Image analysis – Applications – Biomedical applications
Reexamination Certificate
1998-07-17
2001-08-21
Bella, Matthew C. (Department: 2621)
Image analysis
Applications
Biomedical applications
C382S133000, C382S224000, C382S132000
Reexamination Certificate
active
06278793
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to computer aided detection of abnormal anatomical regions in radiographs, and, more particularly, to optimization of such computer aided image detection schemes based on overall image quality.
2. Background of the Invention
Detection and analysis of target objects in digital images are useful and important tasks. For example, detection and diagnosis of abnormal anatomical regions in radiographs, such as masses and microcalcifications in womens' breast radiographs, so-called mammograms, are among the most important and difficult tasks performed by radiologists.
Breast cancer is a leading cause of premature death in women over forty years old. Evidence shows that early detection, diagnosis and treatment of breast cancer significantly improves the chances of survival, reducing breast cancer morbidity and mortality. Many methods for early detection of breast cancer have been studied and tested, among them mammography. To date mammography has proven to be the most cost effective means of providing useful information to diagnosticians regarding abnormal features in the breast and potential risks of developing breast cancer in large populations. The American Cancer Society currently recommends the use of periodic mammography and screening of asymptomatic women over the age of forty with annual examinations after the age of fifty. Mammograms may eventually constitute one of the highest volume X-ray images routinely interpreted by radiologists.
Between thirty and fifty percent of breast carcinomas detected radiographically demonstrate microcalcifications on mammograms, and between sixty and eighty percent of breast carcinomas reveal microcalcifications upon microscopic examination. Therefore, any increase in the detection of microcalcifications by mammography may lead to further improvements in its efficiency in the detection of early breast cancer.
Currently acceptable standards of clinical care are that biopsies are performed on five to ten women for each cancer removed. With this high biopsy rate is the reasonable assurance that most mammographically detectable early carcinomas will be resected. Given the large amount of overlap between the characteristics of benign and malignant lesions which appear in mammograms, computer-aided detection of abnormalities will have a great impact on clinical care.
At present, mammogram readings are performed visually by mammographic experts, that is, physicians and radiologists. Unfortunately, visual reading of mammograms has two major disadvantages. First, it is often possible to miss the breast cancer in its early stages. This is because, unlike many other cancers, there is as yet no clear way to detect premalignant changes in the breast. This results partly from the relative inaccessibility of breast tissue. A second disadvantage of visual reading of mammograms is that these readings are both labor intensive and time consuming. Multiple readings of a single mammogram may be necessary in order to increase the reliability of the diagnosis.
Therefore, it would be advantageous and useful to have computer-assisted or aided detection (CAD) systems to help radiologists and physicians obtain quicker and more precise results when performing visual readings of mammograms. Such CAD systems would aid in cancer detection and improve the efficiency and accuracy of large-scale screening.
Various computer assisted detection systems have been investigated to assist diagnosticians in their diagnosis of breast cancer.
U.S. Pat. application Ser. No. 08/352,169, filed Dec. 1, 1994, which is hereby incorporated herein by reference, describes a CAD system for finding abnormal regions (masses or microcalcifications) in digital mammograms using topographical extraction techniques. The system described therein finds suspected regions and then uses several top view layers of each suspected region to determine whether or not that region looks like an abnormal region. A rule-based selection of changes in features is used to select and evaluate suspicious regions.
The topographical system is based on stacking several top-view layers of a suspected region and then evaluating whether that region looks like an abnormal region. This approach is similar to generating topographic maps of a surface and deciding, based on those maps, whether a region on the surface is a mountain, based on the rate of change in elevation for all directions.
To date, other research in computerized detection systems for masses and microcalcifications in digital mammograms has largely been based on thresholding methods or neural networks. One other method, described in Doi, U.S. Pat. No. 4,907,156, uses varying threshold levels to detect nodules for enhancement and detection of abnormal anatomic regions in a digital image of a human chest. U.S. Pat. No. 5,289,374 discloses a method and system for analysis of false positive produced by the system of Doi '156 to reduce the false positive found by the latter system.
In another method, Giger et al, U.S. Pat. No. 5,133,020, use a thresholding technique to locate abnormal regions in a digital image of a radiograph, and then, once the regions have been located, uses classification processing to determine whether the detected abnormal region is benign or malignant. Giger's classification is based on the degree of spiculation of the identified abnormal regions.
U.S. Pat. No. 5,319,549 uses texture analysis on a number of small regions of interest in order to determine a classification of normal or abnormal of a patient's lungs.
Doi, U.S. Pat. No. 5,343,390, discloses a method and system for automated selection of regions of interest and detection of septal lines in digital chest radiographs. Doi uses a large number of adjacent regions of interest selected corresponding to an area on a digital image of a patient's lungs. The regions of interest each contain a number of square or rectangular pixel arrays and are selected to sequentially fill in the total selected area of the lungs to be analyzed. The regions of interest are analyzed to determine those exhibiting sharp edges, that is, high edge gradients. A percentage of the sharp edge regions of interest are removed from the original sample based on the edge gradient analysis, the majority of which correspond to rib edge containing regions of interest. After removal of the sharp edge regions of interest, texture measurements are taken on the remaining sample in order to compare such data with predetermined for normal and abnormal lungs.
The above methods and systems have various disadvantages and deficiencies, including that they use absolute measurements only, they are not always robust as a function of image quality variations, and either their sensitivity is too low or they tend to result in too many false positives.
Each radiograph is different in its image and physical characteristics, and some are more difficult to read, interpret or computer analyze than others. A difficult or subtle radiograph may be difficult for either a CAD scheme or a radiologist or both to diagnose. A radiologist will typically spend more time and give more attention to reading and interpreting a difficult image than would be applied to an easy image.
On the other hand, unlike radiologists, CAD systems do not distinguish between difficult and easy images prior to their processing of the images. Accordingly, CAD systems apply the same processing to all images, regardless of their global image characteristics or difficulty as determined by the imaging physics (for example, regardless of the ratio of signal to noise in an image) and breast tissue structure.
For example, the CAD systems referred to above all apply the same processing and rules to all images, regardless of the global characteristics of the image.
Measures of object or target characteristics have been addressed, but only in the context of the difficulty of finding already detected objects or of specific target objects in an image. For example, detection difficulty has bee
Gur David
Zheng Bin
Bella Matthew C.
Choobin M B
Pillsbury & Winthrop LLP
University of Pittsburgh
LandOfFree
Image quality based adaptive optimization of computer aided... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image quality based adaptive optimization of computer aided..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image quality based adaptive optimization of computer aided... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2542102