MUNICH--(BUSINESS WIRE)--The Munich-based start-up Smart Reporting receives millions from European funds for their project "ImageREPORT". ImageREPORT is a joined venture between various European partners. The software project combines intelligent image recognition for the evaluation of medical imaging with semiautomated structured reporting. ImageREPORT aims at significantly improving radiological care by linking image analysis and structured reporting.
Project funding has been awarded by "Eurostars", which is a joint program between EUREKA and the European Commission. The grant will be used develop an intelligent image analysis and reporting software that can be used to analyze X-ray images and computertomography (CT) scans of the thorax. The software will be made available to radiology departments in hospitals. For their project Smart Reporting has partnered up with Thirona, which is a Dutch company focusing on the development of automated medical image analysis, and the Dutch Radboud university medical center in the city of Nijmegen.
Radiologists are facing an increased number of medical imaging data that need to be evaluated. In between 1999 and 2010 the number of medical tomograms has increased tenfold and modern technology has led to stark improvements in image resolution. To date, images are almost exclusively evaluated by visual means, and reports are written as free text, thus, creating a large amount of variability in between reports from different radiologists. Variable quality and ambiguities within such reports often necessitate further communication between the medical professionals and amount to a greater workload. Here is where the ImageREPORT project comes into play.
ImageREPORT is based on two components: a machine learning-based system for image analysis and an intelligent software for structured reporting that processes results for medical reports and eases the writing of such. In the first step X-ray and CT images are analyzed automatically using deep learning-based, computer-aided detection (CAD) algorithms. These algorithms – unlike the human eye – can detect the most inconspicuous anomalies. In a second step image analysis results are fed into the reporting software, which guides the radiologist using a step-by-step approach through the entire report. Eventually, semiautomated reports are generated based on analysis results and the information provided by the radiologist. These reports have a clear structure, they are reproducible and can be used for big data analysis.
Prof. Dr. Wieland Sommer, the founder of Smart Reporting, explains:
"Researchers all over the world work on intelligent image recognition software for medical images that improve diagnosis and therapy in patients. A new era of imaging has started ever since deep learning techniques have emerged. In combination with semiautomated reporting the quality of radiological care and patient well-being can be improved significantly. The more accurate the diagnosis, the better the therapy. Moreover, radiologists benefit from the assistance provided by the intelligent software."