1. INTRODUCTION:

Life expectancy has been increasing worldwide due to significant improvements in healthcare, and medicine, as well as due to growing consciousness about personal and environmental hygiene¹. The cost associated with health care services continues to soar because of the increasing price of prescription drugs, medical instruments, and hospital care². It is an utmost necessity to develop and implement new strategies and technologies in order to provide better health care services at an affordable price to the aging population³.

Artificial intelligence is gradually changing the landscape of healthcare and biomedical research⁴. Artificial Intelligence is a field of science that pursue the goal of creating intelligent application and machine that can be mimic human cognitive functions, such as

learning and problem solving machine learning {NL} and deep learning {DL} are subsets of artificial intelligence {AI}. Machine learning implies⁵. A team of researchers at Google Inc. and collaborating institutions showed that an AI system trained on thousands of images can achieve physician-level sensitivity and specificity in diagnosing referable Diabetic retinopathy (DR)^[6]. AI has recently re-emerged into the scientific and public consciousness, as new breakthroughs and technologies are announced from technology companies and scientists at a breakneck pace⁴

1.2. Suspects of Artificial Intelligence:⁵

I. Machine Learning

II. Deep Learning

Machine Learning implies training Algorithms to solve task independently using pattern recognition. For example, Researchers can apply machine learning Algorithms to radiology by training them to recognize pneumonia in lung scan⁵.

Deep learning solutions rely on neural networks with artificial neurons modeled after a human Brain. These networks have multiple hidden layers and can derive more insides then linear algorithms. Deep learning algorithms are widely used to reconstruct medical images and enhance their quality.⁵

1.3 Historical Overview of Artificial Intelligence in medicine:

Medicine was identified early as one of the most promising application areas for AI. Since the mid-twentieth century, researchers have proposed and developed many clinical decision support systems. Rule-based approaches saw many successes in the 1970s, and have been shown to interpret ECGs, diagnose diseases, choose appropriate treatments, provide interpretations of clinical reasoning and assist physicians in generating diagnostic hypotheses in complex patient cases⁴. It was difficult to implement a system that integrates deterministic and probabilistic reasoning to narrow down relevant clinical context, prioritize diagnostic hypotheses, and recommend therapy⁷.

2.0 Radiology:

Radiology uses radiation to image the insides of bodies for diagnosing and treating disease⁸. Diagnostic radiologists use multiple medical-imaging modalities — the most widely used being X-rays radiography, computed tomography, magnetic resonance imaging (MRI) and positron-emission tomography — to detect and diagnose diseases⁴. Radiologists use a collection of images for disease screening and diagnosis, to identify the cause of illness and to monitor the patient trajectory during the course of a disease⁹.

Figure 1: Historical context for artificial intelligence in radiology. The history of computerized radiology falls at the developmental intersection of radiography and modern computers⁹.

Radiology departments maintain a database of historical images in a Picture Archiving and Communication System, which typically provides thousands of examples to train the neural networks. Computational approaches for radiology diagnoses have been proposed and implemented since the 1960s. With the help of modern machine-learning methods, many radiology applications of AI, such as the detection of lung nodules using computed tomography images, the diagnosis of pulmonary tuberculosis and common lung diseases with chest radiography and breast-mass identification using mammography scans, have reached expert-level diagnostic accuracies⁴.

European Radiology Experimental, the new journal launched by the European Society of Radiology, is placed in the context of trends. It describes three general and seven radiology-specific trends and how experimental research in radiology can play a role. “Today’s research is tomorrow’s practice”¹⁰

Three general trends:

· Population aging

· Personalized and precision medicine

· Information technology development

· Population aging: Population ageing is a human success story, reflecting the advancement of public health, medicine, and economic and social development, and their contribution to the control of disease, prevention of injury, and reduction in the risk of premature death. Population ageing is one of the four “mega-trends” that characterize the global population of today—

· population growth,

· population ageing,

· urbanization and,

· international migration¹¹

· Personalized and precision medicine: The second general trend includes personalized medicine, an approach which considers the individual characteristics of each subject for disease susceptibility, biology, and prognosis of diseases, and response to treatment. When the goal is to create a new taxonomy of human diseases based on molecular biology, we refer to the so-called precision medicine¹². Personalized and precision medicine does not end with the individual patient or patient’s disease¹².

