Sponsored Listings

Orthogate arrow Articles arrow Medical Informatics arrow Application of Artificial Intelligence to Medical Data
Application of Artificial Intelligence to Medical Data PDF Print E-mail
Written by ChrisOliver   
Monday, 14 January 2021
Medicine and surgery provide formidable and stimulating challenges to computer science and the application of artificial intelligence (AI) technology.

AI is computer software that performs tasks we would consider intelligent if done by a person. This includes giving expert advice, understanding “natural” computer languages, speaking like human and recognising complex patterns like handwriting. The popular view of intelligence is that it is associated with high level problem solving; people who can play chess, solve mathematical problems, make complex surgical decisions, and so on, are regarded as intelligent. Intelligence, however, is like an iceberg. The top of the iceberg represents a small amount of processing activity that relates to a high level of problem solving. That is the part that we can reason about and introspect. Much of the “iceberg,” however, is unseen and is devoted to its interaction with the physical environment. With intelligence we are dealing with information from a range of senses; visual, auditory and tactile, and coupling sensing to action, including the use of language, in an appropriate reactive fashion. Some of the unseen intelligence, like the submerged part of the iceberg is not accessible to reasoning and introspection.

The three most useful AI Programs today are expert systems, natural language, and neural networks. An expert system can solve real world problems by following the same IF/THEN rules a human expert follows. A software knowledge engineer interview’s one or several experts and encodes their thinking process into the software knowledge base.

The IF/THEN rules follow become expert software knowledge frames. Expert systems are useful for simple medical diagnosis and problem solving. Natural language software is the branch of AI that focuses on getting computers to understand spoken or typed Language. A neural network is a digitised model of a human brain, simulated in the binary memory of a personal computer. A neural network is made up of artificial neurons, connected to each other by weights indicating the strength of the connection. As a neuron becomes energised by input, it fires, sending a digital message to other neurons. There are hundreds or even thousands of these inter-linked neurons, arranged in layers, and all together they form a neural network, capable of learning from experience. Today, neural networks are the most exciting areas in AI. It is now possible to have real operational computer models of the human brain routinely perform outstanding feats of pattern recognition. Some examples are reading human handwriting at high speed, learning to identify male versus female faces, and even learning to speak almost perfectly.

To perform medical AI a computer must be able to represent typical and routine plans of medical treatment. The model may need to be able to represent preferences of both the patient and the treating physician for different interventions and outcomes. If it is to interact autonomously with a physiological a system, then the correct identification of patient state must precede any control action. Most diagnostic techniques have previously been designed to deal with diagnosis in the largely static context of the traditional medical consultation. The complications of interpreting continuously varying physiological systems in real time stretch such existing diagnostic methods considerably.

The wealth of medical knowledge that is potentially relevant to making a clinical decision makes it imperative that AI reasoning strategies are focused and clinically useful.

Medical AI programs may be use in a variety of ways. Programs may be designed to serve as consultants on complex problems where the medical practitioner knows ahead of time that help may be necessary. They may also serve as more limited reasoning tools to help the user explore particular consequences and models. Alternatively, they can be used as background monitors to detect unusual, possibly dangerous conditions that can then be brought to the attention of the health-care providers.

AI models have been developed of medical imaging, cardiac, electrical, biomechanical behaviours, circulatory dynamics and renal function. Intelligent electronic patient records may evolve in the future. An emerging problem, however is that of describing systems at different levels. Often one part of a model system requires greater detail than the rest of it. Neural networks can be used to describe how these different descriptions coexist and interact. With the development of computational techniques drawn from AI, prototype systems have been developed that advance monitoring in a quite different direction. Rather than simply displaying measurements for clinicians to interpret, the goal is to develop intelligent patient monitors that assist clinicians in the task of interpretation itself.

It is also possible to envisage devices that close the loop between, measurement and treatment. Empowered with the ability to automatically interpret clinical signals, one could construct therapeutic devices that automatically assess a patient’s clinical state and then alter the treatment they deliver. The closed loop control of drug delivery systems, or ventilators that are capable of automatically altering their settings have been developed.

In the future, new technologies may contribute to the development of more robust artificial organs with the ability to sense, intelligently interpret and respond to the varying contexts of the physiologic milieu they are placed in. The motivations for developing intelligent patient monitors are numerous. The most pressing needs arise from the difficulties’ individuals face when they need to monitor and react to continuous data. These human factors include the problems of data overload, vigilance, varying expertise, and human error. It comes as no surprise that clinicians may have difficulty in using the vast amount of information that can be presented to them on current monitoring systems. Not only is the amount of information available greater than can reasonably be assimilated or displayed, but the clinical environment provides sufficient distraction to reduce the effort that can be devoted to signal interpretation. It is not uncommon that current alarm technology floods clinicians with false messages providing unnecessary distraction. The level of expertise that individuals bring to a task like the interpretation of clinical signals varies enormously, and it is not always possible to access more skilled colleagues to remedy such deficits. This frequently leads to errors in diagnosis and selection of inappropriate treatment. Indeed, the majority of complications associated with anaesthesia, for example, result from inadequate training or insufficient experience of the anaesthetist.

