Artificial Intelligence in Pharmacy: A Review

Quazi Kamil  Hafiz Anees Ahemad,*Dr. Majaz Quazi, Quazi Wasil, Dr. G. J. Khan J.I.I.U’S Ali-Allana College of Pharmacy Akkalkuwa, Dist- Nandurbar (425415) Maharashtra, India.

*Correspondence: quazikamil@gmail.com ; Tel.: (7057517747)

INTRODUCTION

Artificial intelligence (AI) refers to any task carried out by a computer or programme that would otherwise require a human to use intellect to complete. Making machines that exhibit intelligence, particularly those that can recognise speech, make decisions, and translate between different languages, is a task of science and engineering. Artificial intelligence (AI) is the simulation of human intelligence by technology, particularly computer systems. Learning, planning, self-correction, problem-solving, knowledge representation, motion, manipulation, and creativity are all included in this. It is a science and a collection of computational methods that draw inspiration from how humans use their bodies and neural systems to feel, remember, reason. .

In reality, AI has already had an impact on our way of life, either directly or indirectly, and is helping to shape the world of tomorrow. Despite the essential uses of digital assistants for mobile phones, driver assistance systems, bots, text and speech translators, and systems that help with product and service recommendations and personalised learning, AI has already become an integral part of our daily lives and has significantly impacted our way of life.(1)

History of Artificial intelligence (AI)

A formal language for logical reasoning was initially introduced by George Boole in 1847. Alan M. Turing's description of the Turing-machine in 1936 marked the next significant turning point in the history of artificial intelligence. The artificial neuron model was developed in 1943 by Warren McCulloch and Walter Pitts, and the theory of decision, which gave a thorough and formal framework for describing agents' preferences, was developed in 1944 by J. Neumann and O. Morgenstern.

The first neural computer was developed by Marvin Minsky and Dean Edmonds in 1951, but Donald Hebb initially proposed a value-changing rule for the connections of artificial neurons in 1949. When John McCarthy coined the term in the summer of 1956, artificial intelligence (AI) was officially born. It was the first time the topic attracted scholars' interest, and it was covered during a symposium held at Dartmouth.

The first generic problem solver was evaluated the following year, and McCarty—regarded as the father of AI—announced the LISP programming language for developing AI software one year later. List processing language, or Lisp, is still widely used today. In 1965, Herbert Simon predicted that "Machines will be able to do any work a man can do within twenty years." Years later, scientists discovered that it is virtually hard to create an algorithm that can perform any task that a person can.

Nowadays, AI has a new definition: developing intelligent agents to make our work more efficient (Russel & Norvig, 2005; McDaniel, 1994; Shirai & Tsujii, 1982; Mitchell, 1996; Schreiber, 1999). Marvin Minsky and Seymour Paper released Perceptron in 1968 as an example of the limitations of basic neural networks. Washington, DC hosted the inaugural International Joint Conference on Artificial Intelligence in 1970. Alain Colmerauer invented PROLOG, a new language for developing AI systems, in 1972. Johnson Laird, Paul Rosenbloom, and Allen Newell finished their SOAR dissertations for CMU in 1983. (2)

APPLICATIONS OF AI

Artificial intelligence (AI), which refers to the use of computer systems to carry out operations that ordinarily need human intellect , is pervasive in 21st-century culture and is widely used in the medical field. According to some estimates, public and private sector investments in artificial intelligence (AI) in healthcare will total $6.6 billion by 2021, with applications for AI in clinical settings spanning from the automation of differential diagnosis to the management of health systems. This issue of AJHP contains a relevant and essential ASHP Statement on the Use of Artificial Intelligence in Pharmacy due to the fast adoption of AI in healthcare and the possibility for its broader application in the medication-use process. In order to employ AI in patient care and optimise drug use, this statement anticipates the roles of pharmacy departments, pharmacists, and pharmacy technicians.

