PhD Scientific Days 2024
Budapest, 9-10 July 2024
Poster Session K - Theoretical and Translational Medicine 2.
Using Contrastive Learning to Create Vector Representations of Adverse Events from Spontaneous Reports, Facilitating Machine-Assisted Pharmacovigilance
Előadó neve
Balogh, Olivér, MSc
Neptun code
HYKT8D
Előadó munkahelye
Semmelweis University Department of Pharmacology and Pharmacotherapy
Előadó telefonszáma
06203685079
Előadó e-mail címe
balogh.oliver.marton@semmelweis.hu
Az előadás címe
Using Contrastive Learning to Create Vector Representations of Adverse Events from Spontaneous Reports, Facilitating Machine-Assisted Pharmacovigilance
Szerző(k) neve és munkahelye
Balogh Olivér Márton1,2, Pétervári Mátyás1,3, Csernák Áron Márk1, Puhl Eszter1, Horváth András4, Ferdinandy Péter1,5, Ágg Bence1,2,5
1: Department of Pharmacology and Pharmacotherapy, Semmelweis University, Budapest
2: HUN-REN–SU System Pharmacology Research Group, Department of Pharmacology and Pharmacotherapy, Semmelweis University, Budapest
3: Sanovigado Kft, Budapest
4: Faculty of Information Technology and Bionics, Pázmány Péter Catholic University, Budapest
5: Pharmahungary Group, Szeged
Bemutatás módja
Poszter
Szekció
Poster Session K - Theoretical and Translational Medicine 2.
Language of the presentation
English
Preferred session
Theoretical and Translational Medicine
Összefoglaló szövege
Introduction
In the pursuit of drug safety, drug monitoring systems play a crucial part, yet the tools of pharmacovigilance face challenges by relying on text-based data that is unsuitable for contemporary machine learning approaches.
Aims
Here, we adapt contrastive learning algorithms, one from natural language processing and one from computer vision, to generate vector representations of adverse events from spontaneous reports to serve as machine-readable resources in pharmacovigilance.
Method
We present comprehensive analyses of the resulting representations through qualitative and statistical evaluation methods, and support our visualizations of the results with cluster analysis. We demonstrate the applicability of the representations by our down-stream classifier model, performing drug-event causality assessment. Model performance was measured on independent benchmarks and summarized by area under the receiver operating characteristic (AUROC) values.
Results
Evaluation of the representations revealed patterns reflecting both functional and causal relations of the adverse events, and demonstrated how they capture drug-safety related information better than existing taxonomies. Using the representations, our classifier model outperformed previous methods (AUROC: 0.88 vs 0.75).
Conclusion
As such, we advance machine-assisted pharmacovigilance by providing interpretable adverse event representations that are flexible and effective feature vectors for training arbitrary models. This enables the application of machine learning approaches to aid the decision making of experts during causality assessment, leading to improved and more cost-effective drug safety.
Funding
The research was financed by the Thematic Excellence Programme (2020-4.1.1.-TKP2020) of the Ministry for Innovation and Technology in Hungary, within the framework of the Therapeutic Development and Bioimaging thematic programmes of the Semmelweis University. This project has received funding from the HUN-REN Hungarian Research Network. Project no. RRF-2.3.1-21-2022-00003 has been implemented with the support provided by the European Union. This study has been supported by a Semmelweis 250+ Excellence Fellowship and by the ÚNKP-23-3-I New National Excellence Program of the Ministry for Culture and Innovation from the source of the National Research, Development and Innovation Fund.
University
Semmelweis University
Supervisor
Ágg Bence
Publication of my abstract
I do not give consent to the publication of my abstract on the website of the congress.
Kind
Szabad
Status
elfogadva
Accepted presentation method
poszter
Előadás fájl jóváhagyás
nem rendelkezett róla
Előadó
7403
Start
16:00
End
16:03