Emerging Role of Artificial Intelligence in Clinical Trials:
Enhancing Accuracy, Efficiency and Safety
Madhuri Sanjay Wable, Adsul Samruddhi Subhash, Omkar Bapurao Bramhane,
Bhand Revannath Narayan, Gayke Sanket Ramesh, Waghmare Sweeti Mohan
Shantiniketan College of Pharmacy, Dhotre, Maharashtra, India.
*Corresponding Author E-mail: ushaadsul99@gmail.com
ABSTRACT:
Clinical studies are increasingly using medical imaging. AI tools and imaging biomarker calculation are later incorporated in the clinical trial setting to decrease radiological reading times and provide objectivity to the assessment of new therapy response. Clinical trial risk assessment is increasingly using artificial intelligence (AI) to increase efficiency and safety. 142 papers published between 2013 and 2024 were examined in this scoping review, with an emphasis on operational (n = 45), efficacious (n = 46), and safety (n = 55) risk prediction. For tasks including phase transition prediction, treatment impact estimation, and adverse drug event prediction, artificial intelligence (AI) approaches such as causal machine learning, deep learning (e.g., transformers, graph neural networks), and classical machine learning are employed. These techniques make use of a variety of data sources, including patient information, scholarly articles, clinical trial protocols, and molecular structures. Applications for large language models (LLMs) have increased recently; in 2023, they were used in 7 out of 33 research. While some models achieve excellent performance (AUROC up to 96%), difficulties are still, including selection bias, limited prospective research, and data quality issues. Notwithstanding these drawbacks, AI-based risk assessment has a lot of potential to revolutionise clinical trials, especially by enhancing risk-based monitoring systems. The final half of the 1.5–2.0 billion USD, 10–15-year development cycle for a single new medicine is spent on clinical trials. As a result, a failed clinical trial costs between 800 million and 1.4 billion USD, including the preclinical development expenses in addition to the trial's investment. Only one out of ten compounds that enter a clinical trial make it to market, which is due to poor recruiting and patient cohort selection strategies as well as ineffective patient monitoring during trials. We describe how current developments in artificial intelligence (AI) can be applied to improve trial success rates by reshaping important clinical trial design phases.
KEYWORDS: Artificial Intelligence, Clinical Trials, Machine Learning Drug Development, Risk Assessment, Patient Recruitment.
INTRODUCTION:
A clinical trial is a systematic study involving human medicine1. Volunteers to assess the safety and efficacy of a novel Clinical trials are a series of evaluations in medical research and medication development that produce safety and efficacy evidence for health interventions in humans2. Clinical trials are started only once adequate information has been obtained about the quality of non-clinical safety3. Approval from the health authority or ethics committee is obtained in the country where the drug's approval is requested4. Clinical trials are systematic investigations involving human subjects to determine the safety, efficacy, and therapeutic potential of new drugs or medical interventions. They represent a critical bridge between laboratory discoveries and real-world patient applications. Typically, clinical trials progress through four structured phases, each designed to assess different parameters such as safety, dosage optimization, therapeutic efficacy, and post-marketing surveillance. The successful completion of these stages ensures that new drugs are both safe and effective before being introduced to the market.Historically, the concept of controlled experimentation dates back to ancient times—King Nebuchadnezzar’s dietary study and James Lind’s 1747 scurvy experiment are early examples of clinical trials. Over the centuries, the development of regulatory frameworks and ethical guidelines, including the formation of the U.S. FDA (1906) and the adoption of randomized controlled trial designs (1920s–1940s), established the foundation for modern clinical research. Despite these advances, the traditional clinical trial model remains expensive, lengthy, and prone to inefficiencies. The average development cost for a single drug is between $1.5 and $2 billion, and only 10% of drug candidates entering clinical trials ultimately reach the market. The primary causes of failure include poor patient recruitment, inadequate trial design, and insufficient real-time monitoring.To address these limitations, Artificial Intelligence (AI)—encompassing machine learning, deep learning, and natural language processing—has become a critical tool for optimizing the clinical trial lifecycle. From automated imaging analysis to predictive modeling and patient selection, AI technologies are transforming how data are analyzed and decisions are made.Clinical trials are essential for the introduction of novel pharmaceuticals to the market (5).
King Nebuchadnezzar carried out the first experiment that was ever recorded that resembled a clinical study6.