By the Declaration on the Use of Artificial Intelligence in Pharmacy, ASHP hopes to have an impact on AI research and development to ensure that patients gain from its usage while reducing errors and unexpected consequences. The ASHP statement covers a number of crucial areas that call for pharmacist involvement, such as identifying the steps in the medication-use process that would benefit from the use of AI, deciding which AI techniques—supervised, unsupervised, or reinforcement learning—are best suited for the tasks at hand, and evaluating AI, including the creation of clinical validation standards. The statement also urges health organisations to create roadmaps for AI integration as they do strategic planning. In order to guarantee that AI is properly included into the medication-use process, pharmacists should also play a leadership role in these planning activities. An ASHP statement of this kind will undoubtedly need to be updated frequently to reflect the development of AI in healthcare given the continuously changing nature of technology in general and AI in particular. (3)

Artificial intelligence (AI) has become more prevalent in society's numerous spheres, particularly the pharmaceutical sector. the application of AI to a variety of pharmaceutical industry fields, such as drug discovery and development, drug repurposing, increasing pharmaceutical productivity, and clinical trials, among others; this usage of AI lessens the workload of human workers while also attaining goals quickly. The networks that make up the fundamental architecture of AI systems have formed the basis for the development of several technologies. The International Business Machine (IBM) Watson supercomputer is one such AI-based tool (IBM, New York, USA). It was created to help with the study of a patient's medical data and its association with a sizable database, ultimately leading to the suggestion of cancer treatment options. The quick diagnosis of illnesses is another application for this technique.

AI in the lifecycle of pharmaceutical products: Since that AI can assist with logical medication design, it is possible to envision its involvement in the creation of pharmaceutical products from the laboratory to the patient's bedside. establish the best course of treatment for a patient, including tailored medications; maintain the clinical data obtained and apply it for future drug development; and assist in decision-making

Artificial intelligence in Drug Discovery

The broad chemical space, which contains more than 1060 molecules, encourages the creation of numerous pharmacological compounds. Yet, the medication development process is constrained by a lack of cutting-edge technologies, making it a time-consuming and expensive endeavour that can be resolved by applying AI. AI has the ability to identify hit and lead compounds, as well as expedite therapeutic target validation and structure design optimization. Figure illustrates many ways that AI is being used in the drug discovery process.

.

Figure 1. AI in drug discovery

AI in advancing pharmaceutical product development: The subsequent inclusion of a novel therapeutic molecule into an appropriate dosage form with the requisite delivery properties is necessary. The traditional method of trial and error can be replaced in this area by AI. Problems with stability, dissolution, porosity, and other formulation design-related concerns can be solved using a variety of computational techniques.

AI in pharmaceutical manufacturing

With the increasing complexities of manufacturing processes together with increased need for efficiency and better product quality, current manufacturing systems are aiming to confer human knowledge to machines, continuously changing the production practise. The pharmaceutical business may benefit from the application of AI in manufacturing. Using the automation of many pharmaceutical activities, tools like CFD use Reynolds-Averaged Navier-Stokes solvers technology to examine the effects of agitation and stress levels in various pieces of equipment (such stirred tanks). Comparable systems, including big eddy simulations and direct numerical simulations, use sophisticated techniques to address challenging flow problems in the industrial industry. (4)

Robots are used in various medical procedures because they are more reliable for doctors, more advanced in their work, and capable of completing any task more quickly and effectively than humans. AI plays a significant role in many areas of pharmacy, including drug discovery, drug delivery formulation development, marketing, management, and marketing. (5)

Artificial Intelligence in Diagnosis and Healthcare

From the late 1980s, artificial intelligence (AI) has advanced quickly. The performance of healthcare datasets has increased, and during the past two decades, articles on AI have advanced exponentially. Nonetheless, the availability of AI devices increased with the arrival of more powerful computing. The two major components of AI are machine learning, which analyses structured data (such as pictures, EP, and genetic data), and natural language processing, which analyses unstructured data. Throughout the past 20 years, both AI devices have undergone extensive method, algorithm, and application improvement.

However, numerous attempts and novel AI techniques have been applied in recent years, and a select number of diseases, including congenital cataract disease, cancer, nervous system disease, cardiovascular disease, and liver disease, were perhaps assessed using AI. Deep learning, a modern technology, has sparked an AI boom, and in the near future, significant changes to diagnostic medical imaging systems including endoscopic diagnosis, pathology, and dermatology are expected. In this article, the authors provide a fundamental technical understanding of prominent techniques, algorithms, and applications that have recently developed in the field of medical diagnosis. (6)

Healthcare

Artificial intelligence (AI) and computer vision developments have the potential to significantly improve healthcare, especially in diagnostic fields like pathology and radiology. There is a lot of discussion about how new technologies will affect physician stakeholders. However, little is known about the viewpoints, interests, and worries of the pathology community as a whole.

