Introduction
One of the biggest challenges that hospitals in the United Arab Emirates face is the shortage of experienced radiologists. Although the number of radiologists has been increasing over the years, the increasing demand for their services surpasses the supply. In the United Kingdom, for instance, a report by Shoham (2022) indicates that only 2% of radiology departments can meet their imaging reporting requirements on time, a problem attributed to a shortage of nearly 122,000 radiologists.
The American College of Radiology reported in 2019 that there has been a consistent 30% growth in demand for radiologists in the United States over the past two years (Shoham, 2022). Ather, Kadir, and Gleeson (2020) reaffirm these findings, noting that the increasing number of cases of lifestyle diseases is driving up the demand for radiology services. However, the supply of the needed experts has relatively stagnated. In some countries, patients must wait for several days or even weeks to receive their radiology results, which compromises a doctor’s ability to understand the current state of a given condition.
Computed tomography (CT-scan) and magnetic resonance imaging (MRI) services are becoming common and popular diagnostics for doctors who are keen on having a thorough understanding of the condition of their patients before they can start targeted treatment. According to Shoham (2022), the demand for MRI and CT-scan services has been going up by 10% each year because of the need for these targeted treatments. However, the number of radiologists has not increased in tandem with the demand. It has become crucial to find a solution to address this problem, thereby enhancing the quality of healthcare services.
The revolution brought about by the digitization in healthcare and imaging diagnostics has led to the emergence of big data in the field of radiology. It is emerging that concepts of artificial intelligence (AI), machine learning, and algorithms can thrive in the field of radiology. Langlotz (2019) explains that machines are more effective than human beings when it comes to performing repetitive tasks. They learn to do it more quickly and effectively once they are trained.
According to Sechopoulos, Teuwen, and Mann (2021), AI gained significant acceptance during the outbreak of coronavirus disease (COVID-19), as there was a substantial increase in demand for radiology services at a time when medical facilities were overwhelmed worldwide. These machines can process numerous variables in a diagnostic within a relatively short period (Mudgal and Das, 2019). They can also detect a condition such as cancer, in its earliest stage possible, when a human eye cannot detect it. AI promises to address the current shortage of radiologists in healthcare institutions worldwide.
The rationale for this study is to determine whether AI has the potential to effectively address the shortage of radiologists in the United Arab Emirates. According to a report by Baltruschat et al. (2021), 1,030 new breast cancer cases were reported in the country in 2020, with the total number of new cancer cases reported to be 4,800. The report indicates that in the same year, 1,896 cancer deaths were reported. These deaths can be avoided if the disease is detected early enough and appropriate interventions are taken. The solution to this problem starts with addressing the dire shortage of radiologists in hospitals.
Background
The concept of AI has gained massive popularity in many fields over the past two decades. According to Winkel et al. (2021), the idea that a machine can be trained and then allowed to perform tasks that traditionally would only be performed by human experts is intriguing. In the healthcare sector, the concept of AI has gained widespread acceptance due to the unique capabilities of these machines.
Through machine learning and big data, it is now possible for AI to process diagnostic data in the radiology department within a short period, enhancing the efficiency, accuracy, and speed of such processes. Taylor-Phillips et al. (2022) observe that AI has significantly reduced the burden of radiologists in hospitals that have successfully introduced the system. A single radiologist, working in a fully equipped laboratory with the right machines and an AI system, can undertake and process numerous radiology tests without strain.
The field of AI has attracted the attention of many scholars over the years as they try to understand and explain its significance and relevance in the field of radiology. According to Mun et al. (2021), existing literature has confirmed the significance of AI in the field of radiology. Existing studies have identified the department of radiology as one where AI can excel, not only because of the obvious shortage of experts but also because of the unique ability of the machine to read radiology results in a more detailed and efficient manner than any human expert would.
There is a consensus that when using AI, radiology laboratories can avoid delays in releasing results and deliver more accurate results, which can help save the lives of millions of patients worldwide. Existing reviews on the topic are available, but there are still contentious issues that are yet to be effectively addressed. Although the benefits of AI to the field of radiology have been established in the existing literature, it remains contentious how these machines can effectively be utilized to address the shortage of radiologists.
Bahl (2020) explains that there is a growing concern that AI may replace radiologists in hospitals. Others are concerned about the legal and ethical implications of using this technology. For instance, there is still no consensus on who should be held accountable if an AI makes a critical mistake that leads to harm to a patient (Sechopoulos & Mann, 2020). A section of the AI community argues that the responsibility for such mistakes lies with the person training the AI, others believe that the person who runs the machine should be blamed, while another group believes the developer of the program should be held accountable. Such conflicts in the existing literature make it necessary to have updates such as this.
Aim and Objectives
The use of AI in the field of medicine has attracted the attention of numerous scholars over the past decade. As mentioned above, literature about this topic already exists. As such, defining the aim makes it possible to state a specific area of the study where one seeks to enhance existing knowledge due to identified gaps, inconsistencies, and conflicting information. Bell, Bryman, and Harley (2018) explain that having a clearly defined aim enables readers to understand what a scholar seeks to achieve in a study. By the end of this study, the researcher seeks to evaluate whether proven technological solutions (AI) can reduce the impact of the shortage of radiology staff.
