This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
The association between Computer Aided Diagnosis (CAD) and medical informatics is more than two decades years old, the first attempts started 30 years ago, these attempts used computers for analyzing mammographs and chest radiographs automatically. This chapter will give summary of presently existing challenges and developments within CAD systems. CAD systems can be helpful to the physician, for example, searching in huge data resulted from computer tomography CT for locating any pathological changes in patients, and also CAD system may track sharp changes in x-ray images, such as in mammography, or may determine suspicious regions in complex images like x-ray images of the thorax.
All Computer Aided Analysis and Diagnosis (CAAD) systems work upon receiving input image from the users, process and analyze the input image, the processing of the input image usually produces a set of important features. These features will be analyzed to generate the required output or what so called diagnosis.
Computer aided diagnosis algorithms must be trained first, secondly validated and then delivered to customers for regular usage. On the other hand, computer-aided diagnosis algorithms are used in corresponding systems all over the world.
This chapter presents definitions and overviews of medical computer aided analysis and diagnosis technology in general, and the Computer Aided Analysis and Diagnosis (CAAD) systems for skin lesions and skin burn in specific. It is important to underline the skin burn injuries in this chapter in order to discover how those injuries evolves, and differentiate between the different skin ulcers. In addition, this chapter discusses the various types of Computer Aided Analysis and Diagnosis (CAAD) systems. Finally, this chapter provides examples of current available Computer Aided Analysis and Diagnosis (CAAD) systems in the world.
2.2 CAAD Systems
Computer Aided Analysis and Diagnosis (CAAD) integrates image information from different medical imaging modalities into new image, which is dynamically performed for serial analysis of the images for a patient, these systems are helpful for making the correct decision for patients during clinical diagnosis and treatment.
The CAAD systems may include the data acquisition, the noise reduction, as well as the basic image processing and analysis, such as the image segmentation, the image registration, the image fusion, the pattern recognition and the image display.
Computer Aided Diagnosis is a series of steps used in medicine field to assist the physician's explanations and findings. They can get a great deal of information when using Imaging techniques in X-ray diagnostics, the radiologist must scan and assess extensively in a short period of time. CAD systems can be used for example with computed tomography CT digital images to search for typical appearances and to determine noticeable sections (possible diseases). CAD is considered quite new interdisciplinary technology that combines fundamentals of artificial intelligence and digital image processing with medical image processing. CAD has been used in the tumors detection. CAD supports the preventive medical checkup in mammography, the detection of lung and colon cancer, As a matter of fact; CAD cannot and may not be substituted with the physician, but it can be as a support. In any case the final diagnosis is the doctor's responsibility.
2.3 Characterizations of Computer Aided Analysis and Diagnosis Systems
As a matter of fact, the concept of CAD is considered extensive, CAD can be applied to all kind of medical imaging modalities, including radiography, MRI, CT, nuclear and ultrasound medicine imaging, applied for all human body organs like the skull, heart, thorax, abdomen and extremities, and all sorts of assessments including skeletal imaging, soft tissue imaging and many more.
Doi (2005) identified the basic technologies involved in certain CAD schemes that define the technology of Computer Aided Analysis and Diagnosis systems, which include:
(1) Image processing for detection and extraction of abnormalities
(2) Image features quantitation for candidates of abnormalities
(3) Data processing for classification image features as normals and abnormals (or benign and malignant).
(4) Quantitative evaluation and retrieval of images similar to those of unknown lesions
(5) observer performance studies using ROC analysis.
According to Acha et. al. (2003), the CAD tool may contain the following stages:
1. Image acquisition, contains tools and equipments for acquiring digital images.
2. Segmentation. The image segmentation algorithms, but none of them can be used as a standard, because most of them are highly application dependent.
3. Classification, the last stage of the CAD, once the image is segmented into readable regions, those regions will classified using the suitable classifier.
Guoyue Chen et. al. 2004, mentioned the procedure of Medical Imaging based CAD which includes the following main steps:
Imaging data acquisition, imaging data preprocessing such as image segmentation, registration, finally the most important step is the pattern recognition and classification.
2.4 Image Processing Technology and CAAD Systems
Image processing of medical images has two purposes. Firstly, the reliability of the physician's diagnosis is greatly enhanced and secondly it is effective, efficient and cost competitive. Each missed diagnosis is a lost opportunity for the patient to recover more quickly. An inefficient image processing method increases the cost of the health care system which is already burdened by high costs. Often, these two competing demands result in tradeoffs that are usually expressed covertly, rather than openly. It is the combination of engineers, radiographers and physicians that can reach the best compromise.
