INTRODUCTION:

A new era of precision medicine innovation and efficiency has begun with the integration of Artificial Intelligence (AI) into drug research procedures. Precision medicine, defined as personalized healthcare suited to individual characteristics, has gained popularity due to its promise to improve treatment outcomes and reduce side effects. By enabling researchers to evaluate sizable datasets, predict drug-target interactions, and develop new therapeutic compounds, artificial intelligence (AI) technologies like machine learning, deep learning, and natural language processing have sped up the drug discovery process.¹

Historical Context and the Medical AI Concept:

AI in medical refers to the use of artificial intelligence methods, algorithms, and technology in healthcare and medicine. It involves evaluating medical data, making judgments, and carrying out duties that are normally completed by human healthcare providers using computer systems and specialized software. By utilizing machine learning, natural language processing, and other AI approaches, artificial intelligence (AI) in medicine aims to improve the precision, effectiveness, and efficiency of medical diagnosis, treatment, and patient care.

Medical image analysis, medication research, personalized treatment planning, disease diagnosis and prediction, virtual health assistants, electronic health record administration, and patient monitoring are just a few of the many uses of AI in medicine. AI systems can assist healthcare professionals in making better decisions, identifying patterns, and forecasting patient outcomes by analyzing vast amounts of patient data and medical literature. This will improve patient care and medical outcomes. Major advancements in both AI and the medical sciences have influenced the development of AI in medicine over the course of a few decades. The establishment of the field of artificial intelligence (AI) in the 1950s by researchers such as Alan Turing, who created the idea of intelligent computers, is one of the major turning points and occasions in the history of AI in medicine². Attempts to mimic human issues are part of early AI research.the ability to solve problems with systems built on rules and rational thought³. The idea of "expert systems" first surfaced in the 1960s, when computer algorithms were designed to incorporate the expertise and knowledge of human specialists to enhance making decisions in specialized field⁴. This opened the door for medical uses of AI. The development of early expert systems, including MYCIN and Dendral, hastened the application of AI in medicine. These developments show how far AI applications in medicine have evolved over the years, opening the door to improved healthcare results and more individualized treatment plans.

Type of AI and their application in medical field:

1. Supervised Learning:

Can be used to classify disorders and identify malignancies.

2. Unsupervised Learning:

This type can be used to evaluate patient data and discover hidden patterns or linkages.

3. Reinforcement Learning:

Has the potential to help optimize treatment regimens or drug dosages.

4. Deep Learning:

It can be used for activities such as image analysis and pathological diagnosis.

5. Natural Language Processing:

It can be used for tasks such as electronic health records analysis or communication^5,6,7.

AI in Discovering New Drugs:

Examples of virtual applications include drug development and discovery, diagnostic support, individualized treatment programs, and virtual health support. AI can aid in more accurate and efficient disease diagnosis. Medical imaging data, including X-rays, CT scans, and MRI images, may be analyzed by AI algorithms to find abnormalities and support the early detection of diseases. AI systems analyze vast databases of chemical structures, biological interactions, and clinical trial data to assist researchers in finding potential medication candidates. This has the ability to produce new medications more quickly and expedites the drug discovery process.⁸

Figure 1: AI's Use in Drug Delivery

AI in drug discovery suggests:

Identification and Validation of the Target:

Drug development has been significantly impacted by artificial intelligence, especially in the areas of target validation and identification. Finding possible biological targets and clarifying their involvement in diseases are the first steps in this procedure. These targets must then be validated to make sure they are directly engaged in a disease mechanism and that altering them is likely to produce therapeutic advantages. Through the analysis of genomic, proteomic, and metabolomic data, it also plays a significant role in determining possible therapeutic targets. Researchers might gain valuable insights for medication development by using machine learning algorithms to sort through large datasets and find the proteins or biological processes implicated in certain diseases⁹. AI aids in target selection as well. Drug able, safe, effective, and commercially viable are the best qualities in a target. However, targets that were previously challenging to treat are now included in evolving disease therapy strategies. Because it verifies that a molecular target is directly involved in a disease process and that altering the target is likely to have a therapeutic effect, target validation is a crucial step in the drug discovery process.¹⁰

Virtual Screening and Drug Design:

“In order to predict how potential therapeutic chemicals would interact with target proteins, AI-powered virtual screening techniques look at their three-dimensional structures. This speeds up the process of developing new drugs and helps scientists find promising treatment concepts for more research¹¹. Virtually drug assessment is a computer method that matches chemical compositions to target molecules using artificial intelligence to forecast how novel drugs will work. Using this technique, scientists can rapidly assess a chemical library's capacity to bind to and block specific receptors or enzyme targets¹². Algorithms using artificial intelligence are able to predict binding affinities, analyze chemical structures, and rank substances for further experimental investigation”.

