PROGNOSTIC

Research Promotion Foundation of Cyprus, (CODEVELOP-AG-SH-HE/0823): "Ultrasound Shear Wave Elastography Driven Prognostic Tool for Precision Oncology." (PROGNOSTIC) June 2024 - May 2026 €600,000

Project details

Principal Investigator: Triantafyllos Stylianopoulos

Project ID: CODEVELOP-AG-SH-HE/0823/0202

Total Funding: €600,000

Call for proposal: HORIZON EUROPE – CO-DEVELOP

Funding scheme: THE RESEARCH AND INNOVATION FOUNDATION PROGRAMMES FOR RESEARCH, TECHNOLOGICAL DEVELOPMENT AND INNOVATION “RESTART 2016 – 2020”

Coordinated in: Cyprus (University of Cyprus)

From June 2024 - May 2026

Overview of the project

Effective cancer treatment increasingly depends on early identification of patients who will (or will not) benefit from a given therapy. In solid tumours, mechanical stiffening of the tumour micro-environment (TME) modulates blood flow, immune infiltration and drug penetration, yet this biomechanical information is largely absent from present-day clinical decision making.

The PROGNOSTIC project tackles this gap by integrating Shear Wave Elastography (SWE) with a machine-learning pipeline that transforms ultrasound stiffness maps into a predictive biomarker of therapeutic response. Our central idea is that:

1. Baseline tumour stiffness and heterogeneity carry mechanistic information about vascular compression, hypoxia and immune exclusion.

2. By learning these SWE signatures, an AI model can forecast response or resistance to both standard-of-care (SoC) and novel treatments (e.g. mRNA cancer vaccines).

3. The same SWE-based tool can be used longitudinally to monitor mechano-therapeutic priming (e.g. losartan) and to adapt therapy if resistance emerges.

To demonstrate clinical utility, PROGNOSTIC combines orthotopic mouse models and an ongoing prospective clinical study in breast and prostate cancer. In pre-clinical work, we couple TME-normalising drugs (losartan) and KRAS-targeted mRNA vaccination to test whether modulating stiffness converts predicted non-responders into responders. In the clinical arm, SWE scans are collected before and during therapy, and model predictions are compared with RECIST outcomes.

Ultimately, PROGNOSTIC aims to deliver a CE-ready software toolkit that gives oncologists a fast, non-invasive “mechanical biopsy”, guiding personalised treatment choices and accelerating the adoption of mechano-immuno-therapies.

Research Objectives

Research Objective 1 (RO1) Biomarker Discovery & Validation
Validate the SWE-based machine-learning model for (i) automatic tumour segmentation and (ii) prediction of response vs non-response to SoC therapy in ≥100 patients with breast or prostate cancer.

Research Objective 2 (RO2) Mechanotherapeutic Priming
Investigate whether losartan-mediated TME normalisation reduces stiffness and converts PROGNOSTIC-labelled non-responders into responders in orthotopic pancreatic and colon cancer models.

Research Objective 3 (RO3) Combination Strategy
Evaluate a triple-combination regimen (losartan + mRNA vaccine + SoC chemotherapy) in animal models, using PROGNOSTIC to adapt dosing schedules and quantify therapeutic synergy.

Research Objective 4 (RO4) Software Translation & Market Readiness
Package the validated algorithm into a clinician-friendly GUI (GUI v0.9 beta), integrate PACS connectivity and CE-ready logging, and complete a SWOT-based commercialisation plan for first-in-human deployment.