COMP3330 History Assignment

Artificial Intelligence Applications In Medicine

Artificial Neural Networks & Genetic Algorithms

Applications Of Artificial Neural Networks & Genetic Algorithms

Neural networks and genetic algorithms form one of the most recent trends in the development of computer-assisted diagnosis. Both neural networks and genetic algorithms must "learn" their knowledge interactively from the user.

The two main areas of application are healthcare support and machine learning. These are discussed below.

Healthcare Support

These types of AI applications have been used in:

treatment of back pain

diagnosis of breast cancer

classification of giant-cell arthritis

classification of acute myocardial infarction

image mapping - detecting heart diseases, skin diseases and bone deficiencies by using images

Some other examples are provided below:

Department of Medical Cybernetics and Artificial Intelligence, University of Vienna, Austria:
The given task is to forecast the intervals between the heartbeats recorded from a foetus. The six tested neural network models combine input windows, hidden layer feedback, and self-recurrent unit feedback in different ways. One of them has additional self-recurrent feedback loops around the units in the state layer, which enable the system to deal with time-warped patterns.
 

Greenebaum Cancer Center, Department of Medicine, University of Maryland School of Medicine:
UM researchers used artificial neural networks to differentiate and diagnose several types of colon tumours. 39 tissue samples were taken from patients who had well-documented cases of sporadic (or "common") colon cancers or the more serious IBD-related growths and cancers.
 
The researchers extracted the DNA from the samples and then used high-tech gene microarray equipment to analyse 8,064 genes to determine the level at which they were present in each colon sample. These "gene expression" levels were translated into numbers, which were processed by artificial neural networks.
 
Using gene information from 27 of the 39 samples, researchers trained the neural network to recognise the two types of colon cancer, and then gave it information from 12 samples it had never seen. It made the correct diagnosis in all 12 cases.
 
Until now, there was no reliable way to discriminate between sporadic (or common) colon cancers and the more serious IBD-related growths and cancers, especially in their early stages. The method could ultimately help doctors to identify the cancers earlier and spare some patients from unnecessary, debilitating surgery.
 

Department of Clinical Physiology, Faculty of Medicine, Lund University, Sweden:
Artificial neural networks have been trained to detect acute myocardial infarction from a patient's (12-lead) ECG. A total of 11,572 ECGs were provided as input - 1,120 ECGs from patients with acute myocardial infarction and 10,452 control ECGs.
 
The performance of the neural networks was better than those of conventional rule-based systems and an experienced cardiologist, indicating that artificial neural networks may be useful in assisting ECG readers with all levels of experience.
 

Department of Anaesthesiology, University of Utah, Salt Lake City, Utah:
Artificial neural networks have been successfully applied to many areas of monitoring and control in anaesthesia. They have become a powerful tool to integrate physiologic signals and extract meaningful information.
 

Laboratory for Research on Neuroscience of Autism, Children's Hospital, San Diego Research Center:
Two-layer and three-layer feed-forward artificial neural networks were trained to predict behavioural performance from single-trial EEG in autistic and normal subjects, in a task involving response to rare stimuli and shifting of attention between vision and audition. Performances of the networks on separate test sets varied across subjects but were usually at least 80%.
 

The University of Pennsylvania Medical Center, Philadelphia:
Researchers have developed a "smart" ICU system that improves the vital-signs monitoring of critically ill patients. A combination of neural network and fuzzy logic technology is used to convert a patient's vital-sign measurements into easy-to-follow visual models to assist physicians and nurses in monitoring patients' physiological parameters.

Machine Learning

Medicine has formed a rich test-bed for machine learning experiments in the past, allowing scientists to develop complex and powerful learning systems. Machine learning techniques usually employ neural networks, but may also encompass a large variety of other methods as well, each with their own particular characteristic benefits and difficulties.

While there has been much practical use of expert systems in routine clinical settings, machine learning systems are generally used in a more experimental way. There are many situations in which they have made, and continue to make, significant contributions:

Developing the knowledge bases used by expert systems:
Given a set of clinical cases that act as examples, a machine learning system can produce a systematic description of those clinical features that uniquely characterise the clinical conditions. A classic example of this type of system is KARDIO, which was developed to interpret ECGs.
 

Exploring poorly understood areas of medicine:
Using techniques such as 'data mining' and 'knowledge discovery', it is possible to use patient data to automatically construct pathophysiological models that describe the functional relationships between the various measurements. These models might be used to look for changes in a patient's condition if used at the time they are created. Alternatively, if used in a research setting, these models can serve as initial hypotheses that can drive further experimentation.
 

Using learning systems to discover new drugs:
The learning system is given examples of one or more drugs that weakly exhibit a particular activity, and based upon a description of the chemical structure of those compounds, the learning system suggests which of the chemical attributes are necessary for that pharmacological activity. Based upon the new characterisation of chemical structure produced by the learning system, drug designers can try to design a new compound that has those characteristics. This application is relatively new, and significant developments can be expected here in the next few years.