 |
||
          |
Raynaud's phenomenon is condition which can cause complete interruption of the blood supply to the extremities when triggered by exposure to cold. For the majority of sufferers this is nothing more than an irritation and can be dealt with simply by wearing warm gloves and socks in winter. For a small number however, Raynaud's is a symptom of one of several underlying diseases, the most serious of which is Scleroderma. Scleroderma causes a variety of symptoms that stem from a thickening of the skin and connective tissue. The damage that it does to the blood vessels leads to severe Raynaud's, and the reduction in circulation can be so extreme that the tissues of the affected area die and amputation is required. Diagnosis of the disease from the various symptoms coupled with blood tests is relatively straight forward, but determining the activity/severity of the disease is more difficult. This is particularly important for clinical trials aimed at evaluating treatments for improving the circulation. One technique that is being explored involves studying the condition of the capillaries in the nail-fold - the skin overlapping the fingernail at its base - with an optical microscope. Here, the capillaries appear as a rows of long thin loops, the thickness and regularity of these loops depending on the severity of the disease. This creates the need for a quick and objective method of quantifying the capillary abnormalities observed, and it is planned to solve this with computer image analysis. At present, a video camera is coupled to a microscope and the capillary dimensions are measured by hand from a single video frame. At the magnification used (x200 - x600), capillaries one blood cell thick can be seen. The red blood cells pass through these fine capillaries in single file with the white blood cells appearing as gaps in the string of red cells. This, coupled with subtle changes in focus mean that not all of the capillary network is visible at any one instant, and so rather than work from a single frame, we intend to integrate the information over successive frames. This will form a composite image from which the capillary dimensions will be measured automatically by a computer system. This requires a system that `understands' the structures present. Our plan is to build on previous work in the Wolfson Unit in which parameterised statistical models of anatomical structures are constructed and used to find examples of the objects they represent in new images. Once the system can recognise the individual capillary loops, a set of measurements will be defined to characterise their appearance and hence condition. It is not yet known exactly which measurements are the most medically significant, and an automated system will facilitate the testing of a wide range of possibilities. To be able to integrate the information over successive video frames, the relative motion between the finger and the microscope must be corrected for. As well as the image of the capillaries, each video frame can contain noise from the camera/video recorder system, specks of dust in the microscope's optical system, and patches of light reflected from the oil used to render the skin more transparent. Accurate registration therefore, requires the identification of image pixels belonging to the capillary network alone. The most effective solution to this problem has been to employ software developed in the Wolfson Unit for the investigation of breast cancer, which detects pixels belonging to linear structures of a specified width. This isolates the capillary pixels which are then `thinned' to form a skeletal representation of the capillary network one pixel thick. The results of this process are illustrated in the figure. The top half of the figure shows the skeletal representations of two nail-fold images, one red, and the other blue. Using the computer a `best fit' for the two sets of points is found - the bottom of the figure shows the results of combining the original video frames based on this information. This project is funded by The Raynaud's and Scleroderma Association, and is a collaboration between The University of Manchester Rheumatic Diseases Centre at Hope Hospital, and the WIAU. P. D. Allen: phillip.allen@.man.ac.uk
Research Courses Clinical Radiology Industrial liaison Contact Us Join Us Search Home Contact
the Website Administrator
with comments or queries about this website. |
|
