Digital by design approach to develop a universal deep learning AI architecture for automatic chromatographic peak...

Area of Science:

• Pharmaceutical Science
• Healthcare Technology
• Artificial Intelligence

Background:

• Chromatographic data processing faces scrutiny, evidenced by FDA citations, necessitating advanced technological solutions.
• The healthcare sector's digital transformation is hindered by slow and complex technology implementation processes.
• Existing methods for chromatography peak integration require robust technological improvements.

Purpose of the Study:

• To present a "Digital by Design" managerial approach for implementing AI in healthcare's chromatography peak integration.
• To develop and propose an AI-based solution for accurate chromatography peak integration.
• To establish a GxP framework for AI utilization in the healthcare industry.

Main Methods:

• Adapted pharmaceutical Quality by Design (QbD) principles into a "Digital by Design" managerial framework.
• Utilized a convolutional neural network (CNN) model for predicting analytical variability in chromatography peak integration.
• Proposed a GxP framework encompassing data management, model management, and human-in-the-loop processes.

Main Results:

• Demonstrated the application of a CNN model for predicting analytical variability in chromatography peak integration.
• Outlined a comprehensive GxP framework for the safe and effective use of AI in healthcare.
• The analytical variability prediction component shows potential for enabling Industry 4.0 objectives.

Conclusions:

• The "Digital by Design" approach facilitates the integration of AI for chromatography peak integration in healthcare.
• AI-powered analytical variability prediction can enhance real-time release testing, automated quality control, and continuous manufacturing.
• A robust GxP framework is crucial for the successful adoption of AI in regulated healthcare environments.