoration in central auditory structures.

Longitudinal risk prediction for pediatric glioma with temporal deep learning.

Pediatric glioma recurrence can cause morbidity and mortality; however, recurrence pattern and severity are heterogeneous and challenging to predict with established clinical and genomic markers. Resultingly, almost all children undergo frequent, long-term, magnetic resonance (MR) brain surveillance regardless of individual recurrence risk. Deep learning analysis of longitudinal MR may be an effective approach for improving individualized recurrence prediction in gliomas and other cancers but has thus far been infeasible with current frameworks. Here, we propose a self-supervised, deep learning approach to longitudinal medical imaging analysis, temporal learning, that models the spatiotemporal information from a patient's current and prior brain MRs to predict future recurrence. We apply temporal learning to pediatric glioma surveillance imaging for 715 patients (3,994 scans) from four distinct clinical settings. We find that longitudinal imaging analysis with temporal learning improves recurrence prediction performance by up to 41% compared to traditional approaches, with improvements in performance in both low- and high-grade glioma. We find that recurrence prediction accuracy increases incrementally with the number of historical scans available per patient. Temporal deep learning may enable point-of-care decision-support for pediatric brain tumors and be adaptable more broadly to patients with other cancers and chronic diseases undergoing surveillance imaging.

Ellipsoid Zone Integrity and Visual Function in Dry Age-Related Macular Degeneration

In this longitudinal retrospective image analysis, conducted on patients diagnosed with dry age-related macular degeneration (AMD) and 5 years of follow-up imaging data, the study aimed to investigate the relationship between ellipsoid zone (EZ) integrity on spectral domain optical coherence tomography (SD-OCT) and visual acuity (VA). Using a machine learning-enabled feature extraction tool, quantitative EZ parameters were derived from SD-OCT images. The analysis revealed significant correlations between EZ integrity metrics and VA. Eyes with excellent VA (≥20/25 Snellen) exhibited higher EZ integrity, including less EZ attenuation, thicker ellipsoid zone-retinal pigment epithelium (EZ-RPE) thickness, and higher EZ intensity, in contrast to eyes with worse VA (≤20/40 Snellen). Additionally, eyes with geographic atrophy (GA) in the foveal region displayed compromised EZ integrity compared to those without GA. Notably, baseline EZ integrity metrics were predictive of future VA loss. These findings suggest that quantitative SD-OCT measurements of EZ integrity could potentially detect early changes in dry AMD and serve as valuable indicators for predicting future functional outcomes. Furthermore, these measurements hold promise for use in clinical trial screenings, offering insights into the progression of the disease and its impact on visual acuity. This study underscores the importance of EZ integrity assessment in understanding and managing dry AMD.

Longitudinal study investigating the influence of COMT gene polymorphism on cortical thickness changes in Parkinson's disease over four years

Parkinson's disease (PD) is a progressive neurodegenerative disorder affecting over 3% of those over 65. It's caused by reduced dopaminergic neurons and Lewy bodies, leading to motor and non-motor symptoms. The relationship between COMT gene polymorphisms and PD is complex and not fully elucidated. Some studies have reported associations between certain COMT gene variants and PD risk, while others have not found significant associations. This study investigates how COMT gene variations impact cortical thickness changes in PD patients over time, aiming to link genetic factors, especially COMT gene variations, with PD progression. This study analyzed data from 44 PD patients with complete 4-year imaging follow-up from the Parkinson Progression Marker Initiative (PPMI) database. Magnetic resonance imaging (MRI) scans were acquired using consistent methods across 9 different MRI scanners. COMT single-nucleotide polymorphisms (SNPs) were assessed based on whole genome sequencing data. Longitudinal image analysis was conducted using FreeSurfer's processing pipeline. Linear mixed-effect models were employed to examine the interaction effect of genetic variations and time on cortical thickness, while controlling for covariates and subject-specific variations. The rs165599 SNP stands out as a potential contributor to alterations in cortical thickness, showing a significant reduction in overall mean cortical thickness in both hemispheres in homozygotes (Left: P = 0.023, Right: P = 0.028). The supramarginal, precentral, and superior frontal regions demonstrated significant bilateral alterations linked to rs165599. Our findings suggest that the rs165599 variant leads to earlier manifestation of cortical thinning during the course of the disease. However, it does not result in more severe cortical thinning outcomes over time. There is a need for larger cohorts and control groups to validate these findings and consider genetic variant interactions and clinical features to elucidate the specific mechanisms underlying COMT-related neurodegenerative processes in PD.

