Scholarlink

Objective <p>Artificial intelligence (AI) was used to assess the compatibility of Memantine hydrochloride (MH) and Galantamine hydrobromide (GH) with formulation-specific excipients, namely PLGA 75:25, PLGA–mPEG block copolymer, and Poloxamer 407. This study explores the use of the Drug-Excipient (DE) Interaction platform for biodegradable polymers used in long-acting drug delivery systems and experimentally assesses computational predictions.</p> Methods <p>The DE-Interact artificial neural network (ANN) model was used to predict drug-excipient interactions with molecular descriptors calculated based on PubChem fingerprints using the PaDEL-descriptor tool. Fourier Transform Infrared Spectroscopy (FTIR), Differential Scanning Calorimetry (DSC), and High-Performance Thin-Layer Chromatography (HPTLC) were employed to prepare and characterize binary physical mixtures of drugs and excipients to confirm or refute the computational predictions.</p> Results <p>All drug-excipient combinations that were studied with the ANN model were found to be compatible. Those findings were supported by experimental analyses because no significant changes in spectral characteristics, thermal behavior, or chromatographic profiles were observed. The stability and compatibility of the chosen components were ensured by the fact that there were no detectable interactions or degradation.</p> Conclusion <p>AI-based prediction combined with experimental validation is an effective, fast, and low-cost method to use in compatibility screening in the initial stages of the process. Such a strategy can simplify the formulation development process and reduce the workload in experiments. More research on larger-scale data will be suggested to enhance the effectiveness and broad usage of AI-supported compatibility prediction.</p> Graphical Abstract <p></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

Artificial Intelligence-Assisted Prediction and Experimental Assessment of Drug–Excipient Compatibility for Early Formulation Development

  • 期刊论文

  • 发表日期：

    2026年7月31日

查看全文
  • Prachi Rabari,
  • Hardik B. Rana,
  • Vaishali Thakkar,
  • Tejal Gandhi,
  • Swayamprakash Patel

摘要

Objective

Artificial intelligence (AI) was used to assess the compatibility of Memantine hydrochloride (MH) and Galantamine hydrobromide (GH) with formulation-specific excipients, namely PLGA 75:25, PLGA–mPEG block copolymer, and Poloxamer 407. This study explores the use of the Drug-Excipient (DE) Interaction platform for biodegradable polymers used in long-acting drug delivery systems and experimentally assesses computational predictions.

Methods

The DE-Interact artificial neural network (ANN) model was used to predict drug-excipient interactions with molecular descriptors calculated based on PubChem fingerprints using the PaDEL-descriptor tool. Fourier Transform Infrared Spectroscopy (FTIR), Differential Scanning Calorimetry (DSC), and High-Performance Thin-Layer Chromatography (HPTLC) were employed to prepare and characterize binary physical mixtures of drugs and excipients to confirm or refute the computational predictions.

Results

All drug-excipient combinations that were studied with the ANN model were found to be compatible. Those findings were supported by experimental analyses because no significant changes in spectral characteristics, thermal behavior, or chromatographic profiles were observed. The stability and compatibility of the chosen components were ensured by the fact that there were no detectable interactions or degradation.

Conclusion

AI-based prediction combined with experimental validation is an effective, fast, and low-cost method to use in compatibility screening in the initial stages of the process. Such a strategy can simplify the formulation development process and reduce the workload in experiments. More research on larger-scale data will be suggested to enhance the effectiveness and broad usage of AI-supported compatibility prediction.

Graphical Abstract