Radiology is well positioned for mapping cancer heterogeneity in the individual patient and guiding the adaptive therapy, especially if molecular imaging, radiomics, radiogenomics and habitat imaging techniques are used.¹² In the context of personalized and precision medicine, the research on contrast materials and tracers will play a big role¹².

· Information technology development: Radiology was the leading discipline in the medical digital era. Information technology (IT) development is one of the most important trends impacting individual and social life worldwide. From current applications on mobile phones and tablets to applications of artificial intelligence (AI) to medicine, our future will be strongly influenced by IT.

Advanced techniques of quantitative imaging, AI applications to clinical decision support systems, and big data analysis are only a few examples of radiology specific trends coming from IT development¹².

Fig. 2: General and specific trends influencing radiology and radiological research. MDT, multidisciplinary team; IT, information technology; AI, artificial intelligence¹².

2.1. Seven radiology-specific trends:

· Subspecialties versus unity of discipline: The tension between subspecialties and the unity of the discipline is increasing. Radiology subspecialties have a long history, the radiologist embedded in a clinical team seems to be the best model of radiology consultation, with a non-negligible trade-off paid to productivity. To have radiologists sub-specialized in specific fields is an obligatory way to answer clinical needs, to maintain our central role in multidisciplinary teams, and last but not least, to guide both experimental and clinical research¹².

To retain the unity of the discipline is not an old conservative academic viewpoint but a current need are as follows:

1. General previous training can be an advantage in comparison to clinicians who practice imaging in their field¹³.

2. Modern imaging techniques commonly explore the body of the patient also outside the area of interest, moreover, in multidisciplinary cancer teams, radiologists are frequently asked for their opinion about diagnosis¹⁴.

3. The distinction among subspecialties is blurred, important fields are cross-bordered, and the identity of subspecialties is always changing and evolving while imaging techniques migrate from one subspecialty to another;

4. Organizational aspects favour a central radiology department due to the impossibility to have radiologists dedicated exclusively to one subspecialty; additionally, a central radiology department able to manage large and expensive equipment allows a more cost-effective work-flow;

5. Some hybrid systems require the combination of radiology and nuclear medicine expertise, suggesting a unified training program as has already been initiated under the unique “Radiology” denomination in 2014 in The Netherlands¹⁵.

· Patient safety: A patient-centered approach also includes patient safety, in particular radioprotection, Radiologists should acknowledge the efforts made by the industry for a reduction of x-ray exposure, especially for computed tomography (CT)¹⁶.

Radiation protection is a key aspect of maintaining the safety of patients in diagnostic and interventional radiology. The three fundamental principles of radiation protection of patients are

i. justification,

ii. optimisation, and

iii. the application of doses As Low As Reasonably Achievable (ALARA)¹⁷.

· Quantitative imaging: Imaging procedures will provide more and more output, not only images but also clinical data, numbers, indices, the core of the so-called quantitative imaging. Digital images are intrinsically data. In certain cases, data. In certain cases, data can be more important than images. Examples are bone mineral densitometry and trabecular bone score through dual energy x-ray absorption-metry, where reproducibility defines the smallest detectable difference and the time to follow-up. This is crucial for imaging bio-markers to be used for radiomics and radiogenomics, in particular for MRI-derived parameters¹².

· Standardized and structured reporting: Radiologic reporting is evolving towards standardized descriptors and diagnostic categories, in the context of structured reports. The Breast Imaging Reporting and Data Systems (BI-RADS), more than two decades after its first introduction in 1993, has been imitated in many other fields of diagnostic imaging and the practice of radiology will follow this trend to facilitate the information transfer to patients and clinicians, including other radiologists¹².

· Search for a higher level of evidence: We will be increasingly asked to demonstrate that radiology works in favour of patients, not only in terms of diagnostic performance but also at higher levels of the evidence-based medicine hierarchy, which implies impact on treatment, patient outcome, and societal effects¹⁸.