VC-resist glioblastoma cell state: vessel co-option as a key driver of chemoradiation resistance

Glioblastoma (GBM) is a highly lethal type of cancer. GBM recurrence following chemoradiation is typically attributed to the regrowth of invasive and resistant cells. Therefore, there is a pressing need to gain a deeper understanding of the mechanisms underlying GBM resistance to chemoradiation and its ability to infiltrate. Using a combination of transcriptomic, proteomic, and phosphoproteomic analyses, longitudinal imaging, organotypic cultures, functional assays, animal studies, and clinical data analyses, we demonstrate that chemoradiation and brain vasculature induce cell transition to a functional state named VC-Resist (vessel co-opting and resistant cell state). This cell state is midway along the transcriptomic axis between proneural and mesenchymal GBM cells and is closer to the AC/MES1-like state. VC-Resist GBM cells are highly vessel co-opting, allowing significant infiltration into the surrounding brain tissue and homing to the perivascular niche, which in turn induces even more VC-Resist transition. The molecular and functional characteristics of this FGFR1-YAP1-dependent GBM cell state, including resistance to DNA damage, enrichment in the G2M phase, and induction of senescence/stemness pathways, contribute to its enhanced resistance to chemoradiation. These findings demonstrate how vessel co-option, perivascular niche, and GBM cell plasticity jointly drive resistance to therapy during GBM recurrence.

Brain structural co-development is associated with internalizing symptoms two years later in the ABCD cohort.

About 1/3 of youth spend more than four hours/day engaged in screen media activity (SMA). This investigation utilized longitudinal brain imaging and mediation analyses to examine relationships among SMA, brain patterns, and internalizing problems. Data from Adolescent Brain Cognitive Development (ABCD) participants with baseline and two-year follow-up structural imaging data that passed quality control (N = 5,166; 2,385 girls) were analyzed. Joint and Individual Variation Explained (JIVE) identified a brain co-development pattern among 221 brain features (i.e., differences in surface area, thickness, or cortical and subcortical gray-matter volume between baseline and two-year-follow-up data). Generalized linear mixed-effect models investigated associations between baseline SMA, structural co-development and internalizing and externalizing psychopathology at two-year follow-up. SMA at baseline was related to internalizing psychopathology at year 2 (β=0.020,SE=0.008,P=0.014) and a structural co-development pattern (β=0.015,SE=0.007,P=0.029), where the co-development pattern suggested that rates of change in gray-matter volumes of the brainstem, gray-matter volumes and/or cortical thickness measures of bilateral superior frontal, rostral middle frontal, inferior parietal, and inferior temporal regions were more similar than those in other regions. This component partially mediated the relationship between baseline SMA and future internalizing problems (indirect effect = 0.020, P-value = 0.043, proportion mediated: 2.24%). Greater youth engagement in SMA at ages 9-10 years statistically predicted higher levels of internalizing two years later. This association was mediated by cortical-brainstem circuitry, albeit with relatively small effect sizes. The findings may help delineate processes contributing to internalizing behaviors and assist in identifying individuals at greater risk for such problems.

Evaluation of silk fibroin-based urinary conduits in a porcine model of urinary diversion.

Background: The primary strategy for urinary diversion in radical cystectomy patients involves incorporation of autologous gastrointestinal conduits into the urinary tract which leads to deleterious consequences including chronic infections and metabolic abnormalities. This report investigates the efficacy of an acellular, tubular bi-layer silk fibroin (BLSF) graft to function as an alternative urinary conduit in a porcine model of urinary diversion. Materials and methods: Unilateral urinary diversion with stented BLSF conduits was executed in five adult female, Yucatan mini-swine over a 3 month period. Longitudinal imaging analyses including ultrasonography, retrograde ureteropyelography and video-endoscopy were carried out monthly. Histological, immunohistochemical (IHC), and histomorphometric assessments were performed on neoconduits at harvest. Results: All animals survived until scheduled euthanasia and displayed moderate hydronephrosis (Grades 1-3) in reconstructed collecting systems over the course of the study period. Stented BLSF constructs supported formation of vascularized, retroperitoneal tubes capable of facilitating external urinary drainage. By 3months post-operative, neoconduits contained α-smooth muscle actin+ and SM22α+ smooth muscle as well as uroplakin 3A+ and pan-cytokeratin + urothelium. However, the degree of tissue regeneration in neotissues was significantly lower in comparison to ureteral controls as determined by histomorphometry. In addition, neoconduit stenting was necessary to prevent stomal occlusion. Conclusion: BLSF biomaterials represent emerging platforms for urinary conduit construction and may offer a functional replacement for conventional urinary diversion techniques following further optimization of mechanical properties and regenerative responses.