The study aims to determine whether it is indeed possible to utilize AI to address the perennial shortage of radiology staff in hospitals. It will be a major milestone if local healthcare institutions can use the technology to address this problem. It will help the hospitals to manage emerging lifestyle diseases, especially the problem of cancer. To help achieve the above aim, the researcher seeks to achieve the following specific objectives:
- To explain the reasons for the shortage of radiology staff.
- To explore the advancement, obstacles, and implementation of AI in radiology.
- To demonstrate the impact of implementing AI in radiology.
Research Questions
AI has emerged as a valuable technological tool in the field of radiology and healthcare in general. It is important to address specific questions in line with the aim and objectives above to ensure that only relevant data is collected from the field. Aityan (2021) advises that it is essential to clearly define research questions to ensure that a researcher collects data that is relevant and addresses specific questions. The researcher seeks to answer the following research questions through this systematic literature review:
- Would AI reduce the impact of the shortage of radiologists and radiographers?
- What are the reasons for the shortage of radiology staff?
- What are the advancements, obstacles, and implementation of AI in radiology?
- What is the impact of implementing AI in radiology?
Methodology and Protocol
The study aims to investigate whether AI can mitigate the impact of the radiologist shortage. Defining the method used to collect and process data is one of the most important stages of conducting a study. It outlines the deliberate steps taken to respond to research questions and achieve the study’s aim, using data collected from various sources.
Aityan (2022) argues that the validity and reliability of a study can only be attained if the method used is capable of addressing specific issues in the research. In this chapter, the focus is on discussing the design of the research, the research strategy employed, the selection strategy, and critical appraisal issues. The chapter then discusses the ethical appraisal, data abstraction employed, and the analysis approach used to process data obtained from the identified sources.
Design
When selecting the appropriate research design, a researcher must clearly define the study’s aim and objectives. According to Bougie and Sekaran (2020), the chosen research design should enable a researcher to achieve the aim and objectives by directly and comprehensively responding to the research questions. In this study, it was of interest to determine whether AI could reduce the impact of the radiologist shortage. The nature of this study necessitates a detailed explanatory approach to understand how AI may impact the healthcare sector, with a specific focus on addressing the radiologist shortage in healthcare institutions.
The researcher considered qualitative research to be the most effective research design for this study. According to Cooper (2018), qualitative research refers to a method in which one analyzes non-numeric data to understand how and why a phenomenon occurred or would occur in a given manner. It utilizes unstructured questions to deliberately allow respondents to provide detailed explanations when answering each question.
It also allows a researcher to request a further explanation during the interview if a response is unclear or significantly different from the majority of the data collector’s expectations. The goal is to develop themes that explain how AI can directly address the problem of radiologist shortages in hospitals. It is essential to note that, although this is a qualitative study, some statistics are included to help explain the level of shortage of radiologists in the country and how the problem can be addressed with the aid of AI.
Search Strategy
The process of addressing a research question requires a clearly defined search strategy. Data used in this study were primarily collected from secondary sources, as De and Kammerlander (2020) recommend. Secondary data was obtained from peer-reviewed journal articles, books, and reliable online sources. The researcher conducted database searches, specifically focusing on medical journal databases, including EMBASE, CINAHL Plus, MEDLINE, PubMed, and the Cochrane Library. Keywords and search terms, such as AI, Big Data, machine learning, algorithm, radiology, and AI in Medicine, were used to conduct online research and identify relevant materials for the study.
The search was subjected to specific restrictions to ensure that only data relevant to the study’s goal was obtained. As such, the search was strictly limited to the impact of AI on addressing the shortage of radiologists. Other benefits of AI in the health sector were excluded from this study. It was also necessary to ensure that the sources used were as recent as possible, as noted by Eden, Nielsen, and Verbeke (2019).
As such, the search was limited to studies conducted within the past five years. Sources older than 5 years were only included if they addressed a specific issue in the study. The journals were hand-searched to ensure a detailed understanding of their contents. Most of these articles and books were obtained from online sources. The researcher employed a straightforward internet search strategy, visiting the databases as mentioned earlier and searching for relevant materials using the specified keywords and phrases.
The outcome of the search process revealed a strong relationship between the use of AI and addressing the shortage of radiologists, as discussed in the next chapter. The search enabled the study to address its major research questions. The Prisma flowchart shown in the figure below identifies the method used in this systematic review to select specific sources.
The use of a Prisma flowchart was justified because of the need to identify and review materials that specifically addressed the issue of concern. As Machado and Davim (2020) observe, researchers often encounter vast resources when searching for materials, some of which may not directly address the study’s goal. As such, one has to eliminate the least relevant resources. The above process also enables the identification of the most recent studies in the field of investigation.