Prediction of Drug Properties:

Because they have an indirect impact on the drug's pharmacokinetic qualities and target receptor family, physicochemical parameters like as the drug's solubility, partition coefficient (logP), degree of ionization, and intrinsic permeability must be taken into account while developing a new drug^13,14. Numerous AI-powered techniques are available for predicting physicochemical features. For example, massive data sets produced during the prior compound refining are used to train the ML software¹⁵.Drug design algorithms use molecular descriptors including SMILES strings, potential energy measurements, electron density around the molecule, and 3D atom locations to construct viable compounds and apply DNN to forecast their attributes.

Prediction of Bioactivity:

A medicinal compound's affinity for the target protein or receptors determines how effective it is. The therapeutic response cannot be produced by drug molecules that lack interaction or affinity for the targeted protein. Rarely, produced pharmaceutical substances may bind with undesirable proteins or receptors and cause toxicity. Predicting drug-target interactions so depends on drug target binding affinity (DTBA). By considering a drug's characteristics or resemblance to its target, AI-based methods can ascertain the drug's binding affinity. In order to create feature vectors, characteristic-based interactions determine the chemical components of the drug and the target. Conversely, similarity-based interactions assume that similar compounds will interact with the same targets, taking into consideration the drug's and the target's similarities ¹⁶.

Prediction of toxicity

Any medical molecule's toxicity must be predicted in order to prevent negative effects. The cost of drug research is increased by the frequent use of cell-based in vitro assays as exploratory studies, which are followed by animal experiments to ascertain a compound's toxicity. LimTox, pkCSM, admetSAR, and Toxtree are a few web-based solutions that can help reduce the expense¹⁷.

Optimizing drug discovery and drug development using AI

AI solutions are continuously being proposed to solve a number of nanotechnology-related issues. The development of nanosystems, nanocomputing, and artificial intelligence techniques to apply the concepts of nanoscale modeling are crucial fields, with a focus on effective measurement of parameters, anticipating, and system modeling as well as the reduction of processing time¹⁸. “A range of procedures, including as medication formulation, manufacturing methods, storage of bioactive ingredients, and transportation to the designated locations, are necessary for drug delivery to have a more significant therapeutic impact. Therefore, increasing the physiochemical features of the medicine and overcoming the instability of bioactive molecules—which can significantly degrade pharmacokinetic qualities—require modifying drug delivery systems¹⁹.”

Use of ML and AI:

Advances in Computational Approaches:

“AI and Machine Learning The demand for data reliability, as well as the rising complexity of the studies outlined above, has resulted in the use of AI in pathology^32,33,34,35. AI is a broad scientific field that trains machines to extract features or information beyond human visual perception using algorithms^36,37,38. AI methods are made to train a machine classifier for a certain segmentation after first extracting pertinent visual representations.

Either supervised or unsupervised methods can be used to complete a diagnostic or prognostic task. AI's capacity to analyze vast amounts of data quickly could significantly speed up the identification of new histopathological characteristics that could improve our comprehension of or capacity to predict how a patient's sickness will develop and how they will respond to a particular treatment”^39,40.

Digital Pathology:

Whole slide image (WSI) scanners and complex, new imaging systems have been developed and implemented as a result of efforts to overcome some of the challenges related to traditional pathology procedures. This has enabled pathology to transition into the digital age, also referred to as digital pathology. WSI scanners take many pictures of entire tissue sections on the slide in a matter of minutes. These images are then digitally stitched together to create a WSI that a pathologist can view on a computer screen^41,42.

Digital image analysis is used in pathology to quantitatively assess histological features, morphological patterns, and physiologically significant regions of interest, as well as to rapidly and accurately identify and quantify particular cell types^43,44.

Additionally, data from tissue slides that might not be accessible during manual examination using normal microscopy can be captured thanks to statistical image analysis methods^45,46.

