Insight

Rami Hajri: Gustave Roussy Team Advances pCR Prediction Using Machine Learning

chuv.ch

Aug 14, 2025, 06:36

Rami Hajri: Gustave Roussy Team Advances pCR Prediction Using Machine Learning

Rami Hajri, Attending Radiologist, Breast and Gynaecologic Imaging at CHUV, shared a post on LinkedIn:

“Imagine being able to predict, before the first infusion whether breast cancer will completely respond to treatment (pCR)?

During my time at Gustave Roussy, our team undertook the challenge of investigating this possibility with machine learning, combining:

Clinical data (age, molecular subtype, genetic mutations, TNM stage, receptor status, histological grade) from 235 patients
Qualitative MRI features (tumor size, T2 signal, peritumoral edema)
200+ radiomics descriptors from T2 & T1-DCE MRI sequences

The result? Our top model reached an AUC of 0.80 for predicting pathological complete response (pCR) in triple-negative breast cancer.

Why it matters: As indications for neoadjuvant therapy in breast cancer continue to expand, accurately predicting pathological complete response could transform patient care, enabling more precise treatment tailoring, reducing unnecessary toxicity for non-responders, and, in selected cases, even sparing surgery pending confirmation from ongoing prospective multicentric trials. This paves the way for truly personalized, AI-guided therapeutic strategies!

Our study was also highlighted by AuntMinnie.com.”

Title: Prediction of Breast Cancer Response to Neoadjuvant Therapy with Machine Learning: A Clinical, MRI-Qualitative, and Radiomics Approach

Authors: Rami Hajri, Charles Aboudaram, Nathalie Lassau, Tarek Assi, Leony Antoun, Joana Mourato Ribeiro, Magali Lacroix-Triki, Samy Ammari, Corinne Balleyguier

Read the Full Article.

Rami Hajri