We can conceptualise several distinct layers in the construction of an intelligent monitoring and control system, spanning the range from raw patient signals to control decisions. Each layer requires a number of different technologies. At the signal level, the development of novel transducers and signal processing techniques will continue to provide better ways of measuring physiological state. As important as the development of new sensors is the identification of those sensors that provides the most clinically relevant information. Fewer, more appropriate sensors are preferable to a proliferation of physiological signals that have to be correlated and interpreted before they give value.

After physiological signals are generated they need to be validated. Automating the process of signal validation has probably attracted less attention than it deserves, and this has contributed to the proliferation of false alarms from monitoring systems. At present it is often up to the clinician to ascertain whether a measurement accurately reflects a patient’s status, or is in error. However, validation is not necessarily a straightforward task. While in many situations, signal error is clear from the clinical context, it can also manifest itself as subtle changes in the shape of a waveform. The development of intelligent alarm systems that automatically validate signals prior to generating alarms remains an important focus.

Once validated signals are available, and noise and artefact have been filtered, pattern recognition technologies come into play, detecting regularities within the signal to suggest particular clinical conditions. Neural networks provide a way of matching signal to symbol. At the next level we move beyond the signal to symbol layer, and perform inferences with the symbols extracted from clinical data. Central to this activity is the notion of model-based reasoning, where patho-physiological knowledge is encoded as some form of computational model. The inferences that we derive from these models may be diagnoses, explanations of observed behaviour, predictions about future patho-physiological states, or control actions. The automated diagnosis of clinical conditions can assist clinicians by identifying rare diseases or complex cases. For example, intelligent monitors can track several diagnostic hypotheses over time. By comparing the dynamic behaviour of measured patient variables with the behaviours expected of each hypothesis, they can assist clinicians in narrowing down the range of possible causes.

Finally, there is the task layer, in which the intelligent monitoring system models the activities of the clinician. These models capture knowledge about the needs of and constraints operating upon different individuals, the knowledge they bring to a particular task, and the type of dialogue the system is likely to have with them. By explicitly capturing notions of clinical task we enhance the likelihood that intelligent systems meet clinical need.

At the heart of the challenge to medical AI is its ability to identify and satisfy fundamental clinical needs. There is little advantage in developing a complex system that mimics skills that most clinicians possess and use effectively. Rather, we should attempt to provide support for cognitive functions that clinicians perform poorly. The long lag in the introduction of computerised decision support into medicine is probably as much due to failure on this point as it is to the limitations of current technology. It is because the work in intelligent monitoring and control aims to satisfy perceived needs of working clinicians that it has a good chance of making a significant and positive contribution to health care. Even if clinicians do not routinely need support with patient diagnosis, there is still a case for assistance with machine diagnosis. Often there is a lack of specialised mechanical knowledge needed to work with complex machines, rather than any inherent difficulty in the logic of the task. It is a common medico-legal problem that, close to half of all anaesthetic mishaps associated with human error is related to operating the anaesthetic machine and ancillary equipment. Another smaller but significant portion is related to machine failures. It may be that systems that can self-diagnose faults, or suggest solutions to technical problems during surgery will have real clinical use.

Complex decision support systems introduce their own models of operation for the user to absorb, and if these are counterintuitive, imprecise or idiosyncratic they can make system use difficult. Any mismatch between the cognitive models used by a clinician and those designed into a decision support system means that the clinician has to alter the way he or she thinks before obtaining the benefit of the system. Although such exhortations for rational design may seem obvious in the light of the considerable work in human-computer interaction, adhering to them is not a trivial task. In contrast to the introduction of unwanted complexity, oversimplification of conceptual models may also have its own price to pay. Reducing the need for the clinician to carry detailed models of patient state can reduce cognitive load, but may leave them in a worse position if they are caught with an unexpected problem. While a clinician may initially expend effort to build up an accurate picture of a patient’s state, the accuracy of all models decays with time.

The underlying motivation in the development of AI must be to improve clinical outcome, whether it be through the reduction in error rates, or the optimisation of treatment delivery. Future challenges exist in the definition of the clinical role that such systems should fulfil, and the way in which they will be used by clinicians. It is only through a constant referral back to clinical motivation that one can hope to effect clinical outcome and meet genuine clinical needs. There has been a large amount of research in AI in medicine over the past twenty-five years. Many AI systems at present remain largely in the proof of concept stage. Their introduction being hampered by medico-legal and ethical considerations. AI systems will become more available as computer processing power increases and costs decrease. It is most probable in the next decade AI systems will become very valuable tools and be seen frequently in the clinical workplace.
Last Updated ( Saturday, 16 July 2020 )
Submit Community News
If you have some orthopedic news that you would like to share with the community, please submit your story, article, announcement or review here.