The researcher found it necessary to contact specific organizations and topic experts to help in explaining the concept being investigated. In that respect, two major hospitals offering radiology services were identified. In these institutions, the researcher identified a total of 4 experts to help in understanding the issue. Although this was a systematic review, the researcher considered it necessary to contact organizations and experts on this topic to verify the findings made in this review. The process also enabled the identification of some gaps and conflicting information in the secondary data. Information obtained from the identified sources formed part of the findings presented in the next chapter.
Study Selection
The researcher defined specific inclusion criteria that had to be considered before including material in the study. They had to specifically address the problem of the shortage of radiologists, the use of AI in the radiology department, and how AI complements the work of radiologists. Studies had to be published within the last 5 years to ensure that they are current. Similarly, the researcher employed the PICO framework to define the exclusion criteria. Materials that do not specifically address the problem of this study (P), those with irrelevant issues (I), those that could not facilitate meaningful comparison (C), and those that could not link the impact of AI on addressing the problem of the shortage of radiologists (O) were excluded.
The criteria were used to ensure that the data collected directly responded to the study’s aim and objectives. It helps to eliminate cases where a researcher deviates from the study’s focus, wasting time collecting information that adds little value to the research, as Lee and Saunders (2017) caution. The outcome of the selection process was the identification of specific sources that addressed the problem being investigated. The strategy is justified due to the need for relatively small but highly relevant studies that can enable the researcher to address the research goal.
Critical Appraisal
In a systematic review, the quality of a study is directly dependent on the quality of the sources used. As such, it was essential to ensure that all the materials used in this study met the specific inclusion criteria discussed above. One of the criteria used to determine the quality of a study was whether it was published in a peer-reviewed journal. Pirozzoli and Sengupta (2019) believe that peer-reviewed journals tend to be of higher quality due to the large number of experts involved in their production and review.
The number of participants involved in these articles was another criterion. The researcher mostly selected articles that included more than 50 participants in the data collection process. As mentioned earlier, it was also essential to use recent studies to ensure that the information provided was as up-to-date as possible. This process of appraising the studies made it possible to select a few high-quality materials that enabled a response to the research questions. The outcome of the process was the collection of quality materials that directly responded to the aim and objectives of this study.
Ethical Appraisal
Ethical concerns must not be overlooked when conducting a systematic review of the literature. Mukherjee (2020) explains that ordinarily, the focus would be placed on maintaining ethical concerns when collecting data from people. However, it is equally important to embrace ethics when data is obtained from secondary sources. The process used to establish ethical aspects of the review was multifaceted.
First, the researcher attempted to avoid personal bias when collecting data. As such, the materials used in the study were primarily selected based on their quality, as outlined in the inclusion criteria. The temptation to only include materials published by authors from specific institutions was avoided. Instead, the focus was placed on how well each addressed the study’s aim, as well as the validity and reliability of their findings. The researcher also attempted to ensure that all sources used explained how they addressed ethical concerns during data collection.
The principle used to judge the ethical quality of sources used was the methodology each used to collect and process their data. The researcher ensured that all sources used had clearly defined methods and that their findings were based on the outcomes of these methods. Studies that did not raise ethical concerns were excluded from the study. The outcome of the ethical appraisal process was the collection of unbiased data that responded to the research questions. The strict inclusion criteria were justified due to the need for accurate information to inform policymakers’ decisions.
Data Abstraction
Data abstraction is essential when conducting a systematic literature review because it enables the provision of only essential information, which, in this case, helps to understand the impact of AI in overcoming the problem of the shortage of radiologists. One of the processes used in data abstraction involved having piloted forms. These forms were used to help summarize information on the sources used to respond to specific research questions and objectives. The researcher also sought the help of reviewers to verify the collected data. Saunders, Lewis, and Thornhill (2019) explain that even in cases where a study relies solely on secondary data, it may still be possible to seek reviewers who can help verify the data.
The researcher took into consideration various variables when seeking data from the identified sources. One of the variables was the characteristics of AI that made it possible to address the shortage of radiologists in hospitals. It was of interest to determine the unique ways in which AI can be applied to perform functions that radiologists conventionally perform. The efficiency of these machines in performing the identified functions as effectively as humans or even better was another variable considered.
AI can only address the problem if it is capable of delivering the same or better quality as a radiologist would. The quantity of work that the AI can perform within a given period was another essential variable. The current problem is caused by a situation where the amount of work that radiologists are expected to perform is overwhelming. The speed at which AI can perform these tasks effectively is, therefore, of great significance. Any missing information was collected from the few experts who were involved in the study.
Analysis
When the materials have been collected and screened, the next phase is to conduct the analysis. The goal of the analysis is to directly respond to the study’s objectives, as outlined by Sekaran and Bougie (2016). In this qualitative research, narrative analysis was deemed the most suitable approach for collecting data. According to Spalek (2019), when conducting a narrative analysis, one can employ structural, functional, dialogical, or thematic approaches. The choice of narrative analysis that a researcher selects should always be based on the primary goal of the study.
