Volume 29, Issue 7 (10-2022)
RJMS 2022, 29(7): 189-203 |
Back to browse issues page
Research code: 0
Ethics code: 0
Clinical trials code: 0
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Shalchian A, Esmailkhani A, Razavi S, Zahedi bialvaei A. A Review of the Role of Artificial Intelligence in the Diagnosis and Treatment of Infectious Diseases. RJMS 2022; 29 (7) :189-203
URL: http://rjms.iums.ac.ir/article-1-7225-en.html
URL: http://rjms.iums.ac.ir/article-1-7225-en.html
PhD, Microbial Biotechnology Research Center, Iran University of Medical Science, Tehran, Iran , abedzahedi@gmail.com
Abstract: (2801 Views)
Background & Aims: Infectious diseases can be caused by the direct or indirect transmission of microorganisms such as viruses, bacteria, parasites or fungi. The spread of these diseases and infections may cause a global pandemic such as COVID-19. Establishing and using artificial intelligence can help scientists predict infectious diseases to prevent the spread of epidemics, understand the behavior of microorganisms, and discover drugs to control disease faster. Today, artificial intelligence is on the verge of evolving the health care system through disease-focused analysis and interventions to promote faster, more reliable, and more cost-effective solutions to human well-being. Artificial intelligence systems use cognitive computing, deep learning, variable neural networks, and machine learning and can play an important role in diagnosing, screening, monitoring, reducing the workload of caregivers, and predicting new treatments. This article reviews the potential applications of artificial intelligence in the field of infectious diseases that can help the health institutions of the global community in combating the increase of infectious diseases.
Methods: To achieve the goals mentioned above, almost all articles about application of artificial intelligence in the field of medical sciences and infectious diseases was evaluated in PubMed, and Scopus databases.
Results: Among the available analytical tools, artificial intelligence (AI) is recognized as the most powerful and promising tool for the human race (3). AI is the output of input sources: big data that needs to be refined, structured, and integrated. What we call big data can be defined by volume, speed, variety, variability, accuracy, and complexity. These terms refer to the amount of data, the speed of data entry and exit, the range of data types and sources, and accuracy, respectively. However, the volume and speed of data in today's healthcare are generally not high enough to require large data. However, in the context of omics, which generate hundreds of thousands of data-related topics on gene polymorphism, gene expression, metabolic, lipidomic, and proteomics, there is a need to develop better tools for identifying specific cases of general data orientation. Artificial intelligence can not only provide instant insights into the spread of disease by processing large volumes of data, but can also help predict new outbreaks. Population contact and movement tracking can be analyzed through artificial intelligence models to enable preventive measures, detection and understanding of epidemic outbreaks. In the current era of technological advancement, artificial intelligence models are widely playing an important role in analyzing massive data from various sources of infectious diseases such as national surveillance systems, reporting and monitoring systems, genome databases, outbreak reports, vaccination reports and human dynamics information. With the influx of huge volumes of data, the integration of data under primary data management and knowledge extraction enables the AI program to reveal latent trends. Artificial intelligence helps to model pandemics and simulate diseases in the field of diseases so that policy makers can take effective health care measures. AI plays an important role in controlling the new Covid-19 virus and helps global influenza tracking systems by predicting new influenza outbreaks in various parts of the world, and provides immediate insight by analyzing social media communications to track potential outbreaks. Artificial intelligence applications can help prevent preventive behaviors from spreading to infectious diseases.
Despite its good ability to diagnose malaria and possibly improve its diagnosis in the near future, there is a serious problem with resistance to antibacterial and antiparasitic drugs (20). Artemisinin-based combination therapy guidelines, approved 20 years ago, are now being challenged by the emergence of Plasmodium falciparum parasites with reduced susceptibility to these therapies. Mathematical modeling using pharmacokinetic-pharmacodynamic relationships specific to the parasite host stage predicted that artemisinin resistance was due to the resistance of the ring stages to the effect of the drug.
Recent studies have also shown the use of machine learning in effectively identifying the potential antimicrobial capacity of antibiotic-candidate compounds (23). Shen et al. have developed a decision support system that can suggest a patient-specific antibiotic treatment based on factors such as body temperature, site of infection, signs / symptoms, side effects, antibacterial spectrum, and even contraindications and select drug interactions with other drugs. This was possible thanks to the significant set of data used to build the model. This system includes 507 infectious diseases and their treatment methods in combination with 332 different places of infection, 936 symptoms related to the gastrointestinal tract, reproductive system, nervous system and other systems, 371 types of complications, 838407 types of bacteria, 341 types of antibiotics. There were 1504 pairs of reaction models (antibacterial spectra) between antibiotics and bacteria, 431 pairs of drug interactions and 86 pairs of specific contraindications of specific populations of antibiotics. In another study, models were developed to reduce the use of antibiotics. Infants can experience Systemic Inflammatory Response Syndrome (SIRS), which is a symptom caused by sepsis or non-infectious agents. Because it is difficult to make clear and rapid diagnoses using classical laboratory tests, it has been shown that the best set of predictors of non-laboratory and early variables can be identified using a random Forrest method (28).
The Internet of Things (IoT) gathers a wealth of information about our habits. We can predict that medicine will also benefit greatly from the IoT. Most clinical laboratory tests are performed automatically, and the complexity of the data generated can be increased. Strategies for implementing AI in health care institutions have yet to be developed (34). The first goal is to set up an advanced data management system. While most hospitals and clinics have such systems, these systems are often obsolete because they are not compatible with the type of data we produce today. In the ongoing follow-up of infectious diseases, hospitals must have a systematic way to predict the onset of nosocomial infections. However, while AI is widely seen as a threat to "shared" occupations, it should also be seen as an opportunity. Hence, hospitals, clinics and other regulatory bodies should see it as an opportunity. Recent work has shown the advantage of combining AI approaches for better detection. For example, ultrasound has been shown to be a useful tool for confirming the diagnosis of lung infection or pneumonia. This diagnosis depends on two factors: the operator's expertise and the potential bias when interpreted by the physician. Pattern-based diagnosis and image analysis have been used for automatic pneumonia grading (35). The neural network has correctly identified pneumatic infiltration (sensitivity higher than 90% and specificity 100%). In addition, geographic information about infectious diseases should be consistent with medical records and patient histories (36). It is important to immediately determine the relationship between the situation and other characteristics of patients such as professional activities, family environment, type of housing, contact with animals, etc. More preparedness for epidemics in the hospital according to the number of beds, and activation of measures Specifically tested to prevent the spread of infection.
Conclusion: As discussed in this study, the use of AI and machine learning is very promising. While access to medicine in developing countries is still a (financial) constraint, the use of AI to break the transmission chain may be the best cost-effective solution in the long run. Data structure is an essential part that must be defined upstream to enable the integration and analysis of such data sharing. Increasing commodity exchanges and travel does not reduce the risk of spreading infectious diseases unless global strategic decisions are made about implementing big data architectures and integrating them into AI-based solutions.
Type of Study: review article |
Subject:
Infectious Disease
Send email to the article author
