Skip to main content

Development of a Soft-Copy Reading Environment for Digital Mammography in Breast Cancer Screening

Objective

The lack of an adequate soft-copy reading environment is a major technological obstacle hindering the transition to digital mammography in Europe's breast cancer screening programmes. A Soft-Copy REading ENvironment (SCREEN) that can replace film-based reading in screening programmes will be developed and its system design constraints specified. Problems posed by current digital image display technologies and by the strong requirements on diagnostic throughput and quality in screening mammography need to be solved. The research and development will focus on computer-aided lesion interpretation; new applications of computer-aided detection; digital image standardisation; intelligent computer-aided workflow management; computer-assisted training and in-service monitoring of radiologists; lesion data-mining in image databases; and workstation design. The resulting prototype SCREEN system will enable an experienced screening radiologist to read one hundred cases an hour. The lack of an adequate soft-copy reading environment is a major technological obstacle hindering the transition to digital mammography in Europe's breast cancer screening programmes. A Soft-Copy REading ENvironment (SCREEN) that can replace film-based reading in screening programmes will be developed and its system design constraints specified. Problems posed by current digital image display technologies and by the strong requirements on diagnostic throughput and quality in screening mammography need to be solved. The research and development will focus on computer-aided lesion interpretation; new applications of computer-aided detection; digital image standardisation; intelligent computer-aided workflow management; computer-assisted training and in-service monitoring of radiologists; lesion data-mining in image databases; and workstation design. The resulting prototype SCREEN system will enable an experienced screening radiologist to read one hundred cases an hour.

OBJECTIVES
The main objective of the SCREEN project is to develop and specify the design of a Soft-Copy REading ENvironment that can replace film-based reading in European screening mammography programmes. The lack of such an adequate environment is the main remaining technological obstacle hindering the transition from film-based to digital mammography. The underlying technological problems are (1) insufficient luminosity and spatial and grey scale resolution of current digital image displays, (2) lack of suitable and robust image processing and visualisation algorithms, and (3) insufficient understanding of how to integrate these in a workflow that will itself need significant adaptation to soft-copy reading. The objective of the research and development in the SCREEN project is to understand and solve these problems. It is widely expected that the integration of soft-copy reading in European screening mammography will increase the health-care quality and possibly reduce costs.

DESCRIPTION OF WORK
The composition of the participants is essential to achieving the goals within the time frame. They will bring to the project (1) the best available computer-aided diagnosis system for the detection of breast lesions, (2) a prototype soft-copy reading workstation that will be a good starting point and basis for the proposed research and development (R&D), (3) understanding and experience of all aspects of film-based screening mammography, (4) experience of image processing, image enhancement, image visualisation and image standardisation in mammography and (5) experience of computer-aided teaching and training in mammography.

The only way to overcome current display technology problems and to achieve the diagnostic throughputs and accuracy required in screening mammography is with algorithms and software that exploit the new possibilities that arise in soft-copy reading. Therefore the work will focus on: (1) R&D to optimise the basic hardware and software for the handling of the 50 Mbyte or more of data per case that need to be transferred to the display. (2) R&D and implementation of an intelligent user-interface that will require a minimum of image manipulation. (3) R&D of new methods to do in-service monitoring of radiologists and acute performance and to provide feedback. (4) Development of a computer-assisted training system for knowledge renewal and for the training of new radiologists. (5) The creation of a digital image database of 1000+ screening cases (of which 500 are pathological) that will generate new image analysis research. (6) The understanding and modelling of the digital acquisition devices and displays, which is needed for (7) the development of algorithms that will work on a wide range of acquisition and display systems. (8) R&D on segmentation and characterisation of lesions. (9) Development of lesion data-mining algorithms for feature-based search in image databases. (10) The continuous testing of components with screening radiologists. sultations with two research institutions.

Funding Scheme

CSC - Cost-sharing contracts

Coordinator

MEVIS - CENTRUM FUER MEDIZINISCHE DIAGNOSESYSTEME UND VISUALISIERUNG GMBH
Address
Universitaetsallee 29
28359 Bremen
Germany

Participants (4)

BARCO NV
Belgium
Address
President Kennedypark 35
000 Kortrijk
MEVIS TECHNOLOGY GMBH & CO. KG
Germany
Address
Universitaetsallee 29
28359 Bremen
STICHTING KATHOLIEKE UNIVERSITEIT
Netherlands
Address
Geert Grooteplein-noord 9
6525 EZ Nijmegen
THE CHANCELLOR, MASTERS AND SCHOLARS OF THE UNIVERSITY OF OXFORD
United Kingdom
Address
University Offices, Wellington Square
OX1 2JD Oxford