Unsupervised machine learning algorithms identify expected haemorrhage relationships but define unexplained coagulation profiles mapping to thrombotic phenotypes in hereditary haemorrhagic telangiectasia

Abstract

Hereditary haemorrhagic telangiectasia (HHT) can result in challenging anaemia and thrombosis phenotypes. Clinical presentations of HHT vary for relatives with identical casual mutations, suggesting other factors may modify severity. To examine objectively, we developed unsupervised machine learning algorithms to test whether haematological data at presentation could be categorised into sub‐groupings and fitted to known biological factors. With ethical approval, we examined 10 complete blood count (CBC) variables, four iron index variables, four coagulation variables and eight iron/coagulation indices combined from 336 genotyped HHT patients (40% male, 60% female, 86.5% not using iron supplementation) at a single centre. T‐SNE unsupervised, dimension reduction, machine learning algorithms assigned each high‐dimensional datapoint to a location in a two‐dimensional plane. k‐Means clustering algorithms grouped into profiles, enabling visualisation and inter‐profile comparisons of patients’ clinical and genetic features. The unsupervised machine learning algorithms using t‐SNE and k‐Means identified two distinct CBC profiles, two iron profiles, four clotting profiles and three combined profiles. Validating the methodology, profiles for CBC or iron indices fitted expected patterns for haemorrhage. Distinct coagulation profiles displayed no association with age, sex, C‐reactive protein, pulmonary arteriovenous malformations (AVMs), ENG/ACVRL1 genotype or epistaxis severity. The most distinct profiles were from t‐SNE/k‐Means analyses of combined iron‐coagulation indices and mapped to three risk states – for venous thromboembolism in HHT; for ischaemic stroke attributed to paradoxical emboli through pulmonary AVMs in HHT; and for cerebral abscess attributed to odontogenic bacteremias in immunocompetent HHT patients with right‐to‐left shunting through pulmonary AVMs. In conclusion, unsupervised machine learning algorithms categorise HHT haematological indices into distinct, clinically relevant profiles which are independent of age, sex or HHT genotype. Further evaluation may inform prophylaxis and management for HHT patients’ haemorrhagic and thrombotic phenotypes.

Functional filter for whole-genome sequencing data identifies HHT and stress-associated non-coding SMAD4 polyadenylation site variants >5 kb from coding DNA

“…All three individuals with variants 1 and 2 had clinically confirmed HHT. 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 35 , 66 After identification of the SMAD4 variants, recruiting clinicians also reported SMAD4- compatible clinical phenotypes: The first-degree relatives with variant 2 had no other identified cause to HHT. They each experienced daily nosebleeds and had classical HHT telangiectasia, and one had pulmonary arteriovenous malformations requiring treatment and hemihypertrophy (left-right axis defect).…”

“…For this, the annotated WGS files were retrieved for the 34 validation set DNAs where a causative variant had already been identified in known HHT genes through clinical pipelines. 13 , 14 , 15 , 16 , 34 , 35 Variant information was collected through unique variant IDs consisting of chromosome number, variant starting position, reference sequence, and altered sequence (e.g., chr1:111_C/TTT). To confirm that no two variants were represented by the same variant ID, the full list was compared to a set where only unique values were stored, and the two lists were identical.…”

Pathogenic Variant Frequencies in Hereditary Haemorrhagic Telangiectasia Support Clinical Evidence of Protection from Myocardial Infarction

“…Case notes were reviewed with ethical approval from the Hammersmith and Queen Charlotte’s and Chelsea Research Ethics Committee (LREC 2000/5764). Key findings in this cohort referring to individual clinical features corresponding to the respective genetic variants were published previously [ 57 , 71 , 72 , 85 ]. In the current study, we focused on the variants and the number of times each variant was reported.…”

“…Large prospectively-accrued series have identified environmental factors that contribute to more severe clinical manifestations once AVMs are present, particularly anaemia aggravating high-output states [ 49 ] and associated with earlier mortality [ 50 , 51 ]. Additionally, we have identified that in HHT, iron deficiency is associated with venous thromboemboli [ 52 ] and ischaemic strokes through pulmonary AVMs [ 53 ]; iron treatments can be associated with worsening nosebleeds in approximately 1 in 20 patients [ 54 , 55 ]; higher serum iron levels and intravenous iron treatments are independent risk factors for cerebral abscess due to pulmonary AVMs [ 56 ]; and, most recently, that unsupervised machine learning identifies relevant heterogeneity in blood indices [ 57 ]. Despite HHT clinical risks, recent European studies have shown that life expectancy is normal, or near-normal, in HHT [ 50 , 58 , 59 , 60 ].…”

“…factors for cerebral abscess due to pulmonary AVMs [56]; and, most recently, that unsupervised machine learning identifies relevant heterogeneity in blood indices [57]. Despite HHT clinical risks, recent European studies have shown that life expectancy is normal, or near-normal, in HHT [50,[58][59][60].…”