He commanded people to follow a diet of solely meat and alcohol because he thought it would keep them healthy7. Instead, he allowed the doubters to eat just water and beans for ten days, following which he would evaluate their health8. The king permitted the meat-loving folks to continue their diet after the studies were over since they seemed to be better fed than the required meat eaters9.
Fig no 1.1: British physician Dr. James Lind was first to perform a parallel-arm medical experiment 10
Lind discovered that of 6 therapies, oranges and lemons had the greatest positive effect on the sailor's health 11.
International Clinical Trials Day is celebrated on 20th May
· In 1906: FDA was founded
· In1923: concept of randomisation
· In 1943: the UK Medical Research Councils (MRC) trial to treat the common cold is the first double blind controlled study
Year/Period |
Milestone |
Description |
|
Ancient Era |
King Nebuchadnezzar’s dietary trial |
First documented controlled experiment |
1747 |
James Lind’s scurvy trial |
First systematic medical experiment |
1906 |
Establishment of FDA |
Regulation of drug safety begins |
1923 |
Randomization introduced |
Basis for modern trial methodology |
1943 |
First double-blind study |
MRC’s common cold trial |
2000s Present |
AI and digital integration |
Data-driven and automated clinical trials |
1. Treatment Trials
2. Prevention Trials
3. Diagnostic Trials
4. Screening Trials
5. Quality of life trials
1. Preclinical (animal) testing
2. Investigational New Drug Application (IND)
3. Phase 0 (Microdosing Study)
4. Phase I (Human Pharmacology and Safety)
5. Phase II (Therapeutic Exploration and Dose Ranging)
6. Submission of New Drug Application (NDA) is the formal step asking the FDA to consider a drug for marketing approval
7. FDA reviewers approve the application or find it either “approvable” or “not approvable”
8. Phase IV (post marketing surveillance)
|
Preclinical Data |
|
↓ |
|
IND Filling |
|
↓ |
|
IND Approval |
|
↓ |
|
Phase I |
|
↓ |
|
Phase II |
|
↓ |
|
Phase III |
|
↓ |
|
NDA filled |
|
↓ |
|
NDA Approval |
|
↓ |
|
Marketing Permission |
|
↓ |
|
Phase IV |
Fig1.Clinical Drug Development Phase
|
●Exploratory CT’s (Proof of Concept) |
Phase 0 Phase 1 Phase 2 |
Development of CTs (Pre-marketing Phase Of Clinical Drug Development |
|
●Definitive CT’s |
Phase 3 |
Post marketing Surveillance |
Extensive preclinical or laboratory research is necessary before testing in humans may begin12. Typically, research involves years of testing on human and animal cells13. If the testing is successful at this point, the sponsor gives the data to the FDA and asks for permission to start human testing14. An investigational new drug (NDA) application is what this is known as15. If authorised by the FDA, human testing is started using a properly documented and authorised prototype16.
Sr.No. |
Phase |
Objective |
Sample Size |
AI Application Example |
1 |
Phase 0 |
Microdosing, pharmacokinetics |
<20 |
Predicting optimal drug candidates |
2 |
Phase I |
Safety and dosage |
20–80 |
AI-based toxicity prediction |
3 |
Phase II |
Efficacy and side effects |
100–300 |
Biomarker-based efficacy modeling |
4 |
Phase III |
Comparison with standard therapy |
1,000+ |
AI-assisted patient stratification |
5 |
Phase IV |
Post-marketing surveillance |
Thousands |
Real-world adverse event prediction |
Phase 0: Microdosing Phase 0 is the study of novel drugs in microdoses to determine PK information in humans prior to starting phase 1 research17. A microdose is less than one hundredth of the maximum dose of a test drug that is determined to have a pharmacological effect, with a maximum dose of fewer than 100 micrograms18. The goal is to get first pharmacokinetic information19. Preclinical information: two routes of administration for a subacute toxicity study in a single species20.
Phase I: Initial phase of human subject testing Phase 1 trials are the initial phase of human subject testing21. A small (20–80) sample of healthy volunteers is often chosen22. Trials intended to evaluate a drug's safety, tolerability, pharmacokinetics, and pharmacodynamics are included in this phase23. To determine the right dose for therapeutic usage, phase 1 trials typically generally include dose range, also known as dose escalation studies24.