AI in Diagnosing Uncommon Diseases:

AI collaborates with humans to combine and analyze a variety of data. Medical professionals’ profit from diagnostic decision support systems because they provide a list of pertinent differential diagnoses. Numerous well-known application scenarios have seen the effective implementation of these systems in the past. It has been used to track the demographic, clinical, and epidemiological traits of patients in order to identify and diagnose coronavirus illness 2019 (COVID-19) in recent years. In order to identify new cases, algorithms have been created and are presently being utilized to gather networks and document data from patients with rare diseases⁴⁷.

Genetic Data-Based Artificial Intelligence Systems:

Numerous AI systems have proven to be effective in analyzing data and making accurate diagnoses in the areas of phenotypic and genomic analysis. Since over 80% of RDs are inherited, artificial intelligence has a lot of promise in this area. A range of RDs have been subjected to various tools. PhenIX employs the Disease-Associated Genome, which identifies Mendelian illnesses by combining genetic information with phenotypic ideas. PhenIX efficiently analyzes the genetic sequence of the patient, identifies variants, and chooses them according to phenotypic similarity and pathogenicity.

AI in cancer Treatment:

A variety of disease-specific data, including examinations, image post-processing, and interpretation for clinical reporting and treatment planning, are necessary for high-quality cancer imaging. In order to evaluate tumor response to treatments, spot issues, and estimate the likelihood of tumor recurrence, the imaging techniques are intended to detect tumor entities and metastatic patterns as well as to gather extra clinical data from different imaging modalities⁴⁸.

In order to better understand the in vivo effectiveness of therapeutic drugs, recent research has concentrated on merging biological imaging and drug administration to visually monitor the drug delivery process in real time. To help with molecular imaging and controlled drug release for optimal therapeutic response, researchers are working to create a multipurpose theranostic probe for targeted drug delivery. Additionally, molecular imaging provides more accurate data for designing candidate medications with ideal specificity for the target and pharmacodynamic effectiveness, accordingly, while traditional approaches struggle to analyze pharmacokinetic characteristics and pharmacodynamic information⁴⁹.

Applying AI to biomarker imaging could enhance cancer detection and therapy because AI systems have proven to be able to decipher large datasets with complex patterns. Additionally, using AI to forecast how NPs will interact with the target medication, the tumor site, and the membranes of cells can yield more details on drug incorporation and releasing kinetics, enabling the development of more effective nanomedicine formulations for the treatment of cancer.

Colorectal Cancer and Artificial Intelligence:

“Overall, there have been some encouraging results in CRC from integrating genetic testing and AI algorithms. 53 colon cancer patients with the assistance of the Union for International Cancer Control (UICC) II were split into two groups using gene expression in a simulation experiment: those who experienced a relapse following surgery and those who did not. Three neural networks were tested by the researchers for classification accuracy: S-Kohonen (91%), Back-propagation (66%), and SVM (70%). They suggested that colon cancer classification is a better fit for the S-Kohonen neural network.

To identify differentially expressed genes (DEGs) for the purpose of categorizing patients at high risk of colon cancer recurrence, a study employing SVM analysis was conducted. Intriguingly, they found a 15-gene signature that could be useful in predicting the prognosis and risk of recurrence for patients with colon cancer.”

Clinicians can now predict the prognoses of CRC patients thanks to AI systems. To train prediction models for assessing disease-free survival, overall survival, radiochemotherapy response, and recurrence, a researcher used gene identifiers⁵⁰.

Advanced AI-Based Applications:

AI-Powered Medicine Delivery Nanorobots:

“New businesses arise more quickly as a result of the rapid advancements in materials science, molecular biology, mechanical dynamics, artificial intelligence, and other sciences. Over time, micro/nanorobots began to push the limits of science. The concept of employing microrobots to treat medical ailments was first proposed by Richard Feynman in 1959. Then came the idea of nanorobots. Micro/nanorobots, a fast-evolving technology, are widely employed in the medical industry for tasks such as drug delivery, medical diagnosis, and auxiliary operation. Micro/nanorobots can move independently, which enables us to deliver controlled nanoparticles to hard-to-reach places, in contrast to conventional drug delivery techniques that depend on blood circulation to reach the target.”

