Artificial Intelligence for Biomedicine and Healthcare

ONC0261D Artificial intelligence applied to diagnostic imaging and reporting systems

Artificial intelligence applied to diagnostic imaging and reporting systems

Academic year 2024/2025

Sommario del corso

• Course objectives
• Results of learning outcomes
• Program
• Course delivery
• Learning assessment methods
• Teaching tools
• Course card

Course objectives

The educational objectives are to provide both theoretical and practical skills in the field of artificial intelligence applied to diagnostic imaging. The course is designed to align with the specific educational goals of the Artificial Intelligence for Biomedicine and Healthcare master’s program. The course will guide the students through the main steps required for the development of radiomic signatures, from image segmentation to classification, including the use of deep learning techniques. A simplified case study will be proposed to address practical challenges related to multi-center datasets and clinical validation.

Results of learning outcomes

Upon completion of the course, the student shall be able to understand the steps of a radiomics pipeline and choose appropriate tools and strategies for segmentation, feature extraction, features selection and classification and analysis of results. The student will also gain awareness of the validation procedures necessary for clinical deployment.

Program

Fundamentals of AI in imaging

• Overview of AI and radiomics in diagnostic imaging
• Introduction to radiomic signatures: from image to model
• Overview of machine learning and deep learning methods

Segmentation and feature extraction

• Manual, semi-automatic and deep learning-based segmentation
• Feature extraction: hand-crafted and learnable features
• Pre-processing and harmonization in multi-center datasets

Classification and model evaluation

• Machine learning classifiers for radiomic data
• Model performance metrics (ROC, sensitivity, specificity, etc.)
• Pitfalls in model validation: overfitting, data leakage, etc.
• Clinical validation strategies and deployment considerations

Hands-on project

• Guided development of a simplified radiomic pipeline
• Application to multi-center datasets
• Interpretation and discussion of results

Course delivery

Oral lectures and labs, offering interaction with the lecturer in presence mainly for managing digital images and clinical reports. Lectures' slides and other learning material will be made available through the Campusnet page in the Moodle platform.

Learning assessment methods

The students will be evaluated and graded based on written and/or oral tests. The marks will be averaged with those of the other Modules according to the number of credits of this Module.

• Teaching material
• Exams and finals
• Class schedule
• Go to Moodle
• Register for the course
• Enrolled students
• Invia email agli studenti registrati
• Export table