Artificial Intelligence in Pharmacy: A Comprehensive Review of Applications, Drug discovery and development, Challenges

Article classification of World Journal of Pharmaceutical Research The industry that artificial intelligence is transforming is pharmacy (1). In recent days, AI has made its way in transforming the way drugs are discovered, developed, and dispensed, with the promise of a significant level of precision, efficiency, and patient care (2). An increasing adoption of AI technologies is being integrated into pharmacy practices from drug development to personalized medications, which includes machine learning, NLP, and data analytics (3). By predicting how different compounds may behave in the human body, AI has the potential to make the drug discovery process faster (4). AI systems utilize vast data such as genomics, medical literature, and clinical trials to obtain insights that would manually be impossible to be seen by human beings (5). AI-based algorithms are already being used in pharmacies in areas such as inventory management and supply chain optimization, while also improving patient safety through the evaluation of potential drug interactions or adverse drug reactions (6).  Modern healthcare has learnt AI as the tool of the trade. AI can assess patient data at the individual level so medications can be tailored according to genetics, environment, and lifestyle, thereby getting perfect treatment (7). Besides, AI can automate several routine tasks in a pharmacy, like checking and dispensing prescriptions, allowing more time for patient consultations and clinical care. Overall, applications of AI are multifaceted in pharmacy-bettering drug development, personalized therapy, and operational efficiencies-all directing towards improved the patient outcomes (8).

 METHODOLOGY -

The systematic approach of finding, curating, and finally extracting current literature, which as regards AI applications has been on-the-job training in the pharmacy practice is depicted in Figure 1. In this manner, advancement, challenges, and new trends in such a field are reviewed adequately and up to date (9).

Fig 1- Methodology Stages.

To embrace the most recent advancements, challenges, and applications of AI-related drug discovery and pharmacy, has been well investigated. AI-driven literature survey sources were examined thoroughly, such as PubMed, Scopus, Web of Science, and Semantic Scholar also chosen standpoints or sources, such as arXiv.net and ResearchGate, were used in the literature search initiated by modern search engines (10). AI search algorithms rolled out for the selection of the most up-to-date literature available in all categories of peer-reviewed journal publications and preprint servers (11).

2.Multidisciplinary Databases -

The whole data scanning process was quite systematic with information gathering from different multidisciplinary databases pertinent to medical sciences, pharmaceutical sciences, and computer sciences (12). This step assured that divergent thoughts have been applied for AI applications in pharmacy The search strings used consisted of keywords and Boolean operators alike AI in drug discovery'', machine learning in pharmacy'', challenges in AI for pharmaceuticals'', and modern AI applications. These challenge data are systematically scoped out to extract information relevant to applications, methodology, and challenges.
Machine learning in pharmacy is synonymous to artificial intelligence in pharmaceuticals, concerns that AI poses to the drug industry, and current applications of AI.  Key information was systematically extracted regarding the applications, techniques, and challenges associated (13).

3.Data visualization and synthesis in modern times—

data are collected, synthesized, and visualized into next-generation AI-driven visualization platforms. In a similar way, this graph representation of information and knowledge about specific trends indicates their particular entry points for a variety of information (14).

3.AI in Pharmacy -

Pharmaceutical care is an irreplaceable and important function of health care. It provides for the safe and effective use of medications to affect positive patient health outcomes (15). The application of AI in pharmacy will open numerous avenues in enhancing all aspects of pharmaceutical practice, starting from drug design all the (16). AI has stood up for the "way to medication management and patient care". It has made a potent entry into drug design and development .  AI Algorithms study huge biomimetic datasets such as alphanumeric data or genetic information, molecular structure data, clinical trial data, to discover possible drug candidates and predict their efficacy and safety profiles . All this contributes to the faster and less expensive discovery of drugs with maximum success in their further development (17). Pharmacists will get enriched assistance from AI in medicine management by customizing the medication therapies for specific patients (18). By considering patient data, such as medical history, genetic data, and drug interactions, algorithms would involve personalized medication dosing, adverse drug reaction identification, and prevention of medication errors (19). AI systems can also offer monitoring of drug adherence and interventions/reminders for improved compliance (20). Another impact that AI has in the pharmacy field is automating operations and offering improved workflow strategies. Robotic systems can expertly count and dispense edications, thus keeping accidental human errors low (21). In respect of inventory management, forecasting drug needs, stock level assessment, and waste minimizations, the smart systems assist in order packing, alerting on possible drug interactions, and verification of medication order entry for accuracy. This AI development will also help to conduct later consultations of the patient as well as in educating the patients on-site in the clinics (22). Virtual assistants or chatbots would provide patients with evidence-based information on medicines in answering their questions. The trained professionals will offer instructions on how to use the medication, explain its probable side effects, and detail possible interactions with other medicines or food, thus making patients fully empowered in their decisions related to their health (23). But on the other hand, the infusion of AI into pharmacy raises issues and concerns- the privacies and security of data, regulatory compliance, ethical application of AI, and so on. These are eminent considerations that need to be addressed. Another is the balance that must be maintained between AI-dominated automation and the indispensable human side of patient care delivered by the pharmacies (24).

