Italiano - English

System for automatic disease localization and scoring in videocapsule

With Prof.Guido Costamagna's research team at Università Cattolica del Sacro Cuore-Rome, we have studied and developed an innovative software for analysis and scoring of celiac disease signs in images from capsule endoscopy.

We had preliminary studied algorithms using Matlab Image Processing Toolbox.

In a second step the algorithms have been packed in a standard software for Windows. The software can analyses the full video from videocapusle (about 8 ours length) in few minutes and it returns a celiac disease chart.

Videocapsule result is a video up to 8hr lenght. Our software processes full video and returns a chart that scoring presence of celiac disease signs in the video. The software returns a celiac disease scoring map: where are disease signs ? at what level ? Physicians can easily review images, confirm or reject. This might provide drastic time reduction in video capsule medical reporting.

In order to evaluate the software accuracy, we performed one in blind clinical trail. About 3milion frames have been analysed. Results have been compared with both gold standard and diagnosis from expert physician.

In our trial, the software has provided sensitivity and specificity close to related gold standard requiring just 2% of time compared to trained physicians. Because aim of the software isn't automatic diagnosis but it should to aid physicians to perform diagnoses, sensibility is more valuable than specificity.

Thanks to our software, physicians might have a good tools to improve and speed up their daily work on inspecting video-capsule.

The study has been founded by video capsule manufacturer and developed at Endoscopy Unit, Università Cattolica del Sacro Cuore, Rome, Italy.

Authors:PkLab, Dr.Phd G.Spera, Dr.Prof M.Ingrosso, Dr G.Pirozzi, Dr.E.Riccioni, Dr C.Spada, Dr Prof.G.Costamagna.

For more information please contact us or refer to Prof.Guido Costamagna at Università Cattolica del Sacro Cuore, Rome.