Meritxell ARENAS PRAT, Victor HERNANDEZ MASGRAU

Bárbara Malavé

Radiomics of pre-treatment CT and bpMRI for predicting biochemical recurrence in localized prostate cancer patients

This research focuses on radiomic analysis using imaging and clinical data from a historical cohort of patients with similar characteristics who were previously treated with radiotherapy. Specifically, the study evaluates whether radiomic features can improve the prediction of biochemical recurrence in prostate cancer. T2-weighted and diffusion-weighted (DWI) magnetic resonance images are selected because they are the sequences most commonly used by radiologists for risk group stratification, which is closely associated with recurrence risk. Computed tomography (CT) images are also included to determine whether they provide additional predictive value beyond MRI. Objectives: o To create a database of radiomic features extracted from CT and MRI (T2-weighted and DWI sequences) in patients with localized prostate cancer. o To identify the most relevant radiomic and clinical features for the study objectives. o To apply machine learning techniques for model training and validation. o To determine the model with the highest predicting performance. Radiomic features are extracted not only from the prostate gland itself but also from the seminal vesicles and from the combined prostate–seminal vesicle volume, allowing a more comprehensive approach of tumour extension and surrounding structures. This study will be conducted in collaboration with the Radiotherapy Oncology Service of the Sant Joan University Hospital of Reus.

04-MAR-2026

Radiòmica, models predictius, oncologia

Recomanem realitzar pràctiques a l'Hospital Sant Joan de Reus per entendre millor el context d'aquest TFG