Eu Resist : An integrated system for management of antiretroviral - - PowerPoint PPT Presentation

IST-2004-027173

EuResist: An integrated system for management of antiretroviral drug resistance

Francesca Incardona (Informa s.r.l.)

2 Arevir 2008

EuResist : to support clinicians treating HI V patients

The EuResist project aims at developing an integrated system for prediction of response to antiretroviral treatment

Started: January 1st 2006 Will end: September 30th 2008 An integrated and comprehensive genotype-response database has been created. Several distinct prediction engines have been developed and combined into the EuResist Prediction System. Novel approach: viral genotype data integrated with clinical data. Focus is on genotype - response correlation A critical amount of resistance data is needed.

3 Arevir 2008

EuResist consortium

IBM Haifa Research Lab. Informa s.r.l. Coordinator University of Siena Scientific coorrdinator Max Plank Institute Karolinska Institute University hospital of Cologne Kingston University RMKI (Hungary) Roma 3 University Subcontractor

4 Arevir 2008

EuResist objectives

The project specific aims were:

• To create the EuResist Integrated DataBase by

merging three large resistance data sets: ARCA (Italy), AREVIR (Germany) and Karolinska’s (Sweden)

• To define a ‘standard datum’ aimed at determining

the minimum number of variables that maximise the information

• To study different methods to build the predictive

engines

• To compare and combine the different methods into

the final EuResist Predictive System To make the final System freely available on the Web

5 Arevir 2008

ARCA

AREVIR

KI

EuResist System schem a

6 Arevir 2008

Figures as of March 2008 (simplified schema) PATIENTS Patient ID Gender Year of birth Country of origin HCV status HBV status THERAPY PatientID Treatment regimen Date of start Date of stop Reason for change/stop GENOTYPE PatientID Date Sequence Method CD4 PatientID Date CD4/mmc CD4% HIV RNA PatientID Date Copies/ml <LLD (undetectable) Method AIDS EVENTS PatientID Event Date STATUS PatientID Date Followed Lost Died

6 4 .8 6 4 1 8 .4 6 7 3 0 4 .8 3 8 2 2 .0 0 6 2 4 0 .7 9 5

The I ntegrated EuResist DB

Integrates clinical and virological data from multiple sources

7 Arevir 2008

The I ntegrated EuResist DB

Integrates clinical and virological data from multiple sources

A contribution to health information standards has been defined (HL7) based on EuResist DB data structure

8 Arevir 2008

The Forum for Collaborative HIV Research

time Genotype Treatment switch Viral load 0 to 12 weeks Short-term model: 4-12 weeks Viral load Pre-therapy HIV RNA Reason for change CD4 Patient demographics (age, gender, race, route of infection) Past genotypes Past treatments Past AIDS diagnosis

… plus “derived” features (e. g. the “genetic barrier” defined as the probability not to develop resistance to the drugs included in the regimen)

EuResist “Classical” Standard Datum

9 Arevir 2008

The prediction engines

An array of independent prediction engines based on different models has been realised: Instance Based Reasoning (IBR): local fitting procedure which selects compact subsets of predictive variables: large amount of data is crucial Generative-discriminative engine: global fitting method employs first a generative model that uses all data and then applies Kernel method (or Support Vector Machines) for prediction Evolutionary model: includes genetic evolutionary information into derived features (not in the SD) and uses different machine learning techniques for prediction Fuzzy logic: an existing predictor retrained on the EuResist IDB to generate derived features (not in the SD) for the IBR engine

10 Arevir 2008

Success/ failure prediction performance

AUC Accuracy = (1 – Error rate)

Train Test Train Test

I BM

0.747 (0.027) 0.744 0.745 (0.024) 0.724

MPI

0.766 (0.030) 0.768 0.754 (0.031) 0.748

I nforma/ RM3

0.758 (0.019) 0.745 0.748 (0.031) 0.757

I BM

0.768 (0.025) 0.76 0.752 (0.028) 0.757

MPI

0.789 (0.023) 0.804 0.780 (0.032) 0.751

I nforma/ RM3

0.762 (0.021) 0.742 0.754 (0.030) 0.757

Maximal Feature Set Minimal Feature Set

Maximal feature set performs better than minimal

11 Arevir 2008

The EuResist system prediction results

Engines com bination: after exploring several methods, simple mean combination has been chosen Results

• Mean combiner

learns faster than

single engines

• Performs better

than current state of the art (comparison with Stanford HIVDB)

12 Arevir 2008

The EVE evaluation study

13 Arevir 2008

The EVE evaluation study

14 Arevir 2008

The EuResist Web interface

I nput

15 Arevir 2008

The EuResist Web interface - input

I nput

16 Arevir 2008

The EuResist Web interface

Output

17 Arevir 2008

The EuResist Web interface - Output

18 Arevir 2008

The future

EuResist Web based free decision support system on-

line

Clustering: EuResist is already expanding from the 3

initial databases and entering new collaborations based on reliable and fair rules

EuResist network GEIE: A European Grouping to

deploy project results

19 Arevir 2008

19

Clustering

EuResist Network partner in upcoming CHAIN

project (VIIFP Health Programme)

Luxembourg clinic joined the EuResist IDB

Join us!

if you want to collaborate please contact me or visit the web site www.euresist.org

Cooperation with Virolab project

(www.virolab.org)

20 Arevir 2008

20

Clustering

EuResist provided data to Virolab for projects based on

Virolab+ EuResist data:

Risk factors of accumulation of resistance during failing treatment

Influence of primary resistance mutations or substitutions

• n CD4+ T-cell count evolution among HIV-1 positive

patients while naïve to antiretrovirals

Evaluation of the predictive performance of fitness landscapes for therapy outcome of baseline estimated fitness and genetic barrier towards resistance Quantification of virological and immunological response

• f decision support systems

Virolab is next to provide data to EuResist

21 Arevir 2008

Rules for participation

22 Arevir 2008

EuResist Network GEI E

ESTABLI SHED! A European Grouping (Informa, UniSiena, Max Plank,

Karolinska - UniKoeln next to join) to deploy project results, maintain and update the IDB and the prediction system

It will (hopefully!) collect financial support from

private companies and/or governmental institutions to carry on EuResist activities

The Grouping is a partnership without profit goals It will give free services to the public

23 Arevir 2008

Main results at now

Technical The IDB with more than 18.000 patients The EuResist prediction system performing better than current state of the art (Stanford HIVDB) Scientific

Clinical, immuno-virologic, therapeutic and socio-demographic

features in addition to viral genotype, as well as derived features, improve prediction results

Prediction results seem not to be significantly improved just by

further increasing the training data size, given models and features.

Standard datum to be reformulated with long-term model?

Strategic

EuResist Network GEIE Clustering

24 Arevir 2008

Aknowledgements to:

• Maurizio Zazzi (University of Siena)
• Andre Altmann (Max Plank Inst.)
• Mattia Prosperi (Informa s.r.l.- University of Roma3)
• Monica Merito (Informa s.r.l.)
• All EuResist team

Thank you

f.incardona@informacro.info – www.euresist.org