Here, we provide the findings from a survey of 487 pathologists who work in 54 different countries and were asked to share their thoughts on the use of AI in clinical practise. Notwithstanding these drawbacks, which included challenges assessing answer bias and confirming the identities of survey participants, numerous intriguing discoveries were made. On general, respondents had good attitudes regarding AI, with approximately 75% expressing enthusiasm or interest in using AI as a diagnostic tool to boost pathology workflow efficiency and quality assurance. It's interesting that a sizable proportion of respondents, even among the more upbeat cohort, expressed worries about AI, particularly the possibility of job replacement and displacement. In total, around 80% of respondents anticipated that within the next ten years, AI technology would be implemented in pathology labs. Using Kolmogorov-Smirnov (KS) testing, attempts to uncover statistically significant demographic traits (such as age, sex, type/location of practise) predictive of views towards AI identified various relationships. Prior to the widespread application of AI in pathology, respondents made significant comments on the necessity to intensify efforts towards physician training and to address medical-legal issues.(7)

Artificial Intelligence in Dosage Form Development

A promising strategy for speeding up the development of pharmaceutical products is artificial intelligence (AI)-based formulation development. AI is a flexible tool with several algorithms that can be used in many situations. Among the most popular ways of administration are solid dosage forms, which include tablets, capsules, powder, granules, etc. Critical material attributes (CMAs) and processing parameters can influence product qualities during the product development process, including dissolving rates, physical and chemical stabilities, particle size distribution, and the dry powder's aerosol performance

Nevertheless, the traditional method of developing products by trial and error is ineffective, tiresome, and time-consuming. Nowadays, AI has been acknowledged as a cutting-edge technique for developing pharmaceutical formulations that is attracting a lot of interest. These are some key observations offered by this review: (1) a general overview of artificial intelligence (AI) in pharmaceutical sciences and key regulatory agency directives, (2) methods for creating a database for solid dosage formulations, (3) knowledge on data preparation and processing, (4) a brief introduction and comparison of AI algorithms, and (5) details on applications and case studies of AI as applied to solid dosage forms. We’ll also talk about the potent method known as deep learning-based image analytics and its uses in the field of medicine. Using new AI technology, researchers.

The pharmaceutical sector has used AI-based drug development extensively, and it is seen as a powerful alternative to the traditional process. Chemistry, material science, chemical engineering, computer science, computer vision, and machine learning are just a few of the disciplines that are combined in the AI method. Pharma 4.0 is a framework for utilizing cutting-edge digital technologies to address some persistent challenges in the pharmaceutical industry.

Tablets:

 Due to their patient compliance and ease of use, solid dosage forms such tablets, powders, and granules continue to be the most popular drug product form. Most likely, these medications will keep controlling the pharmaceutical business in the coming years. In the 1990s, scientists and researchers started looking into solid dosage forms of AI applications. The number of publications on AI in solid dosage forms has increased annually by 100% since 2015,  Tablets are the solid dosage form that receives the greatest attention overall and account for more than 60% of all solid dosage form development linked to AI . We compiled the most recent AI applications to present a comprehensive overview to help people better understand how AI algorithms can be used to various solid dosage formulations. This section will cover a number of uses of AI in tablet formulations, including as forecasting drug release, optimising crucial production processes, and spotting tablet flaws.

Predicting Drug Release

In vitro and in vivo drug release studies are two of the most important pre-clinical tests carried out throughout product development. Critical material properties and essential processing factors have an impact on medication release characteristics. For instance, even little adjustments to the compaction parameters, such pressure and tablet geometry, or other  elements, like drug loading, may have a big impact on dissolving rates.

In addition, specialised tools such apparatuses, UV-visible spectrophotometers, and USP-approved vessels are needed for a traditional in vitro drug release investigation. These traditional studies require a lengthy analysis process in its entirety. using the aid of artificial intelligence.

Developing 3D Printed Tablets Using AI

Figure 2. 3D Printer

One of the most cutting-edge methods for personalised medicine is three-dimensional (3D) printing, which has the potential to create tablets taking into account physiology. genetic profiles and patient medication responses Various techniques, including fused filament fabrication, binder jetting, selective laser sintering, pressure-assisted microsyringe, and stereolithography, have been utilised to create customised 3D-printed tablets. The quality of the finished products is tightly regulated during the production process by factors like nozzle temperature, platform temperature, and printing speed. These factors may also have an impact on the drug's in vitro and in vivo release profiles Consequently, AI technologies show significant promise to be implemented into this technique and discover the best parameters to enhance the 3D printing process and reduce the experimental effort with various variables.

