Ai AI and machine learning ML have been widely used in drug discovery and clinical research in the biopharmaceutical industry. These technologies also have the potential to influence many other aspects of drug development and manufacturing. In recent years, because AI / ML has the potential to simplify and accelerate this process, such applications in formulation prescription development have been increasing, especially for challenging APIs, such as those with poor solubility and bioavailability.
To determine the optimal prescription, it involves extensive, time-consuming and expensive in vitro and in vivo physical experiments and screening. Using the AI / ML algorithm can identify the most likely successful formulation prescriptions, thereby reducing the time and resources required for physical evaluation. These techniques also allow the discovery of new materials with excipient characteristics and methods for new formulation prescription development that have never been considered before.
Has played a key role in the formulation prescription development
Preparations play a key role in stabilizing the drug substance, promoting patient administration, and reducing potential side effects."Unfortunately," said Christine Allen), CEO of biotechnology startup Intrepid Labs at the University of Toronto. " Although preparations directly improve safety and effectiveness, the huge potential of preparations is often underestimated and has a low priority in drug development. She noted that due to safety concerns and lack of efficacy, about 80 percent of drugs fail in clinical studies, and those with better formulations are likely to succeed.”
Poor solubility / bioavailability poses many challenges to R & D
According to Allen, 70% to 90% of new chemical entities (NCE) have solubility problems. Added Sanjay Konagurthu, Senior Director of Science and Innovation, added, " A formulation that is well in improving solubility and bioavailability but is inadequate in safety, efficacy or patient acceptability fails to meet its therapeutic potential. Therefore, one of the biggest challenges is for the researchers to carefully balance these factors and develop formulation prescriptions that can not only improve drug delivery but also meet the broader criteria needed for successful patient treatment.”
Speed is another persistent challenge in the journey from molecules to medicine."While maintaining safety and efficacy, scientists are looking for a faster way to overcome barriers such as poor solubility and bioavailability and the trial-and-error prescription development process to deliver therapy to patients as soon as possible," Konagurthu explains.
Since poor API solubility generally corresponds to lower bioavailability that affects potency, there is a negative chain effect from discovery to clinical testing. From the molecular characteristics to the characteristics of the solubility and bioavailability of API."For this reason, identifying the most effective combination of API and excipients is a complex and resource-intensive process that requires rigorous experiments, data collection and analysis," he said.
Historically, scientists have hoped to address the problem of poor API solubility with trial-and-error testing, which often leads to longer development times."As pharmaceutical companies focus on providing life-saving therapies for patients as soon as possible, there is an urgent need for streamlined workflows that will enable scientists to effectively find the best ways to improve drug bioavailability," Konagurthu concludes.
The prediction function has many advantages
Advanced prediction algorithms can help to overcome this important challenge."AI / ML provides low-cost predictions for properties such as solubility, dissolution rate, and stability," says Allen."Importantly, advanced AI AI AI methods can no longer rely on ML models that have a strong need for large amounts of data. In addition, the combination of automatic preparation and real-time sensing technology further enhances the data-driven process, " she added.
In addition to predicting the most effective combination of solubility enhancement techniques and excipients for drug development, the use of AI / ML in such applications can help scientists better understand the complex behavior of molecules, increase the speed of API discovery, and improve the accuracy of formulation prescription development."Ultimately, because these new technologies are able to generate high-quality data and insights for the entire drug discovery and development process, their use can have a significant positive impact on budgets, schedules, and resources," Konagurthu said.
Different data types are very useful
Another advantage of using AI / ML in formulation prescription development is the ability to utilize extensive datasets based on predicted targets, route of administration, etc. For example, an ML model trained on a dataset can be used to predict the apparent solubility or dissolution rate. In many cases, computational chemical methods, including quantum mechanics (QM) and molecular dynamics (MD), quantitative structure-activity relationship models, and aspects of ADMET (absorption, distribution, metabolism, excretion, and toxicity) analysis are also used to create customized prescription predictions. To specifically address the problem of poor solubility, many solutions use proprietary algorithms to create predictive models based on substance-specific properties. These tools are then used to generate predictive models for improved solubility and bioavailability; accelerated stability models for shelf life and packaging determination; material science, compaction simulations, and process models; and ADME / PK (pharmacokinetics) models for predicting the impact of API physiochemical properties and pharmacokinetics.