In this study, the researcher considered thematic analysis as the most appropriate narrative analysis for the research. It involves identifying the main topics in the study and discussing them in detail (Smith, 2020). It was considered an effective method for determining whether AI would alleviate the impact of the radiologist shortage. In this case, the researcher identified specific uses of AI in radiology as the specific themes.
Results and Discussion
The previous chapter has explained how the researcher collected and processed data obtained from different sources. In this chapter, the goal is to present results and conduct a detailed discussion of the findings. Based on a systematic review of the literature, it was necessary to determine whether the introduction of artificial intelligence could mitigate the impact of the radiologist shortage in hospitals. Findings from this chapter will inform decisions made by policymakers, particularly those interested in introducing AI as a means of addressing the radiologist shortage in major hospitals across the nation.
The primary question that this systematic literature review aims to answer is whether AI can help alleviate the shortage of radiologists in hospitals. Radiology has become a critical area of medicine, particularly due to the increasing need for accurate diagnosis of medical conditions to facilitate targeted medication (Collado-Mesa, Alvarez, and Arheart, 2018). The preliminary review of the literature revealed an increasing demand for radiology services, but the number of radiologists has not kept pace at the same rate. It is also important to note that it may take some time for a recent college graduate to become proficient in reading diagnostic results. In this section, the focus is on discussing the results of the findings made from the review.
Characteristics of Studies Included in the Systematic Review
The previous chapter explained that this study relies entirely on secondary data sources, although a few experts were consulted to help understand their opinions on this issue, which some consider divisive. The researcher was keen on utilizing peer-reviewed journals for the systematic review. They were considered to be relatively reliable. It is also easy to determine the validity of such scholarly articles by assessing the method they used to collect and process data.
Another major characteristic of the studies used is that they had to be recent. Noguerol et al. (2019) explain that the field of AI has undergone a rapid transformation. The capacity of AI to undertake specific tasks that previously could only be done by humans has greatly improved over the years. In the field of radiology, there is a consensus that AI can interpret MRI or CT scan images more accurately than a radiologist, a relatively recent advancement in the field of AI (Ranschaert, Morozov, and Algra, 2019). As such, the data used had to be as recent as possible to help understand the recent developments.
All the sources used were published within the last five years. Most of the sources used were published in Europe and North America, which is a major concern, as the concept has yet to receive thorough research in the local context. Most studies were conducted in a radiology laboratory at a cancer center, while others took place in colleges. Some of the studies included were field studies, based on data sampled from respondents, while others were systematic literature reviews. The studies were mostly qualitative or quantitative in their design.
Data Extraction
The researcher identified specific peer-reviewed journal articles that directly address the aim and objectives of this study. As mentioned above, these are recent studies published within the last five years that focus on the use of AI as a means of reducing the shortage of radiologists and its impact on the sector. The data extraction table, shown in Table 1, outlines the year, country of authorship, study setting, design used, sample size, and a brief result for each of the primary articles included.
Table 1. Data Extraction Table
Data Synthesis
The systematic review of the literature shows that AI is becoming increasingly essential in the field of radiology. According to Wuni, Botwe, and Akudjedu (2021), AI technology plays a critical role in enhancing efficiency in this field. As shown in Figure 2, one of the benefits of AI in radiology is that it creates more efficient workflows. Technology can define how diagnosis, image reading, interpretation, and printing of results are done. The AI can read and interpret radiology images more quickly than a radiologist. The technology also enables the early detection of disease, allowing for the management of a condition at the earliest stage possible.
When AI is used, there is a high likelihood of improved diagnostic accuracy in the radiology laboratory. Once trained to detect specific abnormalities in an image, these machines can identify them more accurately. They can also diagnose numerous medical conditions from a single image. It means that a patient will not need to undergo several tests for various conditions. This capacity lowers the overall cost of diagnosis and reduces the time needed to conduct such tasks.
Personalized diagnostics are also possible when using AI. This technology eliminates the need to maintain physical patient records, a tradition that has long been a part of the healthcare sector. Instead, data is obtained in a digital format, which minimizes cases of lost information. A patient’s data is stored in a shared database that can be easily accessed by the relevant physician and other medical experts providing care to the patients (Qin et al., 2019). These obvious benefits of AI have made it gain massive popularity in the field of medicine, specifically in radiology.
A systematic review of the literature reveals the growth in the number of AI products in radiology that have been made available since 2006. As shown in Figure 3, there was a sharp drop in the supply of these products from 2006 to 2008. A slight increase was registered in 2009 and 2010, followed by another drop in 2011. Between 2012 and 2018, there was significant growth in the number of these products, although the growth was unpredictable.
In 2019, there was a sharp increase in the number of products, followed by another drop in 2020. The drop in 2020 can be attributed to the COVID-19 pandemic outbreak. The statistics shown in the figure strongly suggest that the use of AI products in radiology is on the rise. It is becoming increasingly evident that technology is becoming critical in this sector due to the numerous benefits it offers.