· Increasing relevance of interventional radiology: A major trend is surely the increasing penetrance of interventional radiology, which is a fundamental asset to improve the clinical profile of radiology. For the next generations it is of importance that we continue to lead the way in device and method innovation in interventional radiology. We should also try to build higher levels of evidence in favour of interventional radiology compared with standard methods, competing with other specialists working in the field. Notably, the innovation of devices and methods is an easier task than building high-level evidence, as “most interventional radiologists lack expertise in the relatively challenging advanced methods used in comparative effectiveness and cost-effectiveness research”. This challenge implies efforts in education and mentoring, beginning with training during post-graduation schools¹².

· Technological evolution: Last but not least, we have to consider the continuous technological evolution of existing imaging methods, the introduction of new imaging methods as well as various effects of IT and AI development in the field of medical imaging.

3.0 Technical challenges in AI developments ¹⁹:

Here we focus on the challenges o AI. For each challenge, required definitions are summarized. In each challenge, we mainly focused on the research and development. The challenges are as follows:

3.1. Problem Identification and Formulation¹⁹

Problem identification and formulation are essential processes that should be implemented in AI-based agents. This issue plays an essential role in organizing HLI-based agents designated in a self-organized manner. Implementing the mentioned processes is very difficult in practice because of some issues explained as follows. As was previously mentioned, we know that human knowledge about its environment can be found in different domains, including known parts, semi-known parts, and unknown parts. There is a set of problems that cannot be formulated in a well-defined format for humans, and therefore there is uncertainty as to how we can organize HLI-based agents to face these problems. a method for solving a problem is to map the problem space into a search space. In this method, detecting the problem, which can be single-state, multiple-state, or contingency, and then formulating it in a well-defined manner are the first steps of AI-based problem-solving.

3.2. Energy Consumption¹⁹

It was shown that these models are costly to train and develop from financial and energy consumption perspectives. During the operation of an HLI-based agent, the agent may use a predefined plan for learning multiple models concurrently to support self-awareness. Therefore, high computational power is required to support more abilities of cognition. This means that enough energy should be supplied for executing the HLI-based agent. In order to mitigate this problem, four solutions were given in the literature, as given below:

· Investing in new paradigms with low energy consumption for HLI, such as quantum computing.

· Finding modern mathematical frameworks to find learning models with lower calculations, which leads to lower energy consumption.

· Sharing models to prevent energy consumption. A researcher can share a model with other researchers around the world.

· In an AI-based system, if there is no way to decrease the load of computations of learning processes, energy harvesting techniques can be used to return the wasted energy.

3.3. Data Issues¹⁹

A type of AI-based agent invests in data-driven methods to construct learning models. In these algorithms, data issues cause various problems. Some of these problems are explained below:

· Cost is one of the main issues of data. Major sources of cost are gathering, preparing, and cleaning the data.

· The size of collected data in a wide range of systems such as IoT (Internet of Things) is another data-related challenge. This huge amount of data leads to a new concept, called big data.

· Data incompleteness (or incomplete data) is another challenging problem in machine learning algorithms which leads to inappropriate learning of algorithms and uncertainties during data analysis. This issue should be handled during the pre-processing phase.

· Bias is a human feature that may affect data gathering and labeling. Sometimes, bias is present in historical, cultural, or geographical data. Consequently, bias may lead to biased models which can provide inappropriate analysis. Despite being aware of the existence of bias, avoiding biased models is a challenging task.

3.4. Robustness and Reliability¹⁹

The robustness of an AI-based model refers to the stability of the model performance after abnormal changes in the input data. The cause of this change may be a malicious attacker, environmental noise, or a crash of other components of an AI-based system.

3.5. Security¹⁹

AI-based systems are widely used to design secure methods. It is obvious that every piece of software, including learning systems, may be hacked by malicious users. In designing intelligent systems, the security challenge is a critical issue that has received much attention.

3.6. Privacy¹⁹

Users’ data, including location, personal information, and navigation trajectory, are considered as input for most data-driven machine learning methods. Data will be the most important thing in different types of AI in the next century. This type of learning algorithm trains an algorithm across multiple decentralized computational resources, thus addressing critical issues such as data privacy, data security, and data access rights. During the design of an HLI-based agent, the definition of privacy issues may be somehow complex and complicated, whether humans accept privacy for other intelligent entities or not.