Moment Evolution Equations and Moment Matching for Stochastic Image EPDiff

Models of stochastic image deformation allow study of time-continuous stochastic effects transforming images by deforming the image domain. Applications include longitudinal medical image analysis with both population trends and random subject-specific variation. Focusing on a stochastic extension of the LDDMM models with evolutions governed by a stochastic EPDiff equation, we use moment approximations of the corresponding Itô diffusion to construct estimators for statistical inference in the full stochastic model. We show that this approach, when efficiently implemented with automatic differentiation tools, can successfully estimate parameters encoding the spatial correlation of the noise fields on the image.

Clinical application of artificial intelligence in longitudinal image analysis of bone age among GHD patients.

This study aims to explore the clinical value of artificial intelligence (AI)-assisted bone age assessment (BAA) among children with growth hormone deficiency (GHD). A total of 290 bone age (BA) radiographs were collected from 52 children who participated in the study at Sun Yat-sen Memorial Hospital between January 2016 and August 2017. Senior pediatric endocrinologists independently evaluated BA according to the China 05 (CH05) method, and their consistent results were regarded as the gold standard (GS). Meanwhile, two junior pediatric endocrinologists were asked to assessed BA both with and without assistance from the AI-based BA evaluation system. Six months later, around 20% of the images assessed by the junior pediatric endocrinologists were randomly selected to be re-evaluated with the same procedure half a year later. Root mean square error (RMSE), mean absolute error (MAE), accuracy, and Bland-Altman plots were used to compare differences in BA. The intra-class correlation coefficient (ICC) and one-way repeated ANOVA were used to assess inter- and intra-observer variabilities in BAA. A boxplot of BA evaluated by different raters during the course of treatment and a mixed linear model were used to illustrate inter-rater effect over time. A total of 52 children with GHD were included, with mean chronological age and BA by GS of 6.64 ± 2.49 and 5.85 ± 2.30 years at baseline, respectively. After incorporating AI assistance, the performance of the junior pediatric endocrinologists improved (P < 0.001), with MAE and RMSE both decreased by more than 1.65 years (Rater 1: ΔMAE = 1.780, ΔRMSE = 1.655; Rater 2: ΔMAE = 1.794, ΔRMSE = 1.719), and accuracy increasing from approximately 10% to over 91%. The ICC also increased from 0.951 to 0.990. During GHD treatment (at baseline, 6-, 12-, 18-, and 24-months), the difference decreased sharply when AI was applied. Furthermore, a significant inter-rater effect (P = 0.002) also vanished upon AI involvement. AI-assisted interpretation of BA can improve accuracy and decrease variability in results among junior pediatric endocrinologists in longitudinal cohort studies, which shows potential for further clinical application.

Abstract 6217: TP53 loss with whole genome doubling mediates heterogeneous intra-patient therapy response in EGFR-driven lung adenocarcinoma: A TRACERx study

