Predicting clinical progression trajectories of early Alzheimer's disease patients

BACKGROUND

Models for forecasting individual clinical progression trajectories in early Alzheimer's disease (AD) are needed for optimizing clinical studies and patient monitoring.

METHODS

Prediction models were constructed using a clinical trial training cohort (TC; n = 934) via a gradient boosting algorithm and then evaluated in two validation cohorts (VC 1, n = 235; VC 2, n = 421). Model inputs included baseline clinical features (cognitive function assessments, APOE ε4 status, and demographics) and brain magnetic resonance imaging (MRI) measures.

RESULTS

The model using clinical features achieved R2 of 0.21 and 0.31 for predicting 2-year cognitive decline in VC 1 and VC 2, respectively. Adding MRI features improved the R2 to 0.29 in VC 1, which employed the same preprocessing pipeline as the TC. Utilizing these model-based predictions for clinical trial enrichment reduced the required sample size by 20% to 49%.

DISCUSSION

Our validated prediction models enable baseline prediction of clinical progression trajectories in early AD, benefiting clinical trial enrichment and various applications.

Predicting Ambulatory Capacity in Parkinson's Disease to Analyze Progression, Biomarkers, and Trial Design

BACKGROUND

In Parkinson's disease (PD), gait and balance is impaired, relatively resistant to available treatment and associated with falls and disability. Predictive models of ambulatory progression could enhance understanding of gait/balance disturbances and aid in trial design.

OBJECTIVES

To predict trajectories of ambulatory abilities from baseline clinical data in early PD, relate trajectories to clinical milestones, compare biomarkers, and evaluate trajectories for enrichment of clinical trials.

METHODS

Data from two multicenter, longitudinal, observational studies were used for model training (Tracking Parkinson's, n = 1598) and external testing (Parkinson's Progression Markers Initiative, n = 407). Models were trained and validated to predict individuals as having a "Progressive" or "Stable" trajectory based on changes of ambulatory capacity scores from the Movement Disorders Society Unified Parkinson's Disease Rating Scale parts II and III. Survival analyses compared time-to-clinical milestones and trial outcomes between predicted trajectories.

RESULTS

On external evaluation, a support vector machine model predicted Progressive trajectories using baseline clinical data with an accuracy, weighted-F1 (proportionally weighted harmonic mean of precision and sensitivity), and sensitivity/specificity of 0.735, 0.799, and 0.688/0.739, respectively. Over 4 years, the predicted Progressive trajectory was more likely to experience impaired balance, loss of independence, impaired function and cognition. Baseline dopamine transporter imaging and select biomarkers of neurodegeneration were significantly different between predicted trajectory groups. For an 18-month, randomized (1:1) clinical trial, sample size savings up to 30% were possible when enrollment was enriched for the Progressive trajectory versus no enrichment.

CONCLUSIONS

It is possible to predict ambulatory abilities from clinical data that are associated with meaningful outcomes in people with early PD. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Enrichment Benefits of Risk Algorithms for Pulmonary Arterial Hypertension Clinical Trials

Rationale: Event-driven primary endpoints are increasingly used in pulmonary arterial hypertension clinical trials, substantially increasing required sample sizes and trial lengths. The U.S. Food and Drug Administration advocates the use of prognostic enrichment of clinical trials by preselecting a patient population with increased likelihood of experiencing the trial's primary endpoint.Objectives: This study compares validated clinical scales of risk (Comparative, Prospective Registry of Newly Initiated Therapies for Pulmonary Hypertension, the French score, and Registry to Evaluate Early and Long-Term Pulmonary Arterial Hypertension Disease Management [REVEAL] 2.0) to identify patients who are likely to experience a clinical worsening event for trial enrichment.Methods: Baseline data from three pulmonary arterial hypertension clinical trials (AMBITION [a Study of First-Line Ambrisentan and Tadalafil Combination Therapy in Subjects with Pulmonary Arterial Hypertension], SERAPHIN [Study of Macitentan on Morbidity and Mortality in Patients with Symptomatic Pulmonary Arterial Hypertension], and GRIPHON [Selexipag in Pulmonary Arterial Hypertension]) were pooled and standardized. Receiver operating curves were used to measure each algorithm's performance in predicting clinical worsening within the pooled placebo cohort. Power simulations were conducted to determine sample size and treatment time reductions for multiple enrichment strategies. A cost analysis was performed to illustrate potential financial savings by applying enrichment to GRIPHON.Measurements and Main Results: All risk algorithms were compared using area under the receiver operating curve and substantially outperformed prediction per New York Heart Association Functional Class. The REVEAL 2.0's risk grouping provided the greatest time and sample size savings in AMBITION and GRIPHON for all enrichment strategies but lacked appropriate inputs (i.e., N-terminal-proB-type natriuretic peptide) to perform as well in SERAPHIN. Cost analysis applied to GRIPHON demonstrated the greatest financial benefit by enrolling patients with a REVEAL score ≥8.Conclusions: This preliminary study demonstrates the feasibility of risk algorithms for pulmonary arterial hypertension trial enrichment and a need for further investigation.

Boosting power for clinical trials using classifiers based on multiple biomarkers

Machine learning methods pool diverse information to perform computer-assisted diagnosis and predict future clinical decline. We introduce a machine learning method to boost power in clinical trials. We created a Support Vector Machine algorithm that combines brain imaging and other biomarkers to classify 737 Alzheimer's disease Neuroimaging initiative (ADNI) subjects as having Alzheimer's disease (AD), mild cognitive impairment (MCI), or normal controls. We trained our classifiers based on example data including: MRI measures of hippocampal, ventricular, and temporal lobe volumes, a PET-FDG numerical summary, CSF biomarkers (t-tau, p-tau, and Abeta(42)), ApoE genotype, age, sex, and body mass index. MRI measures contributed most to Alzheimer's disease (AD) classification; PET-FDG and CSF biomarkers, particularly Abeta(42), contributed more to MCI classification. Using all biomarkers jointly, we used our classifier to select the one-third of the subjects most likely to decline. In this subsample, fewer than 40 AD and MCI subjects would be needed to detect a 25% slowing in temporal lobe atrophy rates with 80% power--a substantial boosting of power relative to standard imaging measures.