I got this piece of news on CAD for Mammograms via Inescporto.
According to the article, the CAD application achieved 100% accuracy in detection of malignant tumors in 1/3 of all cases but I have no access to the research data, methodology etc so I can’t make a pro
INESC Porto (Institute for Systems and Computer Engineering of Porto) and FEUP (the Faculty of Engineering at the University of Porto) developed software to automatically analyse and validate mammograms. The software, Semantic PACS, can be incorporated into hospital IT systems and it categorises each new mammography using automatic descriptive methods and search methods that compare the semantic contents of the images stored. This means that the software learns from the analysis of previous mammograms and automatically provides relevant information that can help diagnose patients.
The software can be used for 1/3 of all cases with 100% accuracy in detecting malignant tumours. The objective of Semantic PACS is to automatically validate cases with clear results allowing doctors to be able to dedicate more time to assessing more complex cases. In addition, when doctors are unable to determine whether the tumour is benign or malignant, the system can search for similar images to help the doctor assess the seriousness of the tumour. Jaime Cardoso, a researcher at INESC Porto and Professor at FEUP, explains that “from the patient’s point of view this software presents advantages such as reducing the time taken to diagnose cases and the emotional stress associated with this wait”.
Semantic PACS is an example of research completed in Portugal with commercial value. It was developed in collaboration with the industrial partner Emílio Azevedo Campos and will be included in the company’s range of archiving and sharing equipment for medical images. According to Pedro Cardoso, the Director of Emílio Azevedo Campos, Semantic PACS “is a solution that presents unique factors and added value for our healthcare clients”.
The Semantic PACS software was created as part of a project with the same name that was co-funded by the National Strategic Reference Framework (NSRF) with a budget of 320 thousand Euros.