Towards the quantitative medicine: data analysis, mathematical modeling, numerical computations

Towards the quantitative medicine:

data analysis, mathematical

modeling, numerical computations

Natalya Kizilova

Interdyscyplinarne Centrum Modelowania Matematycznego i Komputerowego

2

Big data:

volume, velocity, variety

• Mayer-Schönberger V., Cukier K. 'Big Data: A Revolution that Will

Transform how We Live, Work, and Think'. 2013.

• Big data is high volume, high velocity, and/or high variety information assets that require new forms of processing to enable enhanced decision making, insight discovery and process

3

Medical data: from quantity to new quality

• R. Bell (1970) – systems of ODE discribing humoral immune reaction

• Marchuk G. (1977) Mathematical modeling in immunology: model+blood test data= patient specific treatment. (via

Mathematical modeling of cardiovascular system

•

A.Guyton

(1955) - large circulatory model for

long-term control of arterial pressure. Linear

mechano-electric analogy model

•

F.Grodins

(1959) – dynamical model of

homeostasis (regulation of blood pressure,

flow, volume)

•

Sir J.Lighthill

– biofluid dynamics;

physiological flow group at Imperial College

London.

5

Journals on computational medicine and

biology

• Journal of Computational Medicine. Open access journal.

• International Journal of Computational Medicine and Healthcare

• Computers in Biology and Medicine

• Computational Physiology and Medicine

• International Journal of Computational Models and Algorithms in Medicine

• Computational and Mathematical Methods in Medicine. An Interdisciplinary Journal of Mathematical, Theoretical and Clinical Aspects of Medicine

Computational medicine (1)

• is a fast-growing method of using computer models and sophisticated software to figure out how disease develops and how to treat it.

• has begun to leap off the drawing board and land in the hands of doctors who treat patients for heart

ailments, cancer and other illnesses. Using digital tools, researchers have begun to use experimental and clinical data to build models that can unravel complex medical mysteries.

[R. L. Winslow, N. Trayanova, D. Geman, M. I. Miller. Computational Medicine: Translating Models to Clinical Care. Science Translational Medicine, 2012; 4 (158)] Johns Hopkins Institute for Computational Medicine.

7

Computational medicine (2)

• Biology in both health and disease is very complex. It

involves the feed-forward flow of information from the level of the gene to protein, networks, cells, organs and organ systems. This is already complex, and to make matters even more difficult, it also involves feed-back pathways by which, for example, proteins, mechanical forces at the level of tissues and organs, and environmental factors regulate

function at lower levels such as the gene.

• Computational models help us to understand these

complex interactions, the nature of which is often highly complex and non-intuitive.

• The models allow researchers to understand disease

mechanisms, aid in diagnosis, and test the effectiveness of different therapies/surgeries. By using computer models potential therapies can be tested "in silico" at high speed. The results can then be used to guide further experiments to gather new data to refine the models until they are highly predictive.

Computational medicine (3)

• Computational physiological medicine is using computer models to look at how biological systems change over time from a healthy to an unhealthy state. This approach is being used to help develop better treatments for cancer, diabetes and heart disease.

• Computational anatomy uses medical images to detect changes, for example, in the shape of various structures in the brain. Researchers have found shape changes that appear to be associated with

Alzheimer's disease and neuropsychiatric disorders, such as schizophrenia.

• Computational models of electrical activity in the heart are on their way to being used to guide doctors in preventing sudden cardiac death and in diagnosing and treating those at risk for it.

• Advanced mathematical models are allowing researchers to better understand how networks of molecules are implicated in cancer and then use this knowledge to predict which patients are at risk of

9

Biocuration, the activity of organizing,

representing and making biological information

accessible to both humans and computers, has

become an essential part of biological discovery,

biomedical research, diagnostics, and treatment.

What is urgently needed: http://www.biocurator.org/

1) Massive exchange of data between journal publications and databases;

2) Curators, researchers and university

administrations should develop an accepted recognition structure to facilitate community- based curation efforts

3) and increase the visibility and support of scientific curation as a professional career.

Biocuration

• To extract knowledge from published papers

• To interact with researchers to facilitate direct data submissions to open access databases

• To connect information from different sources in a coherent and comprehensible way

• To inspect and correct automatically predicted gene structures and protein sequences to provide high-quality proteomes

• To develop and manage structured controlled vocabularies

that are crucial for data relations and the logical retrieval of large data sets

• To integrate knowledge bases to represent complex systems such as metabolic pathways and protein-interaction networks.

• To correct inconsistencies and errors in data representation

• To help data users to render their research more productive in a timely manner

11

Clinical diagnoses and basic investigations

are critically dependent on the ability to

record and analyze physiological signals.

