Prognostic prediction of primary brain lymphoma using artificial intelligence analysis of MRI scans



Main objectives

This study aims to build and evaluate the prognostic contribution of two artificial intelligence models based on MRI imaging acquired in routine clinical practice. The first model evaluates tumor burden, while the second assesses patient frailty.


Primary cerebral lymphoma (PCL) is a rare pathology whose incidence is constantly evolving. It is histologically homogeneous, but highly heterogeneous in terms of recurrence rate and overall survival. The identification of prognostic factors would make it possible to stratify patients and optimize clinical decision-making strategies. At present, age and scales assessing patient performance are the most important independent prognostic factors in PCL. Magnetic resonance imaging (MRI) has a central role in the diagnosis and evaluation of therapeutic response in PCL, but its prognostic value is currently very low and has been little explored in large multicenter cohorts.

Methodology/Technology used

Tumor burden will be assessed by (i) lesion volume and number, (ii) texture information and (iii) lesion localization. Patient fragility will be assessed by volume and texture analysis of the temporal muscle included in the MRI acquisition field.



Lucia Nichelli ()


AP-HP, Inria

Other projects



US Caractérisation et prédiction de la survenue de formes graves ou létales du COVID-19 à partir des données issues de l’EDS de l’AP-HP

Names of partners involved
AP-HP, Inria & Centrale Supélec

read more