Investigation of expert rule bases, logistic regression, and non-linear machine learning techniques for predicting response to antiretroviral treatment

By Mattia C. F. Prosperi, André Altmann, Michal Rosen-Zvi, Ehud Aharoni, G. Borgulya, F. Bazsó, Anders Sönnerborg, Eugen Schülter, Daniel Struck, G. Ulivi and 3 more. 2008

55 citations recorded by the source corpus.

Background The extreme flexibility of the HIV type-1 (HIV-1) genome makes it challenging to build the ideal antiretroviral treatment regimen. Interpretation of HIV-1 genotypic drug resistance is evolving from rule-based systems guided by expert opinion to data-driven engines developed through machine learning methods. Methods The aim of the study was to investigate linear and non-linear statistical learning models for classifying short-term virological outcome of antiretroviral treatment. To optimize the model, different feature selection methods were considered. Robust extra-sample error estimation and different loss functions were used to assess model performance. The results were compared with widely used rule-based genotypic interpretation systems (Stanford HIVdb, Rega and ANRS). Results A set of 3,143 treatment change episodes were extracted from the EuResist database. The dataset included patient demographics, treatment history and viral genotypes. A logistic regression model using high order interaction variables performed better than rule-based genotypic interpretation systems (accuracy 75.63% versus 71.74–73.89%, area under the receiver operating characteristic curve [AUC] 0.76 versus 0.68–0.70) and was equivalent to a random forest model (accuracy 76.16%, AUC 0.77). However, when rule-based genotypic interpretation systems were coupled with additional patient attributes, and the combination was provided as input to the logistic regression model, the performance increased significantly, becoming comparable to the fully data-driven methods. Conclusions Patient-derived supplementary features significantly improved the accuracy of the prediction of response to treatment, both with rule-based and data-driven interpretation systems. Fully data-driven models derived from large-scale data sources show promise as antiretroviral treatment decision support tools.

Subfields: Infectious Diseases, Virology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Infectious Diseases

In Virology

Citations in the corpus

Combining Kernel and Model Based Learning for HIV Therapy Selection (2017, methodology): “The database has previously been used to build models such as the therapy alignment model, to predict the outcome of a particular therapy [17, 18].”
Improved empirical methods in reinforcement-learning evaluation (2015, methodology): “from the EUResist project (Zazzi et al., 2012; Prosperi et al., 2009).”
Prediction of response to antiretroviral therapy by human experts and by the EuResist data‐driven expert system (the EVE study) (2010, unknown): “However, such additional information has been consistently found to increase in several recent studies using rule-based or data-driven systems [13,18,19].”
Machine Learning and Decision Support in Critical Care (2016, methodology): “Machinelearning methods have been found to predict more accurately the response of patients to drugs in retrospective analysis than do rule-based methods used for the same task [194].”
IDEPI: Rapid Prediction of HIV-1 Antibody Epitopes and Other Phenotypic Features from Sequence Data Using a Flexible Machine Learning Platform (2014, background): “…ARVs, including some protease, integrase, nucleoside reverse transcriptase inhibitors, and co-receptor antagonists, the resistance phenotype is determined by the interaction of many sites [8–12], or the protein tertiary structure [13,14], prompting ongoing methodological development (e.g. [15–17]).”
Predicting Response to Antiretroviral Treatment by Machine Learning: The EuResist Project (2012, methodology): “Finally, the mixed-effects engine considers a number of 2nd-and 3rd-order interactions among variables (drug ! drug, drug ! drug ! drug and drug ! mutation), thus accounting for composite effects [13] .”
History-alignment models for bias-aware prediction of virological response to HIV combination therapy (2012, background): “…regression (Wang et al., 2003; Larder et al., 2007; Deforche et al., 2008; Rosen-Zvi et al., 2008; Bickel et al., 2008; Altmann et al., 2009; Prosperi et al., 2009; Bogojeska et al., 2010; Revell et al., 2010), have emerged for tackling the problem of predicting the virological response to…”
Antiretroviral Therapy Optimisation without Genotype Resistance Testing: A Perspective on Treatment History Based Models (2010, background): “Additional details for the study design, the SD definition and data extraction procedures and constraints have been discussed previously [30,31,34].”