An external shell and an internal payload enable a micro/nanorobot system to actively approach a predetermined target. Either endogenous dynamics—achieved by chemical or biological interactions—or exogenous dynamics—such as magnetic, ultrasonic, and light energy propulsion—are responsible for the propulsion of micro/nanorobots. The systemically circulation-based drug delivery approach now in use lacks the navigational capabilities necessary for precise distribution. Motion-capable micro/nanorobots are an attractive option that can partially satisfy these needs. The main components of nanorobots are sensors, power supply, integrated circuits, and safe backups of data, all of which are powered by computing technologies such as A. They are designed to detect and attach, avoid collisions, identify targets, and then expel from the body. Advances in nano/microrobots enable them to navigate to the desired location according to physiological conditions. For regulated drug and gene delivery, implantable nanorobots must consider factors like sustained release, control release, and dose adjustment. Additionally, drug release must be automated with AI tools like fuzzy logic, NNs, and integrators.

Microchip implants are utilized for both controlled release and locating the implant in the body^51,52,53.

AI and Personalized Medicines:

The practice of customizing medical care to each patient's particular needs is known as personalized medicine. One of the most significant applications of AI in personalized medicine is pharmacogenomics, the study of how genes affect an individual's reaction to drugs. The ideal prescription and dosage can be chosen by using AI algorithms that use a patient's genetic composition to predict how they will react to different treatments.

“For example, AI models have been used to predict patient responses to antidepressants based on genetic variation, allowing clinicians to give patients personalized mental drugs. AI models have been used to predict patient responses to antidepressants based on genetic variation, allowing physicians to prescribe more personalized mental drugs”⁵⁴. For patients with major depressive illness, selecting the right antidepressant is typically challenging and necessitates a trial-and-error approach. Machine learning presents a promising approach to tailoring antidepressant prescriptions. Machine learning presents a promising approach to tailoring antidepressant prescriptions. However, while this is encouraging, the model's clinical application is limited, and models must be developed to account for factors other than effectiveness alone^55,56. AI tailors treatments based on lifestyle, patient preferences, and environmental factors, in addition to genetic data.

AI in Drug Delivery and Formulation:

The pharmaceutical industry has long struggled with the challenges of drug formulation and delivery. Traditional approaches for optimizing formulations and delivery mechanisms frequently involve lengthy and costly trial-and-error operations^57,58. AI-generated predictive models are used to optimize medicine formulations, ensuring that active components reach the target spot in the body as efficiently as possible. For example, AI algorithms can forecast a medication's release profile from a certain formulation, enabling for the development of controlled-release pharmacological formulations with a consistent medicinal impact over time.

AI and Nanotechnology:

Machine learning has a wide range of applications in nanomedicine. Machine learning, for instance, can be used to increase our comprehension of how a nanoparticle's structure affects its properties as well as its interaction with the tissues and cells it targets.

Nanotechnology is evolving quickly and adopting new technical forms like integrated analytical systems and machine learning. In addition to predicting activity in vitro and in vivo, artificial intelligence holds great promise for closing the loop of nanoparticle synthesis, characterisation, refining, and testing. There are still a number of difficulties and roadblocks on the path that started in experimental laboratories⁵⁹.

Prospects & Future Paths for AI-Assisted Drug Discovery:

In the upcoming years, it is anticipated that the worldwide AI-in-drug discovery sector would grow dramatically. Artificial intelligence (AI) advancements such as machine learning, natural language processing, and deep learning have produced complex algorithms that can decipher complex biological data and forecast drug-target interactions.

Large compound libraries are being virtually screened by AI-powered systems, which are more accurate and efficient than conventional techniques at finding potential therapeutic candidates. In order to help choose the most promising candidates for further development, these algorithms can also forecast the pharmacokinetic and toxicological profiles of lead molecules and optimize them. AI is also transforming medication design by making it possible to create novel molecular structures with predetermined characteristics. AI-generated substances can be modified to target particular biological processes and disorders, creating new opportunities for personalized therapy. The market potential for AI in the pharmaceutical sector is anticipated to grow even further as long as regulatory bodies like the FDA continue to embrace AI applications in drug development.

There are still obstacles to overcome, though, like model validation, data privacy issues, and integrating AI into current drug development processes. Furthermore, medication research and development is just one aspect of artificial intelligence's potential. Additionally, it is crucial for clinical trials, patient classification, and real-time therapy reaction monitoring, enabling customized medical interventions that enhance patient outcomes.

