Scholarlink

Purpose <p>Drug-excipient interactions significantly influence the stability, bioavailability, and therapeutic efficacy of pharmaceutical formulations. Although effective, conventional stability studies and advanced analytical techniques are often resource-intensive and time-consuming. This study introduces Drug-Excipient Interaction tool named as “DE-Interact”, an Artificial Intelligence based predictive tool designed to rapidly predict drug-excipient compatibility, thereby expediting formulation development.</p> Methods <p>The Drug-Excipient Interact model was developed using PubChem chemical fingerprints as molecular descriptors and trained with an artificial neural network. A curated dataset of drug-excipient pairs was divided into training, test, and validation sets. The tool’s predictions were experimentally validated through Differential scanning calorimetry, Fourier transform infrared spectroscopy, Thin layer chromatography, High-performance thin layer chromatography, and High-pressure liquid chromatography. Three combinations: Brinzolamide-Polyethylene glycol, Domperidone-Citric Acid, and Olanzapine-Lactose were selected as case studies.</p> Results <p>The Drug Excipient Interact model accurately predicted incompatibilities for the selected combinations, which were confirmed by experimental analyses. Differential scanning calorimetry revealed altered thermal behaviour in drug-excipient mixtures, Fourier transform infrared spectra revealed disrupted functional group signals, and chromatographic studies indicated degradation or altered chemical profiles over time. These validations support the predictive robustness of the tool.</p> Conclusions <p>Drug Excipient Interact demonstrates strong predictive performance in identifying drug-excipient incompatibilities, reducing the reliance on prolonged stability testing. By integrating computational and experimental approaches, the model provides a cost-effective, scalable, and time-efficient solution for early formulation screening. This Artificial Intelligence driven framework has the potential to accelerate pharmaceutical product development and optimize excipient selection strategies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Harnessing the Power of DE-Interact (Drug-Excipient Interaction): An Artificial Intelligence (AI)-Based Predictive Tool for Drug-Excipient Interaction in Formulation Development

  • 期刊论文

  • 发表日期：

    2025年11月21日

查看全文
  • Mangesh Kulkarni,
  • Disha Joshi,
  • Swayamprakash Patel

摘要

Purpose

Drug-excipient interactions significantly influence the stability, bioavailability, and therapeutic efficacy of pharmaceutical formulations. Although effective, conventional stability studies and advanced analytical techniques are often resource-intensive and time-consuming. This study introduces Drug-Excipient Interaction tool named as “DE-Interact”, an Artificial Intelligence based predictive tool designed to rapidly predict drug-excipient compatibility, thereby expediting formulation development.

Methods

The Drug-Excipient Interact model was developed using PubChem chemical fingerprints as molecular descriptors and trained with an artificial neural network. A curated dataset of drug-excipient pairs was divided into training, test, and validation sets. The tool’s predictions were experimentally validated through Differential scanning calorimetry, Fourier transform infrared spectroscopy, Thin layer chromatography, High-performance thin layer chromatography, and High-pressure liquid chromatography. Three combinations: Brinzolamide-Polyethylene glycol, Domperidone-Citric Acid, and Olanzapine-Lactose were selected as case studies.

Results

The Drug Excipient Interact model accurately predicted incompatibilities for the selected combinations, which were confirmed by experimental analyses. Differential scanning calorimetry revealed altered thermal behaviour in drug-excipient mixtures, Fourier transform infrared spectra revealed disrupted functional group signals, and chromatographic studies indicated degradation or altered chemical profiles over time. These validations support the predictive robustness of the tool.

Conclusions

Drug Excipient Interact demonstrates strong predictive performance in identifying drug-excipient incompatibilities, reducing the reliance on prolonged stability testing. By integrating computational and experimental approaches, the model provides a cost-effective, scalable, and time-efficient solution for early formulation screening. This Artificial Intelligence driven framework has the potential to accelerate pharmaceutical product development and optimize excipient selection strategies.