Radiomics

Starting around 2012, the need for diagnostic and prognostic predictive models based on radiological and nuclear medicine imaging provided the fertile ground for the emergence of the field known as Radiomics. Since then, Radiomics has evolved while continuously addressing several critical aspects, including:

• (a) the formats and protocols adopted in radiology for storing medical images,
• (b) the study of signal pre-processing phases,
• (c) the research and development of machine learning techniques to improve predictive performance,
• (d) issues related to model validation, reproducibility, and replicability of results.

In addition to increasing maturity across these areas, Radiomics is also expanding beyond traditional radiology departments, moving toward technologically related domains such as radiation therapy. Here, the integration of Radiomics with theragnomics appears to provide a solid foundation for the emergence of a new discipline: Dosiomics.

The TUTORIAL

This tutorial is meant to be a general introduction to the main concepts and challenges behind radiomics and dosomics. On first, we will introduce theaudience to several key topics in Radiomics, including: the characteristics of data source standards, the concept of imaging biomarkers, the typical analytical pipeline of Hand-Crafted Radiomics (with a hands-on example), and the open challenges in the current state of the field (both technical and clinical) . Subsequently, we will present the clinical rationale underlying the emerging concept of Dosiomics and introduce a technological approach (with a hands-on session), highlighting its potential as well as its limitations. Finally, the tutorial will encourage the participants to share their experience, visions and establish a channel for future cooperation among participants in these exciting fields

Frontal Modules (1st round)

• Opening : General presentation
• DICOM : the protocol and image storage format. DICOM objects of interest, geometries, related common (and uncommon) issues
• Hand Crafted Radiomics a gentle introduction to the field: from the concept of imaging biomarkers to hand-crafted features, feature selection strategies, machine learning modeling, and model validation. Best practices: TRIPOD, IBSI, etc.
• MV4 short hands-on session with an example of a radiomics analysis tool.
• Hand Crafted vs AI Crafted Radiomics vs Deeep Learning Radiomics : many ways for crafting image fetures

Round Table with invited speakers

• Common issues in radiomics (30’): the gap between data analysis and clinical practice. Open challenges: reproducibility, replicability, explainability, biological correlates, technological obsolescence. Radiomics: a promised land or a land of promises?

Frontal Modules (2nd round) How Radiotherapy works : a gentle introduction to Radiation therapy From radiomics to dosiomics : the problem, the idea, the vision Hand Crafted RAdiomics a gentle introduction to the field: from the concept of imaging biomarkers to hand-crafted features, feature selection strategies, machine learning modeling, and model validation. Best practices: TRIPOD, IBSI, etc. Tutorial: MV4Dose short hands-on session with an example of a dosiomics analysis tool Challenges ahead : organ modelling, adaptive RT, DICOM dialets, ...

frontal lectures

Real-world use cases, presented by invited speakers

Round Table, interactive open discussion