2.5 CAAD systems for Skin wounds and lesions
There are two main challenges in CAD systems have to process, the detection of lesions with in an image and the diagnosis of such detected lesions. For physicians, to decide what the cause of wound is. They usually use direct visual observation regarding appearance of that wound, that appearance may contain valuable information about its cause, severity, timewise change of status, and healing diagnosis. This information is readily extracted using digital imaging. The core purpose of such CAAD systems is to aid physicians in analysis procedure, which contains the lesion boundary detection, the quantification of diagnostic features, the classification into different types of lesions, the visualization, the storage, the database management, etc. Stolz et al. (1994) defined a diagnosis idea for dermatoscopic images which is the ABCD rule of dermatoscopy. The four different letters means; Asymmetry, Border, Color and Differential structures. A weighted combination of the mentioned criteria provides a total dermatoscopic score that is used for lesions classification. As a matter of fact, those mentioned criteria had adopted in many CAAD researches related to melanoma.
2.6 Burns an Overview
One of the most catastrophic conditions encountered in medicine are Burns. All of us have experienced the severe pain that even a small burn can bring. The visible physical and the invisible psychological scars are long lasting and often lead to chronic disability. Burn injuries represent a diverse and varied challenge to medical and paramedical staff. Accurate management requires a skilled multidisciplinary approach that addresses all the problems facing a burn patient. As early wound closure shortens hospital stay and duration of illness.
Burns to the skin may occur in different ways. Burns can be caused by dry heat such as fire, moist heat such as steam or hot liquids, Radiation, friction, heated objects, the sun, electricity, or chemicals. Hettiaratchy and Dziewulski (2004) classified incidence of burns by age group, for example the Working age, between 15 -64 has the majority about ( > 60%) of burns incidents. These are mostly due to flame burns, and up to a third are due to work related incidents.
Figure 2.1 shows the possible causes of burns and the rate of burn by age.
Figure 2.1: Causes of burns (left) and rate of burns by age (right) (Hettiaratchy and Dziewulski, 2004)
2.7 Skin burn assessment and classification
Various visual methods used by physicians to assess the burn area. Classic visual observation of tissue colour can be useful (Hansen et. al, 1997), however human vision needs precision and it will be difficult to quantify the slow changes. When assessing a burn injury, depth is essential because it guides treatment and associates with the probability of wound healing (Heimbach et. al., 1992). Even though there are many sophisticated techniques to assess burn depth, including histologic examination, vital dyes, laser Doppler, thermography, and ultrasound, these techniques are time-consuming, expensive, and require special equipment that is not readily available (Heimbach et. al., 1992).
The medical assessment of depth can be estimated by combining information of the mechanism of injury with the wound characteristics outlined in Table 2.1 (Duffy et. al., 2006).
Analgesic and moisturizer
Moderate to severe
Red, white, yellow
Mild to moderate
Topical antimicrobial; may require skin graft
Topical antimicrobial; skin graft
Deep full (subdermal)
Topical antimicrobial; skin graft or flap
+, present; -, absent; +/-, may or may not present
Table 2.1: Modern burn classification system (Duffy et. al., 2006).
For burn area, there are three commonly used methods of estimating burn area (Hettiaratchy and Papini, 2004); the good and quick way of estimating medium to large burns in adults is Wallace rule of nines. The human body is partitioned into areas of 9%, and the total burn area can be calculated but it is not accurate in children.
In burn depth assessment, four elements that should be assessed (Hettiaratchy and Papini, 2004): bleeding on needle prick, sensation, appearance, and blanching to pressure. For the appearance three groups are identified:
A red, moist wound that clearly blanches and then rapidly refills is superficial.
A pale, dry but blanching wound that regains its colour slowly is superficial dermal.
Deep dermal injuries have a mottled cherry red colour that does not blanch (fixed capillary staining).
A dry, leathery or waxy, hard wound that does not blanch is full thickness.
Figure 2.2: Wallace rule of nines (Hettiaratchy and Papini, 2004).
Another study combines the modalities of burn wound depth assessment stated that burn wounds are classified into four categories of increasing depth: epidermal, superficial partial-thickness, deep partial-thickness, and full-thickness (Devgan et. al. 2006).
2.8 Some medical CAAD systems
At present there are many systems that have been used clinically. Although some of these systems are small, however make positive contributions to care. In this section, we will summarize some examples of CAD systems, as a matter of fact, there are various CAD systems for medical use, some of them are used in researches and some other are approved and used globally in many medical institutions.
According to Doi (2005), around 1500 CAD systems are currently used to assist radiologists in the early detection of breast cancer in many hospitals, clinics and screening centers around the world. Following are some CAAD systems approved by the American Food and Drug Administration (FDA):
2.8.1. ImageChecker® system
It is the only CAD system approved in 1998 by the Food and Drug Administration (FDA) for use in screening, diagnostic and digital mammography. Manufactured by R2 Technology Inc. in Los Altos, CA, it has been installed in more than 1400 clinical sites worldwide. ImageChecker CAD identifies (ROI) on mammography images and brings them to the attention of the radiologist in order to decrease false negative readings. (HOLOGIC Co., 2008).