Konagurthu Observe that techniques like QM / MD provide modeling functions that allow insight into physicochemical properties and molecular-level interactions for structural mechanism exploration, free energy assessment, and spectral characterization."It has turned out that combining these insights with AI / ML methods is more accurate and effective than trial-and-error experiments," he added.
Application in formulation prescription development is increasing
Konagurthu Said that the literature has increased substantially over the past two decades on the role of computational chemistry in the pharmaceutical industry, highlighting the ability of AI / ML in formulation prescription development. He noted: " The literature shows an increase in adoption rates and provides evidence for the scientific community of how these new technologies address the long-standing industry challenge: the problem of poor API solubility and bioavailability.
Konagurthu said one way to infer the extent of AI / ML adoption is to see the increase in regulatory submissions involving these methods. Since 2021, the FDA has received more than 100 pharmaceutical and biological applications using AI / ML components, covering the field of drug development from molecules to drugs. He noted that such a surge has prompted the FDA to create dedicated regulatory groups that are both resources for the pharmaceutical industry and promoters of innovation.
Specific to the high API development formulation prescription, Allen observed that the use of AI / ML is less advanced than other drug development areas, but it is growing.
There are still some obstacles yet to be removed
There are many barriers to clear for further formulation prescription development using AI / ML solutions. As with any data-driven application, a amount of relevant high-quality data is required to create and train AI / ML models for formulation prescription development."Alternative ML methods that require less data should also be evaluated. However, a real success in this area requires integrating the AI / ML platform into the existing pharmaceutical framework, " Allen commented.
Other challenges noted by Konagurthu, include model interpretability, intellectual property issues, implementation costs, labor training, and the need for regulators to develop and establish standards that not only help maintain the integrity of these algorithms, but also help regulators assess reporting for AI / ML use in drug development.
According to Allen, these barriers could be overcome by building more interdisciplinary teams, integrating AI / ML expertise with pharmaceutical knowledge, and increasing investment in AI / ML capacity and training in the pharmaceutical industry."Joint efforts in data management, regulatory engagement, technology integration, and training will ensure that the advantages of AI / ML can be fully realized in drug formulation prescription development," Konagurthu agreed.
Partnerships can be very valuable
Konagurthu Said drug developers can also make better use of AI / ML platforms by identifying and working with the CDMO, organizations that have extensive experience and have a good track record of managing the challenges of new technology adoption."For scientists looking to address poor API solubility and bioavailability, working with a CDMO specializing in computational drug development could be a long way. CDMO is committed to staying at the forefront of drug innovation and adopting new technologies focused on speed, scalability, and innovation as part of its end-to-end products, " he said. He also stressed that extensive collaboration between pharmaceutical companies, CDMOs and regulators on next-generation technologies could bring many positive benefits to the pharmaceutical industry and the patients it serves.
The AI / ML is expected to have a significant impact
"Addressing the and bioavailability challenges of solubility using AI / ML technology is a major shift in the pharmaceutical industry. These technologies provide a more streamlined, more accurate and more cost-effective way forward for drug development. Especially for insoluble APIs, they are able to optimize solubility enhancement technologies and selection of excipients and pave the way for more efficient and sustainable processes, " Konagurthu said.
In the short term, Allen expects that the enhanced predictive power provided by the AI / ML platform will lead to a more efficient formulation prescription development process. In the long run, as the pharmaceutical formulation prescribing datasets reach critical quality and scientists become more proficient in AI / ML, continued advances in these technologies will continue to unleash new potential in formulation science that requires constant adaptation and learning within the industry."AI / ML has great potential to accelerate the discovery of new drug preparation prescriptions and innovative drug delivery technologies," she commented.
In fact, Konagurthu believes that AI / ML technology " promises to further accelerate the process from molecules to medicine and lead us to a healthier society. He added: " As the industry continues to evolve, embracing these technological advances will be key to overcoming the long-standing challenges of drug development and unlocking new possibilities in the search for more effective and accessible therapies.”
