Curriculum Vitae

Johan van Soest

Department of Radiation Oncology (MAASTRO)
GROW School for Oncology & Developmental Biology
Maastricht University Medical Centre+ (MUMC+)
Maastricht, The Netherlands

Institute of Data Science
Faculty of Science and Engineering
Maastricht University
Maastricht, The Netherlands

Email: johan.vansoest at maastro.nl or j.vansoest at maastrichtuniversity.nl

Education

2013-2018 Doctor of Philosophy (PhD)
Maastricht University
Clinical Data Science in Radiotherapy
Thesis (DOI): 10.26481/dis.20181128js
2010-2012 Master Medical Informatics
University of Amsterdam
Activities:Member of the course evaluation committee
 IPHIE masterclass
2009-2010 Pre-Master Medical Informatics
University of Amsterdam
Components: Informatics, Care Information, Medicine
2005-2009 Bachelor of ICT
Fontys University of Applied Science
Major: Software Engineering
Minor: Medical Technology

Work experience

June 2017 - present Postdoctoral researcher
1 Department of Radiation Oncology (MAASTRO), GROW School for Oncology and Developmental Biology, Maastricht University Medical Centre
2 Institute of Data Science, Maastricht University
Involved in the following projects:
  • SAGE: Successor of the VATE project (see below). This project focuses on data quality and similarity in a distributed setting (where data stays inside the hospitals and is not being shared).
  • Limburg Meet (LIME) - Personal Health Train: development of a communication platform between (medical) institutes in the Limburg region. Institutes are primary care, general hospitals and specialized care facilities. The goal is to minimize the data transfer and duplication needed, while enabling patients to give control over their data.
  • VWData - Personal Health Train: distributed learning on vertical partitioned data sources. Specifically, enabling privacy-preserved analysis on information from a longitudinal cohort study and the national statistics agency (CBS).
  • Vektis - Personal Health Train: Application of the Personal Health Train and distributed learning concept on healthcare reimbursement information.
  • FAIRSights: Development and investigation into a software system to make clinical data FAIR (findable, accessible, interoperable and reusable).
  • FAIRcomml: Development and investigation into a software system to describe prediction models, and to link these models to FAIR data.
June 2017 - present Chief Technical Officer and co-founder of Medical Data Works B.V.
Translate research software into clinical products and practice.
June 2016 - September 2016 Visiting Research Scholar, University of Toronto
Setting up a collaborating infrastructure to perform distributed machine learning on routine clinical datasets (nightly extraction of datasets). Special focus on cohort differences, and when to decide to re-train a prediction model.
July 2014 - September 2014 Visiting Research Scholar, University of Sydney
Setting up a collaborating infrastructure (equally as used in the VATE project) to perform distributed machine learning on routine clinical datasets (nightly extraction of datasets). Main reason to choose distributed machine learning is to keep datasets (and therefore patient information) in the hospitals. Instead, the candidate prediction models are sent to the local centers, rather than centralizing datasets.
August 2013 - June 2017 PhD Student, Maastro Clinic.
Involved in the following projects:
  • VATE: Validation of High TEchnology using Machine Learning. This project is dedicated to rectal cancer, however the methods used regarding data extraction, representation, and execution of distributed machine learning are related to other distributed learning projects within the Knowledge Engineering group of Maastro Clinic.
  • SeDI: Semantic Dicom; this project aims at developing a semantic PACS. Regarding data representation, the DICOM header is parsed and represented using a (semantically interoperable) ontology. Furthermore, additional image-derived information can be calculated and stored using the same ontological representation.
  • SAGE: Successor of the VATE project. This project focuses on data quality and similarity in a distributed setting (where data stays inside the hospitals and is not being shared).
  • CloudAtlas: Successor of the SeDI project. This project focuses on optimizing the plan creation time in Radiation Therapy, without reductions of treatment plan quality.
  • BBMRI 2.0: Successor of the development tasks of CTMM-TraIT. In this project, we are focussing on the semantically interoperable storage of image (DICOM) derived information and image annotations.
August 2012 - present Technical Application Management, Maastro Clinic.
Technical application manager for the CancerData.org image archive, and the CTMM-TraIT BioMedical Image Archive (bmia.nl). Both archives are used to store research-related DICOM information, which is anonymized before leaving the local hospital.
November 2011 - July 2012 Scientific Research Project (SRP) / Master thesis for the Master Medical Informatics.
Subject: Score Engineering - Optimizing the SOFA score for predicting intensive care mortality using evolutionary strategies.
Description: Scoring systems are ubiquitous in medicine, especially in intensive care departments. We developed an evolutionary algorithm in R to optimize the SOFA score for mortality prediction. Evaluation of newly defined scoring schemes was performed using (bootstrapped) area under ROC (AUC), Brier score, and Hosmer-Lemeshow C statistic.
January 2011 - July 2011 Internship at the Academic Medical Center, Amsterdam. Dept. of Medical Informatics collaborating with dept. of Geriatrics.
Identification of risk factors for recurring falls in elderly patients, comprising of two tasks:
1.: Development of a clinical registry system for medical process control and data collection for research purposes.
2.: Analysis of collected data to identify risk factors for recurring falls using statistical analysis. Furthermore, automation of medication name correction using string similarity measures.
February 2009 - July 2009 Graduation Internship Bachelor of ICT
Internship at UMC St. Radboud Nijmegen, dept. of Nuclear medicine.
Extension of their internally developed framework for application development. Final applications are used as plug-ins for Siemens Syngo e.soft workstation. Assignment was to extend the framework with tools for 3d visualisation and delineation, using IDL (Interactive Data Language).
July 2008 - December 2011 Software developer at Zeegers Advertisings B.V.
All-round software development on different platforms to support their operations. Applications developed on different platforms (C# .NET, Java, PHP), based on their requirements.
Highlight projects: Backoffice administration system, digital reports for elementary schools

Additional courses followed

6-12 March 2015 "Basic Clinical Radiobiology"
ESTRO course
26-28 March 2014 "Protege Short Course"
Standford University
October 2013 "Machine learning"
Andrew Ng, Standford University (Coursera course)
17-21 June 2013 "Managing and Integrating Information in the Life Sciences"
Netherlands BioInformatics Center (NBIC), LUMC
18 April 2013 "Good Clinical Practice"
National Course, MUMC+
21-24 March 2013 "Evidence and new challenges in Rectal Cancer"
ESTRO course

Additional activities

2015 - 2017 Student Editorial Board member
Journal: Methods of Information in Medicine

Membership of Professional Associations

Current and completed awarded grants

2019 - 2021 Varian Medical Systems FAIRcomml Principal Investigator
2018 - 2019 Varian Medical Systems FAIRsights Principal Investigator
2016 - 2018 Netherlands eScience Center path finder Data Quality Co-Principal Investigator

Teaching and supervision

PhD students: MSc students: BSc students:

Published Scientific contributions