Computer Assisted Diagnosis for Endoscopy

By an automated image-based tissue characterization the “EndoCAD®“ system supports physicians with the analysis of endoscopic images in an objective and standardized way. Using innovative means for image analysis and interpretation allows the characterization and detection of malignant and dysplastic tissue much earlier and supplies diagnostic hints during the intervention.

The Challenge

EndoCAD® - Computer-assisted Diagnosis for Endoscopy
© Fraunhofer IIS/Kurt Fuchs
EndoCAD® - Computer-assisted Diagnosis for Endoscopy

If left untreated, malignant neoplasms of the respiratory and gastrointestinal tracts usually have wide-reaching consequences. Treatment of very large tumors that have been detected too late leads to a considerably limited quality of life for patients.

Hence, the objective is early detection of primary tumors and tumor recurrences through suitable preventative exams. Today use of endoscopic systems combined with digital recording and archiving is standard for diagnostic imaging of the aerodigestive tract. With these systems, the morphology and texture of the affected tissue can be analyzed directly.

However, many endoscopic exams are currently still associated with diagnostic uncertainty, since often only inadequate differentiation between various mucous membrane types and lesions is possible. A definitive diagnosis always requires the pathological and histological analysis of tissue samples.

Our Solution: EndoCAD®

EndoCAD® - Esophagus Macroscopic
© Fraunhofer IIS
EndoCAD® - Esophagus Macroscopic

In order for the specialist to be provided not only with different visual impressions but also with objective support in the diagnosis, automated image-based tissue detection is suitable as a supplement to conventional biopsy. With high-resolution cameras, textural, morphological, color and, if applicable, functional parameters are recorded for characterization of the tissue. Through the use of novel image analysis and interpretation procedures, degenerative tissue or preliminary stages of such changes can be detected and a suggested diagnosis given earlier than was possible before. Automatic retrieval of similar reference images based on the image content supports case-based diagnosis during reporting of findings. Via this procedure, suitable reference cases, including the relevant patient history data, can be displayed during the exam.

“EndoCAD®“ has not yet been certified as a medical device. The Fraunhofer IIS is presenting “EndoCAD®“ with the objective of gaining partners for further development, production and marketing.

More Information


Your Development Partner for Medical Image Analysis

Talk to us – together we will find a solution for your medical image processing challenge:  -> Please send an email if you are interested

Software based Image Analysis and Enhancement

Our focus are the domains of Microscopy and Endoscopy.