Modalities

Two relatively recent revolutions have changed the landscape of biomedical imaging irreversibly: a radical digitization of the whole imaging chain and a transition from qualitative to quantitative imaging.

On the one hand, the acquisition of imaging studies has become almost exclusively digitized over the last decade. Production of images has shifted from cumbersome film technology to fully digital storage of data sets. Radiological examinations are stored in PACS systems. Images have become a commodity, and can be transferred through digital networks. There is no need for a direct physical link between the production unit (be it a flat-screen X-ray system, digital mammography unit, CT, MRI, ultrasound or small-animal scanner) and the reader of the images. This digitization and the advancement of IT systems in general, also fostered the continuous improvement of different advanced 3D imaging techniques such as MR, PET and CT, causing an evolutionary change in the use and the possibilities of biomedical (human and small-animal) imaging.

Currently, medical images are most often analyzed by a human expert (e.g. a radiologist). However, manual analysis is cumbersome and time consuming, and therefore expensive. In addition, it is prone to inter and intra observer variation, i.e. different experts yield different results, and even the results of a single expert analyzing the same dataset at different time points may vary. Moreover, the continuous improvements in image acquisition hardware have caused an ever increasing amount of data, to such an extent that the (human) analysis is becoming a limiting factor. We believe there is an urgent need for image processing expertise to accurately analyze and interpret these data sets. The potential of those advanced, quantitative imaging techniques is currently underexploited in hospitals and biomedical research institutions.

It is our vision that imaging studies can benefit substantially from automated advanced quantitative image analysis methods, an aspect that is currently lagging. Advanced methods allow more accurate and less time-consuming quantitative measurements. This yields reproducible measurements or imaging biomarkers allowing better diagnosis and follow-up of the disorder and drug candidate. This is especially beneficial in studies involving comparative measurements at multiple time points, which is standard practice in biomedical research, clinical trials and patient follow-up. The goal of the image analysis is to yield quantitative data. To use the results as reliable imaging biomarkers, the quantification should be as reproducible as possible.