Phase II: Phase II trials are intended to determine whether a medication offers enough activity or promise of efficacy, or whether it offers an additional benefit, to justify additional research in a conclusive phase III trial. When practical, the same phase III endpoint or a validated surrogate endpoint, such as tumor response, time-to-event endpoint, or biomarker, could be used to evaluate efficacy in phase II trials25.Rapid assessment is a common characteristic of phase II primary endpoints26. The Response Evaluation Criteria in Solid Tumors (RECIST) has historically been employed as the primary goal for tumor response evaluation in single-arm phase II oncology trials including a single new cytotoxic drug27. The main indicator of effectiveness is frequently the objective response rate (ORR), which is the percentage of the best overall full partial responses28.
Phase III: Before seeking regulatory approval, a new medication or formulation is tested in a large group of patients in a phase 3 clinical study to verify its effectiveness, track side effects, and compare it to standard therapies. Phase III trials seek to determine which treatment is more effective for a specific kind of cancer and to learn more about its negative effects29. The impact of the treatment on individuals' quality of life
An entirely new therapy various dosages of the same medication, which is essentially the same treatment and frequently a novel approach to administering a conventional treatment (radiotherapy, for instance)30.
Typically, a lot more patients participate in phase 3 trials than in phase 1 or 231. This is due to the possibility of slight variations in success rates. For the trial to demonstrate a difference, a large number of patients are required.
Phase IV: As the name implies, the phase 4 trial, also known as post-marketing surveillance, is carried out after the medication has been sold and made accessible to the general public32. The phase 4 trial's primary goals are to evaluate the medication's effectiveness in real-world situations, investigate the long-term risks and advantages of using it, and find any uncommon side effects33. Any uncommon or long-term effects of the medication can be seen in a phase 4 trial over a considerably longer time span and in a much bigger patient group34. The medication may be taken off the market and no longer prescribed if safety monitoring does in fact show issues with it35.
AI-driven image processing reduces radiological reading times and enhances the accuracy of disease progression tracking. Deep learning models can detect subtle treatment effects and automate biomarker quantification.
AI simplifies and accelerates patient enrollment by analyzing electronic health records (EHRs), genetic data, and medical histories. Automated systems can identify eligible participants and predict dropout risks, reducing recruitment time by up to 90%.
AI systems continuously assess patient data to predict adverse drug reactions and detect protocol deviations. Predictive algorithms improve the efficiency of Risk-Based Monitoring (RBM) systems, allowing early detection of anomalies and ensuring patient safety.
AI models can forecast trial outcomes, optimize dose selection, and even simulate trial scenarios (“digital twins”) to reduce unnecessary human trials.
Since 2023, LLMs like GPT-based systems have been employed for automated data extraction, protocol analysis, and regulatory documentation, significantly accelerating data management processes.
Tab No.1.3 : Advantages and Limitations
|
Sr No |
Advantages |
Limitations/Challenges |
|
1 |
Faster data analysis |
Selection bias |
|
2 |
Reduced trial costs |
Lack of standardized AI validation |
|
3 |
Enhanced patient safety |
Data privacy and security issues |
|
4 |
Improved recruitment |
Limited prospective real-world studies |
|
5 |
Objective decision-making |
Dependence on high-quality data |
Randomisation in clinical trials:
The method of allocating clinical trial participants to treatment groups is known as randomization36. Each participant has a known chance of being assigned to any group thanks to randomization. For randomization to be successful, group assignment must be unpredictable in advance.
The goal of randomization is to rule out the third scenario. Clinical trial statistics are based on the assumption that randomization was used and worked. Valid statistical interpretation of the raw results is made possible by successful randomization37.
Evaluation of AI in clinical trials:
A pioneering project at Stanford in the 1970s, the MYCIN Project managed the creation of a method for identifying bacterial subtypes38. In diagnostic radiography, computer- aided diagnosis gained popularity as a "second opinion" or adjunct later in the 1980s39. These attempts persisted throughout the 1990s and 2000s, but computing was just too sluggish to carry them out40. Because artificial intelligence is developing so quickly, this technology is now more widely available. Deep convolutional neural networks (CNNs) were utilized for the automated classification of skin lesions in one study investigating the dermatologic-level classification of skin cancer. It was demonstrated that an AI image classification system could distinguish between benign seborrheic keratoses and keratinocyte carcinomas with performance (41) comparable to that of all evaluated experts.
