PostDoc Position in Machine Learning–Based Multimodal Data Analysis for Predicting Response and Toxicity in ¹⁷⁷Lu-PSMA Radioligand Therapy in Metastatic Prostate Cancer

Key words: Multimodal database; Machine learning; Lutetium-177; PSMA; metastatic prostate cancer; Prediction; Tumor response; Treatment-related toxicities

Background

Radiopharmaceutical Therapy (RPT) with 177Lu-PSMA is an innovative therapeutic approach for patients with metastatic castration-resistant prostate cancer (mCRPC). It combines a molecule that specifically binds to the Prostate-Specific Membrane Antigen (PSMA), commonly overexpressed on prostate cancer cells, with the 𝛽-emitting radionuclide Lutetium-177 [1,2]. Once administered, the radiopharmaceutical delivers localized ionizing radiation directly to cancer cells, helping to destroy them while minimizing damage to surrounding healthy tissue. The clinical efficacy and safety of 177Lu-PSMA RPT have been demonstrated in several international clinical studies [3,4]. However, reliable predictors of therapeutic response and treatment-related toxicity remain insufficiently characterized, limiting the optimization of patient selection and personalized treatment strategies.

A first monocentric study conducted by our team (CREATIS & Leon-Berard center) proposed an automated pipeline for extracting a specific diagnostic biomarker from 68Ga-PSMA PET/CT images (total tumor volume) and demonstrated its predictive potential for response to 177Lu-PSMA treatment using a machine learning approach [5]. Future work should focus on multicenter validation, incorporation of survival endpoints (OS, PFS), and further refinement of predictive models to maximize clinical utility and facilitate integration into routine practice.

In this context, the UNIRIV project is a French multicenter study designed to identify predictive factors of response and toxicity in patients with metastatic prostate cancer treated with 177Lu-PSMA RPT. A key component of UNIRIV is the development of a national multimodal database, serving as a dedicated research infrastructure, implemented by the data center of the French National Federation of Cancer Centers (UNICANCER). This platform provides a secure digital environment for the centralized collection, management, and analysis of retrospective clinical, biological, functional imaging (SPECT/CT and PET/CT), and dosimetric data from a large cohort of patients treated with 177Lu-PSMA in France.

Building upon this framework, the aim of the present postdoctoral project is to extend our initial work to the UNIRIV database in order to further to refine ML-based predictive models and advance the identification of diagnostic biomarkers through the analysis of multicentric, multimodal data related to 177Lu-PSMA therapy.

Objectives

  1. Get to know UNIRIV: Understand and work with the UNIRIV database and assess the need to develop tools specifically designed for data extraction and processing.
  2. Predictive analysis from ML models: Refine machine learning-based predictive models and advance the identification of diagnostic biomarkers from UNIRIV database.
  3. Scientific output: Disseminate the results through scientific outputs, including peer-reviewed publications and presentations at national and international conferences.

Environment 

The recruited person will work in a multidisciplinary team composed of medical physicists, physicians, researchers and computer scientists of CREATIS laboratory and the LUMEN, the nuclear medicine department of the Léon Bérard Cancer Center (Lyon). 

Supervisors / Contacts:

Dr. David Sarrut, Research Director (CREATIS laboratory) david.sarrut@creatis.insa-lyon.fr

Dr. Jean-Noël Badel, Medical Physicist, PhD (Léon Bérard Cancer Center, CREATIS laboratory) JeanNoel.BADEL@lyon.unicancer.fr

Expected skills and other information:

  • Expected skills: Data analysis, image processing, AI
  • Technical skills: Python (required)
  • English (required) and French (optional)

Salary: around 2 100 euros 

Expected start: as soon as possible

Duration : 1 year

Location : Lyon, CREATIS laboratory, Léon Bérard Cancer Center, France

Bibliography

[1] Kratochwil C et al. PSMA-targeted radionuclide therapy of metastatic castration-resistant prostate cancer with 177Lu-labeled PSMA-617. Journal of Nuclear Medicine. 2016;57:1170–1176.

[2] Fendler WP et al. 177Lu-PSMA radioligand therapy for prostate cancer. Journal of Nuclear Medicine. 2017;58:1196–1200.

[3] Sartor O et al.Lutetium-177–PSMA-617 for Metastatic Castration-Resistant Prostate Cancer. New England Journal of Medicine. 2021;385:1091–1103.

[4] Hofman MS et al. [177Lu]Lu-PSMA-617 versus cabazitaxel in patients with metastatic castration-resistant prostate cancer (TheraP): a randomised, open-label, phase 2 trial. Lancet Oncology. 2021;22(5):650–661.

[5] Rios-Sanchez E et al. Predicting PSA50 response to 177Lu-PSMA therapy using machine learning and automated total tumor volume. EJNMMI Physics. 2025;12:95