4.0 Autonomous robotic surgery:

Minimally invasive surgery (MIS) refers to surgical techniques that minimise or even eliminate surgical incisions on the patient's body. Laparoscopic MIS has been widely adopted over conventional open surgery in the past few decades due to the reduction in trauma and less blood loss leading to shorter hospital stays. However, laparo- scopic surgery also presents significant challenges for surgeons²⁰. In the last 2 decades, considerable interest has been put into developing robotic surgical systems that assist surgeons during MIS operations aiming to improve surgical ergonomics in a limited space²¹. Several well‐known robot‐assisted surgery (RAS) systems, such as the DaVinci surgical system, Raven, and MiroSurge, employed the master‐slave paradigm that aims to replicate a surgeon's motion or share the control of a task²². Robotic systems controlled by AI are routinely used in assembly lines in industry and in many biomedical laboratoriesn. However, the development and adoption of autonomous robots in medical interventions has been considerably slower. For many decades, robotic surgery has been synonymous with robotically assisted surgery — systems that facilitate surgical procedures and enable motions smoother than those achievable by human hands, but that still require a surgeon for movement control. Such systems are designed to translate the surgeon’s hand movements into the movements of instruments inside the patient⁴.

Fig. 3: Revo-i: It consists of a surgeon control console, a four-arm robotic operation cart, a high-definition vision cart, and reusable endoscopic instruments.

As suturing is one of the most common procedures during surgery, autonomous knot-tying robots have been developed²³. The autonomous system had better consistency of suturing, higher quality of anastomosis (measured by the pressure at which the anastomosis leaked), and a fewer number of mistakes that required removing the needle from the tissue than hand-sewn suturing, laparoscopy and robot-assisted surgery with the da Vinci surgical system²³. With the continuing development of pre-programmed, imageguided and teleoperated surgical robots, more robot-assisted or automated intervention methods are expected to be incorporated in surgical practice⁴.

5.0 Role of AI in last decades:

6.0 Recent applications of AI and future aspect:

The past five years marked a revival of research and applications of Artificial Intelligence in Radiology. A research reveals that before 2018, less than 500 annual manuscripts included the terms “artificial intelligence” and “radiology”. However, in 2018, this rate doubled to ~1000 articles, and in 2019 and 2020, this number reached ~2000. It is clear that the use of AI in radiology is gaining momentum, primarily due to its potential to enhance the field. AI has the ability to increase radiologist efficiency, highlight urgent cases, increase diagnostic confidence, reduce workload, and help inform patient prognosis and treatment strategies. AI is positioned to transform the work of radiologists in three major steps when it comes to image analysis: detection, characterisation and monitoring²⁵.

6.0.1 Detection:

Detection is the process of identifying and bounding a specific subregion, likely containing a lesion or abnormality²⁵. This AI pillar is key to improving workflow efficiency and productivity. It is estimated that in some cases, a trained reader has to analyze an image every 3-4 seconds for 8 hours in order to catch up with all excessive workload²⁶. To achieve better workflow efficiency, radiologists are adopting the usage of AI-based CADe (Computer Aided Detection), utilising a pattern-recognition algorithm. This allows evaluation of larger imaging data volumes in a much faster manner²⁵.

6.0.2 Characterisation:

Characterisation refers to identifying specific qualities of a pathologic finding, such as size, extent, and internal texture. These qualities can be used to classify lesions into different diagnostic categories. Current technologies for characterisation include Computer Aided Diagnosis (CADx) systems that, similar to the CADe systems, are based on predefined discriminative features that lack generalisability. This can limit their utility, as qualitative descriptions are often difficult to quantitatively define and measure. This limitation is magnified by the fact that humans are limited in the number of qualitative features they are able to identify by visual exam alone, leading to lack of standardisation and significant variability among readers²⁵.