Al in Patient Recruitment and Selection:
Three criteria may now be used to analyze AI's function and how it will improve the clinical trial design process: patient monitoring, cohort composition, and patient recruiting42. Patients may lose interest in continuing a clinical trial after they are chosen for it. AI systems can help with this by simplifying trial designs, automatically recommending trials, and automatically evaluating eligibility to improve the composition of patient cohorts43. An AI system was found to reduce the trial recruitment workload by 90% in research involving pediatric oncology patients. In addition to optimizing the cohort makeup, these automated procedures assemble the cohort considerably more quickly, which speeds up the recruitment process. The process of recruiting is very time-consuming and labor-intensive. As was previously said, AI technology can automate the screening process by validating and identifying patient eligibility44. The impact of real-time automated patient screening systems for clinical trial eligibility in the emergency room is examined in one study that was published in JMIR medical informatics45.
· Integration of AI with blockchain for transparent data handling
· Real-time adaptive clinical trials using continuous AI feedback
· Personalized trial designs based on patient genomics
· Collaboration between regulatory authorities and AI developers
· Development of ethical frameworks for AI decision-making in healthcare
CONCLUSION:
Although the application of AI in CTs is still relatively new, the field is rapidly developing. We expect the range of applications to expand and the number of implementations to rise quickly as regulators offer more guidelines on the acceptability of AI in particular domains. Significant and continuous advancements in artificial intelligence (AI) guarantee that the technology will be able to significantly impact the critical phases of clinical trial design, from study planning to implementation, in order to increase trial success rates, reduce associated costs, and benefit the medical community by supporting human decision-makers. Artificial Intelligence is redefining the traditional landscape of clinical trials. By improving patient recruitment, optimizing trial design, and enhancing safety monitoring, AI not only reduces development costs but also accelerates the delivery of life-saving therapies. Despite challenges like bias and data quality, the continuous advancement of AI algorithms, coupled with regulatory adaptation, promises a new era of smart, data-driven, and ethical clinical research.
1. Downing NS, Aminawung JA, Shah ND, Krumholz HM, Ross JS. Clinical trial evidence supporting FDA approval of novel therapeutic agents, 2005-2012. Jama. 2014 Jan 22; 311(4):368-77.
2. Umscheid CA, Margolis DJ, Grossman CE. Key concepts of clinical trials: a narrative review. Postgraduate Medicine. 2011 Sep 1; 123(5): 194-204.
3. Andrade EL, Bento AF, Cavalli J, Oliveira SK, Schwanke RC, Siqueira JM, Freitas CS, Marcon R, Calixto JB. Non-clinical studies in the process of new drug development-Part II: Good laboratory practice, metabolism, pharmacokinetics, safety and dose translation to clinical studies. Brazilian Journal of Medical and Biological Research. 2016; 49(12): e5646.
4. Yank V, Rennie D. Reporting of informed consent and ethics committee approval in clinical trials. JAMA. 2002 Jun 5; 287(21): 2835-8.
5. Dorsey ER, Venuto C, Venkataraman V, Harris DA, Kieburtz K. Novel methods and technologies for 21st-century clinical trials: a review. JAMA Neurology. 2015 May 1; 72(5): 582- 8.
6. Sack R. Images of Nebuchadnezzar: The emergence of a legend. Susquehanna University Press; 2004.
7. Willett W, Skerrett PJ, Giovannucci EL. Eat, drink, and be healthy: the Harvard Medical School guide to healthy eating. Simon and Schuster; 2017 Sep 19.
8. Nabhan GP. Coming home to eat: The pleasures and politics of local foods. WW Norton & Company; 2002 Nov 17.
9. Shapiro P. Clean meat: how growing meat without animals will revolutionize dinner and the world. Simon and Schuster; 2024 Apr 9.
10. Collier R. Legumes, Lemons and Streptomycin: A short history of the clinical trial.
11. Baron JH. Sailors' scurvy before and after James Lind–a reassessment. Nutrition Reviews. 2009 Jun 1; 67(6): 315-32.
12. Nair A, Morsy MA, Jacob S. Dose translation between laboratory animals and human in preclinical and clinical phases of drug development. Drug Development Research. 2018 Dec; 79(8): 373-82.
13. DiPaolo JA. Relative difficulties in transforming human and animal cells in vitro. Journal of the National Cancer Institute. 1983 Jan 1; 70(1): 3-8.
14. Van Norman GA. Drugs, devices, and the FDA: part 1: an overview of approval processes for drugs. JACC: Basic to Translational Science. 2016 Apr; 1(3): 170-9.
15. Holbein MB. Understanding FDA regulatory requirements for investigational new drug applications for sponsor-investigators. Journal of Investigative Medicine. 2009 Aug; 57(6): 688- 94.