Powder:

            One of the most common and traditional pharmacological dose forms is powder. They are made up of a dry substance with tiny, separated particles. Many other dosage forms, such as capsules and tablets, are built upon powders. The process control of powder engineering for small molecules and biologics has effectively used AI technologies. Moreover, research has shown that carrier-based dry powder inhalation applications for AI have a lot of potential.

Capsules :

Drugs are contained in capsules, which have gelatin or other materials as the outer shell. Another popular solid dose form, particularly for oral delivery, is the capsule. There is, however, a dearth of literature detailing the use of AI techniques in creating capsule-based formulations. Many types of capsules, including hard gelatin, soft gelatin, modified release, and enteric capsules, have been employed to encapsulate the drug powders in order to get various drug release profiles. Using an improved CNN, Zhou et al. showed that it is possible to spot capsule faults . In this investigation, manually manufactured capsules with various flaws, such as perforations, concave heads, uncut bodies, oil stains, and shrivelled, locked, or nested capsules, were used. L2 regularisation was included in the upgraded CNN.(8)

Artificial Intelligence In Pharmacovigilance

The use of "artificial intelligence" (AI) approaches to medication development and lifecycle drug management, including pharmacovigilance (PV), is generating a lot of enthusiasm. PV is referred as as "all scientific and data collection activities linked to the detection, assessment, and understanding of adverse events" by the US FDA. The FDA's definition of PV is broad and covers a variety of scientific inquiry methods, including registries, clinical pharmacology studies, individual case safety reports (ICSRs), pharmacoepidemiologic investigations, and others. Despite the fact that the FDA is investigating the application of AI in many of these fields, the research in these fields is still insufficiently developed from a regulatory standpoint to consider widespread implementation. Here, we concentrate on the use of AI to analyse data from various sources to identify adverse events (AEs) that must be reported in accordance with regulatory standards, prepare these AEs as ICSRs, and conduct further reporting and evaluation on these AEs.

Researchers outline the ICSR-related procedures and workflows that various AI techniques might be most successfully used in. Following the description of a framework for assessing AI readiness for ICSR processing and evaluation, he had presented examples of the application of AI to ICSR processing and evaluation in industry and the FDA, compare these examples to the readiness framework, and identify unresolved scientific and policy issues that must be resolved before the full potential of AI can be exploited for ICSR processing and evaluation and PV in general. (9)

Artificial intelligence in Pharmacology

Artificial intelligence (AI) is a branch of computer science that simulates aspects of human behaviour, such as intelligent data analysis. Deep and machine learning are integrated with specific algorithms used by artificial intelligence. Living in the digital age can result in the daily production of vast amounts of medical data. The data used in healthcare systems may be learned, comprehended, and analysed using machine learning. The use of artificial intelligence (AI) in pharmaceutical science, particularly pharmacological research, has grown in recent years. Data from preclinical (laboratory animal) and clinical (human) trials can be analysed with its assistance. Personalized treatment, clinical trial research, radiation, surgical robotics, intelligent electronic health records, and the prediction of epidemic outbreaks are just a few of the processes in which AI plays a significant role. The pharmaceutical sciences are greatly impacted by AI. When compared to conventional technologies, the use of diverse analytical techniques including magnetic resonance imaging (MRI), X-ray, electrocardiogram (ECG), and histopathological imaging yields more accurate results with the use of variable sensor and data acquisition systems. By analysing input and stored data, it also enhances healthcare-based data by detecting the retinal and cutaneous response

AI offers medical services related to animal research, such as animal behaviour, mobility, physiological and pathological changes, and the choice of appropriate medications under particular conditions.(10)

Future Aspects

Without a doubt, artificial intelligence (AI) is reshaping the pharmaceutical sector. precision of the diagnosis. So what does the pharmaceutical industry's AI future hold?

The creation of new medications is one of the most intriguing uses of AI in the pharmaceutical sector. Researchers can test prospective drug candidates more quickly and discover new therapeutic targets with the use of AI. AI can also be utilised to create medicinal compounds with increased potency.

