Radiomics: theory and practice

Preliminary programme:

DAY 1

Morning session - theoretical lectures (virtual and in-person)
A brief introduction to radiomics - I Buvat 
Review of basic concepts

Research planning - I Buvat
Step-by-step approach to hypothesis design, data preparation and curation, model evaluation, etc.

Radiomics: feature extraction - F Orlhac
Settings for feature calculation (discretization, spatial resampling), definitions, IBSI, filters

Interactive QA

Afternoon session - practical work (in-person only)
Hands on LIFEx
Reading and manipulating images, segmenting, extraction radiomic features, scripting

Presentation of the participants projects (optional) - session 1 (5 min per participant)
Social dinner

DAY 2

Morning session - theoretical lectures (virtual and in-person)
Reproducibility of radiomic features: need for harmonization - F Orlhac 
Challenge of reproducible radiomics, center and population effect, solutions for harmonization and ComBat

Building a radiomic model - J Mullaert
Classification vs prediction, data curation, feature selection, models

Interactive QA

Afternoon session - practical work (in-person only)
Hands on Shiny
Feature harmonization with ComBat

Hands on R (or another environment for statistical analysis)
Data curation, pre-processing, building a model (part 1)

Presentation of the participants projects (optional) - session 2 (5 min per participant)

DAY 3

Morning session - theoretical lectures (virtual and in-person)
Model evaluation - J Mullaert
Metrics, cross-validation strategies, external validation

Model interpretation and reporting a radiomic study - I Buvat
Clever Hans effect, tools to guide the reporting (CLEAR, METRICS, RQS…)

Interactive QA

Afternoon session - practical work (in-person only)
Hands on R (or another environment for statistical analysis)
Building a model (part 2), evaluation of performance

Using metrics to assess the quality of a radiomic study

Feedback and general discussion with participants