New imaging biomarkers for more efficient cancer treatment
|KUNGLIGA TEKNISKA HÖGSKOLAN - Kungliga Tekniska Högskolan/Skolan för teknik och hälsa
|Funding from Vinnova
|SEK 500 000
|May 2017 - March 2018
|Challenge-Driven Innovation – Stage 1 initiation
Purpose and goal
All 3 goals of the project have been reached: 1. A model study of new AI-based imaging biomarkers predicting the clinical outcome in lung cancer has been carried out with encouraging results. 2. IT infrastructure has been created, consisting of a software platform enabling development, evaluation, marketing and testing (on own data) new imaging biomarkers. 3. We have prepared application of the same approach to other disease groups; next we are planning for an extension to dementia disorders.
Expected results and effects
The model study, which demonstrates the feasibility of our approach, has resulted in a manuscript that is ready to be submitted for publication. The software platform will shortly be opened up for new users. These may include companies who want to market their software products, image analysis researchers who want to develop and evaluate new analysis methods on well defined test data, and medical users who want to explore and test available analysis methods to solve a clinical problem.
Planned approach and implementation
KTH and KI have together carried out the model study with some support from Elekta. KTH has constructed the IT infrastructure in close collaboration with Contextvision, and the modules available in the software platform include software from Contextvision, as well as from other external sources. The close collaboration between academic and commercial partners has facilitated the ambition of creating a common platform of value for both companies, researchers and medical practitioners.