Abstract Introduction: Mutations in the epidermal growth factor receptor (EGFR) are often found in never-smokers who develop lung adenocarcinoma (LUAD) and resistance to receptor-targeted tyrosine kinase inhibitors (TKIs) generally occurs within five years of treatment. Mixed responses, where individual tumor lesions within the same patient respond differently to treatment, contribute to early treatment failure and indicate the involvement of multiple genomic alterations. TP53 loss of function has been associated with both tolerance to chromosomal instability (CIN) and with shortened progression free survival in EGFR-driven tumors. The aim of this study was to investigate the hypothesis that CIN, together with loss of p53 function, may lead to diverse genotypes that underlie the mixed responses observed in EGFR-driven LUAD. Experimental Procedures: We used genetically engineered mouse models (GEMMs), driven by EGFR with or without concomitant Trp53 loss and human isogenic cell lines to investigate cellular evolution and the effect of whole genome doubling on targeted therapy responses and mechanisms of resistance. Next generation sequencing and shallow whole genome single cell analyses, together with longitudinal imaging analysis from the Aura clinical trials (AURA2, AURA3 and the AURA extension cohort, Identifiers: NCT02094261, NCT02151981 and NCT01802632) was used to investigate the effect of p53 loss on tumor evolution. Results: EGFR mutant tumors with clonal Trp53 loss or TP53 pathway disruption displayed increased Weighted Genome Integrity Index (wGII) and higher cell to cell variation in both the mouse and TRACERx data sets. We found that TP53 loss of function increased the incidence of mixed responses and resistance to targeted therapy in both mouse and human tumors leading to early treatment failure and reduced survival. Whole-exome sequencing (median depth of 92x, range: 58-169x) of nine erlotinib-resistant EGFR mouse tumors identified four EGFR bypassing mutations (oncogenic KRAS mutations; Q61H, Q61R, and two G12D mutations) and one likely driver mutation in FGFR2 (C286R). We could only identify one known resistance associated mutation, EGFRT790M, in EGFR mutant tumors with concomitant loss of Trp53. In depth analysis revealed no major copy number differences in treatment naÏve vs resistant EGFR mutant tumors. In contrast, 70% of all tumors with concomitant Trp53 loss had amplified a region of mouse chromosome 7, harboring MET and BRAF. Investigating an isogenic EGFR/TP53 mutant cell model system revealed whole genome doubling as advantageous in overcoming the selection pressure induced by targeted therapy. Conclusion: We find that loss of TP53 in the context of mutated EGFR leads to an altered and plastic genomic landscape, with multiple copy number changes, which in turn facilitates therapy resistance. Citation Format: Sebastijan Hobor, Maise Al Bakir, Marcin Skrzypski, Alexander M. Frankell, Bjorn Bakker, Thomas B. Watkins, Aleksandra Markovets, Jonathan R. Dry, Andrew P. Brown, Jasper van der Aart, Dahmane Oukrif, Marco Novelli, Matthew J. Renshaw, William Hill, Hilda van den Bos, Diana C. Spierings, Juliann Chmielecki, Carl Barrett, Kevin Litchfield, Elza de Bruin, Floris Foijer, Karen H. Vousden, TRACERx consortium, Robert E. Hynds, Crispin T. Hiley, Nnennaya Kanu, Simone Zaccaria, Eva C. Gronroos, Charles Swanton. TP53 loss with whole genome doubling mediates heterogeneous intra-patient therapy response in EGFR-driven lung adenocarcinoma: A TRACERx study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 6217.

Machine Learning–Based Automated Detection of Hydroxychloroquine Toxicity and Prediction of Future Toxicity Using Higher-Order OCT Biomarkers

Despite guidelines for hydroxychloroquine (HCQ) toxicity screening, there are clear challenges to accurate detection and interpretation. In the current report, the feasibility of automated machine learning (ML)-based detection of HCQ retinopathy and prediction of progression to toxicity in eyes without preexisting toxicity has been described. Retrospective, longitudinal cohort study. Subjects on HCQ therapy. This was an institutional review board-approved, retrospective, longitudinal image analysis of 388 subjects on HCQ. Multilayer, compartmental, retinal segmentation with ellipsoid zone (EZ) mapping was used to harvest quantitative spectral-domain (SD)-OCT biomarkers. Using a combination of clinical features (i.e., cumulative HCQ dose and the duration of therapy) and quantitative imaging biomarkers (e.g., volumetric EZ integrity and compartmental measurements), ML models were created to detect toxicity and predict progression based on ground-truth OCT-based toxicity readings by 2 masked retina specialists. Furthermore, 10-fold cross-validation was performed. The model performance was visualized using receiver operator curves and calculating the area under the curve (AUC). The corresponding sensitivity and specificity values were evaluated for the feasibility of HCQ toxicity screening and prediction. The prevalence of HCQ toxicity in this cohort of 388 patients was 9.8% (n= 38). Twenty-one eyes progressed to toxicity during follow-up. OCT-based features (i.e., partial EZ attenuation, EZ volume, outer nuclear layer volume, and compartmental thicknesses) and clinical features (i.e., HCQ daily dose, HCQ cumulative dose, and duration of therapy) showed significant differences between the toxic and nontoxic groups. Percentage area with partial EZ attenuation (i.e., percentage of the macula with an EZ-retinal pigment epithelium thickness of ≤ 20 μm) was the most discriminating single feature (toxic, 35.7 ± 46.5%; nontoxic, 1.8 ± 4.4%; P < 0.0001). Using a random forest model, high-performance, automated toxicity detection was achieved, with a mean AUC of 0.97, sensitivity of 95%, and specificity of 91%. Furthermore, the toxicity progression prediction model had a mean AUC of 0.89, with a sensitivity and specificity of 90% and 80%, respectively. This report described the feasibility of high-performance automated classification models that used a combination of clinical and quantitative SD-OCT biomarkers to detect HCQ retinal toxicity and predict progression to toxicity in cases without toxicity. Future work is needed to validate these findings in an independent dataset.