• ECG and HRV recordings from patients at a high risk of

sudden death;

• BP monitoring for diagnostics and treatment

hyper(hypo)tension;

• Fluctuations of hormones and molecular biological signal messengers and transducers in neuroendocrine dynamics;

• Multiparameter recordings in sleep apnea;

• Long-term repeatitive measurements for control slowly

developed chronic diseases (Parkinson’s, senility dementia, etc)

• Multiparameter recordings in epilepsy;

Holter monitoring (3-8 electrodes, 24-48 h)

•

Detailed information on

electric activity of heart

•

Long-term data storage

•

Data analyses

13

What we have:

• The measured physiological signals represent the processes that are multivariable, complex,

nonstationary and nonlinear;

• Conventional mathematical methods for data

analysis are developed for steady linear processes, and based on normal distributions, etc; like analysis of means, standard deviations, histograms, power- spectrum analysis, correlation analysis;

• Data Analysis as an emerging activity/discipline, one distinct from Mathematical Statistics and requiring its own literature [J. Tukey ‘Future of Data Analysis’

1962.];

• In physiological signals we have a treasure of

important information that remains unrevealed and unused.

What we need:

• Nonlinear signal processing;

• Principal components analysis;

• Independent components analysis;

• Computational harmonic analysis;

• Randomized algorithms;

• Incomplete data analysis;

• FEM modeling and analyses;

15

Multiscale approach

for data analysis

ECG+HRV+BP+

+respiration+hormons+ +blood O₂ saturation

FEM of orthodontic tooth movement

17

FEM is recognized in

•

Orthopedy

(individual design of implants,

prosthesis, insoles, footwear);

•

Vascular surgery

(stents, bypass, plastic

surgery, varicose veins);

19

Modelling the circle of Willis to assess the

effects of anatomical variations and occlusions

on cerebral flows

Digital CVS Model

21

Different topology of arterial beds of the

large intestine

Wave propagation and reflection in vsaculatures

with loops

23

Until now the innovative research of complex

biomedical signals has been hampered

by the lack of

•

Open access data resources,

•

Analytical tools for complex data analysis,

•

Mathematical models,

•

Open source software,

24

PhysioNet

https://www.physionet.org/

• Temporary supported PhysioBank data : well- characterized, carefully, multiply reviewed and corrected data;

• PhysioToolkit software : related open-source

rigorously tested software for analysis and prediction complex systems dynamics;

• Data and software contributed by authors of published

articles + on-line full text access;

• Data Chromatix is a technique for visualizing trends in biomedical signals by bringing memory of the

system's past behavior into the current display window.

[Goldberger A.L., et al. PhysioBank, PhysioToolkit, and PhysioNet : Components of a New Research Resource for Complex Physiologic Signals. Circulation. 2000. 101:

25

PhysioNet is approved as official repository for

Scientific Data by Nature Publishing Group

•

Computing in cardiology :Reducing false

arrhythmia alarms in Intensive Care Unit;

•

Analysis of event-related potentials in

Brain-Computer Interface recordings;

•

Posturographic data for locomotory,

neural, visual, balance disorders;

•

Motion capture and gait analysis data and

software.

PhysioNet databases and toolkit software

• Importing and exporting data

• Signal and time series analysis

• Physiologic models and simulations (Synthesize cardiovascular system variables and ECGs)

• Software (Development and evaluation of ECG analyzers)

27

Data mining

system for

providing

analytical

information on

brain tumors to

public health

decision makers

[R.S. Santos, et al. computer methods and programs in biomedicine. 2013. 269–282.]

Traditional Chinese Medicine (TCM) clinical

data warehouse for medical knowledge

29

‘THE DIGITAL UNIVERSE IN 2020:

Big Data and the Biggest Growth’

• Embedded and medical devices. In the future,

sensors of all types, including those that may be

implanted into the body, will capture vital biometrics, track medicine effectiveness, correlate bodily

activity with health, monitor potential outbreaks of viruses, etc. — all in real time.

• Big data analysis : medical information +

sociological data + geography, weather, solar

activity, ecology, political events, etc – will add new

dimensions in medicine, biology, sociology,

ecology, culture, etc.

[Sponsored by EMC (data storage, information

security, virtualization, analytics, cloud computing and other products and services)]

31

Conclusions

• Big data + mathematical models + computer simulations give a new dimension for medical

diagnosis, planning of therapy/surgery/rehabilitation : from intuitive (experience-based) diagnostics and treatment to expert systems (machine learning

algorithms, neural networks, big data analysis);

• Importance of new mathematical tools, concepts of data measurement, storage and mining will be

crucial for future medicine as inyterdisciplinary science.