Allergic disease trajectories up to adolescence: Characteristics, early-life, and genetic determinants

BACKGROUND

Allergic diseases often develop jointly during early childhood but differ in timing of onset, remission, and progression. Their disease course over time is often difficult to predict and determinants are not well understood.

OBJECTIVES

We aimed to identify trajectories of allergic diseases up to adolescence and to investigate their association with early-life and genetic determinants and clinical characteristics.

METHODS

Longitudinal k-means clustering was used to derive trajectories of allergic diseases (asthma, atopic dermatitis, and rhinitis) in two German birth cohorts (GINIplus/LISA). Associations with early-life determinants, polygenic risk scores, food and aeroallergen sensitization, and lung function were estimated by multinomial models. The results were replicated in the independent Swedish BAMSE cohort.

RESULTS

Seven allergic disease trajectories were identified: "Intermittently allergic," "rhinitis," "early-resolving dermatitis," "mid-persisting dermatitis," "multimorbid," "persisting dermatitis plus rhinitis," and "early-transient asthma." Family history of allergies was more prevalent in all allergic disease trajectories compared the non-allergic controls with stronger effect sizes for clusters comprising more than one allergic disease (e.g., RRR = 5.0, 95% CI = [3.1-8.0] in the multimorbid versus 1.8 [1.4-2.4] in the mild intermittently allergic cluster). Specific polygenic risk scores for single allergic diseases were significantly associated with their relevant trajectories. The derived trajectories and their association with genetic effects and clinical characteristics showed similar results in BAMSE.

CONCLUSION

Seven robust allergic clusters were identified and showed associations with early life and genetic factors as well as clinical characteristics.

Longitudinal clustering analysis and prediction of Parkinson's disease progression using radiomics and hybrid machine learning

BACKGROUND

We employed machine learning approaches to (I) determine distinct progression trajectories in Parkinson's disease (PD) (unsupervised clustering task), and (II) predict progression trajectories (supervised prediction task), from early (years 0 and 1) data, making use of clinical and imaging features.

METHODS

We studied PD-subjects derived from longitudinal datasets (years 0, 1, 2 & 4; Parkinson's Progressive Marker Initiative). We extracted and analyzed 981 features, including motor, non-motor, and radiomics features extracted for each region-of-interest (ROIs: left/right caudate and putamen) using our standardized standardized environment for radiomics analysis (SERA) radiomics software. Segmentation of ROIs on dopamine transposer - single photon emission computed tomography (DAT SPECT) images were performed via magnetic resonance images (MRI). After performing cross-sectional clustering on 885 subjects (original dataset) to identify disease subtypes, we identified optimal longitudinal trajectories using hybrid machine learning systems (HMLS), including principal component analysis (PCA) + K-Means algorithms (KMA) followed by Bayesian information criterion (BIC), Calinski-Harabatz criterion (CHC), and elbow criterion (EC). Subsequently, prediction of the identified trajectories from early year data was performed using multiple HMLSs including 16 Dimension Reduction Algorithms (DRA) and 10 classification algorithms.

RESULTS

We identified 3 distinct progression trajectories. Hotelling's t squared test (HTST) showed that the identified trajectories were distinct. The trajectories included those with (I, II) disease escalation (2 trajectories, 27% and 38% of patients) and (III) stable disease (1 trajectory, 35% of patients). For trajectory prediction from early year data, HMLSs including the stochastic neighbor embedding algorithm (SNEA, as a DRA) as well as locally linear embedding algorithm (LLEA, as a DRA), linked with the new probabilistic neural network classifier (NPNNC, as a classifier), resulted in accuracies of 78.4% and 79.2% respectively, while other HMLSs such as SNEA + Lib_SVM (library for support vector machines) and t_SNE (t-distributed stochastic neighbor embedding) + NPNNC resulted in 76.5% and 76.1% respectively.

CONCLUSIONS

This study moves beyond cross-sectional PD subtyping to clustering of longitudinal disease trajectories. We conclude that combining medical information with SPECT-based radiomics features, and optimal utilization of HMLSs, can identify distinct disease trajectories in PD patients, and enable effective prediction of disease trajectories from early year data.

Longitudinal Imaging-Based Clusters in Former Smokers of the COPD Cohort Associate with Clinical Characteristics: The SubPopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS)

PURPOSE

Quantitative computed tomography (qCT) imaging-based cluster analysis identified clinically meaningful COPD former-smoker subgroups (clusters) based on cross-sectional data. We aimed to identify progression clusters for former smokers using longitudinal data.

PATIENTS AND METHODS

We selected 472 former smokers from SPIROMICS with a baseline visit and a one-year follow-up visit. A total of 150 qCT imaging-based variables, comprising 75 variables at baseline and their corresponding progression rates, were derived from the respective inspiration and expiration scans of the two visits. The COPD progression clusters identified were then associated with subject demography, clinical variables and biomarkers.

RESULTS

COPD severities at baseline increased with increasing cluster number. Cluster 1 patients were an obese subgroup with rapid progression of functional small airway disease percentage (fSAD%) and emphysema percentage (Emph%). Cluster 2 exhibited a decrease of fSAD% and Emph%, an increase of tissue fraction at total lung capacity and airway narrowing over one year. Cluster 3 showed rapid expansion of Emph% and an attenuation of fSAD%. Cluster 4 demonstrated severe emphysema and fSAD and significant structural alterations at baseline with rapid progression of fSAD% over one year. Subjects with different progression patterns in the same cross-sectional cluster were identified by longitudinal clustering.

CONCLUSION

qCT imaging-based metrics at two visits for former smokers allow for the derivation of four statistically stable clusters associated with unique progression patterns and clinical characteristics. Use of baseline variables and their progression rates enables identification of longitudinal clusters, resulting in a refinement of cross-sectional clusters.

Longitudinal deterioration of white-matter integrity: heterogeneity in the ageing population

Deterioration in white-matter health plays a role in cognitive ageing. Our goal was to discern heterogeneity of white-matter tract vulnerability in ageing using longitudinal imaging data (two to five imaging and cognitive assessments per participant) from a population-based sample of 553 elderly participants (age ≥60 years). We found that different clusters (healthy white matter, fast white-matter decliners and intermediate white-matter group) were heterogeneous in the spatial distribution of white-matter integrity, systemic health and cognitive trajectories. White-matter health of specific tracts (genu of corpus callosum, posterior corona radiata and anterior internal capsule) informed about cluster assignments. Not surprisingly, brain amyloidosis was not significantly different between clusters. Clusters had differential white-matter tract vulnerability to ageing (commissural fibres > association/brainstem fibres). Identification of vulnerable white-matter tracts is a valuable approach to assessing risk for cognitive decline.