Pharmaceutical manufacturing processes will be impacted by AI-driven technology, which will enhance process optimization, quality assurance, and predictive maintenance, among other aspects. By enabling more scalable and effective production processes, this solution will save costs and improve product consistency.

Predictive maintenance algorithms will minimize equipment failures and delays, enabling faster and more adaptable production operations, while AI-powered digital twins will replicate and optimize manufacturing processes in real time^60,61.

CONCLUSION:

AI convergence in drug development is transforming the landscape of pharmaceutical innovation, providing unprecedented prospects to expedite processes, lower costs, and improve therapeutic efficacy. By integrating AI into the discovery and clinical stages, the industry is seeing speedier identification of viable drug candidates, more specific disease targeting, and better patient outcomes. Nevertheless, issues such as data quality, model interpretability, and regulatory alignment must be addressed before AI can reach its full capabilities. Looking ahead, continued breakthroughs in AI technologies, together with increased interdisciplinary collaboration, promised to speed up drug discovery and bring more individualized medicines to market, ushering in a new era of precision medicine.

REFERENCE:

1. Raparthi M, Gayam R.S, Nimmagadda S.P.V, Sahu K.M, Putha S, Pattyam P.S, Kondapaka K.K, Kasaraneni P.B, Thuniki P, Kuna S.S. AI Assisted Drug Discovery: Emphasizing Its Role in Accelerating Precision Medicine Initiatives and Improving Treatment Outcomes. Human-Computer Interaction Perspectives. 2022; 2(2): 1-10.

2. Visan L.A, Negut I. Integrating Artificial Intelligence for Drug Discovery in the Context of Revolutionizing Drug Delivery. Life. 2024; 14(233): 2-36.

3. Kaul V, Enslin S, Gross S.A. History of artificial intelligence in medicine. Gastrointest. Endosc. 2020; 92: 807–812.

4. Wojtara M, Rana E, Rahman T, Khanna P, Singh H. Artificial intelligence in rare disease diagnosis and treatment. Artificial Intelligence in Rare Diseases. 2023; 16: 2106–2111.

5. Huang B, Huang H, Zhang S et al. Artificial intelligence in pancreatic cancer. Theranostics. 2022;12(16):6931-6954.

6. Hurvitz N, Azmanov H, Kesler A, Ilan Y. Establishing a second- generation artificial intelligence- based system for improving diagnosis, treatment, and monitoring of patients with rare dis eases. Eur J Hum Genet. 2021; 29(10): 1485-1490.

7. Faviez C, Chen X, Garcelon N, et al. Diagnosis support systems for rare diseases: a scoping review. Orphanet J Rare Dis. 2020; 15(1): 94.

8. DeepChem. Available online: https://github.com/deepchem/deepchem (accessed on 20 October 2023).

9. Bhatt T. K, Nimesh, S. The Design and Development of Novel Drugs and Vaccines: Principles and Protocols; Academic Press: Cambridge, MA, USA, 2021; ISBN 978-0-12-821475-6.

10. Mathai N, Chen Y, Kirchmair J. Validation strategies for target prediction methods. Brief. Bioinform. 2020;21: 791–802.

11. Rifaioglu A.S, Atas H, Martin M.J, Cetin-Atalay R, Atalay V, Dogan T. Recent applications of deep learning and machine intelligence on in silico drug discovery: Methods, tools and databases. Brief. Bioinform. 2019;20 :1878–1912.

12. Zhang O, Zhang J, Jin J, Zhang X, Hu R, Shen C, Cao H, Du H, Kang Y, Deng Y et al. ResGen is a pocket-aware 3D molecular generation model based on parallel multiscale modelling. Nat. Mach. Intell. 2023; 5:1020–1030.

13. Paulz D, Sanapz D, Shenoyz S, Kalyane D, Kalia K, Tekade K.R. Artificial intelligence in drug discovery and development. Drug Discovery Today. 2021;26(1): 80-92.

14. Zang, Q. et al. In silico prediction of physicochemical properties of environmental chemicals using molecular fingerprints and machine learning. J. Chem. Inf. Model. 2017;57: 36–49.

15. Yang, X. et al. Concepts of artificial intelligence for computer-assisted drug discovery. Chem. Rev. 2019; 119: 10520–10594.

