RESUMEN
Alzheimer's disease (AD) is the most prevalent neurodegenerative disease, yet its current treatments are limited to stopping disease progression. Moreover, the effectiveness of these treatments remains uncertain due to the heterogeneity of the disease. Therefore, it is essential to identify disease subtypes at a very early stage. Current data-driven approaches can be used to classify subtypes during later stages of AD or related disorders, but making predictions in the asymptomatic or prodromal stage is challenging. Furthermore, the classifications of most existing models lack explainability, and these models rely solely on a single modality for assessment, limiting the scope of their analysis. Thus, we propose a multimodal framework that utilizes early-stage indicators, including imaging, genetics, and clinical assessments, to classify AD patients into progression-specific subtypes at an early stage. In our framework, we introduce a tri-modal co-attention mechanism (Tri-COAT) to explicitly capture cross-modal feature associations. Data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) (slow progressing = 177, intermediate = 302, and fast = 15) were used to train and evaluate Tri-COAT using a 10-fold stratified cross-testing approach. Our proposed model outperforms baseline models and sheds light on essential associations across multimodal features supported by known biological mechanisms. The multimodal design behind Tri-COAT allows it to achieve the highest classification area under the receiver operating characteristic curve while simultaneously providing interpretability to the model predictions through the co-attention mechanism.
RESUMEN
Polygenic risk scores (PRS) are increasingly used to estimate the personal risk of a trait based on genetics. However, most genomic cohorts are of European populations, with a strong under-representation of non-European groups. Given that PRS poorly transport across racial groups, this has the potential to exacerbate health disparities if used in clinical care. Hence there is a need to generate PRS that perform comparably across ethnic groups. Borrowing from recent advancements in the domain adaption field of machine learning, we propose FairPRS - an Invariant Risk Minimization (IRM) approach for estimating fair PRS or debiasing a pre-computed PRS. We test our method on both a diverse set of synthetic data and real data from the UK Biobank. We show our method can create ancestry-invariant PRS distributions that are both racially unbiased and largely improve phenotype prediction. We hope that FairPRS will contribute to a fairer characterization of patients by genetics rather than by race.