Inflammation vs fibrosis
Redefine disease characterisation by distinguishing inflammatory from fibrotic profiles with integrated multimodal AI.
Redefining fibrosing lung disease with multimodal AI
Patients with fibrosing interstitial lung disease (fILD) can progress through inflammation, fibrosis, or both. PRISM-fILD aims to develop a multimodal AI approach that combines CT images, radiology reports, clinical data and omics to distinguish inflammatory from fibrotic disease profiles, and to support personalised diagnosis and prognosis.
5
Countries with international ILD cohorts4Years of research, 2026 to 2030
10Experts on the international advisory board
Background
What are fibrosing interstitial lung diseases?
Interstitial lung diseases (ILDs) damage the thin tissue around the air sacs, where oxygen passes into the blood. In fibrosing ILD, this tissue is gradually replaced by scar tissue. The lungs become stiff, breathing gets harder, and in severe cases the disease can lead to respiratory failure.
The causes vary, from autoimmune conditions to genetic and environmental factors. The disease is usually driven by inflammation, fibrosis (scarring), or a mix of both, and telling them apart matters for treatment: anti-inflammatory drugs can help when inflammation dominates but may cause harm when fibrosis dominates.
3 millionEstimated number of people worldwide living with idiopathic pulmonary fibrosis, the most widely recognised formMartinez et al., Nat Rev Dis Primers 2017
Up to 76Per 100,000 people in Europe are affected by ILD, the highest reported prevalence estimate (range 6.3 to 76)Olson et al., Adv Ther 2021
MonthsFrequently required to establish a definitive diagnosis, delaying appropriate treatment
Our goal
Estimate how much inflammation and how much fibrosis drives each patient’s disease
PRISM-fILD aims to combine CT images, radiology reports, clinical data and omics, including follow-up over time. Multimodal AI will bring these pieces together to characterise each patient’s disease more precisely and to anticipate how it may develop, supporting doctors in personalised diagnosis and treatment decisions.
- 1Integrate multimodal patient data
- 2Characterise inflammation and fibrosis with AI
- 3Predict the disease course and support personalised decisions
About the project
One diagnosis, many different disease courses
PRISM-fILD stands for Precision Medicine through Redefinition of Diagnosis and Management using Integrated Systems and Multilevel Data in fibrosing Interstitial Lung Diseases.
PRISM-fILD is a four-year research project funded by the Swiss National Science Foundation (SNSF). A multidisciplinary team of clinicians and imaging, pathology, biomarker and AI experts from Switzerland, Germany, Spain, the United Kingdom and the United States works together to advance precision medicine for patients with fibrosing ILD.
Can integrated multimodal analysis distinguish inflammatory from fibrotic disease, improve prognosis, and reveal potentially treatable traits?
Central research question of PRISM-fILDBiological endotypes
Explore omics signatures and biomarkers of inflammation and fibrosis, and identify potentially treatable traits.
Personalised prognosis
Use longitudinal data to predict disease progression and support clinicians in anticipating each patient’s course.
How we work
From cohort to clinic
A research workflow designed to turn multicentre patient data into validated AI decision support.
Collect
International cohorts and prospective Swiss recruitment
Protect
Pseudonymisation and secure, compliant data sharing
Harmonise
Curated, AI-ready multimodal database
Label
MDD consensus, CT annotation and omics
Model
Multimodal AI for characterisation and prognosis
Validate
External, longitudinal and prospective validation
Our approach
Multimodal AI
Each type of patient information will be analysed on its own and then combined, aiming to give a fuller picture of each patient than any single source can provide.
Patient information
Medical imagingLung CT scans
Radiology reportsRadiologists’ written findings
Clinical informationDemographics, symptoms, lung function and lab values
Omics and tissueBlood-based omics and histopathology
Multimodal AI
Combines complementary information into one patient profile
Support for clinicians
Disease characterisationBalance of inflammation and fibrosis
Disease courseHow the disease may develop
Interpretable resultsWhich information shaped each assessment
Trustworthy AI
Trust by design
PRISM-fILD is designed to support clinical judgement, never to replace it.