6.0.3 Monitoring:

Monitoring stands for the follow up of an identified pathology over time, assessing any occurring changes as response to treatment. Often, small changes in lesions are undetectable by the naked eye, and this is where AI algorithms come into play. AI-based monitoring is assisting radiologists by capturing large volume of discriminative features that would otherwise be missed by the radiologist. This allows the AI-based monitoring system to provide a clear picture of the tumour evolution over time²⁵.

6.1 Future of Artificial Intelligence in Radiology:

Market predictions in 2017 suggested an upcoming boom in innovative AI applications in medical imaging by 2022 to 2027. At the 2019 RSNA meeting, an AI Showcase included 123 vendors, nearly double the number exhibiting in 2018. As of early 2020, AI is beginning to bridge the gap between acquiring data and meaningful interpretation of data.

Applying AI to advanced imaging modalities such as CT and MR is in initial phases but already showing great promise, such as early results on differentiating benign from malignant nodules on chest CT scans, along with neurologic and psychiatric uses. Machine learning is being proposed or evaluated for radiation therapy error prevention, treatment planning, and automatic organ segmentation. The use of machine learning and artificial neural networks is likely to increase the growth of AI in health care exponentially.

7.0 CONCLUSION:

AI has enhanced clinical diagnosis and decision-making performance in several medical task domains⁴. Artificial intelligence (AI) is the most recent development in a long series of disruptive technological innovations in radiology. Medical imaging began in the late 1800s after the discovery of the X-ray, but exploded in the late 1900s with the availability of computers to create, analyze, and store digital images⁸. Artificial intelligence (AI) is poised to change much about the way we practice radiology in the near future. The power of AI tools has the potential to offer substantial benefit to patients. The main challenge is to anticipate how rapidly-evolving systems may go wrong or could be abused, and to protect against these possible outcomes (ideally before they happen)²⁷. Once we can conceive of potential harms, we can protect against them, and strive towards the undoubted benefits achievable from AI. It is with this in mind that the radiology community has devoted considerable attention to developing ethical

codes for AI use²⁷. A simplistic view of patient safety in radiology is that the key risk relates to inappropriate radiation exposure. While preventing this is a central part of the responsibility of radiographers and radiologists, there is a much wider range of patient safety aspects of the work of radiology professionals. In this paper, we have not attempted to provide a comprehensive list of all safety issues¹⁶. Our focus area is specific trends in radiology ethical consideration, Autonomous Robotic Surgery, recent applications o Artificial Intelligence and future aspect.

Future speculation about radiology includes widespread AI involvement; however, thus far, translation of AI to clinical radiology has been limited.

8.0 REFERENCE:

1. Majumder Sumit, Mondal Tapas, Deen M. Jamal, Review on Wearable Sensors for Remote Health Monitoring. MDPI, 2017; 1-45.

2. U.S. Health Care Costs Rise Faster Than Inflation. Available online: http://www.forbes.com/ sites/mikepatton/2015/06/29/u-s-health-care-costs-rise-faster-than-inflation/#1384765c6ad2 (accessed on 20 June 2016).

3. Deen, M.J. Information and communications technologies for elderly ubiquitous healthcare in a smart home. Pers. Ubiquitous Comput. 2015, 19, 573–599. [CrossRef]

4. Kun-Hsing Yu, Andrew L. Beam and Isaac S. Kohane, Review on Artificial Intelligence Healthcare. Natural Biomedical Engineering., 2018; 2 ;719-731

5. https://itrexgroup.com/blog/artificial-intelligence-in-radiology-use-cases-predictions/#intro

6. Gulshan, V. et al. Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. JAMA 316, 2402–2410 (2016).

7. Szolovits, P. & Pauker, S. G. Categorical and probabilistic reasoning in medical diagnosis. Artif. Intell. 11, 115–144 (1978).

8. Driver C. Noelle, Bowles S. Bradley, Bartholmai J. Brian and Greenberg-Worisek J. Alexandra, Artificial Intelligence in Radiology: A Call for Thoughtful Application, Clinical and Translational Science 2019; XX,1-3

9. Reed, J. C. Chest Radiology: Plain Film Patterns and Differential Diagnoses (Elsevier Health Sciences, Philadelphia, 2010).

