SUPR
Advanced image analysis of data in SCAPIS
Dnr:

sens2024044

Type:

NAISS SENS

Principal Investigator:

Joel Kullberg

Affiliation:

Uppsala universitet

Start Date:

2024-10-23

End Date:

2025-11-01

Primary Classification:

20603: Medical Image Processing

Allocation

Abstract

The SCAPIS study (Swedish CArdioPulmonary Imaging Study, www.scapis.org) is a unique and large-scale Swedish imaging study for research on cardiovascular and pulmonary disease. A random sample of 30,000 men and women aged 50-64 years have been investigated in six Swedish cities in 2014-2018. A collection of data on cardiovascular risk factors and lung function. The study includes CT scans of heart, lungs and for detailed analysis of body composition. A 10-year follow up scanning is currently ongoing of half the cohort half the cohort. The proposed project aims to apply advanced image analysis techniques (image segmentation, image registration, and deep-regression based techniques) to this large-scale cohort for different types of integration of the imaging and non-imaging data collected. Medical applications include detailed studies of ageing and cardiometabolic disease. The medical applications will focus on causes and consequences of variations in body composition and include following research questions: • How does body composition differ between healthy subjects and those with different diseases such as CVD or diabetes or their risk factors? Are there detectable differences already before disease onset? • How well can a subject's chronological age be predicted from the image data? What regions and features in the images are important for these predictions? Can we estimate a “morphological heart age” from the CT scans of the heart? • How well can the collected image data predict future disease development both without or with inclusion of other clinical information? • Which genetic variations determine human body composition regarding the volume and distribution of fat and lean tissue? How similar is this in males and females? How heterogenous is this in adipose tissue? • Can we, by using Mendelian randomization, determine to what degree body composition causally contributes to the different diseases or to what degree the disease is causing the body composition?