Medical Image Analysis - PDF

Document Sample
Medical Image Analysis - PDF Powered By Docstoc
Description: The present invention relates to systems and methods for analysis of medical images.BACKGROUND OF THE INVENTION Many algorithms for automated analysis of medical images, such as X-ray images, have been suggested. In general, the automated analyses rely upon segmentation and/or comparison to a model. These automated analyses have been applied to analysesof bone mineral density, fractures and lesions. However, previously available automated analyses can be influenced by distortion caused by an angle of incidence of X-ray beams with respect to an imaged bone. FIGS. 9A and 9B show that a change in theangle of incidence from -10 degrees (FIG. 9A) to +30 degrees (FIG. 9B) can cause significant changes in important measured anatomic features. For example, the femur head offset (indicated by the arrow) is 5.2 cm at -10 degrees and 2.1 cm at +30 degrees. The head shaft angles (.theta.) are and, respectively. As a result, even when automated image analyses accurately measure selected image parameters, the automatically measured parameters may be difficult to evaluate. Manual analysis by a radiologist is subject to similar influence from angle of incidence, although to a lesser degree. An expert radiologist may be able to judge an approximate angle of incidence for a particular image and mentally correctmeasurements. However, it may take many years of clinical experience to acquire this level of experience. Also, a radiologist with vast experience in evaluating hip X-rays may have significantly less experience with images of other body portions (e.g.ankle or spine). Alternatively or additionally, an "eyeball" correction tends to be more qualitative than quantitative. It is known to register a 3D image (e.g. CT scan) of an organ from a specific patient onto a 2D image acquired from the same patient (e.g., an X-ray image). This permits information from CT images to be used during interventional procedures byregistering the CT scan to an intra