Isaacson, Richard

AbstractIntroductionWe sought to assess longitudinal electronic health records (EHRs) using machine learning (ML) methods to computationally derive probable Alzheimer's Disease (AD) and related dementia subphenotypes.MethodsA retrospective analysis of EHR data from a cohort of 7587 patients seen at a large, multi‐specialty urban academic medical center in New York was conducted. Subphenotypes were derived using hierarchical clustering from 792 probable AD patients (cases) who had received at least one diagnosis of AD using their clinical data. The other 6795 patients, labeled as controls, were matched on age and gender with the cases and randomly selected in the ratio of 9:1. Prediction models with multiple ML algorithms were trained on this cohort using 5‐fold cross‐validation. XGBoost was used to rank the variable importance.ResultsFour subphenotypes were computationally derived. Subphenotype A (n = 273; 28.2%) had more patients with cardiovascular diseases; subphenotype B (n = 221; 27.9%) had more patients with mental health illnesses, such as depression and anxiety; patients in subphenotype C (n = 183; 23.1%) were overall older (mean (SD) age, 79.5 (5.4) years) and had the most comorbidities including diabetes, cardiovascular diseases, and mental health disorders; and subphenotype D (n = 115; 14.5%) included patients who took anti‐dementia drugs and had sensory problems, such as deafness and hearing impairment.The 0‐year prediction model for AD risk achieved an area under the receiver operating curve (AUC) of 0.764 (SD: 0.02); the 6‐month model, 0.751 (SD: 0.02); the 1‐year model, 0.752 (SD: 0.02); the 2‐year model, 0.749 (SD: 0.03); and the 3‐year model, 0.735 (SD: 0.03), respectively. Based on variable importance, the top‐ranked comorbidities included depression, stroke/transient ischemic attack, hypertension, anxiety, mobility impairments, and atrial fibrillation. The top‐ranked medications included anti‐dementia drugs, antipsychotics, antiepileptics, and antidepressants.ConclusionsFour subphenotypes were computationally derived that correlated with cardiovascular diseases and mental health illnesses. ML algorithms based on patient demographics, diagnosis, and treatment demonstrated promising results in predicting the risk of developing AD at different time points across an individual's lifespan.

AbstractMutations in microtubule‐associated protein tau gene (MAPT) cause frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP‐17). Here, we describe a patient with FTDP‐17 and a novel missense mutation in exon 13 of MAPT, p.E372G. We compare clinicopathologic features of this patient to two previously unreported patients with another exon 13 mutation, p.G389R. The patient with the p.E372G mutation was a 40‐year‐old man with behavioral variant frontotemporal dementia (bvFTD), who subsequently developed agrammatic speech and parkinsonism. One of the FTDP‐17 patients with p.G389R mutation presented at age 24 with agrammatic variant of primary progressive aphasia, and subsequently behavioral dysfunction. The other presented at age 53 with bvFTD, followed by agrammatic speech and corticobasal syndrome. Neuropathologic features of FTDP‐17 due to p.E372G were similar to those of p.G389R, including tau‐immunoreactive Pick body‐like neuronal inclusions and swollen, tapering thread‐like processes in white matter immunoreactive for 3‐repeat and 4‐repeat tau. Biochemical analysis of insoluble tau showed similar isoform compositions in p.E372G and p.G389R. Functional studies of the p.E372G mutation showed marked increase in tau filament formation and its reduced ability to promote microtubule assembly. Together these findings indicate that p.E372G is a pathogenic MAPT mutation that causes FTDP‐17 similar to p.G389R.

A recently identified variant within the fat mass and obesity-associated ( FTO ) gene is carried by 46% of Western Europeans and is associated with an ~1.2 kg higher weight, on average, in adults and an ~1 cm greater waist circumference. With >1 billion overweight and 300 million obese persons worldwide, it is crucial to understand the implications of carrying this very common allele for the health of our aging population. FTO is highly expressed in the brain and elevated body mass index (BMI) is associated with brain atrophy, but it is unknown how the obesity-associated risk allele affects human brain structure. We therefore generated 3D maps of regional brain volume differences in 206 healthy elderly subjects scanned with MRI and genotyped as part of the Alzheimer's Disease Neuroimaging Initiative. We found a pattern of systematic brain volume deficits in carriers of the obesity-associated risk allele versus noncarriers. Relative to structure volumes in the mean template, FTO risk allele carriers versus noncarriers had an average brain volume difference of ~8% in the frontal lobes and 12% in the occipital lobes—these regions also showed significant volume deficits in subjects with higher BMI. These brain differences were not attributable to differences in cholesterol levels, hypertension, or the volume of white matter hyperintensities; which were not detectably higher in FTO risk allele carriers versus noncarriers. These brain maps reveal that a commonly carried susceptibility allele for obesity is associated with structural brain atrophy, with implications for the health of the elderly.