Classification of AI in pharmacy -

When we talk about AI in pharmacy, it can broadly be divided into the various types of combinations according to applications and functions (25).

Here are some prime classifications:

1.Drug Discovery and Development AI:

These are improvised systems based on artificial intelligence, which used to analyze          biological data, predict molecular property, optimize drug design, and simulate drug interaction, all for the sake of accelerating drug discovery and development (26).

2. Clinical Decision Support AI:

In this regard, AI systems are helping healthcare providers to such an extent that they are aiding clinical judgments with respect to patient data, medical images, diagnostic tests, and medical literature, thereby helping to improve the accuracy of diagnosis and treatment planning (27).

3. Healthcare Operations AI:

AI applications of this category optimize processes in the health care continuum such as patient scheduling, resource allocation, stock keeping, and predictive maintenance within the health facilities (28).

4. Pharmacovigilance and Drug Safety AI:

The artificial intelligence systems in the field of pharmacovigilance conduct real-world data analysis for adverse drug reaction monitoring, safety signal detection, and participating in regulatory compliance through assessing potential risks of pharmaceutical products (29).

5. Personalized Medicine AI:

AI Applications in Personalized Medicine - AI is increasingly being used in patient data, genetic information, and biomarker and treatment response data for the development and optimization of personalized treatment plans, drug dosages, and predictions of patient outcomes. The above classifications illustrate a few sample applications that AI caters for within the pharmaceutical space. Some examples of such applications are drug discovery and development, personalized patient care optimization, or healthcare operations optimization (30).

Efficacy of AI in Drug Discovery and Development -

Artificial Intelligence can effectively change the scenario of pharmaceutical discovery and development, which shows numerous applications. And one of the very recently popular applications and introductions are machine learning, deep learning, and AI methods for drug discovery and drug development (31). The pharmaceutical industrial applications include target identification for drug candidates, optimization of clinical trial designs, prediction of drug interactions, and personalization of treatment regimens (32).

1.Drug target Identification and Validation –

The AI processes tons of biological data for identifying novel drug targets (meaning proteins or genes under suspicion in diseases). Classical methods of target identification have been slow and expensive, but AI increases the speed by mining biological data from diverse sources, including genomic, proteomic, and transcriptomic databases. Machine learning algorithms will identify hidden patterns in such complex datasets and then utilize their predictions about which targets will be the most promising (33).
Example: A study published in Nature Biotechnology demonstrated that AI models could predict potential therapeutic targets in cancer cells by analyzing large-scale genomic data (Joulin et al., 2017) (34).

Fig-2 Artificial intelligence in drug discovery and development

2.  Drug screening and lead compound identification

AI algorithms serve to expedite the high-throughput screening (HTS) of chemical compounds to identify interesting drug candidates. Machine-learning models are trained on existing databases of molecular properties and known biological activities for predicting the probability that a compound will exhibit some desired biological effect. Hence this much cheaper and faster than the standard HTS (35).
Example: An example involves the company Atomwise, which uses AI in drug discovery to identify possible inhibitors for Ebola and other diseases by deep learning. Their AI platform analyses the molecular structure of compounds and predicts the efficiency of these compounds in binding to specific targets (36).

3. Predicting drug-drug interactions (DDIS) –

AI models are capable of predicting potential drug-drug interactions (DDIs) through the evaluation of a multitude of chemical, biological, and clinical data. DDIs are a big concern in drug development with adverse effects or reduced efficacy being causation factors. Thus, AI tools may assist in mitigating the potentials of safety problems coming in clinical trials through the early recognition of possible safety interactions during the medication development process(37).

 Example: IBM Watson for Drug Discovery has been utilized to predict and identify DDIs on the basis of available literature and databases to provide an insight into possible interactions of different types of drugs with one another in a body of a patient(38).

4. Optimizing clinical trials –

The fact is that AI occupies the most important position in clinical trial design and management. Using Machine learning can optimize the selection of patients in determining most of the relevant characteristics in genetics, demography, and clinical history (39).

Example: DeepMind, a subsidiary of Alphabet (Google), used AI to optimize the design of clinical trials for Alzheimer's disease. Their AI system helped in identifying patient cohorts that were more likely to respond to certain treatments (40).