16. Hessler G, Baringhaus K.H. Artificial intelligence in drug design. Molecules. 2018;23: 2520.

17. O ¨ztu ¨rk, H. et al. DeepDTA: deep drug–target binding affinity prediction. Bioinformatics. 2018;34: 821–829.

18. Das P.K, Chandra J. Nanoparticles and convergence of artificial intelligence for targeted drug delivery for cancer therapy: Current progress and challenges; Frontiers in medical technology. 2023;3-14.

19. Sacha G.M, Varona P. Artificial Intelligence in Nanotechnology. Nanotechnology. 2013; 24(45):452002.

20. Priyom Bose. Optimizing Drug Delivery Using AI [Internet]. News Medical.net. (2022) [cited 2022 Oct 5] Available from: https:// www.azolifesciences. Com /article /Optimizing-Drug-Deliv ery-UsingAI.aspx

21. Colombo S. Chapter 4– applications of artificial intelligence in drug delivery and pharmaceutical development. In: A Bohr, K Memarzadeh, editors. Artificial intelligence in healthcare. London, UK: Academic Press (2020). p. 85–116. Available at: https://www.sciencedirect.com/science/article/pii/ B9780128184387000046 (cited October 9, 2022).

22. Aliper A, Plis S, Artemov A, Ulloa A, Mamoshina P, Zhavoronkov A. Deep learning applications for predicting pharmacological properties of drugs and drug repurposing using transcriptomic data. Mol Pharmaceutics. 2016; 13(7):2524–2530.

23. Kadurin A, Nikolenko S, Khrabrov K, Aliper A, Zhavoronkov A. druGAN: an advanced generative adversarial autoencoder model for de Novo generation of new molecules with desired molecular properties in silico. Mol Pharmaceutics.2017 ;14(9):3098–3104.

24. Menden M.P, Iorio F, Garnett M, McDermott U, Benes C.H, Ballester PJ, et al. Machine learning prediction of cancer cell sensitivity to drugs based on genomic and chemical properties. PLoS One.2013 ;8(4): e61318.

25. Pivetta T, Isaia F, Trudu F, Pani A, Manca M, Perra D, et al. Development and validation of a general approach to predict and quantify the synergism of anticancer drugs using experimental design and artificial neural networks. Talanta. 2013; 115:84–93.

26. Preuer K, Lewis RPI, Hochreiter S, Bender A, Bulusu KC, Klambauer G. Deepsynergy: predicting anticancer drug synergy with deep learning. Bioinformatics. 2018 ;34(9):1538–46.

27. Zhu H. Big data and artificial intelligence modeling for drug discovery. Annu. Rev. Pharmacol. Toxicol.2020; 60: 573–589.

28. Ciallella H.L. and Zhu H. Advancing computational toxicology in the big data era by artificial intelligence: data-driven and mechanism-driven modeling for chemical toxicity. Chem. Res. Toxicol. 2019 ;32; 536–547.

29. Chan H.S. et al. Advancing drug discovery via artificial intelligence. Trends Pharmacol. Sci. 2019;40 (8): 592–604.

30. Brown N. Silico Medicinal Chemistry: Computational Methods to Support Drug Design. Royal Society of Chemistry 2015.

31. Pereira J.C. et al. Boosting docking-based virtual screening with deep learning. J. Chem. Inf. Model. 2016 ;56: 2495–2506.

32. Baxi V, Edwards R, Montalto M, Saha S. Digital pathology and artificial intelligence in translational medicine and clinical practice; Springer Nature. 2022;35: 23-32.

33. Aeffner F. et al. Introduction to digital image analysis in whole-slide imaging: a white paper from the digital pathology association. J. Pathol. Inf. 2019;10: 9.

34. Heindl A, Nawaz S, Yuan, Y. Mapping spatial heterogeneity in the tumor microenvironment: a new era for digital pathology. Lab. Invest. 2015 ;95: 377–384.

35. Yuan J. et al. Novel technologies and emerging biomarkers for personalized cancer immunotherapy. J. Immunother. Cancer 2016 ;4: 3.

36. Aeffner F. et al. Introduction to digital image analysis in whole-slide imaging: a white paper from the digital pathology association. J. Pathol. Inf.2019; 10: 9

37. Bera K, Schalper K. A, Rimm D. L, Velcheti V, Madabhushi A. Artificial intelligence in digital pathology– new tools for diagnosis and precision oncology. Nat. Rev. Clin. Oncol.2019; 16: 703–715.