16. Kaplan AV, Baim DS, Smith JJ, Feigal DA, Simons M, Jefferys D, Fogarty TJ, Kuntz RE, Leon MB. Medical device development: from prototype to regulatory approval. Circulation. 2004 Jun 29; 109(25): 3068-72.
17. Garner RC, Lappin G. The phase 0 microdosing concept. British Journal of Clinical Pharmacology. 2006 Apr; 61(4): 367.
18. Lappin G, Kuhnz W, Jochemsen R, Kneer J, Chaudhary A, Oosterhuis B, Drijfhout WJ, Rowland M, Garner RC. Use of microdosing to predict pharmacokinetics at the therapeutic dose: experience with 5 drugs. Clinical Pharmacology & Therapeutics. 2006 Sep; 80(3): 203- 15.
19. Banner Jr W. Pharmacokinetics and. Nursing Care of the Critically Ill Child. 2012 May 8: 67.
20. Denny KH, Stewart CW. Acute, subacute, subchronic, and chronic general toxicity testing for preclinical drug development. InA Comprehensive Guide to Toxicology in Nonclinical Drug Development. 2024 Jan 1 (pp. 149-171). Academic Press.
21. Storer BE. Design and Analysis of Phase I Clinical Trials. Biometrics. 1989 Sep 1: 925-37.
22. Almeida L, Azevedo B, Nunes T, Vaz-da-Silva M, Soares-da-Silva P. Why healthy subjects volunteer for phase I studies and how they perceive their participation?. European journal of clinical pharmacology. 2007 Nov; 63(11):1085-94.
23. Thomas AL, Morgan B, Horsfield MA, Higginson A, Kay A, Lee L, Masson E, Puccio-Pick M, Laurent D, Steward WP. Phase I study of the safety, tolerability, pharmacokinetics, and pharmacodynamics of PTK787/ZK 222584 administered twice daily in patients with advanced cancer. Journal of Clinical Oncology. 2005 Jun 20; 23(18): 4162-71.
24. Le Tourneau C, Lee JJ, Siu LL. Dose escalation methods in phase I cancer clinical trials. JNCI: Journal of the National Cancer Institute. 2009 May 20; 101(10):708-20.
25. Conforti F, Pala L, Sala I, Oriecuia C, De Pas T, Specchia C, Graffeo R, Pagan E, Queirolo P, Pennacchioli E, Colleoni M. Evaluation of pathological complete response as surrogate endpoint in neoadjuvant randomised clinical trials of early stage breast cancer: systematic review and meta-analysis. BMJ. 2021 Dec 21; 375.
26. Delgado A, Guddati AK. Clinical endpoints in oncology-a primer. American Journal of Cancer Research. 2021 Apr 15; 11(4): 1121.
27. Ruchalski K, Braschi-Amirfarzan M, Douek M, Sai V, Gutierrez A, Dewan R, Goldin J. A primer on RECIST 1.1 for oncologic imaging in clinical drug trials. Radiology: Imaging Cancer. 2021 May 14; 3(3): e210008.
28. Chen X, Ma K. Neoadjuvant therapy in lung cancer: what is most important: objective response rate or major pathological response?. Current Oncology. 2021 Oct 13; 28(5): 4129- 38.
29. McCormack M, Eminowicz G, Gallardo D, Diez P, Farrelly L, Kent C, Hudson E, Panades M, Mathew T, Anand A, Persic M. Induction chemotherapy followed by standard chemoradiotherapy versus standard chemoradiotherapy alone in patients with locally advanced cervical cancer (GCIG INTERLACE): an international, multicentre, randomised phase 3 trial. The Lancet. 2024 Oct 19; 404(10462): 1525-35.
30. Liu B, Zhou H, Tan L, Siu KT, Guan XY. Exploring treatment options in cancer: tumor treatment strategies. Signal Transduction and Targeted Therapy. 2024 Jul 17; 9(1): 175.
31. Machiels JP, Tao Y, Licitra L, Burtness B, Tahara M, Rischin D, Alves G, Lima IP, Hughes BG, Pointreau Y, Aksoy S. Pembrolizumab plus concurrent chemoradiotherapy versus placebo plus concurrent chemoradiotherapy in patients with locally advanced squamous cell carcinoma of the head and neck (KEYNOTE-412): a randomised, double-blind, phase 3 trial. The Lancet Oncology. 2024 May 1; 25(5): 572-87.