Privacy by design
Patient data are pseudonymised and shared securely under role-based access, in line with Swiss data protection law (FADP), GDPR and HIPAA, and with ethical approval.
Expert reference labels
Multidisciplinary expert consensus, confirmed by the disease course and treatment response, defines the reference standard that the AI learns from and is measured against.
Clinician in the loop
Results will indicate which information shaped each assessment, and AI errors will be analysed systematically. The AI is designed to support clinicians’ decisions and will be evaluated together with them.
Team
Meet the project leads
Pulmonology, AI in medicine and thoracic radiology, backed by an international network of ILD experts.
Prof. Dr. med. Manuela Funke-Chambour
Principal InvestigatorChief Physician, Pneumology · CHUV Lausanne
Prof. Dr. Stavroula Georgia Mougiakakou
Co-Principal InvestigatorHead of the AI in Health and Nutrition Lab (AIHN) · ARTORG Center, University of Bern
Prof. Dr. med. Andreas Christe
Co-Principal InvestigatorUniversity Institute of Diagnostic, Interventional and Pediatric Radiology · Bern University Hospital, InselspitalExpert Advisory Board
Lukas EbnerRadiologySenior Physician · Inselspital Bern and Lucerne Cantonal Hospitallukas.ebner@insel.ch
Sabina BerezowskaPathologyAssociate Professor, Chief Physician · CHUV LausanneSabina.Berezowska@chuv.ch
Anna-Maria Hoffmann-VoldRheumatologyHead of Translational Research, Rheumatology · University Hospital Zurichanna-maria.hoffmann-vold@usz.ch
Markus PolkePulmonologyHead of ILD Centre · Thoraxklinik, University Hospital Heidelbergmarkus.polke@med.uni-heidelberg.de
Katharina BuschultePulmonologyILD Centre · Thoraxklinik, University Hospital Heidelbergkatharina.buschulte@med.uni-heidelberg.de
Michael KreuterPulmonologyDirector, Lung Centre · University Medicine Mainzmichael.kreuter@unimedizin-mainz.de
Gisli JenkinsPulmonologyHead of Fibrosing Lung Disease Centre · Imperial College Londongisli.jenkins@imperial.ac.uk
Maria Molina MolinaPulmonologyHead of ILD Unit · Bellvitge University Hospital, IDIBELLmmolina@idibell.cat
Justin OldhamPulmonologyAssociate Professor, Director of ILD Research · University of Michiganoldhamj@umich.edu
Simon RousseauRespiratory researchAssociate Professor · McGill University, Meakins-Christie Laboratoriessimon.rousseau@mcgill.ca
Resources
Open access, in line with SNSF policyPublications and outputs
Scientific publications, conference contributions, code and software arising from PRISM-fILD, listed as they become available.
Scientific publications
Peer reviewed articles reporting the methods, findings and validation results of the project.
Conference contributions
Abstracts, posters and oral presentations at national and international scientific meetings.
Code and software
Open source tools and trained models developed within the project, released for reuse by the research community.
Publications
2026The role of inflammation in the pathogenesis of pulmonary fibrosis: mechanisms and treatment implicationsNazemi N, Crestani B, Helou DG, Distler O, Gote-Schniering J, Hoffmann Vold AM, Machahua C, Kolb M, Funke-Chambour MEuropean Respiratory Journal · 68(4): 26002582026Prompt-Guided Interactive Segmentation of Interstitial Lung Disease in Thoracic CTDedousis V, Abdur Rahman L, Brigato L, Dack E, Christe A, Roos J, Funke-Chambour M, Christoph F, Huber A, Ebner L, Mougiakakou SMICCAI 2026 Workshops · Thoracic Image Analysis (TIA)News
Latest updates
Event · 16 June 2026
Project partners convene for the PRISM-fILD kickoff meeting
Clinical, imaging and AI partners from Europe and North America convened to formally launch the project.
Project · 1 February 2026
Launch of the four-year SNSF funded PRISM-fILD project
An international collaboration, led from Switzerland, to advance precision medicine in fibrosing ILD.
Contact
Get in touch
Questions about PRISM-fILD or interested in collaborating? Contact the project leads directly.
Share feedback on our software or website