When conducting a systematic review of the literature, it became evident that scholars have been raising concerns about the ethical issues associated with the implementation of AI in the healthcare sector. Although there is a consensus that AI is more efficient and less prone to making mistakes than radiologists, they can also err. When a radiologist commits an error in their line of duty, they will be personally held responsible and can face litigation. However, the same cannot be as straightforward if AI makes a mistake. The current ethical debate is about who should be held accountable for mistakes made by AI.
Different stakeholders have varying opinions about who should bear the responsibility when mistakes can be directly attributed to the use of AI in the radiology department. The European Society of Radiology (2019) conducted a study, and Figure 4 shows the outcome, explaining individuals who should be held accountable when such mistakes occur. A significant section of those interviewed in that study (41.1%) believes that it should be a shared responsibility. They argue that when using technology, it may be challenging to assign all the blame to a single individual. As such, there should be a shared responsibility among all stakeholders, depending on their perceived role in the system’s failure.
The study shows that another significant number of participants (41.1%) believe that radiologists should be fully held responsible for such failures and mistakes. They argue that radiologists have the responsibility of training and monitoring tasks performed by the AI. Although it is possible to fully automate the system with minimal human supervision when AI is used, radiologists are responsible for regularly inspecting the system and the machine to ensure it is working as expected. It means that they cannot fully delegate their responsibility to AI. The new system is designed to simplify and enhance their work efficiency.
The study indicated that developers of AI may also be held accountable (10.2%). They argue that the code and programs of the AI are responsible for some of the mistakes made by AI. In such a case, a radiologist will not be blamed because they do not write code. An error made at the initial stages of developing AI may have devastating consequences when the technology is assigned the role of reading images and processing results. AI may also face a cyber-attack, which may corrupt the software. Developers are expected to find ways to make the system safe enough to resist potential hacking and manipulation.
A minority of respondents noted that insurance companies should be held responsible when mistakes occur due to AI. In such a case, it becomes the responsibility of the management of the relevant institution to ensure that all activities conducted by AI are ensured. Proponents of this strategy argue that doing so will eliminate fear among radiologists and developers, who are responsible for advancing this technology. If fear is instilled in them, then it will kill creativity among developers.
On the other hand, radiologists will deliberately avoid using technology for fear of facing specified consequences. Opponents of this strategy argue that when responsibility is taken away from individuals, it may promote recklessness and dangerous experimentation that can easily lead to death or life-changing injuries. As such, it is essential to hold individuals responsible as a means of ensuring that they prioritize protecting lives in their efforts to make their work easier.
Discussion
Summary of Major Findings
The Use of AI Is Gaining Rapid Popularity in the Field of Radiology
It is viewed as a solution to the increasing shortage of radiologists in hospitals worldwide. As the global population continues to increase and life expectancy is enhanced, the demand for medical services has been on the rise. AI promises to solve problems associated with incorrect disease diagnoses resulting from human errors in interpreting MRI and CT-scan images. It also makes it possible to diagnose major medical conditions in the earliest stages possible.
Size of the Crisis: Shortage of Radiologists
It is essential to begin by defining the scope of the crisis, which is the shortage of radiologists in hospitals nationwide. According to Shoham (2022), only 2% of the radiology departments in the United Kingdom can meet their imaging reporting requirements in time. The situation is particularly dire in developing countries across Asia, South America, and Africa. The study also reveals a consistent 30% growth in demand for radiologists nationwide, which is exacerbated by the fact that approximately 22% of the current workforce is expected to retire within the next five years. The situation in the Middle East, especially in the United Arab Emirates, is worse.
Factors Attributed to the Shortage of Radiologists
One of the primary factors contributing to the shortage of radiologists is the growing global population. As the population increases, so does the demand for medical services. Shoham (202) explains that the increasing number of radiology imaging procedures, such as MRI, CT, and X-ray, is also exerting pressure on radiologists. More doctors are currently demanding proper radiology imaging procedures before they can initiate the treatment process, as a means of ensuring that they administer targeted medication (Pirozzoli & Sengupta, 2019). As the life expectancy continues to increase, radiology services become more essential as senior citizens are more likely to suffer from conditions such as cancer and damage to internal organs that require radiology services.
Market Drivers for the Medical Imaging AI
The shortage of radiologists has been created by market drivers for medical imaging services, which in turn are now creating demand for the use of AI in the field of radiology. Repetitive tasks, such as reading and interpreting images, can be automated by AI to help streamline the system and reduce the time spent in the laboratory. AI is more accurate than human experts, which means that when it is used, it improves diagnostic and prognostic accuracy. The increasing use of imaging biomarkers in clinical practice has also encouraged the use of AI as a tool for radiologists. AI is also evolving, enabling it to go beyond detection to processing data and providing output that specifies the action to be taken to address the condition.
Challenges to Implementing AI in Radiology
It is essential to acknowledge that certain challenges persist, hindering the implementation of AI in radiology. One of the issues is that radiologists are hesitant to adopt new technology for various reasons. Some of them cite legal and ethical concerns, especially when the AI system misinterprets information, while others fear that the technology may render their specialized job obsolete.