Regional Aβ-tau interactions promote onset and acceleration of Alzheimer’s disease tau spreading

Amyloid-beta and tau are key molecules in the pathogenesis of Alzheimer's disease, but it remains unclear how these proteins interact to promote disease. Here, by combining cross-sectional and longitudinal molecular imaging and network connectivity analyses in living humans, we identified two amyloid-beta/tau interactions associated with the onset and propagation of tau spreading. First, we show that the lateral entorhinal cortex, an early site of tau neurofibrillary tangle formation, is subject to remote, connectivity-mediated amyloid-beta/tau interactions linked to initial tau spreading. Second, we identify the inferior temporal gyrus as the region featuring the greatest local amyloid-beta/tau interactions and a connectivity profile well suited to accelerate tau propagation. Taken together, our data address long-standing questions regarding the topographical dissimilarity between early amyloid-beta and tau deposition.

White Matter Structural and Network Topological Changes Underlying the Behavioral Phenotype of MECP2 Mutant Monkeys.

To explore the brain structural basis underlying the behavioral abnormalities associated with Rett syndrome (RTT), we carried out detailed longitudinal noninvasive magnetic resonance imaging analyses of RTT monkey models created by gene-editing, from weaning, through adolescence, till sexual maturation. Here, we report abnormal developmental dynamics of brain white matter (WM) microstructures and network topological organizations via diffusion tensor imaging. Specifically, disrupted WM microstructural integrity was observed at 9months, but recovered thereafter, whereas WM network topological properties showed persistent abnormal dynamics from 9 to 37months. Changes in the WM microstructure and WM network topology were correlated well with RTT-associated behavioral abnormalities including sleep latency, environmental exploration, and conflict encounters. Deleterious and protracted early WM myelination process likely lead to abnormal synaptic pruning, resulting in poor functional segregations. Together, this study provides initial evidence for changes in WM microstructure and network topological organization, which may underlie the neuro-patho-etilogy of RTT.

Increased free water in the substantia nigra in idiopathic REM sleep behaviour disorder

Imaging markers sensitive to neurodegeneration in the substantia nigra are critically needed for future disease-modifying trials. Previous studies have demonstrated the utility of posterior substantia nigra free water as a marker of progression in Parkinson's disease. In this study, we tested the hypothesis that free water is elevated in the posterior substantia nigra of idiopathic REM sleep behaviour disorder, which is considered a prodromal stage of synucleinopathy. We applied free-water imaging to 32 healthy control subjects, 34 patients with idiopathic REM sleep behaviour disorder and 38 patients with Parkinson's disease. Eighteen healthy control subjects and 22 patients with idiopathic REM sleep behaviour disorder were followed up and completed longitudinal free-water imaging. Free-water values in the substantia nigra were calculated for each individual and compared among groups. We tested the associations between posterior substantia nigra free water and uptake of striatal dopamine transporter in idiopathic REM sleep behaviour disorder. Free-water values in the posterior substantia nigra were significantly higher in the patients with idiopathic REM sleep behaviour disorder patients than in the healthy control subjects, but were significantly lower in patients with idiopathic REM sleep behaviour disorder than in patients with Parkinson's disease. In addition, we observed significantly negative associations between posterior substantia nigra free-water values and dopamine transporter striatal binding ratios in the idiopathic REM sleep behaviour disorder patients. Longitudinal free-water imaging analyses were conducted with a linear mixed-effects model, and showed a significant Group × Time interaction in posterior substantia nigra, identifying increased mean free-water values in posterior substantia nigra of idiopathic REM sleep behaviour disorder over time. These results demonstrate that free water in the posterior substantia nigra is a valid imaging marker of neurodegeneration in idiopathic REM sleep behaviour disorder, which has the potential to be used as an indicator in disease-modifying trials.