32. Le Flohic E, Vrijens B, Hiligsmann M. The impacts of undetected nonadherence in phase II, III and post‐marketing clinical trials: an overview. British Journal of Clinical Pharmacology. 2024 Aug; 90(8): 1984-2003.
33. Van Spall HG, Bastien A, Gersh B, Greenberg B, Mohebi R, Min J, Strauss K, Thirstrup S, Zannad F. The role of early-phase trials and real-world evidence in drug development. Nature Cardiovascular Research. 2024 Feb; 3(2): 110-7.
34. Kümpfel T, Giglhuber K, Aktas O, Ayzenberg I, Bellmann-Strobl J, Häußler V, Havla J, Hellwig K, Hümmert MW, Jarius S, Kleiter I. Update on the diagnosis and treatment of neuromyelitis optica spectrum disorders (NMOSD)–revised recommendations of the Neuromyelitis Optica Study Group (NEMOS). Part II: Attack therapy and long-term management. Journal of Neurology. 2024 Jan; 271(1): 141-76.
35. Mascarenhas-Melo F, Diaz M, Gonçalves MB, Vieira P, Bell V, Viana S, Nunes S, Paiva- Santos AC, Veiga F. An overview of biosimilars—development, quality, regulatory issues, and management in healthcare. Pharmaceuticals. 2024 Feb 11; 17(2): 235.
36. Miola AC, Espósito AC, Miot HA. Techniques for randomization and allocation for clinical trials. Jornal Vascular Brasileiro. 2025 Jan 13; 23: e20240046.
37. Hopewell S, Chan AW, Collins GS, Hróbjartsson A, Moher D, Schulz KF, Tunn R, Aggarwal R, Berkwits M, Berlin JA, Bhandari N. CONSORT 2025 explanation and elaboration: updated guideline for reporting randomised trials. BMJ. 2025 Apr 14; 389.
38. Hamamoto R, Komatsu M, Yamada M, Kobayashi K, Takahashi M, Miyake M, Jinnai S, Koyama T, Kouno N, Machino H, Takahashi S. Current status and future direction of cancer research using artificial intelligence for clinical application. Cancer Science. 2025 Feb; 116(2): 297-307.
39. Myslicka M, Kawala-Sterniuk A, Bryniarska A, Sudol A, Podpora M, Gasz R, Martinek R, Kahankova Vilimkova R, Vilimek D, Pelc M, Mikolajewski D. Review of the application of the most current sophisticated image processing methods for the skin cancer diagnostics purposes. Archives of Dermatological Research. 2024 Mar 6; 316(4): 99.
40. Stonebraker M, Pavlo A. What Goes Around Comes Around... And Around... ACM Sigmod Record. 2024 Jul 31; 53(2): 21-37.
41. Reddy S, Giri D, Patel R. Artificial intelligence-based distinction of actinic keratosis and seborrheic keratosis. Cureus. 2024 Apr 21; 16(4).
42. Sedano R, Solitano V, Vuyyuru SK, Yuan Y, Hanžel J, Ma C, Nardone OM, Jairath V. Artificial intelligence to revolutionize IBD clinical trials: a comprehensive review. Therapeutic Advances in Gastroenterology. 2025 Feb; 18:17562848251321915.
43. Sedano R, Solitano V, Vuyyuru SK, Yuan Y, Hanžel J, Ma C, Nardone OM, Jairath V. Artificial intelligence to revolutionize IBD clinical trials: a comprehensive review. Therapeutic Advances in Gastroenterology. 2025 Feb; 18:17562848251321915.
44. Beattie J, Neufeld S, Yang D, Chukwuma C, Gul A, Desai N, Jiang S, Dohopolski M. Utilizing large language models for enhanced clinical trial matching: a study on automation in patient screening. Cureus. 2024 May 10; 16(5).
45. Lee K, Liu Z, Mai Y, Jun T, Ma M, Wang T, Ai L, Calay E, Oh W, Stolovitzky G, Schadt E. Optimizing clinical trial eligibility design using natural language processing models and real- world data: algorithm development and validation. JMIR AI. 2024 Jul 29; 3:e50800.
|
Received on 13.11.2025 Revised on 08.12.2025 Accepted on 28.12.2025 Published on 15.04.2026 Available online from April 18, 2026 Asian J. Pharm. Res. 2026; 16(2):201-206. DOI: 10.52711/2231-5691.2026.00030 ©Asian Pharma Press All Right Reserved
|
|
|
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Creative Commons License. |
|