The vast number of cytology and histology specimens which are sent to pathology departments forms a serious problem. At the laboratory, specimens are processed and routinely stained after which they are ready for examination by skilled pathologists and cyto-technologists, who use optical microscopy. The examination enables a diagnosis, and often some prognostic parameters are also evaluated. This examination remains subjective however meaning that results are not always reproducible. Although the training of pathologists and cyto-technologists within Europe is generally situated at an elevated level, technicians become fatigued particularly if the workload is too high, under which conditions the accuracy of their work slackens.
The NANOSCAN team has automated the analysis and was able to produce more objective and accurate results. The new procedure allows pathology departments to deal effectively with an ever increasing workload. NANOSCAN can be used in any application where the automatic inspection of colour, shape and size is important. The system has already found application in cytology and histology and the food and pharmaceutical industries. It has potential in virtually any area relying on optical microscopy. The system has a wide potential ranging from applications in medicine to the food industry.
It is the use of colour which enables complex structures to be analysed more accurately. In order to analyse any structure, it needs accurate segmentation into its basic sub-components. For simple two-coloured structures, this can easily be achieved by using the traditional black and white techniques. For multi-coloured structures though, the laboratory workers need an advanced colour image system. The NANOSCAN system is based on a patented image segmentation algorithm. Through a comprehensive library of routines, it can support a variety of applications. The patented algorithm enables pathology experts to analyse arbitrarily complex structures in real time. A cluster of standard, networked PCs provides the computer power which is necessary for real time analysis.
The amount of data handled is enormous, and the potential to achieve more objective and much better reproducibility of the diagnosis and prognostic parameters offers further evolution in the field of pathology. One of the three NANOSCAN partners is the Dimac Imaging company, a commercial software developer, specialising in digital image analysis based on a patented colour-classification method. Dimac Imaging has produced the revolutionary colour image analysis package, which is used in NANOSCAN. The project has been developed as a joint venture between Dimac, the University of Copenhagen (Uni-C) and a routine pathology laboratory at Randers Hospital. The Uni-C was acting as the responsible university technology transfer department.
At present, NANOSCAN enables technicians to digitise cytology and histology slides and analyse their morphology on a computer using the existing Dimac colour image analysis package. NANOSCAN makes such analysis more than four times quicker while producing reproducible quantitative results. More information about the NANOSCAN system can be obtained from the Danhit contact person Jørgen Moth. The brochure "Image Processing and Parallel Computing in Medical Applications" equally comprises details on NANOSCAN, as well as on a host of other TTN medical parallel computing projects, which were carried out under the Esprit programme. This publication can be ordered on-line at the all-in price of DM 45,- at AT&M Technologie und Marketing GmbH.