10. Dunnick NR (2003) Report of the 2002 Intersociety Commission meeting: Radiology 2002—today’s research is tomorrow’s practice. AJR Am J Roentgenol 180:925–928

11. Schmid Karoline, Lai Mun Sim, Kamiya Yumiko. World Population Ageing. United Nations, Department of Economic and Social Affairs.2019;1-36.

12. Sardanelli Francesco. European Radiology Experimental. Trends in radiology and experimental research. 2017; 1:1; 1-7

13. Bello J (2012) Turf issues in radiology and its subspecialties. Neuroimaging Clin N Am 22:411–419

14. Flor N, Di Leo G, Squarza SA et al (2013) Malignant incidental extracardiac findings on cardiac CT: systematic review and meta-analysis. AJR Am J Roentgenol 201:555–564.

15. Holscher HC (2016) Hybrid imaging: The Dutch approach. Presentation at the ESR Leadership Meeting. Barcelona, 17 Nov 2016

16. European Society of Radiology (ESR) Eurosafe Imaging. Available at: http://www.eurosafeimaging.org/. Accessed on 4 Jan 2017

17. European Society of Radiology and European Federation of Radiographer Societies. Patient Safety in Medical Imaging: a joint paper of the European Society of Radiology (ESR) and the European Federation of Radiographer Societies (EFRS). European Society of Radiology (ESR) Insights into Imaging. 2019;1-17.

18. Sardanelli F, Hunink MG, Gilbert FJ, Di Leo G, Krestin GP (2010) Evidence based radiology: why and how? Eur Radiol 20:1–15.

19. Saghiri Mohammad Ali, Vahidipour Mehdi S., Jabbarpour Reza Mohammad, Mehdi Sookhak and Agostino Forestiero. Review on A Survey of Artificial Intelligence Challenges: Analyzing the Definitions, Relationships, and Evolutions. Applied Science. 2022; 12. 1-21

20. Jinpei Han1, Joseph Davids, Hutan Ashrafian, Ara Darzi1, Daniel S. Elson, Mikael Sodergren. A systematic review of robotic surgery: From supervised paradigms to fully autonomous robotic approaches. International Journal of Medical Robotics and Computer Assisted Surgery. December 2021, 1-11.

21. Ashrafian H, Clancy O, Grover V, Darzi A. The evolution of robotic surgery: surgical and anaesthetic aspects. Br J Anaesth. 2017; 119(suppl_1): i72‐i84.

22. Leonard S, Wu KL, Kim Y, Krieger A, Kim PC. Smart tissue anasto- mosis robot (STAR): a vision‐guided robotics system for laparoscopic suturing. IEEE Trans Biomed Eng. 2014; 61(4): 1305-1317.

23. Yang, R., Shin, E., Newman, M. W. & Ackerman, M. S. When fitness trackers don’t ‘fit’: end-user difficulties in the assessment of personal tracking device accuracy. In Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing 623–634 (ACM, 2015).

24. Shademan, A. et al. Supervised autonomous robotic soft tissue surgery. Sci. Transl. Med. 8, 337ra364 (2016).

25. Mina S. Makary, Carol A. Vitellas. Artificial Intelligence in Radiology: Current Applications and Future Technologies. HealthManagement.org The Journal. 2021; 21(4): 205-208.

26. Ahmed Hosny, Chintan Parmar, John Quackenbush, Lawrence H. Schwartz, and Hugo J. W. L. Aerts. Artificial intelligence in radiology. National centre of Biotechnology Information. 2018; 18(8): 500-510.

27. Adrian P. Brady and Emanuele Neri. Artificial Intelligence in Radiology—Ethical Considerations. Diagnostics. 2020, 10, 231; 1-9.

Received on 06.03.2023 Modified on 18.03.2023

Res. J. Pharmacognosy and Phytochem. 2023; 15(3):264-270.

DOI: 10.52711/0975-4385.2023.00041

3.1. Problem Identification and Formulation19

3.2. Energy Consumption19

3.3. Data Issues19

3.4. Robustness and Reliability19

3.5. Security19

3.6. Privacy19