Such fears have led to reduced clinical implementation of AI despite the advances in technology that have been made. The technology is still relatively new, and a section of the community feels that it should not be trusted with decisions that can cost a life (Rao, 2021). Some management teams are also unwilling to make the necessary investment to implement the technology due to the fear that such expenses may not be recouped.
Artificial Intelligence Advancements in Radiology
AI capabilities in the radiology field have undergone significant enhancements. Currently, it is easy to detect pneumothorax, consolidation, pleural effusion, pulmonary lesions, and atelectasis through a Chest X-ray AI. It is also possible to automatically measure relevant anatomies and abnormalities, such as the volume of the heart, the volume of coronary calcium, the diameter of the aorta, and the volume of lung nodules, using Chest CT AI. Prostate cancer can be detected at the earliest stage using MRI AI. Shoham (2022) explains that advancements in the field of AI make it possible to understand the stage of a medical condition at the earliest stage possible and even propose the appropriate treatment regimen.
Impact of Implementing AI in Radiology
Implementing AI will not only handle the shortage of radiologists but also increase the accuracy of these services. Mun et al. (2021) note that while radiologists have an average accuracy score of 66% in predicting brain tumor neuropathology, the score for AI is 85%, and it is improving. AI also lacks the human bias that often affects the effectiveness of radiologists. The technology significantly reduces the time it takes to conduct a test, release a result, and interpret it in a way that is useful to physicians. This means that there will be an enhancement in the quality and quantity of processes conducted by radiologists.
Trends in AI Image Evaluation Solution in Radiology
AI technologies are advancing, and their capabilities in radiology are increasing. One of the emerging trends in the transition from narrow, single-purpose algorithms to more comprehensive AI algorithms is the ability to conduct a comprehensive body area scan from a single test. It means that several conditions can be detected from a single MRI or CT scan, which is economical and time-saving. It is becoming a critical tool for screening common chronic conditions among the general population. It is also automating repetitive tasks in the healthcare sector, helping to reduce the demand for radiologists and other specialists.
Original Aim and Objectives of the Study and Situating Findings
It is essential to connect the study’s outcomes to its original aims and objectives. The study aimed to evaluate whether artificial intelligence can mitigate the impact of the radiology staff shortage. The outcome of this investigation has revealed that AI can effectively address the shortage of radiologists and even enhance the quality and quantity of services they offer. The study’s findings have addressed each of the research objectives. It has clarified the causes for the shortage of radiology staff, examined the advancement, obstacles, and implementation of AI in radiology, and revealed the impact of implementing AI in radiology.
It is equally necessary to situate findings within the broader context of the current knowledge. The systematic review conducted shows that the information available in published sources does not reflect the current capabilities of AI (Leeuwen et al., 2021). This study demonstrates that advancements in the field of AI are occurring at a significantly faster pace than scholars have been able to document. As such, it is safe to assume that by the time this document is published, AI technologies in radiology shall have advanced further.
Limitations
It is essential to acknowledge that the review has limitations that future scholars can address through further research. One of the major limitations of this study was that it mainly relied on secondary data sources. A few experts from local universities were consulted to verify some of the information obtained from secondary data. This means that the validity and reliability of the data presented in this study largely depend on the validity and reliability of the sources included in the study. Future scholars should consider collecting primary data from local experts when conducting further research. It is also necessary to note that the majority of the sources used were from Western nations.
As shown in Table 1, the majority of the sources were published in the US and the UK. Only a handful of the sources were published within the Middle East. Although the usage of AI can be universally applied in any country, technological advancements in each country are often one of the defining factors in its implementation. The technological advancements in Western Europe and North America are significantly different from those in the United Arab Emirates and its neighbors in the region. As such, future scholars should consider conducting studies that focus more on the region.
Conclusion
The study aimed to determine if artificial intelligence can help reduce the shortage of radiologists. Based on the findings presented in this paper, it is concluded that AI can reduce the impact of the radiologist shortage. It has been proven that radiologists spend a significant amount of time reviewing images from MRI and CT scans to aid in the diagnosis of various medical conditions in the laboratory. The study has also confirmed that this tedious task of reading and interpreting images can be done more efficiently by AI. The AI machine analyzes such images more quickly and accurately, making it possible to diagnose a medical condition in its earliest stages.
The time that radiologists would have spent reading and interpreting the images can be used in undertaking other tasks. It means that one or two radiologists who utilize AI can perform their tasks more efficiently than 40 radiologists who do not use the technology. Although some ethical issues have been raised, including concerns among radiologists that the technology may render their work obsolete, the technology is intended to make their work more straightforward and efficient. Local hospitals should adopt AI as a means of mitigating the radiologist shortage.