AI can also be used to customise medical care. The AI can choose the ideal medication for a patient by looking into their genome and personal medical history. Moreover, AI can forecast a patients. (11) The electronic footprint of a patient will be mined by AI for key information. At initially this will save time and boost efficiency.

However after sufficient testing, it will also directly influence patient care. Consider a type 2 diabetic patient who is having a consultation. At the moment, a doctor must spend a lot of time reading outpatient letters, examining blood work, and getting clinical guidelines from many disjointed systems. Contrarily, AI could prepare the most significant dangers and actions automatically given the patient's clinical record. Additionally, it might automatically create a summary letter for the doctor to review and approve from the consultation's recorded chat. Both of these applications would save considerable time and could be adopted relatively quickly because they assist clinicians rather than replace them.

AI could proactively suggest consultations when it determines that the patient’s risk of developing a particular diabetic complication warrants intervention. In contrast, it would be impractical to task a human being with the responsibility of closely monitoring every test result and appointment of every diabetic patient in a practice in real time.

AI-based healthcare systems will also integrate advanced diagnostic knowledge into primary care. Images might be taken at a GP office and submitted to a specialised dermatological AI system for immediate analysis if an image of a skin lesion is sufficient to accurately detect its cause. Individuals designated as low risk would receive rapid reassurance while high-risk patient’s would experience decreased referral waiting times because clinics would only be receiving chosen cases. This idea is not just applicable to skin lesions; AI has showed promise in analysing a wide range of imaging data, including retinal scans, radiography, and ultrasound, among others.  A lot of these pictures can be taken using reasonably priced, easily accessible equipment.

Future AI research should focus on a small number of carefully chosen tasks that broadly match the trends mentioned in this article. Building a relationship where AI may benefit clinicians by increasing their efficiency or cost-effectiveness and doctors can provide AI with the crucial clinical exposure it needs to master complex clinical case management is necessary for integrating these technologies into clinical practise. The public's reluctance to adopt an increasingly divisive technology will be the biggest barrier to AI's mainstream adoption, thus it will be crucial to ensure that AI does not hide the human face of medicine throughout the process. (12)

CONCLUSION

The conclusion is that AI is a rapidly developing field in every industry, including pharmacy, and that it requires greater advancement for both updating the status and for new research. As medication-use domain specialists, pharmacists play a crucial role in developing and evaluating AI in healthcare. To collaborate with data scientists and evaluate AI's role in patient care critically, as clinical practise continues to advance, it is essential to have a solid grasp of its fundamental ideas.  Artificial Intelligence requires a large amount of healthcare data to train and learn from in order to provide more accurate clinical decision and increased treatment efficiency. There are different types of Artificial. Intelligence techniques that are applied to analyse structured and unstructured data from healthcare datastores. These techniques provide a more accurate and efficient diagnosis for a patient, and the faster and more targeted the diagnosis is, the sooner a patient can recover. AI has been applied in many areas in the medical field, including managing healthcare records and data, drug creation, treatment design etc.

ACKNOWLEDGEMENT

We are thankful to the Principal and Management of JIIU’s Ali-Allana College of Pharmacy  Akkalkuwa,Dist- Nandurbar for providing moral support and necessary facilities during completion of this work.

REFERENCES:

1)    Harkut DG, Kasat K. Introductory chapter: artificial intelligence-challenges and applications. Artificial Intelligence-Scope and Limitations. 2019 Mar 19.

2)    Benko A, Lányi CS. History of artificial intelligence. InEncyclopedia of Information Science and Technology, Second Edition 2009 (pp. 1759-1762). IGI Global.

3)    Cobaugh DJ, Thompson KK. Embracing the role of artificial intelligence in the medication-use process. American Journal of Health-System Pharmacy. 2020 Nov 16;77(23):1915-6.

4)    Paul D, Sanap G, Shenoy S, Kalyane D, Kalia K, Tekade RK. Artificial intelligence in drug discovery and development. Drug discovery today. 2021 Jan;26(1):80.

5)    Patel SS, Shah SA. Artificial intelligence: Comprehensive overview and its pharma application. Asian Journal of Pharmacy and Technology. 2022;12(4):337-48.

6)    Kamdar JH, Jeba Praba J, Georrge JJ. Artificial intelligence in medical diagnosis: methods, algorithms and applications. Machine Learning with Health Care Perspective: Machine Learning and Healthcare. 2020:27-37.