Multi-Resemblance Multi-Target Low-Rank Coding for Prediction of Cognitive Decline With Longitudinal Brain Images.

An effective presymptomatic diagnosis and treatment of Alzheimer's disease (AD) would have enormous public health benefits. Sparse coding (SC) has shown strong potential for longitudinal brain image analysis in preclinical AD research. However, the traditional SC computation is time-consuming and does not explore the feature correlations that are consistent over the time. In addition, longitudinal brain image cohorts usually contain incomplete image data and clinical labels. To address these challenges, we propose a novel two-stage Multi-Resemblance Multi-Target Low-Rank Coding (MMLC) method, which encourages that sparse codes of neighboring longitudinal time points are resemblant to each other, favors sparse code low-rankness to reduce the computational cost and is resilient to both source and target data incompleteness. In stage one, we propose an online multi-resemblant low-rank SC method to utilize the common and task-specific dictionaries in different time points to immune to incomplete source data and capture the longitudinal correlation. In stage two, supported by a rigorous theoretical analysis, we develop a multi-target learning method to address the missing clinical label issue. To solve such a multi-task low-rank sparse optimization problem, we propose multi-task stochastic coordinate coding with a sequence of closed-form update steps which reduces the computational costs guaranteed by a theoretical convergence proof. We apply MMLC on a publicly available neuroimaging cohort to predict two clinical measures and compare it with six other methods. Our experimental results show our proposed method achieves superior results on both computational efficiency and predictive accuracy and has great potential to assist the AD prevention.

Cortical inflammation and brain signs of high-risk atherosclerosis in a non-human primate model.

Atherosclerosis is a chronic systemic inflammatory disease, inducing cardiovascular and cerebrovascular acute events. A role of neuroinflammation is suspected, but not yet investigated in the gyrencephalic brain and the related activity at blood−brain interfaces is unknown. A non-human primate model of advanced atherosclerosis was first established using longitudinal blood samples, multimodal imaging and gene analysis in aged animals. Non-human primate carotid lesions were compared with human carotid endarterectomy samples. During the whole-body imaging session, imaging of neuroinflammation and choroid plexus function was performed. Advanced plaques were present in multiple sites, premature deaths occurred and downstream lesions (myocardial fibrosis, lacunar stroke) were present in this model. Vascular lesions were similar to in humans: high plaque activity on PET and MRI imaging and systemic inflammation (high plasma C-reactive protein levels: 42 ± 14 µg/ml). We also found the same gene association (metabolic, inflammatory and anti-inflammatory markers) as in patients with similar histological features. Metabolic imaging localized abnormal brain glucose metabolism in the frontal cortex. It corresponded to cortical neuro-inflammation (PET imaging) that correlated with C-reactive protein level. Multimodal imaging also revealed pronounced choroid plexus function impairment in aging atherosclerotic non-human primates. In conclusion, multimodal whole-body inflammation exploration at the vascular level and blood−brain interfaces identified high-risk aging atherosclerosis. These results open the way for systemic and central inflammation targeting in atherosclerosis in the new era of immunotherapy.