Recommendations
Practice and Management
The findings of this study strongly suggest that AI, as a technology, is unavoidable in the radiology sector. Its obvious benefits will help reduce pressure on radiologists while at the same time enhancing the diagnosis of life-threatening diseases such as cancer. The following recommendations should be considered by the relevant stakeholders:
- When introducing AI in a hospital’s radiology department, it is critical to involve all the relevant stakeholders, including radiologists, from the earliest stage possible.
- Some form of training would be necessary to ensure that radiologists understand how to work effectively with AI.
- Fears and ethical concerns associated with the use of AI should be addressed to ensure that radiologists can effectively adopt and utilize this new technology.
- The management of hospitals should be willing to set aside funds needed to support research and development of the radiology department once AI technology is introduced.
Education and Research
Educationists and future scholars will find the field of AI interesting and critical in their research. As technology continues to advance, it is essential to acknowledge that some changes are likely to emerge, necessitating further research. The following recommendations should be taken into consideration by future scholars and educationists:
- Future scholars should focus on addressing ethical and legal concerns associated with the use of AI in the field of radiology.
- Further studies are needed to clearly define the role of radiologists in the world of AI and how these human experts can work alongside the machines.
- Educationists, especially those in institutions of higher learning, should provide both moral and financial support for further investigation into how to effectively utilize AI in the field of radiology.
Reference List
Ahuja, A. (2019) ‘The impact of artificial intelligence in medicine on the future role of the physician’, Peer J, 4(7), pp. 4-19.
Aityan, S. (2021) Business research methodology: research process and methods. Singapore: Springer Nature.
Aityan, S. K. (2022) Business research methodology: research process and methods. Cham: Springer.
Ather, S., Kadir, T. and Gleeson, F. (2020) ‘Artificial intelligence and radiomics in pulmonary nodule management: current status and future applications’, Clinical Radiology, 75(1), pp.13-19.
Bahl, M. (2020) ‘Artificial intelligence: a primer for breast imaging radiologists’, Journal of Breast Imaging, 2(4), pp. 304-314.
Baltruschat, I. et al. (2021) ‘Smart chest X-ray worklist prioritization using artificial intelligence: a clinical workflow simulation’, European Radiology, 31(1), pp. 3837-3845.
Bell, E. Bryman, A., and Harley, B. (2018) Business research methods (5th ed). Oxford: Oxford University Press.
Bougie, R. and Sekaran, U. (2020) Research methods for business: a skill-building approach. Hoboken: John Wiley & Sons.
Collado-Mesa, F., Alvarez, E. and Arheart, K. (2018) ‘The role of artificial intelligence in diagnostic radiology: a survey at a single radiology residency training program’, Journal of the American College of Radiology, 15(12), pp. 1753-1757.
Cooper, D. R. (2018) Business research: a guide to planning, conducting, and reporting your study. New York: SAGE Publications.
De, M. and Kammerlander, N. (Eds.) (2020) Handbook of qualitative research methods for family business. Northampton: Edward Elgar Publishing.
Eden, L., Nielsen, B. and Verbeke, A. (Eds.) (2019) Research methods in international business. Basingstoke: Palgrave Macmillan.
European Society of Radiology. (2019) ‘Impact of artificial intelligence on radiology: a Euroaim survey among members of the European Society of Radiology’, Insights Imaging, 10(105), pp. 1-18.
Geis, J. et al. (2019) ‘Ethics of artificial intelligence in radiology: summary of the joint European and North American multi-society statement’, Canadian Association of Radiologists Journal, 70(4), pp. 329-334.
Gong, B. et al. (2019) ‘Influence of artificial intelligence on Canadian medical students’ preference for radiology specialty: a national survey study’, Academic Radiology, 26(4), pp. 566-577.
Hardy, M. & Harvey, H. (2020) ‘Artificial intelligence in diagnostic imaging: impact on the radiography profession’, The British Journal of Radiology, 93(1108), pp. 7-21.
He, J. et al. (2019) ‘The practical implementation of artificial intelligence technologies in medicine’, Nature Medicine, 25(1), pp. 30-36.
Hosny, A. et al. (2018) ‘Artificial intelligence in radiology’, Nature Review Cancer, 18(1), pp. 500-510.
Kwee, T. and Kwee, M. (2021) ‘Workload of diagnostic radiologists in the foreseeable future based on recent scientific advances: growth expectations and role of artificial intelligence’, Insights into Imaging, 88(1), pp. 9-24.
Langlotz, C. (2019) ‘Will artificial intelligence replace radiologists’, Radiology Artificial Intelligence, 1(3), pp. 6-18.
Lee, B. and Saunders, M. (2017) Conducting case study research for business and management students. New York: SAGE Publications.
Leeuwen, G. et al. (2021) ‘How does artificial intelligence in radiology improve efficiency and health outcomes’, Pediatr Radiol, 4(1), pp. 3-9.
Liew, C. (2018) ‘The future of radiology augmented with Artificial Intelligence: a strategy for success’, European Journal of Radiology, 102(5), pp.152-156.
Machado, C. and Davim, J. (2020) Research methodology in management and industrial engineering. Cham: Springer.