Challenges Of AI In Drug Discovery –

Fig -3 Challenges of AI In Drug Discovery

AI has various applications in hospital health care systems for the organization of dosage forms meant for an individual patient with the selection of appropriate or available administration routes or treatment policies (41).

1.Maintenance of medical record:
Organizing, storing, normalizing, and tracing data would be a very complicated task. Implementation of AI systems would make it easier. The management of patients' medical records does indeed represent a real challenge(42). AI systems could be utilized for organizing, storing, normalizing, and tracing records. This Google project-in-dirt subpoenas the criteria for an important medical record-excavate within a short span of time. It's pretty much faster and better to make it a well-improved health care project. This project is supported by Moor Fields Eye Hospital NHS, improving the treatment of the eye(43).

2.Treatment plan designing:

Effective AI technologies help to design treatment plans. In a situation where a patient is in critical need of treatment and the appropriate treatment plan cannot be decided, then it becomes imperative to control this situation with AI. Under this techno-centric advice, it will formulate appropriately all treatment plans based on all past history, reports, clinical judgment, etc aids healthcare providers in providing optimal treatment options to patients (44). Relates patient information and conditions to thousands of historical cases abstracted from experience gained over thousands of hours working together with the Memorial Sloan Kettering Cancer Center physicians offering both treatment options to help the oncology clinicians meet the patients' needs. Treatments are well-concretized from the literature from Memorial Sloan Kettering, well over 300 medical journals, and around 200 textbooks, getting close to about 15 million pages of text (45).

3.Assisting in repetitive tasks:

Artificial intelligence aids in the detection and diagnosis for small repetitive tasks like screening X-ray imaging, radiology, ECHO, ECG, etc. Medical Sieve (an algorithm introduced by IBM) is a "cognitive assistant" or proficient analyst and reasoner (46). Improvements are to be made to the conditions of patients with deep medical learning. Then there is a computer program specifically meant for that part of the body which is used in particular diseases. All kinds of imaging analysis could become a possibility with deep learning, including X-ray, CT and MRI (47).

4. Medication Support and Assistance:

supporting health and providing aid for medications. Molly is one such virtual nurse created by a start-up; she has a voice and face that are pleasant and friendly. The intention is that patients can be assisted in their treatment (48). Similar treatment guidance through visits to a doctor is also applicable for chronic patients with the same support from Molly. AI Cure is an app that watches patients, using a smartphone camera. so they can better cope with their conditions. It is useful to patients with difficult-to-manage medications and for those enrolled in clinical trial studies (49).

5.Accuracy of medicine:

AI finds its benefits in areas like genomic research and genetic engineering, with visible effects. Deep Genomics, an AI framework, will to trying and pick up the genetic codes by heart, as the genetic codes can tell one's own chosen disease better than a medical record can the disease's name, records to identify mutations and their correlations with diseases (50).This in turn gives relevant information on the happenings in one cell when DNA mutates due to genetic variation. Craig Venter, regarded as the father of the Human Genome Project, designed an algorithm that tells about the visible traits of patients reducible to their DNA (51). AI technology of "Human Longevity" assists in early-stage identification along with the exact location marking for cancer and vascular diseases (52).

6. Drug creation:

 Developing or inventing a drug takes a long time, usually over a decade, and costs several billions of rupees. "Atomwise": gene therapy discovered through Windows of supercomputers to find therapy from a database of molecular structures (53). Cast a virtual screen program on safe and effective therapeutic intervention against Ebola using existing drugs. It found two drugs that were first associated with an Ebola infection. Within less than a day, it completed its analysis, whereas manual analyses would take months or even years (54). A biopharma company in Boston collects big data about managing patients in terms of reserving data to look for triggers that make some patients thrive in their illnesses. They used biological data from patients that incorporated AI technology to analyze differences between a healthy setting and a pathology-friendly environment. This benefits drug discovery and design, part of health care and addressing conditions (55).
7. AI helping people in the Health care system:

Open AI Framework: 2016-the promise of the open AI ecosystem ranks amongst the top ten promising innovation technologies. Good technology compares and collects social awareness algorithms data (56). The medical firms keep mass data about patients-from childhood treatment history to the present, all could be said to be there for the patients, and then the whole data is going to have to be examined and eventually will offer a recommendation on life (57).
8. Analysis of health systems:

Digitalizing these data makes their retrieval easy for medical firms. Ninety-seven percent of Dutch hospital billing data is electronically accessible (58). These bills contain treatment data, physician names, and hospitals' names and are seamlessly retrievable. Companies such as Zorgprisma Publiek use IBM Watson cloud technology in this process.They are prompt in recognizing the incident and providing the right solution. Thus, the situation improves and avoids patient hospitalization (59).

Fig-4 Application of AI in Pharmacy