Trajectories, bifurcations, and pseudo-time in large clinical datasets: applications to myocardial infarction and diabetes data.

Sergey E Golovenkin,Emmanuel Barillot,Andrei Zinovyev,Alexander N Gorban,Evgeny M Mirkes,Alexander Chervov,Yuliya V Orlova,Jonathan Bac

doi:10.1093/gigascience/giaa128

Abstract

BackgroundLarge observational clinical datasets are becoming increasingly available for mining associations between various disease traits and administered therapy. These datasets can be considered as representations of the landscape of all possible disease conditions, in which a concrete disease state develops through stereotypical routes, characterized by “points of no return" and “final states" (such as lethal or recovery states). Extracting this information directly from the data remains challenging, especially in the case of synchronic (with a short-term follow-up) observations.ResultsHere we suggest a semi-supervised methodology for the analysis of large clinical datasets, characterized by mixed data types and missing values, through modeling the geometrical data structure as a bouquet of bifurcating clinical trajectories. The methodology is based on application of elastic principal graphs, which can address simultaneously the tasks of dimensionality reduction, data visualization, clustering, feature selection, and quantifying the geodesic distances (pseudo-time) in partially ordered sequences of observations. The methodology allows a patient to be positioned on a particular clinical trajectory (pathological scenario) and the degree of progression along it to be characterized with a qualitative estimate of the uncertainty of the prognosis. We developed a tool ClinTrajan for clinical trajectory analysis implemented in the Python programming language. We test the methodology in 2 large publicly available datasets: myocardial infarction complications and readmission of diabetic patients data.ConclusionsOur pseudo-time quantification-based approach makes it possible to apply the methods developed for dynamical disease phenotyping and illness trajectory analysis (diachronic data analysis) to synchronic observational data.

Highlights

Large observational datasets are becoming increasingly available, reflecting physiological state of observed individuals,2 | GigaScience, 20xx, Vol 0, No 0
We develop a methodology of clinical data analysis, based on Golovenkin et al | 3 modeling the multi-dimensional geometry of a clinical dataset as a “bouquet" of diverging clinical trajectories, starting from one or several quasi-normal clinical states
We considered that the small number of continuous numerical variables is not enough to apply the methodology of Categorical Principal Component Analysis (CatPCA) [28]

Summary

Key Points

Large-scale observational clinical datasets represent landscapes of the variety of disease states Diachronic clinical trajectories can be approximated from the multi-dimensional geometry of synchronic data and used for disease dynamical phenotyping ClinTrajan: Python package for finding and analyzing clinical trajectories using elastic principal graph method their lifestyles, exposure to environmental factors, received treatments and passed medical exams. We applied a non-parametric estimator of the cumulative hazard rate function (see Methods) in order to quantify lethal risks along ten identified clinical trajectories in the myocardial infarction complication dataset (Figure 6,A) As a root node in this case, we selected the middle node of one of the internal segments of the principal tree (segment #3 in Figure 7,A), which was characterized by the shortest times spent in the hospital, smallest number of all procedures, no history of inpatient stays or emergency calls in the preceding year, normal predicted (not measured) value of HbA1C, absence of any medication This area of the principal tree was considered as corresponding to quasi-normal state in terms of diabetes treatment. The stratification of patients into different clinical trajectories demonstrates that this is not a universal effect, and that one can distinguish other patient clusters characterized by heavy forms of diabetis characterized by relatively high rates of early readmission

Discussion

Findings

Methods