Moloney, B. et al. (2017) ‘Teaching radiology to medical students: there is a need for change to better prepare students for clinical practice’, Medical Student Education, 24(4), P506-513.
Mudgal, K. and Das, N. (2019) ‘The ethical adoption of artificial intelligence in radiology’, The British Journal of Radiology, 2(1), pp. 1-11.
Mukherjee, S. P. (2020) A guide to research methodology: an overview of research problems, tasks and methods. New York, NY: CRC Press.
Mun, S. et al. (2021) ‘Artificial intelligence for the future radiology diagnostic service’, Frontiers in Molecular Biosciences, 4(28), pp. 12-18.
Neri, E. et al. (2020) ‘Artificial intelligence: who is responsible for the diagnosis’, La Radiologia Medica, 125(1), pp. 517-521.
Noguerol, M. et al. (2019) ‘Strengths, weaknesses, opportunities, and threats analysis of artificial intelligence and machine learning applications in radiology’, Journal of the American College of Radiology, 16(9), pp. 1239-1247.
Pesapane, F. et al. (2018) ‘Artificial intelligence as a medical device in radiology: ethical and regulatory issues in Europe and the United States’, Insights Imaging 9(1), 745-753.
Pirozzoli, S. and Sengupta, T. K. (Eds.) (2019) High-performance computing of big data for turbulence and combustion. New York, NY: Springer.
Qin, Z. et al. (2019) ‘Using artificial intelligence to read chest radiographs for tuberculosis detection: a multi-site evaluation of the diagnostic accuracy of three deep learning systems’, Scientific Reports, 9(1), 4-11.
Ranschaert, R., Morozov, S. and Algra, R. (Eds.) (2019) Artificial intelligence in medical imaging: opportunities, applications and risks. New York: Springer.
Rao, M. (2021) Project management process. New Delhi: Blue Rose Publishers.
Rimmer, A. (2017) ‘Radiologist shortage leaves patient care at risk, warns royal college’, British Medical Association, 359(1), pp. 1-12.
Rodríguez-Ruiz, A. et al. (2018) ‘Detection of breast cancer with mammography: effect of an artificial intelligence support system’, Radiology, (1), pp. 1-9.
Rodriguez-Ruiz, A. et al. (2019a) ‘Can we reduce the workload of mammographic screening by automatic identification of normal exams with artificial intelligence? A feasibility study’, European Radiology, 29(2), pp. 4825-4832.
Rodriguez-Ruiz, A. et al. (2019b) ‘Stand-alone artificial intelligence for breast cancer detection in mammography: comparison with 101 radiologists’, Journal of the National Cancer Institute, 111(9), pp. 916-922.
Saleh, S. (2022) ‘Number of new cancer cases in the general population of the United Arab Emirates as of 2020, by type’, Statistica, pp. 1-3.
Saunders, M., Lewis, P. and Thornhill, A. (2019) Research methods for business students. New York, NY: Pearson.
Sechopoulos, I. and Mann, M. (2020) ‘Stand-alone artificial intelligence: the future of breast cancer screening’, The Breast, 49(1), pp. 254-260.
Sechopoulos, I., Teuwen, J. and Mann, R. (2021) ‘Artificial intelligence for breast cancer detection in mammography and digital breast tomosynthesis: state of the art’, Seminars in Cancer Biology, 72(4), pp. 214-225.
Sekaran, U. and Bougie, R. (2016) Research methods for business: a skill-building approach. Chichester: John Wiley & Sons.
Shaheen, M. (2021) ‘AI in healthcare: medical and socio-economic benefits and challenges’, ScienceOpen Preprints, 4(1), pp. 1-9.
Shoham, A. (2022) ‘The radiologist shortage and the potential of AI’, European Congress of Radiology, 4(1), pp. 1-4.
Smith, M. (2020) Research methods in accounting. Los Angeles: SAGE.
Spalek, S. (ed.) (2019) Data analytics in project management. Boca Raton: Taylor & Francis Group.
Syer, T. et al. (2021) ‘Artificial intelligence compared to radiologists for the initial diagnosis of prostate cancer on magnetic resonance imaging: a systematic review and recommendations for future studies’, Cancers, 13(13), pp. 7-19.
Taylor-Phillips, S. et al. (2022) ‘UK National Screening Committee’s approach to reviewing evidence on artificial intelligence in breast cancer screening’, The Lancet Digital Health, 4(7), pp. 558-565.
Wayme, Q. et al. (2019) ‘Impact of the rise of artificial intelligence in radiology: what do radiologists think’, Diagnostic and Interventional Imaging, 100(6), pp.327-336.
Winkel, L. et al. (2021) ‘Impact of artificial intelligence support on accuracy and reading time in breast tomosynthesis image interpretation: a multi-reader multi-case study’, European Radiology, 31(1), 8682-8691.
Wuni, A., Botwe, B. and Akudjedu, T. (2021) ‘Impact of artificial intelligence on clinical radiography practice: futuristic prospects in a low resource setting’, Radiography, 27(1), pp. 69-73.