2.8.2. MEDx is well optimized, portable software for the visualization, processing, and analysis of medical images. Designed especially for medical imaging researchers and system developers, MEDx can perform multi-modality, multi-dimensional image processing functionality to the medical imaging researchers. With MEDx, users display images and control menus using the industry-standard XWindow display system and Tcl/Tk interface language. MEDx transforms your computer into a powerful image processing workstation that displays, analyzes, and processes your medical image data in a myriad of innovative ways. (Medical Numeric, 2008).
2.8.3. EXINI Diagnostics
It is a set of Computer Aided Diagnosis systems dedicated for nuclear medicine images. EXINI systems for the diagnosis of bone metastases in whole-body bone scans and coronary artery disease in myocardial perfusion images exist. (EXINI Diagnosis, 2008).
It is a Computer-Aided Detection (CAD) system that helps radiologists who read mammograms to highlight areas that might be suspicious for breast cancer and might otherwise be missed. The system uses computing equipment and a special x-ray film scanner that changes the image on the mammogram into digital data. The device uses these data to locate areas suspicious for cancer of the breast. (MammoReader 2009)
2.8.5. RapidScreen™ RS-2000:
It is a computer-aided detection (CAD) system, used to recognize and mark regions of interest (ROI) on digitized frontal radiographs of chest. It can identify features associated with solitary pulmonary nodules from 9 to 30 mm in size, which could represent early-stage lung cancer. The device can be used by the physicians to perform initial diagnoses of the radiograph. It uses multiple-stage classification processes like heuristic decision rules, artificial neural network, and fuzzy logic for accurate classification. It is approved by FDA in 12th of July 2001. (RapidScreen RS-2000 2009)
2.8.6 Lorad Digital Breast Imager:
The Lorad Digital Breast Imager (LDBI) generates digital mammographic images that can be used for screening and diagnosis of breast cancer. The Lorad Digital Breast Imager is intended for use in the same clinical applications as traditional screen-film mammographic systems. The LDBI includes an image acquisition system and hardcopy display, it also includes a workstation computer with a monitor, keyboard, mouse, interface electronics, and storage devices. The device is used to screen and diagnose breast cancer just as regular analog (film) mammography. It is approved by FDA in March 15, 2002. (Lorad Digital Breast Imager 2009)
Developed at the Massachusetts General Hospital. It is used to aid in diagnosis, taking a set of clinical results including signs, symptoms, and laboratory data and then produces a diagnoses list. It gives explanation for each of dissimilar diagnosis, and suggests more examinations. The system contains a data base of crude probabilities for over 4,500 clinical manifestations that are associated with over 2,000 different diseases.
DXplain is being used routinely in a number of hospitals and medical schools, primarily for clinical training, but is also available for clinical consultation. It also has a role as an electronic medical textbook. It can provide a description of more than 2,000 different diseases, focusing on the signs and symptoms that occur in each disease and provides recent references appropriate for each specific disease. (MGHLCS division, 2008).
It diagnoses the results of pulmonary function tests. Puff went into production at Pacific Presbyterian Medical Center in San Francisco in 1977, making it one of the very earliest medical expert systems in use.
It is still in routine use; several hundred copies have been sold and are in use around the world. The PUFF system for automatic interpretation of pulmonary function tests has been sold in its commercial form to hundreds of sites world-wide (Snow et al., 1988).
2.8.9. Second Look™:
It is a computer-aided detection system for mammography used to recognize and mark regions of interest on standard mammographic views to be observed by the radiologist after the initial reading has been completed. Thus, the system assists the radiologist in minimizing observational oversights by identifying areas on the original mammogram that may need a second review. This devise uses image processing and pattern recognition algorithms hosted on a personal computer to detect potential areas of concern for more consideration by the radiologist. It has been approved by FDA since 2002. (Second Look™ 2009)
As seen from considering of existing CAD's, those CAD systems for the assistance to the physicians in making diagnosing, which may shorten the time to make the correct diagnosis and may reduce the number of diagnostic errors. At the same time, physicians may obtain the information on the symptoms of each of the diseases and pathologic syndromes contained therein.
This chapter has presented a condensed research background of the fundamental elements used in skin burn CAD systems. Computer Aided Diagnosis systems receive user's image inputs via digital camera or digitized slides, convert the input image into readable and understandable regions, which are then being analyzed and matched against a set of pre-trained patterns or images stored in the system's database.
CAD has become one of the main research subjects in medical imaging and diagnostic in many medical fields. Actually, a large number of CAD systems have been used to assist physicians in the early detection of various skin lesions.
Computer Aided Diagnosis systems are defined through the type of images used (MRI, CT, Mammography, etc...), the nature of algorithm used to segment the image and the feature classifier type. From the application viewpoint, the benefits of using CAD derive from providing a second opinion to the medical staff.
CAD is a concept combines roles of physician and computer; therefore CAD is definitely different from automated computer diagnosis. CAD will be used as a functional tool for diagnostic examinations in daily clinical work.