38. Vamathevan J. et al. Applications of machine learning in drug discovery and development. Nat. Rev. Drug Discov. 2019;18: 463–477.

39. Serag A. et al. Translational AI and deep learning in diagnostic pathology. Front. Med.2019; 6: 185.

40. Barsoum I, Tawedrous E, Faragalla H, Yousef G. M. Histo-genomics: digital pathology at the forefront of precision medicine. Diagnosis. 2019;6: 203–212.

41. Pantanowitz L. et al. Validating whole slide imaging for diagnostic purposes in pathology: guideline from the College of American Pathologists Pathology and Laboratory Quality Center. Arch. Pathol. Lab. Med. 2013 ;137: 1710–1722.

42. Zarella, M. D. et al. A practical guide to whole slide imaging: a white paper from the Digital Pathology Association. Arch. Pathol. Lab. Med. 2019;143: 222–234.

43. Bera K, Schalper K. A, Rimm D. L, Velcheti V, Madabhushi A. Artificial intelligence in digital pathology– new tools for diagnosis and precision oncology. Nat. Rev. Clin. Oncol.2019; 16: 703–715.

44. Tumeh P. C et al. Liver metastasis and treatment outcome with anti-PD-1 monoclonal antibody in patients with melanoma and NSCLC. Cancer Immunol. 2017; 5: 417–424.

45. Barisoni L, Lafata K. J, Hewitt S. M, Madabhushi A, Balis U. G. J. Digital pathology and computational image analysis in nephropathology. Nat. Rev. Nephrol. 2020; 16: 669–685.

46. Neltner J. H et al. Digital pathology and image analysis for robust high throughput quantitative assessment of Alzheimer disease neuropathologic changes. J. Neuropathol. Exp. Neurol. 2012 ;71: 1075–1085.

47. Visibelli A, Roncaglia B, Spiga O, Santucci A. The impact of ar tificial intelligence in the odyssey of rare diseases. Biomedicine. 2023;11(3):887.

48. Schlemmer HP, Bittencourt LK, D’Anastasi M, Domingues R, Khong PL, Lockhat Z, et al. Global challenges for cancer imaging. JGO. 2018; 4:1–10.

49. Niu G, Chen X. The role of molecular imaging in drug delivery. Drug Deliv (Lond). 2009 ;3: 109–13.

50. Serrano R. D, Luciano C. F, Anaya J. B, Molina G, Ongoren B, Kara A. et al. Artificial Intelligence (AI) Applications in Drug Discovery and Drug Delivery: Revolutionizing Personalized Medicine. Pharmaceutics. 2024; 16: 1328.

51. Hassanzadeh P. et al. The significance of artificial intelligence in drug delivery system design. Adv. Drug Delivery Rev. 2019;151: 169–190.

52. Luo M et al. Micro-/nanorobots at work in active drug delivery. Adv. Funct. Mater. 2018;28.

53. Fu J, Yan H. Controlled drug release by a nanorobot. Nat. Biotechnol.2012;30: 407–408.

54. Sheu Y. H, Magdamo C, Miller M, Das S, Blacker D, Smoller, J.W. AI-assisted prediction of differential response to antidepressant classes using electronic health records. NPJ Digit. Med. 2023;6: 73.

55. Arnold P.I.M, Janzing J.G.E, Hommersom A. Machine learning for antidepressant treatment selection in depression. Drug Discov. Today 2024:29.

56. Liu X, Read S.J. Development of a multivariate prediction model for antidepressant resistant depression using reward-related predictors. Front. Psychiatry 2024, 15, 1349576.

57. Walsh D, Serrano D.R, Worku Z.A, Madi A.M, O’Connell P, Twamley B et al. Engineering of pharmaceutical cocrystals in an excipient matrix: Spray drying versus hot melt extrusion. Int. J. Pharm. 2018;551: 241–256.

58. Lamy B, Tewes F, Serrano D.R, Lamarche I, Gobin P, Couet W et al. Marchand, S. New aerosol formulation to control ciprofloxacin pulmonary concentration. J. Control. Release 2018; 271: 118–126.

59. Egorov E, Pieters C, Rechtman K. H, Shklover J, Schroeder A. Robotics, microfluidics, nanotechnology and AI in the synthesis and evaluation of liposomes and polymeric drug delivery systems. Drug Delivery and Translational Research. 2021;345-352.