Midbrain atrophy related to parkinsonism in a non-coding repeat expansion disorder: five cases of spinocerebellar ataxia type 31 with nigrostriatal dopaminergic dysfunction

BackgroundSpinocerebellar ataxia type 31 (SCA31) is caused by non-coding pentanucleotide repeat expansions in the BEAN1 gene. Clinically, SCA31 is characterized by late adult-onset, pure cerebellar ataxia. To explore the association between parkinsonism and SCA31, five patients with SCA31 with concomitant nigrostriatal dopaminergic dysfunction (NSDD) development, including three cases of L-DOPA responsive parkinsonism, were analyzed.MethodsTo assess regional brain atrophy, cross-sectional and longitudinal imaging analyses were retrospectively performed using magnetic resonance imaging (MRI) planimetry. The midbrain-to-pons (M/P) area ratio and cerebellar area were measured on midsagittal T1-weighted MRI in five patients with SCA31 with concomitant NSDD (NSDD(+)), 14 patients with SCA31 without NSDD (NSDD(−)), 32 patients with Parkinson’s disease (PD), and 15 patients with progressive supranuclear palsy (PSP). Longitudinal changes in the M/P area ratio were assessed by serial MRI of NSDD(+) (n = 5) and NSDD(−) (n = 9).ResultsThe clinical characteristics assessed in the five patients with NSDD were as follows: the mean age at NSDD onset (72.0 ± 10.8 years), prominence of bradykinesia/akinesia (5/5), rigidity (4/5), tremor (2/5), dysautonomia (0/5), vertical gaze limitation (1/5), and abnormalities on 123I-ioflupane dopamine transporter scintigraphy (3/3) and 3-Tesla neuromelanin MRI (4/4). A clear reduction in the midbrain area and the M/P area ratio was observed in the NSDD(+) group (p < 0.05) while there was no significant difference in disease duration or in the pons area among the NSDD(+), NSDD(−), and PD groups. There was also a significant difference in the midbrain and pons area between NSDD(+) and PSP (p < 0.05). Thus, mild but significant midbrain atrophy was observed in NSDD(+). A faster rate of decline in the midbrain area and the M/P area ratio was evident in NSDD(+) (p < 0.05).ConclusionThe clinical characteristics of the five patients with SCA31 with concomitant NSDD, together with the topographical pattern of atrophy, were inconsistent with PD, PSP, and multiple system atrophy, suggesting that SCA31 may manifest NSDD in association with the pathomechanisms underlying SCA31.

Longitudinal diffusion MRI analysis using Segis-Net: A single-step deep-learning framework for simultaneous segmentation and registration

This work presents a single-step deep-learning framework for longitudinal image analysis, coined Segis-Net. To optimally exploit information available in longitudinal data, this method concurrently learns a multi-class segmentation and nonlinear registration. Segmentation and registration are modeled using a convolutional neural network and optimized simultaneously for their mutual benefit. An objective function that optimizes spatial correspondence for the segmented structures across time-points is proposed. We applied Segis-Net to the analysis of white matter tracts from N=8045 longitudinal brain MRI datasets of 3249 elderly individuals. Segis-Net approach showed a significant increase in registration accuracy, spatio-temporal segmentation consistency, and reproducibility compared with two multistage pipelines. This also led to a significant reduction in the sample-size that would be required to achieve the same statistical power in analyzing tract-specific measures. Thus, we expect that Segis-Net can serve as a new reliable tool to support longitudinal imaging studies to investigate macro- and microstructural brain changes over time.

Applying artificial intelligence to longitudinal imaging analysis of vestibular schwannoma following radiosurgery

Artificial intelligence (AI) has been applied with considerable success in the fields of radiology, pathology, and neurosurgery. It is expected that AI will soon be used to optimize strategies for the clinical management of patients based on intensive imaging follow-up. Our objective in this study was to establish an algorithm by which to automate the volumetric measurement of vestibular schwannoma (VS) using a series of parametric MR images following radiosurgery. Based on a sample of 861 consecutive patients who underwent Gamma Knife radiosurgery (GKRS) between 1993 and 2008, the proposed end-to-end deep-learning scheme with automated pre-processing pipeline was applied to a series of 1290 MR examinations (T1W+C, and T2W parametric MR images). All of which were performed under consistent imaging acquisition protocols. The relative volume difference (RVD) between AI-based volumetric measurements and clinical measurements performed by expert radiologists were + 1.74%, − 0.31%, − 0.44%, − 0.19%, − 0.01%, and + 0.26% at each follow-up time point, regardless of the state of the tumor (progressed, pseudo-progressed, or regressed). This study outlines an approach to the evaluation of treatment responses via novel volumetric measurement algorithm, and can be used longitudinally following GKRS for VS. The proposed deep learning AI scheme is applicable to longitudinal follow-up assessments following a variety of therapeutic interventions.