60. Fleming N. How artificial intelligence is changing drug discovery. Nature 2018; 557: S55–S57.

61. DiNuzzo, M. How artificial intelligence enables modelling and simulation of biological networks to accelerate drug discovery. Front. Drug Discovery. 2022;2: 1019706.

Received on 11.01.2025 Revised on 19.02.2025

Accepted on 25.03.2025 Published on 09.05.2025

Available online from May 12, 2025

Res. J. Pharma. Dosage Forms and Tech.2025; 17(2):107-114.

DOI: 10.52711/0975-4377.2025.00016

This work is licensed under a Creative Commons Attribution-Non Commercial-Share Alike 4.0 International License. Creative Commons License.

Current AI Approaches	Focus Area	Advantage
AI Based API System²⁰	Absorption, distribution, metabolism, and excretion (ADME)	Predicting extravasation in tumor tissue. AI techniques may combine data from numerous sources, simplifying better drug delivery methods by experimental and in silico system models.
Passive AI ²¹	Molecular entity features of the drugs	These artificial intelligence models can help incorporate new molecular entity properties, such as frequently known chemicals, to predict efficacious treatment based on the local concentration or bioavailability of a chemical for improved therapeutic results.
Deep Neural Networks	Drug repurposing	At the route level, DNNs are capable of effectively classifying intricate drug action mechanisms. Furthermore, the model might categorize medications as useful, therapeutic, or harmful²².
Generative Adversarial Network (GAN)	De Novo generation of new molecules with suitable molecular properties in silico	Molecular fingerprints with preset anticancer qualities can be created by GANs. Large molecular data sets can also be processed well by this model, which makes it possible to find new chemicals for the treatment of cancer²³.
Artificial Neural Network (ANN	Predicting the synergy of anticancer drugs.	ANN using a back-propagation method was used to forecast and assess anticancer drug synergy. The data helped identify the medication concentrations and cytotoxicity²⁴.
Feed-Forward Multi-Layer Perceptron	Predicting drug sensitivity using cell line screening data to identify dose response.	The approach may predict cancer cell susceptibility to drug molecules using variables derived from cells and medications, which include genetic and chemical information, respectively²⁵.
Deep Synergy (A Deep Learning Feed Forward Neural network)	Predicting anticancer drug synergy.	Predictions of drug interactions based on genetic and chemical information. Accurately predicting the drug combination's synergy and comparing the approaches while differentiating between cancer cell lines are the objectives²⁶.

AI Tools / Platform	Details
DeepChem	MLP model which employs a Python-based AI system to identify a viable candidate in drug research²⁷.
DeepTox	Software that estimates the toxicity of up to 12000 medicines²⁸.
DeepNeuralNetQSAR	Python-based software powered by computational techniques for detecting the chemical activity of substances²⁹.
ORGANIC	A molecular generation tool for creating compounds with certain attributes³⁰.
DeltaVina	A scoring function for reassessing drug-ligand binding affinity.
Chemputer	Helps to report chemical synthesis procedures in a uniform style.
Hit Dexter	Machine learning, CNNs, and artificial neural networks are used to anticipate compounds that may respond to biochemical assays.
GastroPlus	AI and predictive modeling are being employed in numerous animal models to develop medicinal goods (dosage forms).
Neural graph fingerprint	Assists in predicting the properties of new compounds.
PotentialNet	Uses neural networks to estimate the binding affinity of compound³¹.

Type	Energy	Penetration	Move Ability	Persistence
Exogenous Power	Magnetic Fields	Excellent; it can function in a weak magnetic field.	Precise 3D mapping in fluids with magnetic fields that rotate.	Good micro/nanorobots can continue to move under the direction of outside influences.
	Electric energy	The electric field is relatively modest; increase its intensity.	Movement in a horizontal direction brought on by the interaction of electric and other energies.
	Light energy	The ability to pass through of different light (visible light, UV, NIR, etc.) differs.	Typically acts as a trigger for additional reactions and can produce directed movement.
	Ultrasound energy	Good, with powerful penetrating abilities.	Directional movement is usually achieved by combining it with a magnetic field.
Endogenous power	Chemical energy	Not relevant.	With the ability to move, but must be positioned by external factors (such as magnetic attraction).	Unfortunately, depletion of chemical energy can affect the motion performance of micro/nanorobots over time.